Removal of dead cells is essential in the maintenance of tissue homeostasis, and efficient removal prevents exposure of intracellular content to the immune system, which could lead to autoimmunity. The plasma protease factor VII–activating protease (FSAP) can release nucleosomes from late apoptotic cells. FSAP circulates as an inactive single-chain protein, which is activated upon contact with either apoptotic cells or necrotic cells. The purpose of this study was to investigate the role of FSAP in the release of nucleosomes from necrotic cells.
Necrotic Jurkat cells were incubated with serum, purified 2-chain FSAP, and/or DNase I. Nucleosome release was analyzed by flow cytometry, and agarose gel electrophoresis was performed to detect DNA breakdown.
Incubation with serum released nucleosomes from necrotic cells. Incubation with FSAP-deficient serum or serum in which FSAP was inhibited by a blocking antibody was unable to release nucleosomes from necrotic cells, confirming that FSAP is indeed the essential serum factor in this process. Together with serum DNase I, FSAP induced the release of DNA from the cells, the appearance of nucleosomes in the supernatant, and the fragmentation of chromatin into eventually mononucleosomes.
FSAP and DNase I are the essential serum factors that cooperate in necrotic cell DNA degradation and nucleosome release. We propose that this mechanism may be important in the removal of potential autoantigens.
In multicellular organisms, cell death is essential for the control of tissue homeostasis. Dying cells are efficiently removed by phagocytic cells in order to prevent release of potentially harmful cytotoxic and immunogenic cellular content into the microenvironment (). Insufficient removal of dying cells or the inability to mask and digest nuclear material may lead to autoantibody formation, inflammation, and development of autoimmune diseases, such as systemic lupus erythematosus (SLE) ([2-4]). SLE is a chronic inflammatory autoimmune disease that is characterized by unpredictable exacerbations and remissions, with variable clinical manifestations. One of the main characteristics is the formation of autoantibodies against ubiquitous intracellular antigens, for example, double-stranded DNA (dsDNA), histones, and nucleosomes. During the course of SLE, nucleosomes are found in the blood of patients, where they can bind antibodies to form immune complexes that can potentially be deposited in the glomeruli ([5-8]).
Napirei et al () have shown that incubation of necrotic cells with mouse serum leads to chromatin breakdown as well as the removal of nucleosomes, and they observed that serum DNase I plays a crucial role in this process. Moreover, the physiologic importance of DNase I is demonstrated by the fact that decreased activity of serum DNase I has been shown in patients with SLE ([10, 11]). However, Napirei et al showed that serum DNase I was essential, but insufficient by itself, for chromatin breakdown and that the contribution of a serum serine protease was required. This protease was identified as plasmin, based on the observation that protease inhibitors specific for plasmin were able to inhibit nucleosome generation in necrotic cells ().
We have previously shown that serum causes the release of nucleosomes from late apoptotic cells (). The serum factor responsible for this nucleosome release is factor VII–activating protease (FSAP) (). FSAP, which is also known as plasma hyaluronic acid binding protein 2, is a serine protease that circulates in the plasma as an inactive single-chain molecule of 64 kd. It is converted proteolytically into its active 2-chain form, which consists of a 50-kd heavy chain and a 28-kd light chain connected by a disulfide bond (). There are several plasma inhibitors of FSAP, including α2-antiplasmin () and tissue factor pathway inhibitor (), which are inhibitors of plasmin as well. FSAP has been reported to be activated upon contact with either late apoptotic cells or necrotic cells () and circulating histones ().
Based on the role of FSAP in the release of nucleosomes from late apoptotic cells, we investigated in the present study the role of FSAP in the release of nucleosomes from necrotic cells upon incubation with serum.
MATERIALS AND METHODS
Mouse monoclonal antibodies anti–FSAP-4 (), anti–α2-antiplasmin (AAP-20), AAP-11, anti–plasmin 1 (AP-1) (), and antinuclear antibody 58 (ANA58) and ANA60 () were prepared at our department. Iscove's modified Dulbecco's medium (IMDM) was obtained from BioWhittaker Europe. Fetal calf serum (FCS) was obtained from Bodinco. Penicillin and streptomycin were obtained from Gibco Invitrogen. RNase A, DNase I, and β-mercaptoethanol were obtained from Sigma. High-performance enzyme-linked immunosorbent assay (ELISA) buffer and poly–horseradish peroxidase (HRP)–labeled streptavidin were obtained from Sanquin, and 3,5,3′,5′-tetramethylbenzidine was obtained from Merck. A QIAamp DNA Blood Mini kit was obtained from Qiagen. NuPAGE 12% polyacrylamide gels, sodium dodecyl sulfate (SDS) sample buffer, a SilverQuest kit, and a Dynal CD4+ cell isolation kit were obtained from Invitrogen. Lymphoprep was obtained from Axis-Shield.
