Selective cleavage of nucleolar autoantigen B23 by granzyme B in differentiated vascular smooth muscle cellsInsights into the association of specific autoantibodies with distinct disease phenotypes.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 50, No. 1, January 2004, pp 233–241 DOI 10.1002/art.11485 © 2004, American College of Rheumatology Selective Cleavage of Nucleolar Autoantigen B23 by Granzyme B in Differentiated Vascular Smooth Muscle Cells Insights Into the Association of Specific Autoantibodies With Distinct Disease Phenotypes Danielle B. Ulanet,1 Nicholas A. Flavahan,2 Livia Casciola-Rosen,1 and Antony Rosen1 Conclusion. These data demonstrate that the cleavage of B23 by GB in intact cells is dependent upon both cell type and phenotype. The susceptibility of this autoantigen (which is associated with a distinct pulmonary vascular phenotype in scleroderma) to GBmediated proteolysis selectively in vascular smooth muscle cells suggests that the GB-cleavable conformation of autoantigens may occur selectively in the target tissue, and may play a role in shaping the phenotypespecific autoimmune response. Objective. To investigate the association of specific autoantibodies with distinct disease phenotypes. The association of autoantibodies to nucleophosmin/ B23 with pulmonary hypertension in scleroderma, and the susceptibility of autoantigens to cleavage by granzyme B (GB), provided a focus for these studies. Methods. Intact cells were subjected to cytotoxic lymphocyte granule–induced death, and the susceptibility of autoantigens to cleavage by GB was addressed by immunoblotting and/or by a novel immunofluorescence assay. Results. B23 was cleaved efficiently by GB in vitro, but was highly resistant to cleavage by GB during cytotoxic lymphocyte granule–mediated death of many intact cell types. In contrast, this molecule was highly susceptible to GB-mediated proteolysis exclusively in differentiated vascular smooth muscle cells. Topoisomerase I and several other GB substrates did not show this striking change in cleavage susceptibility in different cell types. The hallmarks of systemic sclerosis (scleroderma) include widespread vascular abnormalities (1), excess tissue fibrosis (2), and a high-titer autoantibody response to ubiquitously expressed nuclear/nucleolar proteins (3). While the etiology of the autoimmune response and accompanying vascular pathology is currently unknown, clues may be derived from the observation that different scleroderma-associated autoantibodies serve as markers for specific clinical phenotypes within the disease. For example, antibodies to centromere proteins are characteristic of limited scleroderma (4,5), while antibodies to topoisomerase I are most commonly found in diffuse scleroderma in association with pulmonary fibrosis (6,7). Recently, antibodies to the nucleolar phosphoprotein, nucleophosmin/B23, were identified in scleroderma patient sera, and were found to occur in association with pulmonary hypertension (8), similar to antifibrillarin antibodies (9,10). Interestingly, the antiB23 autoantibodies could be detected in patient sera prior to detection of this pulmonary phenotype. It has been suggested that autoantibodies may be acting as “immunologic imprints” of the events that induced their generation (11). The striking association of different Supported by the Scleroderma Research Foundation, an institutional grant from the Maryland Chapter of the Arthritis Foundation, and by the Donald Stabler Foundation. Dr. Casciola-Rosen’s work was supported by a grant from the NIH (AR-44684). Dr. Rosen’s work was supported by a Burroughs Wellcome Fund Translational Research Award and by grants from the NIH (DE-12354 and HL56091). 1 Danielle B. Ulanet, PhD (current address: University of California–San Francisco Diabetes Center), Livia Casciola-Rosen, PhD, Antony Rosen, MD: Johns Hopkins University School of Medicine, Baltimore, Maryland; 2Nicholas A. Flavahan, PhD: Heart and Lung Institute, Ohio State University, Columbus. Address correspondence and reprint requests to Antony Rosen, MD, Johns Hopkins University School of Medicine, 720 Rutland Avenue, Ross 1059, Baltimore, MD 21205. E-mail: email@example.com or firstname.lastname@example.org. Submitted for publication October 3, 2002; accepted in revised form September 11, 2003. 233 234 ULANET ET AL autoantibodies with distinct phenotypes implies that ubiquitously expressed autoantigens may exhibit specific features in the microenvironment in which initial injury and subsequent disease propagation occurs. Hence, defining the specific alterations that self-proteins may display in the disease-relevant target tissues may shed some light on important pathogenic events. One property that unifies most autoantigens targeted across the spectrum of autoimmunity, which is not shared by non-autoantigens, is susceptibility to cleavage by the cytotoxic lymphocyte granule protease, granzyme B (GB) (12–18). During cytotoxic lymphocyte granule– mediated cell death, proteolysis by GB allows for the generation of autoantigen fragments not observed during other forms of cell death. Of note, all of the scleroderma