Systemic Lupus Erythematosus Patients Contain Significantly Less IgM against Mono-Methylated Lysine than Healthy Subjects

Post-translational modifications on proteins are important in biological processes but may create neo-epitopes that induce autoimmune responses. In this study, we measured the serum IgG and IgM response to a set of non-modified or acetyl- and methyl-modified peptides corresponding to residues 1–19 of the histone 3 N-terminal tail in systemic lupus erythematosus (SLE) patients and healthy subjects. Our results indicated that the SLE patients and healthy subjects produced antibodies (Abs) to the peptides, but the two groups had different Ab isotype and epitope preferences. Abs to the non-modified form, H31–19, were of the IgG isotype and produced by SLE patients. They could not recognize the scrambled H31–19, which contained the same amino acid composition but a different sequence as H31–19. In comparison, healthy subjects in general did not produce IgG against H31–19. However, about 70% of the healthy subjects produced IgM Abs against mono-methylated K9 of H31–19 (H31–19K9me). Our further studies revealed that ε-amine mono-methylated lysine could completely inhibit the IgM binding to H31–19K9me, but lysine had no inhibitory effect. In addition, the IgM Abs could bind peptides containing a mono-methylated lysine residue but with totally different sequences. Thus, mono-methylated lysine was the sole epitope for the IgM. Interestingly, SLE patients had much lower levels of this type of IgM. There was no obvious correlation between the IgM levels and disease activity and the decreased IgM was unlikely caused by medical treatments.We also found that the IgM Abs were not polyreactive to dsDNA, ssDNA, lipopolysaccharide (LPS) or insulin and they did not exist in umbilical cord serum, implying that they were not natural Abs. The IgM Abs against mono-methylated lysine are present in healthy subjects but are significantly lower in SLE patients, suggesting a distinct origin of production and special physiological functions.


Introduction
Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by the production of various autoantibodies (autoAbs). Many of these antibodies (Abs) are reactive to components in the cell nuclei, such as dsDNA, nucleosomes, Sm, La, Ro, etc. [1,2]. As the major components of nucleosomes, histones are also common autoantigens in SLE. These autoAbs form immune complexes with the antigens, which leads to excessive activation of complement and phagocytes, resulting in severe inflammation and multiorgan damage.
Histones can be grouped into linker histones (H1), core histones (H2A, H2B, H3 and H4) and other variants. The core histones are small basic proteins consisting of a globular domain and a flexible N-terminal tail, which comprises about 25-30% of the mass of the individual histones. Two copies of each core histone are assembled into an octamer that has 146 base pairs of DNA wrapped around to form the nucleosome core. The histone N-terminal tails protrude from the nucleosome and are exposed on the surface [3][4][5]. So far, autoAbs to all five of the histones have been found in SLE patients and lupus mouse models [2,6]. Most of these autoAbs recognize epitopes exposed on nucleosome surfaces, such as the 22-42 residues of H19, 1-25 residues of H2B, 1-21 residues of H3, and 1-29 residues of H4. There are fewer autoAbs against epitopes that are not located on histone tails, such as residues 65-85 of H2A and residues 40-55 of H3 [7][8][9][10][11].
Histones can be extensively modified by acetylation, methylation, phosphorylation, ubiquitination, sumoylation, ADP-ribosylation, deimination and proline isomerization [12][13][14][15][16]. Most of the post-translational modifications are found on the histone Nterminal tails. These modifications influence the overall structure of chromatin and play fundamental roles in many biological processes that are involved in the manipulation and expression of DNA, including cell development, differentiation, proliferation and apoptosis. Post-translational modifications on histones or other molecules may also create neo-self epitopes that can be recognized by antigen receptors on T cells or B cells and contribute to the development of autoimmunity in genetically predisposed individuals [17,18]. Plaué et al. reported that ubiquitinated H2A and H2B were targets for SLE autoAbs [19]. van der Vlag et al. reported that apoptosis-associated histone modifications generated the epitopes H2BK12ac, H3K27me3, H4K8ac3, H4K12ac3 and H4K16ac3 that were recognized by autoAbs in SLE patients and lupus mice [20][21][22].
