Low levels of IgG recognizing the alpha-1-antitrypsin ( A 1 AT ) 50-63 1 peptide and its association with Taiwanese women with primary Sjögren ' s 2 syndrome

61 The aim of this study was to examine oxidative stress and low level of alpha62 1-antitrypsin (A1AT) in primary Sjögren's syndrome (pSS), and evaluate the 63 associated autoreactivity against unmodified and their 4-hydroxy-2-nonenal (HNE)64 modified peptides with pSS. Two differentially expressed proteins, alpha-1-acid 65 glycoprotein 1 (A1AG1) and A1AT, exhibited 2-fold differences, and their HNE 66 modifications were identified by depleted-albumin and immunoglobulin G (IgG) 67 serum protein, in-solution digestion, in-gel digestion, and nano-LC-MS/MS from pSS 68 patients and age-matched healthy controls (HCs). Furthermore, levels of proteins, 69 confirmation of HNE modifications, HNE-protein adducts and autoreactivity against 70 unmodified and their HNE-modified peptides were further validated. Levels of the 71 HNE-protein adduct and A1AG1 were significantly higher in pSS patients than HCs, 72 but levels of A1AT were significantly lower in pSS patients compared to HCs. Only 73 the HNE modification of A1AT was confirmed. Further, concentrations of anti74 A1AT50-63 IgG and anti-A1AT50-63 HNE IgA were significantly lower in pSS patients 75 than HCs. Our study suggests that elevated HNE-protein adduct, oxidative stress, 76 level [odds ratio (OR) 4.877, p = 0.003], lowered A1AT level (OR 3.910, p = 0.010) 77 and a decreased level of anti-A1AT50-63 IgG (OR 3.360, p = 0.010) showed an 78 increased risk in pSS patients compared to HCs, respectively. 79 80 Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 30 October 2017 doi:10.20944/preprints201710.0180.v1


Introduction
Primary Sjögren's syndrome (pSS) is a chronic inflammatory autoimmune disease characterized by dysfunction of the exocrine glands leading to dryness of the mouth and eyes [1].Patients with pSS feature the presence of autoantibodies mainly against the ribonucleoprotein complex SS-related antigen A (SSA, Ro) and SS-related antigen B (SSB, La) [1].In 2000~2008, the prevalence of pSS was 16.0 (females, 28.8, males, 3.7; female: male ratio, 7.9) per 100,000 persons; the incidence rate of pSS was 10.6 (females, 18.5, males, 2.9; female: male ratio 6.3) per 100,000 personyears; and the mortality from pSS was 1304.7 (females 987.4,males 3444.2;ageadjusted female: male ratio 0.4) per 100,000 person-years in Taiwan [2].The etiology and pathogenesis of Sjögren's syndrome are not clearly understood [3].Jonsson and Brun proposed etiopathogenic events prior to a diagnosis of SS including a genetic predisposition, environmental triggers, autoantibodies, pathological injury, clinical disease, and clinical presentation [4].
Norheim et al. reported that patients with pSS have high levels of oxidative stress compared to healthy controls (HCs) [5].Wakamatsu et al. found an increase in 4-hydroxy-2-nonenal (HNE)-protein adducts, marker of oxidative stress, in the conjunctiva of SS patients that may play a role in the pathogenesis of dry-eye disease [6,7].HNE is one of the lipid peroxidation products that has an alkene bond and an aldehyde group which react with amino acid residues that form HNE-protein adducts via types of Michael addition and Schiff-base adducts, respectively [8].Amino acid residues that can react with HNE include cysteine (C), histidine (H), lysine (K), arginine (R), glutamine (Q), alanine (A), and leucine (L) [8][9][10].The HNE-protein adduct is an autoantigen and can elicit specific autoantibody formation [11,12].Breit et al. indicated that several immune-mediated diseases were associated with an alpha-1-antitrypsin (A1AT) deficiency including rheumatoid arthritis (RA), anterior uveitis, systemic lupus erythematosus (SLE), and asthma in which A1AT may play roles as an anti-inflammatory and immune regulator [13].A1AT is a serine protease inhibitor [14].Further, two cases were reported with an A1AT deficiency in patients with pSS in which the A1AT level of plasma declined by 1.28~2.10-fold[15,16].
In the present study, our aim was to investigate whether a low level of serum A1AT occurs in Taiwanese women with pSS and then identify the HNE modification on A1AT using depleted-albumin and immunoglobulin G (IgG) serum, in-solution digestion, one-dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (1-D SDS-PAGE), in-gel digestion, and label-free nano-liquid chromatography tandem mass spectrometry (nano-LC-MS/MS) from pSS patients vs. HCs.Further, we also assessed associations of autoantibody isotypes against A1AT 50-63 and their HNEmodified peptides with pSS patients compared to HCs.

Identification and validation of differentially expressed serum proteins by insolution digestion and LC-MS/MS
Enrichment of depleted-albumin and IgG serum protein samples from a single pair of each of nine pooled serum samples (patients with pSS vs. HCs) was analyzed in triplicate by in-solution digestion coupled to nano-LC-MS/MS (Table 1).In total, 255 proteins were detected, of which 28 differentially expressed proteins significantly varied, as shown in Table 1 and Supplementary Table 2.There were seven upregulated proteins and 21 downregulated proteins; relative to the HC serum pools, two of the identified proteins, alpha-1-acid glycoprotein 1 (A1AG1) and A1AT, differed by a 2-fold increase or decrease in patients with pSS serum pools, and 26 proteins differed by a 1.7~1.9-foldincrease or decrease (Table 1).
Next, we validated the LC-MS/MS data of A1AG1 and A1AT, and protein levels of A1AG1 (~48 kDa) and A1AT (~55 kDa) were analyzed by Western blotting (Figure 1).A1AG1 expression levels were significantly higher in pSS samples by 1.53-fold (p = 0.0001) than in HCs, but A1AT levels in pSS samples were significantly lower than those in HCs by 1.84-fold (p = 0.0071, Figure 1A, B, right upper panel).Equal amounts of serum proteins in these experiments were examined (Figure 1A, B, right bottom panel).The AUC value, sensitivity, and specificity of serum A1AG1 and A1AT in pSS samples vs. HCs were calculated based on these results and plotted on an ROC curve.The Western blot results of A1AG1 showed that the AUC was 0.75, the sensitivity was 85.0%, and the specificity was 62.5% for pSS detection at an average densitometric cutoff of 19,994.82; the AUC was 0.67, the sensitivity was 77.5%, and the specificity was 60.0% for pSS detection by A1AT at an average densitometric cutoff of 104,087.25 (Figure 1C).