autoantigens examined to date (topoisomerase I, NOR-90, RNA polymerase II, CENP-B, fibrillarin) are efficiently cleaved by GB in vitro. Since autoantibodies to these molecules are associated with distinct clinical phenotypes in scleroderma, we have hypothesized that this association may be a function of altered susceptibility of autoantigens to unique fragmentation by GB in the disease-relevant microenvironment. In scleroderma, this microenvironment likely includes the small vessels that become damaged during disease initiation and propagation (19). The association of anti-B23 autoantibodies with pulmonary vascular disease (8) focused our attention on the differential susceptibility of this autoantigen to cleavage by GB in cells of the vasculature. We found that while B23 was efficiently cleaved by GB in vitro, its cleavage by GB was not observed in most intact cell types undergoing cytotoxic lymphocyte granule–induced death. In contrast, B23 cleavage by GB was highly efficient in differentiated vascular smooth muscle cells (VSMCs) undergoing cytotoxic lymphocyte granule– induced death. This differential susceptibility to cleavage was not a feature of several other scleroderma autoantigens. We propose that unique features of autoantigens in the disease-relevant microenvironment may influence their susceptibility to cleavage by GB and thereby favor their targeting by the autoimmune response. MATERIALS AND METHODS Isolation of granule contents from the natural killer (NK) cell line, YT, and purification of GB were as previously described (20). The polyclonal rabbit antibody raised against intact B23 (R3434) was generated as previously described (8) and affinity-purified against recombinant B23 that was electrophoresed and transferred onto Immobilon-P (Millipore, Bedford, MA) according to the method of Olmsted (21). The B23 conformation-specific antibody (R3956) was generated in rabbits by immunization with an 8–amino acid peptide of B23 (comprising the LAAD161 GB cleavage site) conjugated to keyhole limpet hemocyanin (KLH) via an N-terminal cysteine residue (KLH-KKVKLAAD). The R3434 antibody preferentially recognizes intact B23 by immunoblotting (with some recognition of GB-induced fragments), while R3956 specifically immunoblots the 22-kd N-terminal GB cleavage fragment of B23 generated by cleavage after Asp161. In vitro cleavage of endogenous nucleolar autoantigens by GB in HeLa lysates. HeLa cells were grown using standard tissue culture procedures in 10% heat-inactivated newborn calf serum. HeLa cells were lysed in Nonidet P40 (NP40) lysis buffer (1% NP40, 20 mM Tris [pH 7.4], 150 mM NaCl, 1 mM EDTA) containing the protease inhibitors pepstatin, antipain, leupeptin, and phenylmethylsulfonyl fluoride and incubated at 37°C for 60 minutes with either purified GB or YT cell granule contents (as indicated in the various figure legends) and 2 mM iodoacetamide (IAA). The amount of granule contents used was based on the GB activity (1 l granule contents contained approximately the same activity as 1 l [3.7 pmoles] of purified GB). Reactions were electrophoresed on 12% sodium dodecyl sulfate (SDS)–polyacrylamide gels and transferred to nitrocellulose for immunoblotting with a patient antiserum recognizing topoisomerase I, or with the polyclonal rabbit antibody recognizing B23 (R3434). Proteins were detected using horseradish peroxidase–labeled secondary antibodies (Pierce, Rockford, IL) and chemiluminescence. Cleavage of endogenous nucleolar autoantigens after in vivo incubation of intact cells with YT cell granule contents or lymphokine-activated killer (LAK) cells. LAK cells were prepared and incubated with K562 Fas-deficient target cells (at a 3:1 effector:target ratio), as previously described (20). Fibroblasts were obtained from human skin biopsies and grown, using standard tissue culture procedures, in Dulbecco’s modified Eagle’s medium (high glucose) with 20% heat-inactivated fetal calf serum. Human umbilical vein endothelial cells (HUVECs) and aortic smooth muscle cells (AoSMCs) were grown with Clonetics EGM-2 and SmGM-2 BulletKits, respectively (BioWhittaker, Walkersville, MD). AoSMCs were differentiated in culture by growth in serum-free medium containing an insulin–transferrin–selenium-X supplement (Gibco BRL, Gaithersburg, MD) for 3 days, followed by growth for an additional 3 days in the same medium supplemented with 25 M MnTMPyP (a superoxide dismutase mimic) and 2 mM N-acetylcysteine (Alexis Biochemicals, Florence, Italy), as previously described (22). For granule content experiments, cells were washed with Ca⫹⫹-free Hanks’ balanced salt solution (HBSS; Gibco BRL), lifted off 10-cm dishes, and resuspended at 8 ⫻ 106 cells/ml in Ca⫹⫹-free HBSS containing 10 mM HEPES (pH 7.4), 0.5M MgCl2, and 0.4M MgSO4. Cells were incubated in the absence or presence of granule contents for 5 minutes on ice. A final concentration of 1.5 mM CaCl2 was then added to each reaction and cells were incubated at 37°C for 2 hours. Autoantigen cleavage was monitored by immunoblot analysis following SDS–polyacrylamide gel electrophoresis and transfer to nitrocellulose. Patient antisera were used to immunoblot