Abstract
Complement activation leads to membrane attack complex formation, which can lyse not only pathogens but also host cells. Histones can be released from the lysed or damaged cells and serve as a major type of damage-associated molecular pattern, but their effects on the complement system are not clear. In this study, we pulled down two major proteins from human serum using histone-conjugated beads: one was C-reactive protein and the other was C4, as identified by mass spectrometry. In surface plasmon resonance analysis, histone H3 and H4 showed stronger binding to C4 than other histones, with KD around 1 nM. The interaction did not affect C4 cleavage to C4a and C4b. Because histones bind to C4b, a component of C3 and C5 convertases, their activities were significantly inhibited in the presence of histones. Although it is not clear whether the inhibition was achieved through blocking C3 and C5 convertase assembly or just through reducing their activity, the outcome was that both classical and mannose-binding lectin pathways were dramatically inhibited. Using a high concentration of C4 protein, histone-suppressed complement activity could not be fully restored, indicating C4 is not the only target of histones in those pathways. In contrast, the alternative pathway was almost spared, but the overall complement activity activated by zymosan was inhibited by histones. Therefore, we believe that histones inhibiting complement activation is a natural feedback mechanism to prevent the excessive injury of host cells.
Introduction
The complement system forms a major part of the host response to infection and cellular injury (1, 2). This system is intricately involved in these processes and consists of a cascade of more than 50 proteins participating in three activation pathways, namely the classical pathway (CP), mannose-binding lectin (MBL) pathway, and alternative pathway (AP) (2). The CP recognizes Ag–Ab complexes contained on the surface of pathogenic factors, including Gram-negative bacteria, viruses, and damaged cells (3). The MBL pathway binds mannose-containing pathogenic surfaces (4). Finally, the AP directly targets surface carbohydrate regions on pathogens such as viruses, bacteria, and fungi (5). The terminal pathway of complement activation by different stimuli is the formation of C3 and C5 convertases, which leads to assembly of the C5b-9 complex, the membrane attack complex (MAC) that can lyse pathogens. Many products generated during complement activation are also able to opsonize damaged cells or pathogens to facilitate phagocytosis (1, 2). In this way, complement activation enhances the ability of Abs and phagocytic cells to clear invading pathogens and cellular debris from the circulation (6).
Complement activation not only kills pathogens but also damages host cells during an inflammatory reaction, and excessive activation contributes to inflammation-driven tissue injury (7). Host cell lysis will release cell breakdown products, including DNA and histones, and those damage-associated molecular patterns have been demonstrated to play important roles in disease progression and host immune responses (8–10). Histones, the most abundant and important damage-associated molecular patterns, can be detected in blood taken from many critical illnesses, such as severe trauma (11), severe sepsis (12, 13), and necrotizing pancreatitis (14). Histones are positively charged proteins and have high affinity for negatively charged phospho-groups in DNA or cell membranes. Histone binding to cell membrane allows ions, particularly Ca2+ influx, into cells to cause harmful effects to cells contacted (11). In addition, histones are also the ligands of TLR2, TLR4, and TLR9, which trigger immune responses, including inflammasome activation and cytokine release (15–17). Histones also interact with coagulation factors in the circulation to promote thrombin generation, fibrin deposition, and systemic coagulation activation (18–22). In animal models, extracellular histones have been shown to mediate multiple-organ injury and even death in sepsis (13, 23). Clinically, correlation between circulating histone levels and organ injury as well as disease severity has also been demonstrated (12).
Recently, we found that extracellular histones interact with C4 protein. C4, coded by both C4A and C4B genes, is synthesized into a single peptide (precursor) and then cleaved into α- (98 kDa), β- (73 kDa), and γ-chains (33 kDa) (24, 25). Upon complement activation, C4 is cleaved by C1s enzyme into C4a and C4b; the latter mainly participates in classical and MBL pathways by forming C3 and C5 convertases, whereas C4a as an anaphylatoxin enhances smooth muscle contraction, histamine release, and vascular permeability as well as serving as a chemotaxis and inflammatory mediator (26, 27). In many disease conditions, particularly in sepsis, complement activation (28) and histone release (12) coexist. The outcome of histones binding to C4 appears important and this study aims to understand the pathophysiology related to the complement system and extracellular histones.