There are many sites on the 1-19 residues of the H3 N-terminal tail that can be modified to generate neo-epitopes. Although autoAbs to the non-modified H3 N-terminal tail have been reported [8], the existence of autoAbs to the modified H3 1-19 Nterminal tail is unknown. One study reported H3K9me3S10ph in an isolated case from a patient who had discoid lupus erythematosus and chronic lymphocytic leukemia [23].
In this study, we tested the reactivity of IgG and IgM from SLE patients and healthy subjects against a set of synthetic peptides corresponding to the 1-19 residues of the H3 Nterminal tail with or without methylation and acetylation. We found that both SLE patients and healthy controls could produce Abs to the peptides, but the two groups had different isotype and epitope preferences. IgG Abs were mainly produced by SLE patients with specificities to the non-modified peptide, whereas the healthy controls contained mostly IgM Abs recognizing epitopes containing mono-methylated lysine. Interestingly, SLE patients had significantly lower levels of this type of IgM than the healthy subjects.

Ethics Statement
This study was performed in accordance with the Declaration of Helsinki and approved by the ethic committees of Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Approval ID 012-2012, Institutional Review Board of IBMS, CAMS). All the personal privacy was well protected throughout the work. All the serum samples used in this work were leftover samples after clinical examinations of patients or routine health check of healthy people. As these samples were treated as abandoned samples and used in anonymous codes, the informed consent was exempted. Patient's clinical information was obtained through a doctor who had the master key of coding and he would strictly abide by the confidentiality agreement. The umbilical cord blood samples were treated as medical waste and could be disposed by the hospitals. They were also used anonymously and coded without identifiable information.  the Table S1. All of the patients met at least 4 classification criteria from the American College of Rheumatology [24]. Patients who were undergoing treatments when their blood was collected were listed in the Table S2. 36 juvenile idiopathic arthritis (JIA) patients (8 males and 28 females; mean age 10.8662.75 yrs old) and 26 patients with other rheumatoid diseases (RD), including juvenile ankylosing spondylitis (2 males and 2 females; mean age 9.5062.08 yrs old), Henoch-Schonlein purpura (2 males and 9 females; mean age 8.3663.01 yrs old), idiopathic thrombocytopenic purpura (1 female; 9 yrs old), Kawasaki disease (1 female; 1 yr old), mixed connective tissue disease (1 female; 10 yrs old), juvenile dermatomyositis (1 male and 5 females; mean age 11.8361.47 yrs old) and Behcet's disease (2 females; mean age 11.5060.71 yrs old) were recruited between 2001 and 2012 from the Department of Pediatric Rheumatology, Capital Institute of Pediatrics. The healthy controls were people who had undergone routine health checkups. None of the controls had any rheumatologic conditions when recruited. Umbilical cord blood was collected from newborns in the Maternity Hospital of Sanhe, Langfang City, Hebei Province, China.

Patients and Healthy Controls
Sera from SLE patients and healthy controls were obtained from whole blood and stored at 280uC until use.
Detection of Anti-peptide Abs by ELISA 96-well microtiter plates (Kelongda Institute, Beijing, China) were coated with H3 peptides cross-linked to BSA (1 mg/ml according to the BSA concentration, 100 ml/well) at 4uC overnight in coating buffer containing 15 mM Na 2 CO 3 , 35 mM NaHCO 3 , pH 9.6. After washing with PBS containing 0.05% Tween-20 (PBST), the plates were blocked with 200 ml of PBST containing 2% BSA for 2 h at room temperature. Then, 100 ml/well of the serum samples (1:100 diluted in PBST containing 2% BSA) were added and incubated at room temperature for 1 h. After washing, the mouse anti-human IgG monoclonal Ab (mAb) KT47 or mouse anti-human IgM mAb KT16 (Absea) diluted at 1:1000 was added. After incubating at room temperature for 1 h, the plates were washed with PBST and HRP-conjugated goat anti-mouse IgG (Product No. A2554. Sigma-Aldrich Co, St. Louis, MO, USA) diluted at 1:5000 was added. After incubating at room temperature for 1 h, the wells were washed and the substrate 2, 29-azino-bis-3-ethylbenzothiazoline-6-sulphonic acid (ABTS, Product No. 0400. Amresco, Solon, OH, USA) was added. Color was developed for 20 min (for IgM) or 40 min (for IgG), and the OD values were read at 405nm.