poly(ADP-ribose) polymerase (PARP), nuclear mitotic appa- CLEAVAGE OF AUTOANTIGEN B23 BY GRANZYME B 235 ratus protein (NuMA), La, U1–70 kd, Ufd-2, fibrillarin, and Mi-2. Calculation of relative catalytic constant (kcat/Km) values. The percent cleavage of each substrate was determined by densitometry. Together with the known enzyme concentration and incubation time, these values were fitted to the first-order rate equation, % substrate cleaved ⫽ 100 ⫻ (1 ⫺ e⫺kcat 䡠 [E]/Km 䡠 time), to calculate nominal kcat/Km. To calculate the difference in cleavage efficiency of different substrates in undifferentiated and differentiated SMCs, kcat/Km values in differentiated cells were divided by those in undifferentiated cells. Site-directed mutagenesis to identify GB cleavage site. A complementary DNA (cDNA) clone encoding B23 was used as a template to generate clones with a selected Asp3 Ala substitution in the putative GB substrate P1 position. Mutagenesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Incorporation of the desired point mutation without additional changes to the original sequence was confirmed by sequencing. The mutated construct was subjected to in vitro transcription/translation in a rabbit reticulocyte lysate-based system (Promega, Madison, WI) to generate 35S-methionine–labeled protein, and the 35 S-methionine–labeled polypeptide was treated with GB as described above to measure for inhibition of cleavage. For detection of oligomeric B23, in vitro–translated samples were electrophoresed on 10% SDS-polyacrylamide minigels in sample buffer containing 0.1% SDS and 5 mM dithiothreitol without prior boiling. Detection of GB-induced B23 fragments by immunofluorescence on AoSMCs treated with granule contents. AoSMCs were plated on coverslips (#1 thickness; Bellco, Vineland, NJ) and grown in either normal growth medium or serum-free medium for differentiation. Cells were washed several times in Ca⫹⫹-free HBSS, and coverslips were incubated in the absence or presence of granule contents in Ca⫹⫹-free HBSS containing HEPES and Mg⫹⫹ for 10 minutes at 4°C. An equal volume of the same buffer, supplemented with 3 mM CaCl2, was then added to the coverslips (for 1.5 mM final concentration) and cells were incubated in a humidified chamber at 37°C for 40 minutes. For immunofluorescence, cells were fixed in 4% paraformaldehyde for 5 minutes at 4°C and permeabilized in cold acetone for 20 seconds. The affinitypurified R3956 antibody was used at 10 g/ml and visualized with a fluorescein isothiocyanate–conjugated secondary antibody (Jackson ImmunoResearch, Avondale, PA). Coverslips were mounted on glass slides with Permount (Lipshaw, Pittsburgh, PA) containing DABCO (Sigma, St. Louis, MO) prior to viewing on an Axioskop fluorescent microscope (Zeiss, Thornwood, NY) and photography with a DC290 Zoom digital camera (Kodak, Rochester, NY). RESULTS Cleavage of endogenous B23 and topoisomerase I by GB in HeLa lysates and during cytotoxic lymphocyte granule–mediated death of intact cells. The susceptibility of B23 and topoisomerase I to cleavage by GB in Figure 1. Demonstration, by immunoblotting, that B23 is not cleaved during granule-induced death of intact target cells, but is cleaved in cell lysates. A, HeLa cell lysates containing 2 mM iodoacetamide were incubated in the absence or presence of granzyme B (GB) or YT granule contents (GC) as follows: topoisomerase I (topo I), 17.2 nM GB, B23, 1.0 l granule contents (⬃172 nM GB). Protein (50 g) was electrophoresed in each gel lane, and autoantigens were detected by immunoblotting using R3434 (for B23 detection) and monospecific patient serum (topo I). B, Intact HeLa cells were incubated in the absence or presence of 2.5 l YT granule contents for 2 hours at 37°C, as described in Materials and Methods. After termination of the reactions, 2 ⫻ 105 cells were electrophoresed in each lane and the samples were immunoblotted as described in A. Solid arrows indicate intact antigens; open arrows indicate GB-cleaved fragments. HeLa cell lysates was addressed by incubating lysates in the absence or presence of GB or YT cell granule contents (containing GB), and autoantigen cleavage was detected by immunoblotting. HeLa lysates were supplemented with 2 mM IAA to inhibit the activity of endogenous caspases. B23 was well cleaved by GB in this system, with the production of 2 major fragments of 20 and 22 kd (Figure 1A). To address the fate of B23 in intact cells undergoing cytotoxic lymphocyte granule– induced death, the cleavage of B23 was examined in HeLa cells exposed to YT cell granule contents (Figure 1B). B23 was highly resistant to cleavage in this system, as demonstrated by a lack of/loss of intact B23 in cleaved samples, and a complete absence of B23 fragments. An overexposed image is shown to highlight this (Figure 1B). This was in marked contrast to topoisomerase I, which is efficiently cleaved by GB under these conditions. 