Materials and Methods
Human plasma and serum
Citrated plasma and serum were isolated from whole blood drawn from critically ill patients, according to the protocol granted by Liverpool Adult Ethical Committee (reference no.: 13/NW/0089). Normal human serum was purchased from CompTech.
Fractionation of human serum and plasma by ultracentrifugation
Citrated plasma and serum (1 ml) from patients were fractionated by ultracentrifugation at 40,000 rpm (4°C) for 1 h and then six layers of equal-volume fractions (166 μl/fraction) were collected. Histones and histone–DNA complexes were then measured in each fraction by Western blotting and ELISA (Cell Death Detection ELISA PLUS; Roche, West Sussex, U.K.), respectively, as previously described (11).
Isolation of histone-binding protein from plasma and mass spectrometry analysis
Isolated citrated plasma was diluted with 2 × PBS (v/v) and centrifuged to eliminate insoluble contents. The harvested supernatant was then precleared using blank Sepharose resin and then loaded on a CNBr-activated Sepharose 4B (GE Healthcare, Little Chalfont, U.K.) column conjugated with calf thymus histones (Roche). After a high-stringency wash with PBS + 0.5% (v/v) Tween 20 (Sigma-Aldrich, Dorset, U.K.) followed by PBS, histone-binding proteins were eluted and separated by gel electrophoresis. Gel slices from SDS-PAGE were washed (two times for 30 min each) with 50% acetonitrile and 0.2 M ammonium bicarbonate (pH 8.9) and then dried in a rotary evaporator. The slices were rehydrated in rehydration buffer (2 M urea, 0.2 M ammonium bicarbonate, pH 7.8) containing 0.2 μg trypsin and incubated at 37°C overnight. Excess rehydrate buffer was then removed and peptides were extracted from the gel slices with 60% acetonitrile and 0.1% TFA. The total peptide extract was concentrated in a rotary evaporator and then desalted using C18 ZipTips according to the manufacturer’s instructions. Mass spectrometry analysis was performed using a MALDI-TOF instrument (Waters Micromass) with a saturated solution of α-cyano-4-hydroxycinnamic acid in 50% acetonitrile/0.1% trifluoroacetic acid.
Detection of histone–C4 complexes by ELISA
Histone–C4 complexes were detected in normal and critically ill patient plasma using Cell Death Detection ELISA PLUS kit (Roche) with modification. In brief, normal plasma was preincubated with different concentrations of calf thymus histones for 10 min. Histones in plasma were first captured by biotinylated anti-histone Ab immobilized on streptavidin-coated 96-well plates. After extensive washing, rabbit anti-human C4 Ab (Abcam) followed by anti-rabbit–HRP Ab were used to probe histone–C4 complexes. Arbitrary units were calculated based on the absorbance (450 nm) to represent the relative levels of the complexes.
Western blotting using HRP-conjugated C4
To double confirm the interaction of C4 with individual histones by different assays, 2 μg of H2A, H2B, H3, and H4; 4 μg of H1; and 6 μg of S100P (as control) were subjected to Western blotting with HRP-conjugated C4 protein. C4 was purchased from Fitzgerald Industries International and conjugated using Lighting-Link HRP Conjugation Kit (Innova Biosciences, Babraham, Cambridge, U.K.) according to the procedure recommended by the manufacturer.
Surface plasmon resonance measurements
The binding parameters of C4 to individual histones, including the equilibrium Kd and affinities’ on-rates (kon) and off-rates (koff), were measured by surface plasmon resonance (SPR) analysis on a ProteOn XPR36 System (Bio-Rad). Chips coated with 20 μg/ml streptavidin (GLH, Bio-Rad), which could directly interact with histones (29), were used for immobilizing individual histones and measuring binding affinities to C4. Running buffer (10 mM HEPES pH 7.4, 150 mM NaCl [0.05% Tween 20]) and regeneration buffer (0.1 M glycine pH 2.2) were used throughout the assay. Five micrograms per milliliter of each recombinant histone (H1, H2A, H2B, H3, or H4) in running buffer was captured only on the surface of flow cell (Fc) 2 with Fc1 set as blank. For kinetics, a concentration series of C4 was injected at a flow rate of 10 μl/min over both captured histone surface and reference surface (blank) at 20°C. KDs were calculated using software provided by the manufacturer.