Measurement of Serum IgM Concentrations by Sandwich ELISA
96-well microtiter plates (Kelongda) were coated with 100 ml/ well of KT16 as a capture Ab (2.5 mg/ml) at 4uC overnight. The plates were then washed with PBST and blocked with 200 ml/well of PBST containing 2% BSA for 2 h at room temperature. The samples diluted in PBST containing 2% BSA (1:1600 for adult serum IgM; 1:800 for pediatric serum IgM) were then added. After incubating at room temperature for 1 h, the plates were washed and HRP-conjugated mouse anti-human IgM mAb KT38 (Absea) diluted at 1:1000 was added as the detection Ab. After incubating for 1 h at room temperature, the plates were washed and the color was developed for 30 min using ABTS as the substrate.

Polyreactivity Tests
The tests were performed as described by Tiller et al. [26]. Briefly, 96-well microtiter plates (MaxiSorp, Product No. 449824. Nunc, Thermo Fisher Scientific, Roskilde, Denmark) were coated with 100 ml/well of dsDNA, ssDNA and lipopolysaccharide (LPS) (Product Nos. D1501, D8899, L2630. Sigma-Aldrich) in PBS at 10 mg/ml and insulin (Product No. 090-03446. Wako, Osaka, Japan) in PBS at 5 mg/ml. After incubating at 4uC overnight, the plates were washed with PBST and blocked in 2 mM EDTA/ PBST (200 ml/well) for 1 h at room temperature. 1:200 diluted eluates from the tandem purification in 2 mM EDTA/PBST were added and incubated for 2 h. The plates were washed, and 100 ml of KT47 or KT16 in 2 mM EDTA/PBST was added and incubated for 1 h at room temperature. After washing, HRPconjugated goat anti-mouse IgG diluted in 2 mM EDTA/PBST was added. After incubating at room temperature for 1 h, the plates were blocked with 2 mM EDTA/PBST (200 ml/well) again. Color development was performed using ABTS as the substrate.
The Reactivity of anti-H3 1-19 K9me IgM to Histones 96-well microtiter plates (Kelongda) were coated with histones (20 mg/ml, 100 ml/well) at 4uC overnight in coating buffer. After washing with PBST, the plates were blocked in 200 ml of PBST containing 2% BSA for 1 h at room temperature. 100 ml/well of the purified IgM anti-H3 1-19 K9me Abs (1:20 dilution) that were incubated in the presence or absence of lysine or mono-methylated lysine (5 mg/ml) were added and incubated at room temperature for 1 h. KT16 was used as the primary Ab and HRP-conjugated goat anti-mouse IgG was used as the secondary Ab.

Statistics
All of the data were analyzed using the GraphPad Prism software (Version 5.01). For data with a normal distribution, an unpaired t test was performed to analyze the differences between two groups and a paired t test was used to compare differences within a group. The Mann-Whitney test was used to compare data with a non-normal distribution. Percentages were compared using the Chi-square test or Fisher's exact test. Pearson test was used for calculating correlation between data with normal distribution. P,0.05 was considered significant.
For Ab purification, 4 or 5 serum samples within a subgroup were pooled together. Equal volumes (1.5 ml) of the pooled sera were absorbed twice with H3 1-19 beads. The unbound Abs in the sera were then absorbed with H3 1-19 K9me beads. The beads were washed with 1 M NaCl to remove the non-or low-specific Abs bound on the beads. The Abs were then separately eluted and analyzed by Western blot and ELISA.