236 ULANET ET AL PARP cleavage in the K562 target cells could be readily assessed. In contrast to B23, PARP was well cleaved in the target cells (Figure 2, upper panel). To precisely define the GB cleavage site(s) in B23, the B23 cDNA was used as a template for sitedirected mutagenesis of potential P1 aspartic acid residues in B23 that contained the GB tetrapeptide consensus sequence recently described by Thornberry and colleagues (23,24). Cleavage of B23 by GB was completely abolished when the aspartic acid residue D161 was mutated to alanine (Figure 3), defining LAAD161 as the GB cleavage site. Cleavage of B23 in VSMCs during cytotoxic lymphocyte granule–mediated death. Since autoantigens targeted across the spectrum of systemic autoimmune diseases are GB substrates and are expressed ubiquitously in most cell types, while only subsets of these autoantigens are targeted by the autoimmune response in a particular disease, we hypothesized that the susceptibility of autoantigens to cleavage by GB may be favored in the disease-relevant microenvironment. The cleavage of B23 and other autoantigens was examined during the cytotoxic lymphocyte granule–mediated death of several different cell types that are of potential relevance in scleroderma. The cell types chosen included the major cellular constituents of blood vessels (fibroblasts, endothelial cells, and SMCs), potential sites of disease initiation and propagation in scleroderma. Ex- Figure 2. Lymphokine-activated killer (LAK) cell–induced death of K562 cells. LAK cells were incubated with K562 target cells, as described in Materials and Methods. After termination of the reactions, 3 ⫻ 105 LAK cells, 1 ⫻ 105 K562 cells, or 3 ⫻ 105 LAK cells plus 1 ⫻ 105 K562 cells were electrophoresed in each gel lane. B23 and poly(ADP-ribose) polymerase (PARP) were detected by immunoblotting with R3434 (for B23) and a monospecific patient serum (for PARP). B23 cleavage in a cell–cell killing assay was also evaluated using K562 target cells incubated with LAK cells. Since K562 cells are Fas-negative, killing of these cells occurs exclusively via the granule exocytosis pathway. In this intact cell-killing assay, B23 cleavage was entirely absent (Figure 2, lower panel). The proteolysis of PARP was also assayed to confirm that the K562 target cells were indeed undergoing apoptosis. Since PARP is not expressed in LAK cells, the efficiency of Figure 3. Cleavage of B23 by granzyme B (GB) at LAAD161. Wildtype (wt) and mutated B23 (D1613 A) 35S-methionine–labeled polypeptides, containing a point mutation in the P1 position(s) of the predicted GB cleavage site (LAAD161), were generated. The polypeptides were incubated in the absence or presence of granule contents (GC) for 60 minutes at 37°C. After termination of the reactions, the gel samples, containing 0.1% sodium dodecyl sulfate, were electrophoresed. Polypeptides were detected by fluorography. Depicted is the oligomeric form of B23, in which cleavage by GB occurs exclusively after D161. CLEAVAGE OF AUTOANTIGEN B23 BY GRANZYME B 237 Figure 4. Generation of B23 fragments during cytotoxic lymphocyte granule–mediated death of vascular cell types. A, Intact fibroblasts and undifferentiated (U) or differentiated (D) aortic smooth muscle cells (AoSMCs) were incubated in the absence or presence of 2.5 l YT granule contents (GC) for 2 hours at 37°C. After termination of the reactions, 2 ⫻ 105 cells were electrophoresed in each lane, and the samples were immunoblotted with R3434 to detect cleavage of B23 or with a patient antiserum recognizing topoisomerase I (topo I). B, Undifferentiated or differentiated AoSMCs were incubated in the absence or presence of granule contents as described above. Cleavage of autoantigens was detected by immunoblotting with patient antisera. NuMA ⫽ nuclear mitotic apparatus protein. periments were performed by incubating intact cells with granule contents and evaluating autoantigen cleavage by immunoblotting. In fibroblasts, B23 remained refractory to cleavage upon treatment with granule contents, while topoisomerase I was well cleaved (Figure 4A). Minimal cleavage of B23 could be detected both in HUVECs (results not shown) and in undifferentiated AoSMCs, as demonstrated by the generation of the 22-kd GB fragment (Figure 4A) and a decrease in levels of the intact protein (clearly noted on lighter autoradiograph exposures; results not shown). To more closely replicate an in vivo microenvironment, AoSMCs were induced to undergo differentiation in culture using a method that serves to increase oxidant activity and the expression of several differentiation proteins (see Materials and Methods and ref. 22). The differentiation status of these cells was confirmed by the expression of high levels of smooth muscle ␣-actin as determined by immunoblotting (results not shown). These differentiated cells were treated with cytotoxic lymphocyte granules, and the cleavage of a panel of autoantigens was assessed. In striking contrast to the complete resistance of B23 to cleavage in other cell types, B23 was efficiently cleaved in these differentiated SMCs (Figure 4A) and was accompanied by the concomitant generation of the signature 22-kd GB cleavage fragment. B23 was cleaved ⬃38-fold more efficiently in differentiated cells (Table 1). Examination of B23 cleavage during cytotoxic lymphocyte granule–induced cell death of keratinocytes