Complement activity assay
The effect of histones on complement activity in the three pathways was measured using COMPL300 Total Complement Functional Screen kit from Wieslab. Briefly, mixtures of the reaction were added to strips of wells for CP estimation that were precoated with IgM, and strips for AP determination were coated with LPS, whereas MBL strips were coated with mannan. Normal human sera were diluted 1/101 (CP and MBL) and 1/18 (AP) assay in specific kit buffers to ensure that only the pathway in question was activated (30). After 1 h of incubation at 37°C and then washing the strips, alkaline phosphatase–conjugated anti-human C5b-9 was added before incubation at room temperature for 30 min. Additional washing was performed, the substrate was added, and the wells were incubated for 30 min. Finally, absorbance values were read at 405 nm. In each assay, standard positive and negative control sera provided in the kit were used. The complement activity for each pathway was expressed as a percentage of the activity of the calibrating serum. C3a and C5a were measured using C3a and C5a ELISA kits (eBioscience). C5b-9 induced by zymosan (CompTech) was measured using an ELISA kit from Quidel.
Ab and heparin blocking assay
An antihistone reagent (non-anticoagulant heparin 20 μg/ml) was incubated with 20 μg/ml H1, H2A, H2B, H3, or H4 proteins or 20 μg/ml antihistone H4 Ab was incubated with H4 prior to complement activation using Wieslab COMPL CP310 kit. Percentage changes were calculated by comparing with untreated (100%).
C4 cleavage assay
C1s (50 μg/ml; CompTech) was incubated with C4 (250 μg/ml) in the presence or absence of histones (100 μg/ml) at 37°C for 30 min, then 2× SDS loading buffer was added and boiled for 10 min prior to SDS-PAGE. The gel was stained with Coomassie Brilliant Blue or subjected to Western blotting with anti-C4a Ab (CompTech). The C4a band intensities were measured using GeneSnap 7.05 software from Syngene, and fold changes were calculated.
Cell viability assay
Viability was assessed using a WST-8 cell proliferation assay kit (Enzo Life Sciences), as described previously (31). Briefly, 5 × 104 cells were seeded into each well of a 96-well plate and grown until fully confluent (24 h). Cells were treated with histones at 100 μg/ml with and without different concentrations of C4 (10–300 μg/ml) for 1 h. After treatment, the medium was changed to a fresh 100-μl growth medium and 10 μl of WST-8 dye was added to each well, followed by further incubation for 2 h. Viability was assessed by measuring the absorbance at 450 nm against a reference 650 nm using a microplate reader (Multiskan Spectrum; Thermo Fisher). Viability of untreated cells was set as 100% for comparison.
Statistical analysis
Intergroup differences were analyzed using ANOVA followed by Student–Newman–Keuls test. Two group comparisons with or without treatment used Student t test unless otherwise specified.
Results
Free histones exist in circulation and can form complexes with C4
Although it is known that nucleosomes can be released after cell death or neutrophil extracellular trap formation (32–36), it is not clear whether circulating histones are still exclusively in the form of histone–DNA complexes. Using ultracentrifugation to fractionate plasma or serum with high circulating histone levels into six fractions, we found that histones were detectable in all six fractions (Fig. 1A, upper panel). However, DNA–histone complexes (most likely nucleosomes) were in fraction 6 only (bottom fraction) (Fig. 1A, lower panel). No difference was found between plasma and serum. This experiment demonstrated that DNA free histones exist in circulation.
Identification of C4 as a histone-binding protein. (A) Critically ill patient plasma was separated into six fractions (1 = upper fraction; 6 = lower fraction) based on density using ultracentrifugation. Circulating histones (top panel) were measured by Western blot, and histone–DNA complexes (bottom panel) were quantified by ELISA (n = 4). (B) Using histone-conjugated Sepharose, a few proteins were pulled down. Among them, there were two major bands on Coomassie Blue–stained gel: one was C4 and the other was CRP, as identified by mass spectrometry. (C and D) The typical spectra of the two major proteins are presented. (E) Histone–C4 complexes were detected by ELISA following the addition of different concentrations of histones to normal plasma. (F) Histone–C4 complexes are elevated in critically ill patient plasma compared with normal plasma (n = 3). *p < 0.001 compared to normal.