Plasma-derived 2-chain FSAP was obtained by purification as described previously (). The construction, production, purification, and activation of the thermolysin-activatable FSAPR313Q mutant has been described in detail elsewhere (). Blood was collected from donors and allowed to clot for 30 minutes at room temperature. After centrifugation at 1,300g for 10 minutes, serum was removed and stored at –20°C. All healthy donors were homozygous for the wild-type form of FSAP. All sera and buffy coats from healthy donors were obtained as anonymized samples from the diagnostic laboratory, and were handled in accordance with the Dutch rules and regulations for the use of human materials.
Sera received from anonymous healthy donors were tested for nucleosome-releasing activity, FSAP activation, and FSAP antigen levels (). Serum from 1 donor was not functional in releasing nucleosomes from late apoptotic cells. In addition, no FSAP–inhibitor complexes were detected upon activation of FSAP, and the serum was negative for FSAP antigen by ELISA. Subsequently, DNA was isolated from the serum of this FSAP-negative donor as well as from a healthy donor, using a QIAamp DNA Mini kit (Qiagen). Sequence analysis was performed as described by Kuijpers et al (). We found a homozygous nonsense mutation in which an arginine with a stop codon (c.607C>T p.R203X) had been substituted, which led to a truncated protein lacking activity.
Cell culture, isolation of human CD4+ T lymphocytes, and induction of necrosis
Human peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient centrifugation on Lymphoprep from buffy coats obtained from healthy donors. From the PBMCs, CD4+ cells were purified with a Dynal CD4+ isolation kit. Jurkat cells were cultured in culture medium consisting of IMDM containing 5% (volume/volume) FCS, 100 IU/ml of penicillin, 100 μg/ml of streptomycin, and 50 μM β-mercaptoethanol. Necrosis was induced by incubation of Jurkat cells or isolated T cells (2 × 106 cells/ml) for 30 minutes at 56°C. After induction of necrosis, cells (1 × 106 cells/ml) were cultured in FCS-free medium for 48 hours at 37°C in an atmosphere of 5% CO2 in the presence of human serum, purified 2-chain FSAP, and/or DNase I.
Analysis of nucleosome release by flow cytometry
Necrotic cells were washed with HN buffer (50 mM HEPES, 100 mM NaCl, pH 7.4). To examine the release of nucleosomes, cells were stained with propidium iodide (PI) that had been diluted in HN buffer (1 μg/ml), and samples were analyzed by flow cytometry. The median fluorescence intensity of PI was measured and quantified using FACSDiva software (Becton Dickinson). In our assay, the FSAP content of serum was set at 100 arbitrary units (AU) per milliliter. Dilutions of serum (expressed as % serum) were compared with the same amount of arbitrary units per milliliter of purified FSAP.
Nucleosome levels were determined by an ELISA, as described elsewhere ([12, 19]). Briefly, monoclonal antibody ANA60, which recognizes histone H3, was used as a capture antibody. Biotinylated F(ab′)2 fragments of ANA58, which recognizes an epitope exposed on complexes of histone H2A, histone H2B, and dsDNA, in combination with poly-HRP was used for detection.
Analysis of DNA degradation by agarose gel electrophoresis
After induction of necrosis, cells were cultured in the presence of purified FSAP and/or DNase I for 24 hours at 37°C in an atmosphere of 5% CO2. Samples were treated with RNase (10 units/ml) for 30 minutes at 37°C. DNA was isolated using a QIAamp DNA Blood Mini kit, and samples were loaded on a 0.9% agarose gel containing 0.5 μg/ml of ethidium bromide. The gel was run in Tris–acetate–EDTA buffer.