that had been similarly grown in a serum-free medium for 6 days did not lead to an increased susceptibility to proteolysis by GB (results not shown). Cleavage of numerous additional autoantigens in undifferentiated and differentiated SMCs was also evaluated. Several findings were noteworthy. Although equal protein amounts were loaded in each gel lane, expression levels of several autoantigens were diminished in the differentiated SMCs (e.g., NuMA, U1–70 kd, and Ufd-2), while levels of other autoantigens were unchanged in the undifferentiated and differentiated cells (e.g., Mi-2). While B23 was resistant to cleavage in undifferentiated SMCs but was very efficiently cleaved in the differentiated cells, all these other autoantigens were efficiently cleaved even in undifferentiated cells, with little difference in cleavage efficiency being observed in differentiated and undifferentiated cells (Figure 4B and Table 1). 238 ULANET ET AL Table 1. Susceptibility of autoantigens to cleavage by GB in undifferentiated and differentiated AoSMCs* Autoantigen B23 Fibrillarin Topo I La U1–70 kd NuMA Ufd-2 Mi-2 Disease Subcellular localization Cleaved by GB in undifferentiated AoSMCs SSc SSc SSc Non-SSc SSc overlap Non-SSc Non-SSc Non-SSc Nucleolar Nucleolar Nuclear Nuclear Nuclear Nuclear Cytoplasmic Nuclear No No Yes Yes Yes Yes Yes Yes Cleaved by GB in differentiated AoSMCs Fold increase in cleavage efficiency (kcat/Km diffentiated cells ⫼ kcat/Km undifferentiated cells) Yes Yes Yes Yes Yes Yes Yes Yes 38 19 2 3 1.5 0.6 2 1.6 * GB ⫽ granzyme B; AoSMCs ⫽ aortic smooth muscle cells; SSc ⫽ systemic sclerosis; topo I ⫽ topoisomerase I; NuMA ⫽ nuclear mitotic apparatus protein. Interestingly, our previous studies have shown that autoantibodies recognizing B23 are frequently associated with autoantibodies to fibrillarin (8). Fibrillarin was also completely resistant to proteolysis during cytotoxic lymphocyte granule–induced death of undifferentiated VSMCs, but was much more efficiently proteolyzed (⬃19-fold) during granule-induced death of differentiated VSMCs (Figure 4B). Unfortunately, all available antibodies exclusively recognized the intact form of fibrillarin and not the GB-induced fragments, precluding definitive analysis of the mechanism of proteolysis at this time. However, the data indicate that fibrillarin may well behave similarly to B23 under these circumstances. The differential susceptibility of another nonscleroderma autoantigen, PARP, to cleavage by GB in undifferentiated versus differentiated AoSMCs could not be assessed due to the minimal expression of this autoantigen in the differentiated cells (results not shown). It is of interest that the cleavage of U1–70 kd in differentiated SMCs exclusively generates the GB cleavage fragment of this molecule, and not the caspasegenerated 40-kd fragment, which is more prominent in undifferentiated cells (Figure 4B). As an additional readout of the cleavage of B23 by GB in intact cells, HeLa and AoSMCs (undifferentiated and differentiated) were grown on coverslips, treated with granule contents, and cleavage was assayed by immunofluorescence with an antibody that selectively recognizes GB-cleaved B23 (R3956). While this antibody was highly specific for the GB fragment of B23 by immunoblotting of denatured protein (results not shown), faint nucleolar staining of AoSMCs could be detected with R3956, even in the absence of treatment with granule contents (Figure 5). Consistent with the finding that B23 was poorly cleaved in cytotoxic lymphocyte granule content–treated undifferentiated AoSMCs, no significant difference was seen between R3956 staining of untreated versus granule content–treated undifferentiated cells. Similar findings were observed in HeLa cells exposed to the same experimental protocol (results not shown). In contrast, identical treatment of the differentiated AoSMCs resulted in a large increase in Figure 5. Selective detection of granzyme B (GB) fragment of B23 by immunofluorescence in nucleoli of differentiated aortic smooth muscle cells (AoSMCs) treated with YT cell granule contents (GC). Undifferentiated or differentiated AoSMCs were incubated in the absence (top) or presence (middle and bottom) of YT cell GC. Generation of GB fragments of B23 was detected by staining with the R3956 antibody and examination by fluorescence microscopy. Bar ⫽ 70 m. (Original magnification of the images in the bottom row is 2⫻ that in the top and middle rows.) CLEAVAGE OF AUTOANTIGEN B23 BY GRANZYME B the intensity of nucleolar staining with R3956 in ⬃30% of cells after a 40-minute incubation, again consistent with the biochemical findings of selective B23 cleavage by GB in these differentiated cells. Staining of the above cell types with a polyclonal antibody to full-length B23 (R3434) did not reveal any alterations in the distribution of B23 within the different cell types either in the absence or the presence of granule content treatment (results not shown). DISCUSSION In these studies, we have begun to investigate whether unique features of ubiquitously expressed autoantigens are specifically observed in the microenvironment in which ongoing tissue