Histone-conjugated Sepharose beads were then used to pull down human plasma proteins. Following extensive washing, proteins bound to histone beads were eluted. Multiple proteins were visualized on Coomassie Blue–stained gels with two major protein bands at ∼70 and 25 kDa (Fig. 1B). Following liquid chromatography–mass spectrometry analysis, C4 and C-reactive protein (CRP) were identified (Fig. 1C, 1D). CRP has been reported to be a major histone-binding protein that neutralizes histone toxicities (37). As to C4, we could detect histone–C4 complexes in normal plasma spiked with calf thymus histones (Fig. 1E) and also in plasma from critically ill patients with high circulating histone levels (Fig. 1F), confirming that histones form complexes with C4 in vivo. In this study, we further investigated the interaction of histones with C4 and its potential biological roles and significance.
Individual histones bind to C4 with different affinity
To determine the relative binding extents of individual histones to C4, equal molar concentrations of individual histones were subjected to gel overlay assay (Fig. 2A, upper) with Coomassie Blue–stained gel, which demonstrated equal loading (Fig. 2A lower). Fig. 2A shows that H3 and H4 predominantly bound to C4 and, to a lesser extent, H1 and H2B, with H2A–C4 binding undetectable using this method. To determine the comparative binding strengths under physiological conditions, SPR was used (Fig. 2B–F). Table I shows that H3 (KD = 0.76 ± 0.12 nM) and H4 (KD = 0.91 ± 0.07 nM) had much higher binding affinity than H1 (KD = 7.26 ± 0.80 nM) and H2B (KD = 9.45 ± 1.43 nM), with weak binding to H2A (KD = 12.67 ± 0.59 nM).
C4 binds to individual histones. (A) Two micrograms of H2A, H2B, H3, and H4; 4 μg of H1; and 6 μg of S100P as a control were subjected to SDS-PAGE. One gel was transferred onto polyvinylidene difluoride membrane and probed with HRP-conjugated C4 protein (upper panel). The other gel was stained with Coomassie Brilliant Blue (lower panel). (B–F) SPR analysis. Individual histones were first immobilized on streptavidin surfaces. Different concentrations of C4 were applied onto each surface and typical binding curves are presented.
Histones dramatically inhibit classical and MBL pathways but not AP
To investigate functional consequence, we used a Complement Functional Screen Kit to measure the effects of histones on the activation of the CP, MBL, and AP pathways. Preincubation of different concentrations of calf thymus histones with human serum significantly reduced the production of MAC by activated classical and MBL pathways. Significant reduction could be detected at 10 μg/ml histones, and only trace amounts of MAC could be formed in the presence of 50 μg/ml histones (Fig. 3A). In contrast, histones showed much less effect on the AP, and 50 μg/ml histones only reduced MAC ∼20%. To evaluate the overall effect of histones on complement activation in human serum, zymosan was used to activate complement in the presence or absence of histones. We found that histones at 50 μg/ml could significantly inhibit the production of MAC induced by zymosan (Fig. 3B). We also assessed the role of individual histones in the classical (Fig. 3C, 3D) and MBL (Fig. 3E, 3F) pathway activation and found that 20 μg/ml individual histones started to significantly reduce activation of both classical and MBL pathways, with H4 and H2B showing the most significant effects.
Histones inhibit complement activation. (A) Classical, MBL, and alternative complement pathways were activated by IgM, mannan, and LPS, respectively, in the absence or presence of different concentrations of calf thymus histones (0–50 μg/ml). MAC was detected by anti-human C5b-9 Abs. The complement activity of control wells without histones was set up as 100%. The mean ± SD of relative activities were presented. (B) Shows the mean ± SD of relative activities activated by zymosan (activating different pathways) in the presence of different concentrations of calf thymus histones (0–50 μg/ml). The mean ± SD of relative activities of classical (D) and MBL (F) pathways in the presence of 20 μg/ml individual histones. The mean ± SD of relative activities of classical (C) and MBL (E) pathways in the presence of different concentrations of individual histones (0–50 μg/ml). Mean ± SD were calculated from at least three independent experiments. ANOVA test, *p < 0.05 compared with untreated.