Plasmin–α2-antiplasmin complex ELISA
Plasmin–α2-antiplasmin complexes were measured using ELISA. Monoclonal antibody AAP-11, which is directed against inactivated and complexed α2-antiplasmin, was used as a capture antibody, and monoclonal antibody AP-1, which is directed against plasmin, in combination with poly-HRP was used for detection. Plasma incubated with urokinase was used as a standard, with a plasmin–α2-antiplasmin complex concentration of 2,100 nmoles/liter. Results were expressed in nmoles/liter ().
FSAP–α2-antiplasmin complex ELISA
Levels of FSAP–α2-antiplasmin complex were determined by ELISA, as recently described (). Briefly, a monoclonal antibody (AAP-20) that recognizes α2-antiplasmin was used as a capture antibody. Biotinylated anti–FSAP-4, which recognizes the light chain of FSAP, in combination with poly-HRP was used for detection.
SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining
After induction of necrosis, cells (5 × 106 cells/ml) were cultured in the presence or absence of plasma-purified FSAP or thermolysin-activated recombinant FSAPR313Q () for 48 hours at 37°C in an atmosphere of 5% CO2. Cells were washed with HN buffer, taken up in SDS-PAGE sample buffer, and samples were applied to NuPAGE 12% polyacrylamide gels. After electrophoresis, silver staining was performed using a SilverQuest silver staining kit.
Results are presented as mean ± SEM. In case of normal distribution, comparison between groups was performed by parametric testing using a t-test. P values less than 0.05 were considered statistically significant.
To test whether human serum was able to release nucleosomes from necrotic cells, we incubated necrotic Jurkat cells or isolated T cells with serum, and the remaining cellular DNA content was measured by flow cytometry after PI staining of the cell remnants. Untreated cells retained all DNA in the cells, as shown by the positive PI peak in Figure 1A. When the cells were incubated with serum from a healthy donor, a shift in the PI peak was observed, indicating that, indeed, part of the cells had lost their DNA (Figure 1A). When FSAP in the serum was blocked with a neutralizing antibody, no decrease in PI staining was observed, indicating that FSAP is in fact involved in the release of nucleosomes from necrotic cells (Figure 1B). The role of FSAP in nucleosome release was confirmed by analysis of serum from an FSAP-deficient donor, and again, no nucleosome release could be detected (Figure 1C). The results of the quantitative analysis of the nucleosome release are shown in Figure 1D.
We incubated necrotic cells with purified FSAP alone or in combination with DNase I to test whether FSAP alone is sufficient for nucleosome release from necrotic cells or whether, as indicated by Napirei et al (), DNase I is also required for this process. When the cells were incubated with plasma-derived purified FSAP, DNA was retained in the cell, a result comparable to that seen with untreated cells (Figure 2A). This indicated that there is no nucleosome-releasing activity of purified FSAP in necrotic cells. Incubation of DNase I alone also did not lead to a decrease in the PI signal (Figure 2B). When cells were incubated with DNase I in combination with purified FSAP, a shift in the PI peak was observed, indicating that FSAP cooperates with DNase I in releasing nucleosomes from necrotic cells (Figure 2C). Results of the nucleosome release assay are summarized in Figure 2D.
To confirm that the cells release their DNA in the form of nucleosomes, an ELISA was performed to measure nucleosomes in the supernatant. Indeed, high levels of nucleosomes were detected in the supernatant of cells that had been incubated with serum or with purified FSAP and DNase I, whereas no nucleosomes or only low levels of nucleosomes were measured in the supernatant of untreated cells or cells that had been incubated with either purified FSAP or DNase I alone (Figure 3).
DNA fragmentation by FSAP and DNase I
To analyze DNA breakdown of necrotic cells, DNA was isolated and agarose gel electrophoresis was performed. Cells were incubated with purified FSAP and/or DNase I. Necrotic cells cultured in the absence of both components showed no DNA laddering, as shown in Figure 4. Moreover, necrotic cells incubated with FSAP alone did not show DNA fragmentation, whereas cells incubated with DNase I alone demonstrated a smear of high molecular weight DNA, indicating random cleavage of the DNA (Figure 4). Notably, necrotic cells incubated with FSAP in combination with DNase I showed DNA fragmentation of high molecular weight DNA, which yielded to a band pattern that was characteristic of nucleosomes.