Western blot was performed to detect the Ab isotypes and amounts of IgG and IgM in the eluates from each subgroup. The results are shown in Figure 3. For the Abs purified using the H3  Figure 1 was due to low-or non-specific binding.
The 25 raw serum samples randomly selected from aSLE patients and adult controls also reacted against GGKme, and the differences between the aSLE and control sera were more obvious because the serum may have contained less non-specific binding to GGKme than to H3 1-19 K9me ( Figure S2).
It was surprising to observe that IgM were specifically produced against an epitope with only one amino acid, i.e., mono methylated lysine. It is known that poly2/auto-reactive IgM Abs play important roles in the protection against autoimmune diseases [27,28]. Genetically manipulated mice that were deficient in secretory IgM but not other Ab classes had an increased propensity to the spontaneous development of IgG anti-DNA Abs and the renal deposition of IgG and complement [29,30]. Lupusprone mice treated with murine IgM anti-dsDNA Abs exhibited a delayed onset of proteinuria and a reduced degree of renal pathology, which resulted in significantly improved survival [31]. The expression of ppc1-5, a natural IgM autoAb, in MRL-lpr mice prevented proteinuria and reduced kidney immune complex formation [32]. MRL/lpr mice deficient in activation-induced deaminase produced high levels of autoreactive IgM but lacked autoreactive IgG. These mice showed a significant reduction in glomerulonephritis and a dramatic increase in survival [33]. In humans, it is known that IgG Abs against dsDNA are involved in the pathogenesis of SLE glomerulonephritis, but this disease is rare in SLE patients with IgM against dsDNA. There was a negative correlation between anti-dsDNA IgM Abs and glomerulonephritis [34]. SLE patients with low disease activity tended to have higher levels of polyreactive IgM Abs [35]. Anti-phosphorylcholine IgM was significantly higher in patients with low disease activity and less organ damage, and anti-cardiolipin and anti-dsDNA IgM were significantly higher in patients without renal disease [36]. IgM against mono-methylated lysine found in this work could be classed as an autoAb because mono-methylated lysine is present on histones in eukaryotic cells [37]. Other proteins, such as p53, NFkB and STAT3, also have mono-methylated lysine [38]. Although we presently do not know the exact effector functions of these Abs, the fact that healthy subjects have significantly higher levels of these Abs than SLE patients suggests the possibility that these Abs are required for healthy conditions, and reduced levels may lead to SLE (-like) autoimmune diseases.
Several mechanisms have been elucidated for IgM autoAbs in preventing autoimmune diseases. IgM autoAbs are efficient in binding and neutralizing autoantigens like dsDNA. The immune complexes are phagocytosed rather than deposited on the glomerular basement membrane [31]. IgM autoAbs facilitate the removal of apoptotic cells, and thus the reduction of IgM may result in an impaired clearance of apoptotic cells and cell debris, which may stimulate the production of pathogenic IgG autoAbs. Although the functions of IgM against mono-methylated lysine are not known, it is possible that histones are the major targets of these Abs, as histones are the most abundant proteins associated with lysine methylation, and only a few methylated non-histone proteins have been found [38]. Because it is difficult for IgM to  Microtiter plates were coated with histones. IgM Abs purified from the H3 [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] K9me beads with or without prior absorption with lysine or mono-methylated lysine were tested. Background wells were not added with purified IgM. KT16 anti-human IgM was used as the primary Ab and HRP-conjugated goat anti-mouse IgG was used as the secondary Ab. Data are expressed as mean+SEM. The results are representative of three separate experiments. doi:10.1371/journal.pone.0068520.g008 penetrate cells, the possible effector sites for IgM Abs against mono-methylated lysine would be on the cell surface or with cell debris, where the Abs could bind histones released from damaged or apoptotic cells and subsequently activate complement to facilitate the removal of unwanted histones.