Antihistone reagents can rescue complement activation
To demonstrate the specificity of histones on complement activation, antihistone H4 and non-anticoagulant heparin, which have been shown to specifically inhibit histone toxicity both in vitro and in vivo (11, 38), were used. Heparin could reverse the inhibition of both classical and MBL pathways by all individual histones (Fig. 4A, 4B), whereas antihistone H4 could significantly rescue the H4-inhibited complement activation of both pathways (Fig. 4C, 4D). Those data demonstrated that the effect of histones on complement inactivation was specific to histones.
Antihistone treatment rescues complement activation. (A and B) Non-anticoagulant heparin (20 μg/ml) was used to rescue complement activities of classical (A) and MBL (B) pathways inhibited by individual histones (20 μg/ml). (C and D) Antihistone H4 Ab (12 μg/ml) was used to rescue complement activities of classical (C) and MBL (D) inhibited by H4 (20 μg/ml). The mean ± SD of relative activities from at least three independent experiments were presented. ANOVA test, *p < 0.05 when compared with untreated, ‡p < 0.05 when compared with that treated with histone alone.
Excess C4 protein only partially rescues histone-inhibited complement activation but significantly reduces cytotoxicity of histones
Using C4 protein up to 300 μg/ml, only one third of the maximal complement activity of classical and MBL pathways could be recovered in the presence of calf thymus histones up to 20 μg/ml (Fig. 5A, 5B). However, the zymosan-induced complement activation could be recovered by 300 μg/ml C4 from 25 to 70% of total activity in the presence of 20 μg/ml histone H4 protein (Fig. 5C). This observation suggests that histones may also target other components of the complement system rather than C4 alone. In contrast, C4 protein could significantly reduce the cytotoxicity of histones to cultured endothelial cells (Fig. 5D).
Effect of C4 protein on histones. (A and B) C4 protein affects histone-inhibited complement activity. Adding C4 protein (0–300 μg/ml) rescued 20 μg/ml histone H4-inhibited activation of CP (A) and MBL (B) pathway. (C) Zymosan activated complement activity in the absence or presence of histone H4 20 μg/ml and C4 protein 300 μg/ml (histone H4/C4 molar ratio = 1:1). (D) Human endothelial cell line, EA.hy926, was treated with 100 μg/ml calf thymus histones in the presence 0–300 μg C4 proteins for 1 h. The percentage of cell viability was detected using WST-8 cell viability kit. Mean ± SD from three independent experiments are presented. ANOVA test, *p < 0.05 when compared to control, ‡p < 0.05 when compared to histone alone.
Histones do not affect C4 cleavage but significantly reduce C3 and C5 convertase activity
To clarify the molecular mechanism of histone-inhibited complement activation through interaction with C4, the effect of histones on the cleavage of C4 to C4b and C4a by C1s, a process of C4 activation, was investigated. We found that histones showed no effect on the production of C4a (Fig. 6A, 6B), indicating histone binding does not affect the ability of C1s to cleave C4 protein. Further investigation showed that histones bind to C4b but not C4a (Fig. 6C). However, in the presence of histones, the production of C3a and C5a were significantly reduced in the classical and MBL pathways but not the AP (Fig. 6E, 6F), suggesting that histone-bound C4b is not as efficient as C4b alone in forming active C3 and C5 convertases. The overall C3a, C5a, and C5b-9 production induced by zymosan (Figs. 6E, 6F) was significantly reduced by histones because of the suppression of both classical and MBL pathways, which are summarized in Fig. 7.