No plasminogen activation upon incubation with necrotic cells
Since plasmin has also been described as functioning together with DNase I in the degradation of chromatin and the release of nucleosomes from necrotic cells (), we tested whether plasminogen is activated after incubation with necrotic cells. Serum was incubated in the presence or absence of necrotic cells, and plasmin–α2-antiplasmin complexes were measured by ELISA. No plasmin–α2-antiplasmin complexes could be detected upon incubation of serum with necrotic cells, indicating that plasminogen was not activated in plasmin (Figure 5). However, FSAP was activated when serum was incubated with necrotic cells, as shown by the formation of FSAP–α2-antiplasmin complexes in the serum (Figure 5).
Histone H1 degradation by FSAP
Napirei et al () have shown that there is a loss of histone H1 upon incubation of necrotic cells with DNase1–/– serum and that this effect is inhibited by the addition of aprotinin or plasminogen activator inhibitor 1. To test whether histone H1 was cleaved by FSAP, we incubated necrotic cells with purified plasma-derived FSAP. We found that histone H1 was degraded after FSAP incubation, as shown in Figure 6A. Western blotting with an antibody against histone H1, followed by mass spectrometry analysis, confirmed that indeed the band that disappeared on gel analysis was histone H1 (Figure 6B). To exclude the possibility that a contamination in the plasma-purified FSAP is responsible for the histone H1 degradation, necrotic cells were also incubated with activated recombinant FSAPR313Q (). Again, the cleavage of histone H1 was seen upon incubation with recombinant FSAPR313Q, confirming the observation that FSAP indeed cleaves histone H1.
We previously showed that the plasma serine protease FSAP is the factor in serum that can release nucleosomes from late apoptotic cells (). DNA release from necrotic cells seems to be a multistep process. In the present study, we showed that FSAP cooperates with DNase I in the degradation and removal of DNA from necrotic cells.
Impaired clearance of dead cells can lead to the induction of inflammation, the formation of autoantibodies, and the development of autoimmune diseases such as SLE ([2-4]). In the absence of serum, nucleosomes remain bound to late apoptotic cells (). Circulating phagocytes may not be very efficient in taking up these large fragments of cellular material. The removal of nucleosomes from dead cells may therefore help in the phagocytosis of the cell remnants and prevent exposure of the immunogenic intracellular content. Moreover, nucleosomes that are released into the circulation are rapidly cleared by hepatocytes (). When nucleosomes are not removed from the cells and stay attached to the cells, they might persist longer and can play an immunogenic and pathogenic role in the development of autoimmune diseases such as SLE.
In contrast to apoptosis, the integrity of the cell membrane is distorted upon necrosis, without cleavage of DNA by intracellular nucleases. Degradation of the DNA of necrotic cells by serum DNase I, followed by removal of nucleosomes, might be an important event in recognition by phagocytes, thereby preventing the development of SLE. The physiologic importance of DNase I has been demonstrated by the decreased activity of serum DNase I in patients with SLE ([10, 11]). Furthermore, Macanovic et al () demonstrated that lupus-prone NZB/NZW mice had significantly lower serum and urine concentrations of DNase I than did normal mice and that this reduction was not related to the presence of the DNase inhibitor actin or anti-DNA antibodies. Napirei et al () showed that mouse serum DNase I functions together with the serine protease plasmin in the generation of nucleosomes in necrotic cells. This identification was based on the observation that protease inhibitors specific for plasmin were able to inhibit nucleosome generation in necrotic cells.
FSAP is a plasma serine protease with structural and functional similarities to plasmin. Both proteases are serine proteases that are inhibited by aprotinin. There are several plasma inhibitors of FSAP, including α2-antiplasmin () and tissue factor pathway inhibitor (), which are also inhibitors of plasmin. Moreover, FSAP has been reported to be inhibited by plasminogen activator inhibitor (). This is a serine protease inhibitor that functions as the principal inhibitor of tissue plasminogen activator (tPA) and urokinase plasminogen activator (uPA).