damage occurs in distinct autoimmune phenotypes. We chose to study the nucleolar autoantigen B23, which is a target of a high-titer autoantibody response in patients with scleroderma, and which is strongly associated with pulmonary hypertension in this disease (8). B23 is cleaved by GB in vitro; such susceptibility to cleavage at a specific site by GB appears to be an exclusive feature of autoantigens (12). While most of these potential GB substrates are ubiquitously expressed, and cytotoxic lymphocyte granule– mediated apoptosis occurs routinely during immunosurveillance by NK and cytotoxic T cells, autoantibody responses to specific self-molecules are associated with distinct clinical phenotypes. We have thus hypothesized that in vivo, GB-mediated cleavage of phenotypespecific autoantigens may be regulated and potentially favored in the relevant disease microenvironment. We demonstrate in this study that B23 exhibits a selective susceptibility to cleavage by GB during the cytotoxic lymphocyte granule–mediated death of VSMCs induced to undergo differentiation in culture. The procedure used for differentiation of these SMCs has previously been shown to induce a selective increase in smooth muscle differentiation markers (smooth muscle forms of ␣-actin, SM1/SM2 myosin, and calponin) and a resumption of the contractile activity of the cells (22). Thus, these cells are phenotypically similar to contractile VSMCs, which are present in the media of blood vessels. In contrast, cultured proliferating SMCs have low expression of differentiation markers and do not display contractile activity (22). Despite the fact that B23 is efficiently cleaved by GB in vitro, and GB has been demonstrated to rapidly accumulate in the nucleolus of target cells (25), B23 is refractory to cleavage by GB during the cytotoxic lymphocyte granule–mediated death of all other intact cell types studied to date. In 239 addition to being cell-type specific, the striking enhancement of GB-mediated proteolysis of B23 and fibrillarin in differentiated SMCs is also autoantigen specific, because the efficiency of cleavage of several other autoantigens studied was not dramatically changed in these cells. The differential susceptibility of B23 to cleavage by GB during cytotoxic lymphocyte granule–mediated death suggests that GB proteolysis of autoantigens can be regulated. It is notable that SMCs are major components of vessels, a proposed site of primary injury in scleroderma. Reversible vasospasm (19), resulting in ischemia–reperfusion injury, is thought to play a role in the development of the vascular lesions in scleroderma (including endothelial cell apoptosis, and a thickening of the intima due to the proliferation of SMCs and excess collagen deposition) (26). The underlying vasospastic activity in scleroderma has been proposed to be caused in part by increased contractility of VSMCs to ␣2adrenergic receptor stimulation (27). Thus, alterations in the phenotype and proliferation status of SMCs likely contribute to the underlying vascular defects in scleroderma, including the vasoconstriction and vascular remodeling that occur in pulmonary hypertension. It is therefore particularly interesting that cleavage of B23 is enhanced in VSMCs demonstrating contractile activity, and that the specific immune response for this molecule is a marker of a subset of scleroderma patients with pulmonary hypertension (8). It is tempting to speculate that during the initial injury leading to the dysregulation of these cells, an active inflammatory response promotes the unique fragmentation of B23 by GB, allowing for the selection of this molecule by the autoimmune response, and hence the association of antibodies against B23 with the pulmonary hypertension phenotype. Fibrillarin is another nucleolar autoantigen in scleroderma which is a GB substrate, and which is also the target of autoantibodies in patients with isolated pulmonary hypertension (9,10). It is therefore particularly interesting that fibrillarin appears to behave similarly to B23 in terms of its enhanced sensitivity to proteolysis during granuleinduced death of differentiated SMCs. It will be important to define whether the striking correlation between exclusive expression of a cleavable form of an autoantigen in relevant cells of the target tissue and autoantibodies to that antigen is also observed in other autoimmune rheumatic diseases. The finding of GB fragments of autoantigens specifically in the relevant diseased tissue would help verify that GB proteolysis of autoantigens is of potential 240 relevance in vivo. The R3956 anti-B23 antibody may be a valuable tool in this endeavor. By immunoblotting, this antibody exclusively recognizes an N-terminal GB cleavage fragment of B23 (but not the intact molecule). Most notably, R3956 staining of granule content–treated cultured differentiated SMCs, but no other cell types examined, revealed a dramatic increase in the intensity of nucleolar staining compared with untreated cells. This antibody may thus be a useful probe for selectively staining cells in tissue sections obtained from relevant diseased human tissue. Since GB cleavage