Histones show no effect on C4 cleavage but significantly reduce C3 and C5 convertase activities. (A) In vitro cleavage of C4 by C1s in the presence or absence of histones. C4 (250 μg/ml) was incubated with C1s (50 μg/ml, active enzyme to cleave C4 into C4a and C4b) ± calf thymus histones (100 μg/ml) at 37°C for 30 min and subjected to 8–18% gradient SDS-PAGE along with calf thymus histones, C4a, C4b, C4, and C1s proteins. A typical Coomassie Brilliant Blue–stained gel is presented. (B) A typical Western blot with anti-C4a Ab is presented (upper panel). Fold changes were calculated by setting up C4a intensity without histones as 1. The relative fold changes of cells treated with calf thymus histones from three independent experiments are presented (lower panel). Student t test, p = 0.2. (C) Two micrograms of C4, C4b, C4a, and S100P (as a control) were subjected to blotting with HRP-conjugated calf thymus histones. A typical blot is presented. (D–F) Complement in serum was activated by IgM (CP), mannan (MBL), LPS (AP), or zymosan in the absence or presence of calf thymus histones treated (50 μg/ml) for 1 h at 37°C. Then, the C3a (D), C5a (E), or MAC levels (F) were detected by ELISA. Mean ± SD from at least three independent experiments are presented. ANOVA test, *p < 0.05 when compared with that without histones.
Schematic representation of the effect of histones in the complement pathway.
Discussion
Complement activation generates MAC to lyse cells and leads to cell death and content release, including histones. The inhibitory effects of histones on complement activation could form a physiological feedback loop to prevent overproduction of MAC and excessive tissue damage. This finding is novel and, with evidence that histone–C4 complexes exist in the circulation of critically ill patients, adds relevance to filling the unknown gap on communication between targeted cells and complement (Fig. 7).
C4 is activated by C1s cleavage to produce C4a and C4b. The C4b is the essential component of both C3 and C5 convertases, a common step of both classical and MBL pathways (Fig. 7). Histones strongly bind to C4 but do not affect C4 activation because there is no difference in C4a production in the presence or absence of histones. Histones bind to C4b but not C4a; therefore, their major effect is to reduce the activity of C3 and C5 convertases, as indicated by reduction of C3a and C5a, the products of C3 and C5 activation. One mechanism could be the interruption of the convertase formation and the other could just affect the catalytic activity even though the complexes are formed. Because the lifetime of C3 and C5 convertases in solution are very short, it is difficult to distinguish the two potential mechanisms. In contrast, histones have minimal effect on AP, in which C4b is not required. However, the overall effect is the significant reduction of C3 and C5 activation as well as the MAC formation. This finding suggests that C4 is a major target of the circulating histones. However, excess of C4 could not fully restore the complement activation in the presence of histones. This finding suggests that histones may have more targets on those pathways, such as C1 or C2 (Fig. 7), and further investigation is required.
It is known that circulating C4 is ∼0.4 mg/ml, but no histones could be detected in blood from heathy donors (39). In critical illness (for example, sepsis), histones could surge up to 100–200 μg/ml (12), but C4 was reported to decrease because of consumption (39). Therefore, the high levels of histones are sufficient to inhibit both classical and MBL pathways. Although high levels of C4 could efficiently detoxify histones in vitro, the low levels of C4 in sepsis may not be sufficient to neutralize high levels of histones. In noncritical illness, such as chronic inflammatory diseases with complement activation, the circulating histones could be very low, but the local concentration of histones released from lysed cells may be high and sufficient to suppress further complement activation and prevent excessive injury of host tissues. Further laboratory experiments and clinical investigation are required to clarify those points.
Disclosures
The authors have no financial conflicts of interest.
Footnotes
This work was supported by the Medical Research Council (G0501641), the British Heart Foundation (PG/14/19/30751 and PG/16/65/32313), and the National Institute for Health Research (II-FS-0110-14061, NIHR-BRF-2011-026). Y.Q. was sponsored by The Higher Committee for Education Development in Iraq.
Abbreviations used in this article:
- AP
- alternative pathway
- CP
- classical pathway
- CRP
- C-reactive protein
- MAC
- membrane attack complex
- MBL
- mannose-binding lectin
- SPR
- surface plasmon resonance.
- Received May 30, 2017.
- Accepted April 16, 2018.
- Copyright © 2018 by The American Association of Immunologists, Inc.
References
In this issue