Plasminogen, the proenzyme of plasmin, is present in the circulation. Activation of plasminogen into plasmin occurs when plasminogen activators cleave a unique bond in the serine protease domain, resulting in 2 chains, which are linked by 2 disulfide bonds. Similar to plasminogen, FSAP circulates in its inactive single-chain form and has to be activated by proteolytic cleavage, which yields a heavy chain and a light chain that are connected by a disulfide bond. We recently showed that FSAP is activated upon contact with late apoptotic cells as well as necrotic cells (). In contrast, plasminogen activation requires the presence of plasminogen activators, which need to be released at the site of action (e.g., tPA from the endothelium) and is not activated by contact with necrotic cells. Since no additional factor besides the contact with necrotic cells is required for FSAP activation, the role of FSAP in the release of nucleosomes from necrotic cells might be physiologically relevant. This is supported by the finding that FSAP-deficient serum or serum in which FSAP is inhibited by an inhibiting antibody is not able to release nucleosomes from necrotic cells.
C1q has also been demonstrated to release nucleosomes from necrotic cells in cooperation with serum DNase (). However, C1q-deficient plasma showed identical results as normal plasma in our system (data not shown).
Whereas FSAP is essential for the induction of the removal of nucleosomes from necrotic cells, purified FSAP is not able to induce nucleosome release. We have shown in these studies that FSAP needs DNase I to release nucleosomes from the necrotic cells, in contrast to late apoptotic cells, where FSAP alone is sufficient to release nucleosomes. We previously reported that serum was not able to induce nucleosome release from necrotic cells. However, those observations were based on a system in which the cells were incubated for 30 minutes with serum (). Since Napirei et al () demonstrated that mouse serum was able to generate nucleosomes in necrotic cells after a longer incubation period, we changed our system to a 48-hour incubation period. With this longer incubation period, we showed that serum can indeed cause the release of nucleosomes from necrotic cells. The increased incubation time may be required for efficient DNA cleavage by serum DNase I after necrosis, whereas intracellular endonucleases already confer DNA cleavage during the process of apoptosis.
The mechanism by which DNase I promotes FSAP-induced nucleosome release is not yet entirely clear. DNase I was able to cleave the DNA of necrotic cells in a random manner, as illustrated by a high molecular weight DNA smear after DNA isolation and agarose gel electrophoresis. However, this DNA was not released from the cells in the form of nucleosomes. One interpretation could be that the DNA fragments are too large to be released from the cells or that the DNA is still linked to the cell by a DNA binding protein. When the isolated DNA from necrotic cells treated with a combination of FSAP and DNase I was analyzed on agarose gels, low molecular weight bands were visible, which is a feature typical of nucleosomes. This indicates that in the presence of FSAP, internucleosomal cleavage occurs via DNase I.
Various explanations for this specific cleavage pattern can be hypothesized. One could be that the FSAP cleaves a DNA binding protein, thereby changing the conformation of the chromatin, resulting in increased accessibility for DNase I. This hypothesis is supported by the fact that histone H1 is degraded upon incubation of necrotic cells with FSAP. Another explanation could be that inhibitors of DNase I are cleaved by FSAP, resulting in increased activity of DNase I. Further investigation is needed to elucidate the mechanism by which FSAP and DNase I function together in the degradation and release of DNA from necrotic cells. In view of the fact that nucleosomes play a central role in the antinuclear antibody response in SLE ([5-8]) and that FSAP acts together with DNase I in the release of nucleosomes from necrotic cells, it would be of interest to investigate whether the nucleosome-releasing activity of FSAP is disturbed in patients with SLE.
In conclusion, this study showed that the plasma protease FSAP cooperates with DNase I in the degradation and removal of DNA from necrotic cells. This mechanism may be important in the clearance of potential autoantigens, thereby preventing the induction of an autoantigen-driven autoimmune response in SLE.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Zeerleder had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Stephan, Marsman, Bulder, Stavenuiter, Aarden, Zeerleder.
Acquisition of data. Stephan, Marsman, Bakker, Bulder, Aarden, Zeerleder.
Analysis and interpretation of data. Stephan, Marsman, Bakker, Bulder, Stavenuiter, Aarden, Zeerleder.
We would like to thank A. R. van der Horst (Sanquin Diagnostic Services, Sanquin Blood Supply) for the Marburg I analysis of all healthy donors, M. de Boer and K. van Leeuwen (Department of Blood Cell Research, Sanquin Blood Supply) for performing sequence analysis on the FSAP-deficient donor, and A. B. Meijer and C. van der Zwaan (Department of Plasma Proteins, Sanquin Blood Supply) for performing the mass spectrometry analysis.