of autoantigens may be particularly pertinent during the initial phase of the autoimmune response, such in vivo studies may prove challenging in scleroderma, where the relevant microenvironment (i.e., pulmonary vessels) is quite inaccessible, and tissue biopsy samples are frequently only available from patients with later or end-stage disease. Similar studies may hold more promise in elucidating the etiology of the autoimmune response in diseases such as cancer, in which the immunizing tissue is self-renewing as the tumor grows, and from which biopsy samples are frequently obtained. Of note, several of the nucleolar scleroderma autoantigens (B23, fibrillarin, and NOR-90) can also be targeted by the autoimmune response in ⬃10% of patients with hepatocellular carcinoma (HCC) (28). In this disease, the onset of HCC is preceded by chronic hepatitis and cirrhosis. Also, work by Zhang et al (29) has demonstrated that during the transition from chronic liver disease to HCC, patients can develop autoantibodies that were not present during the preceding chronic liver disease phase. Since the stages of disease progression of HCC are well defined, the use of this model will allow for the study of changes in the GB cleavage susceptibility of nucleolar autoantigens during tumorigenesis. The mechanism underlying the selective susceptibility of B23 to cleavage by GB in differentiated SMCs is currently unclear. It is possible that the uptake of GB and/or trafficking of GB to the nucleolus could vary in the different cell types (30,31). Preliminary comparison of undifferentiated and differentiated AoSMCs in terms of their levels of expression of the cation-independent mannose-6–phosphate receptor by immunoblotting did not reveal increased levels of this receptor for GB in the differentiated cells (results not shown). Further, since other autoantigens such as topoisomerase I and La were well cleaved in the same cells in which B23 was resistant, it is highly unlikely that the uptake and/or activity of GB is compromised in these cell types. It will be interesting to monitor the localization of GB during cytotoxic ULANET ET AL lymphocyte granule–mediated cell death of the different cell types as well as the relative sensitivity of these cells to this form of apoptosis. It is also possible that undifferentiated cells express a GB-interacting serpin, perhaps enriched in the nucleolus. Immunoblotting studies with an antibody against PI-9 (a proteinase inhibitor of GB in human cells) demonstrated that PI-9 expression was not detectable in either undifferentiated or differentiated VSMCs, while it was expressed at high levels in HUVECs and cytotoxic lymphocytes (results not shown). Other antigen-specific features that might serve to obscure the GB cleavage site (e.g., posttranslational modifications, complex formation) could also be influencing the GB cleavage susceptibility of B23 in the different cell types. Defining the changes in B23 binding partners or modifications between undifferentiated and differentiated SMCs may also provide significant insights. Additionally, understanding how the cleavage site is obscured in intact cells and how this affects processing of B23 may provide important insights into the relevance of the GB cleavage site and/or cleavage in determining the immunogenicity of this molecule. Finally, while these studies have analyzed the cleavage of B23 in distinct differentiation states of VSMCs, they did not address whether B23 was differentially susceptible to cleavage in different populations of vascular endothelial cells. Future studies addressing this question are clearly indicated. These studies demonstrate that B23, which is targeted by patient autoantibodies in scleroderma (in association with pulmonary hypertension), is selectively susceptible to GB-mediated proteolysis in differentiated VSMCs. The observation that phenotype-specific autoantigens express unique susceptibility to cleavage by GB in cells of the potential target tissue focuses attention on tissue-specific antigen conformation and protease cleavability as potentially important parameters in the selection of targets for high-titer, phenotype-specific autoantibody responses. From these data taken together, we propose that the immunizing microenvironment might play a central role in shaping the specific autoimmune response in human autoimmune diseases. REFERENCES 1. Herron GS, Romero LI. Vascular abnormalities in scleroderma. Semin Cutan Med Surg 1998;17:12–7. 2. Jimenez SA, Hitraya E, Varga J. Pathogenesis of scleroderma: collagen. Rheum Dis Clin North Am 1996;22:647–74. 