Autoreactive IgM can be divided into natural or immune Abs. The natural autoAbs (NAA) are produced without exogenous antigen stimulation, as they are present in the cord blood from newborn humans and mice as well as in mice housed in germ-free conditions and fed an antigen-free diet [32]. B1 cells are the major source of NAA, but B2 cells, including marginal zone and follicular B cells, can also produce NAA [27,32,39]. The origin of IgM Abs against mono-methylated lysine is not presently known. It was originally thought these Abs were NAA produced against self-mono-methylated lysine on histones or other proteins. If this was the case, these IgM should have been produced before birth in the absence of foreign antigens. However, we did not observe any IgM Abs against mono-methylated lysine in the umbilical cord sera where substantial amounts of IgM exist (Figure 9). In addition, the Abs did not react against dsDNA, ssDNA, LPS and insulin, which is the general property of polyreactive NAA [40]. Furthermore, most NAA have low affinities to their antigens, but the IgM Abs bound mono-methylated lysine with high affinity, as washing with1 M NaCl did not disassociate these Abs from the H3 1-19 K9me beads. Therefore, these IgM might be 'immune' Abs which are acquired in response to foreign antigens after birth and somehow gain the ability to react against self-mono-methylated lysine through mechanisms such as molecular mimicry [41]. As histone methylation is very common in eukaryotic cells, humans can easily access mono-methylated lysine from countless microorganisms through various infections [42]. In this study, IgM could be detected in children as early as 2 yrs old. It would be interesting to determine when babies start to produce these Abs. To determine whether the Abs against mono-methylated lysine exist in other species, we measured IgM against mono-methylated lysine in BALB/c and C57BL/6 mice, but the levels were undetectable (unpublished observations). This may be because the animals were kept in specific pathogen free conditions where they were unable to contact the necessary antigens to evoke a response against the mono-methylated lysine.
Both the healthy subjects and patients with autoimmune diseases produced autoAbs. However, the relationship between the autoAbs produced in the healthy subjects (mostly the IgM isotype) and the pathogenic autoAbs produced in the patients (mostly the IgG isotype) is still unclear. It is speculated that natural poly2/auto-reactive Abs may occasionally promote autoimmunity by serving as a template for the high affinity pathogenic autoAbs. These poly2/auto-reactive Abs may also suppress autoimmunity by regulating excessive autoimmune and inflammatory responses [32]. It is unknown if the autoreactive IgM produced in the healthy status is a major source of pathologic IgG autoAbs in disease or if the autoAbs in disease have different origins. Although NAA are mainly produced by B1 cells, there is little evidence in humans that B1 cells are the sources of pathogenic autoreactivity, although there are murine models of systemic lupus where B1 cells are the source of pathogenic autoAbs [43]. However, cells of the B2 origin, including marginal zone and follicular B cells, are more likely linked to pathologic autoAbs. An analysis of healthy people reported that about 40% of the new emigrant B cells from the bone marrow (mainly of the B2 subset) and about 20% of the mature naïve B cells had features of autoreactivity and reacted to HEp-2 cells [44]. Although most of these cells will not go into the plasma cell pools, they have the potential to become autoAb secreting cells. In addition, somatic hypermutation creates autoreactive IgG + memory B cells [45]. It has been observed that SLE patients are defective in removing autoreactive B cells from new emigrant B cells, and these cells may have acquired autoreactivity from somatic hypermutations [46][47][48]. In our study, we were unable to find any links between anti-H3 [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] IgG and anti-Kme IgM, although both of these Abs recognize the H3 N-terminal peptides. We did observe that SLE patients produced little but detectable anti-H3 1-19 K9me IgG ( Figure 3). However, we do not know whether they were from antimono-methylated lysine IgM by class switching or from a different origin.
In conclusion, we have identified mono-methylated lysine as a novel epitope for IgM autoAbs in a majority of healthy subjects, and the levels of these IgM Abs in SLE patients are significantly lower than that in the healthy controls. These IgM autoAbs are   acquired after birth and can recognize peptides or proteins containing mono-methylated lysine residues but not di-or trimethylated lysine. Although the serum IgM levels against monomethylated lysine do not associate with disease activity, they may still be useful in discriminating SLE patients from healthy persons.