3. Weiner ES, Earnshaw WC, Senecal J-L, Bordwell B, Johnson P, Rothfield NF. Clinical associations of anticentromere antibodies CLEAVAGE OF AUTOANTIGEN B23 BY GRANZYME B 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. and antibodies to topoisomerase I: a study of 355 patients. Arthritis Rheum 1988;31:378–85. Tan EM, Rodnan GP, Garcia I, Moroi Y, Fritzler MJ, Peebles C. Diversity of antinuclear antibodies in progressive systemic sclerosis: anticentromere antibody and its relationship to CREST syndrome. Arthritis Rheum 1980;23:617–25. Steen VD, Ziegler GL, Rodnan GP, Medsger TA Jr. Clinical and laboratory associations of anticentromere antibody in patients with progressive systemic sclerosis. Arthritis Rheum 1984;27: 125–31. Steen VD, Powell DL, Medsger TA Jr. Clinical correlations and prognosis based on serum autoantibodies in patients with systemic sclerosis. Arthritis Rheum 1988;31:196–203. Greidinger EL, Flaherty KT, White B, Rosen A, Wigley FM, Wise RA. African-American race and antibodies to topoisomerase I are associated with increased severity of scleroderma lung disease. Chest 1998;114:801–7. Ulanet DB, Wigley FM, Gelber AC, Rosen A. Autoantibodies against B23, a nucleolar phosphoprotein, occur in scleroderma and are associated with pulmonary hypertension. Arthritis Rheum 2003;49:85–92. Sacks DG, Okano Y, Steen VD, Curtiss E, Shapiro LS, Medsger TA Jr. Isolated pulmonary hypertension in systemic sclerosis with diffuse cutaneous involvement: association with serum antiU3RNP antibody. J Rheumatol 1996;23:639–42. Okano Y, Steen VD, Medsger TA Jr. Autoantibody to U3 nucleolar ribonucleoprotein (fibrillarin) in patients with systemic sclerosis. Arthritis Rheum 1992;35:95–100. Tan EM. Autoantibodies in pathology and cell biology. Cell 1991;67:841–2. Casciola-Rosen L, Andrade F, Ulanet D, Wong WB, Rosen A. Cleavage by granzyme B is strongly predictive of autoantigen status: implications for initiation of autoimmunity. J Exp Med 1999;190:815–26. Gahring L, Carlson NG, Meyer EL, Rogers SW. Granzyme B proteolysis of a neuronal glutamate receptor generates an autoantigen and is modulated by glycosylation. J Immunol 2001;166: 1433–8. Matsumura S, van de Water J, Kita H, Coppel RL, Tsuji T, Yamamoto K, et al. Contribution to antimitochondrial antibody production: cleavage of pyruvate dehydrogenase complex-E2 by apoptosis-related proteases. Hepatology 2002;35:14–22. Lee KJ, Dong X, Wang J, Takeda Y, Dynan WS. Identification of human autoantibodies to the DNA ligase IV/XRCC4 complex and mapping of an autoimmune epitope to a potential regulatory region. J Immunol 2002;169:3413–21. Nagaraju K, Cox A, Casciola-Rosen L, Rosen A. Novel fragments of the Sjögren’s syndrome autoantigens ␣-fodrin and type 3 muscarinic acetylcholine receptor are generated during cytotoxic lymphocyte granule–induced cell death. Arthritis Rheum 2001;44: 2376–86. 241 17. Casciola-Rosen LA, Pluta AF, Plotz PH, Cox AE, Morris S, Wigley FM, et al. The DNA mismatch repair enzyme PMS1 is a myositis-specific autoantigen. Arthritis Rheum 2001;44:389–96. 18. Schachna L, Wigley FM, Morris S, Gelber AC, Rosen A, CasciolaRosen L. Recognition of granzyme B–generated autoantigen fragments in scleroderma patients with ischemic digital loss. Arthritis Rheum 2002;46:1873–84. 19. Campbell PM, LeRoy EC. Pathogenesis of systemic sclerosis: a vascular hypothesis. Semin Arthritis Rheum 1975;4:351–68. 20. Andrade F, Roy S, Nicholson D, Thornberry N, Rosen A, Casciola-Rosen L. Granzyme B directly and efficiently cleaves several downstream caspase substrates: implications for CTLinduced apoptosis. Immunity 1998;8:451–60. 21. Olmsted JB. Affinity purification of antibodies from diazotized paper blots of heterogeneous protein samples. J Biol Chem 1981;256:11955–7. 22. Su B, Mitra S, Gregg H, Flavahan S, Chotani MA, Clark KR, et al. Redox regulation of vascular smooth muscle cell differentiation. Circ Res 2001;89:39–46. 23. Thornberry NA, Ranon TA, Peterson EP, Rasper DM, Timkey T, Garcia-Calvo M, et al. A combinatorial approach defines specificities of members of the caspase family and granzyme B: functional relationships established for key mediators of apoptosis. J Biol Chem 1997;272:17907–11. 24. Harris JL, Peterson EP, Hudig D, Thornberry NA, Craik CS. Definition and redesign of the extended substrate specificity of granzyme B. J Biol Chem 1998;273:27364–73. 25. Trapani JA, Browne KA, Smyth MJ, Jans DA. Localization of granzyme B in the nucleus: a putative role in the mechanism of cytotoxic lymphocyte-mediated apoptosis. J Biol Chem 1996;271: 4127–33. 26. Kahaleh MB, LeRoy EC. Autoimmunity and vascular involvement in systemic sclerosis (SSc). Autoimmunity 1999;31:195–214. 27. Flavahan NA, Flavahan S, Liu Q, Wu S, Tidmore W, Wiener CM, et al. Increased ␣2-adrenergic constriction of isolated arterioles in diffuse scleroderma. Arthritis Rheum 2000;43:1886–90. 28. Imai H, Ochs RL, Kiyosawa K, Furuta S, Nakamura RM, Tan EM. Nucleolar antigens and autoantibodies in hepatocellular carcinoma and other malignancies. Am J Pathol 1992;140:859–70. 29. Zhang JY, Zhu W, Imai H, Kiyosawa K, Chan EK, Tan EM. De-novo humoral immune responses to cancer-associated autoantigens during transition from chronic liver disease to hepatocellular carcinoma. Clin Exp Immunol 2001;125:3–9. 30. Motyka B, Korbutt G, Pinkoski MJ, Heibein JA, Caputo A, Hobman M, et al. Mannose 6-phosphate/insulin-like growth factor II receptor is a death receptor for granzyme B during cytotoxic T cell-induced apoptosis. Cell 2000;103:491–500. 31. Jans DA, Jans P, Briggs LJ, Sutton V, Trapani JA. Nuclear transport of granzyme B (fragmentin 2): dependence on perforin in vivo and cytosolic factors in vitro. J Biol Chem 1996;271: 30781–9.