Novel HNE modification identification of serum proteins by in-gel digestion and LC-MS/MS
In addition to serum protein levels, we further identified HNE modifications of A1AG1 and A1AT.The average coverage of amino acid sequences in A1AG1 and A1AT were estimated to be 52% and 70%, respectively.No HNE modification was identified on A1AG1 (data not shown).Novel HNE modifications of A1AT were identified by manual examination of the modified spectra using the PeaksPTM module in PEAKS 7 software.Further, HNE modifications of A1AT were confirmed in the two pooled serum samples (patients with pSS vs. HCs) through IP-Western blotting, which detected signals of approximately 55 kDa (Figure 2).Because low coverage of A1AG1 was identified, we also confirmed HNE modifications of A1AG1 using IP-Western blotting, but no signal was detected (data not shown).
MS/MS spectrum data of HNE-modified peptides on A1AT are presented in Supplementary Figure 1B and Supplementary Table 3.The peptide 50 -ITPNLAEFAFSLYR-63 was used to identify A1AT as pSS-specific and was found to have an HNE modification at A58. Identification of the peptide moieties was based on the presence of b-and y-series ions derived from the peptide, and HNE-modified residues were confirmed by an unmodified b8 ion followed by a modified y6 ion that corresponded to a mass increase of 156.11504Da (Supplementary Figure 1C, upper panel).The peptide 360 -AVLTIDEK-367 was used to identify A1AT as HC-specific and was found to have an HNE modification at A360.Identification of the peptide moieties was based on the presence of b-and y-series ions derived from the peptide, and HNE-modified residues were confirmed by an unmodified y7 ion followed by a modified b1 ion that corresponded to a mass increase of 138.10446Da (Supplementary Figure 1C, bottom panel). 50-63and A1AT 50-63 HNE peptides Serum samples were assessed with autoantibody isotypes against A1AT 50-63     and A1AT 50-63 HNE peptides by an ELISA.The level of the anti-A1AT 50-63 IgG antibody in RA was significantly higher than that of HCs by 1.80-fold (p = 0.0002), that of SLE vs. HC was 2.41-fold higher (p < 0.0001), that of RA vs. pSS was 2.52fold higher (p < 0.0001), that of SLE vs. pSS was 3.38-fold higher (p < 0.0001), and that of SLE vs. RA was 1.34-fold higher (p = 0.0321); however, that of pSS was significantly lower than that of the HCs by 1.40-fold (p = 0.0488, Figure 3A, left panel).Anti-A1AT 50-63 IgM expression levels did not significantly differ among patients with pSS, RA, SLE, and HCs, except that of RA was significantly higher than that of pSS by 1.57-fold (p = 0.0143, Figure 3A, middle panel).Levels of the anti-A1AT 50-63 IgA antibody did not significantly differ among patients with pSS, RA, SLE, and HCs (Figure 3A, right panel).

Autoreactivity against A1AT
The level of the anti-A1AT 50-63 HNE IgG antibody in RA was significantly higher than that of HCs by 2.10-fold (p < 0.0001), that of SLE vs. HC was 2.70-fold higher (p < 0.0001), that of RA vs. pSS was 2.69-fold higher (p < 0.0001), that of SLE vs. pSS was 3.48-fold higher (p < 0.0001), and that of SLE vs. RA was 1.29-fold higher (p = 0.0469); however, that of pSS did not significantly differ from that of the HCs (Figure 3B, left panel).Anti-A1AT 50-63 HNE IgM expression levels did not significantly differ among patients with pSS, RA, SLE, and HCs, except that of SLE was significantly higher than that of HCs by 1.36-fold (p = 0.0433) and that of SLE vs. pSS was 1.47-fold higher (p = 0.0088) (Figure 3B, middle panel).Levels of the anti-A1AT 50-63 HNE IgA antibody did not significantly differ among patients with pSS, RA, SLE, and HCs, except that of pSS was significantly lower than that of the HCs by 1.15-fold (p = 0.0484) and that of RA vs. pSS was 1.17-fold higher (p = 0.0346) (Figure 3B, right panel).

Determination of HNE-protein adducts
The level of the HNE-protein adduct can present the oxidative stress status and plays important pathogenic roles in several diseases including cancer, and neurodegenerative, chronic inflammatory, and autoimmune diseases [28].As shown in Supplementary Table 1, serum levels of the HNE-protein adduct in pSS patients were significantly higher compared to those of the HCs (1.27-fold, p = 0.0004).

Discussion
This is the first study to investigate the association between decreased serum levels of autoantibody isotypes against A1AT 50-63 and A1AT 50-63 HNE peptides and the risk of low A1AT levels in pSS patients.In the present study, two differentially expressed proteins, A1AG1 and A1AT, had 2-fold differences in depleted-albumin and IgG serum protein pools of nine pSS patients vs. nine HCs, identified in triplicate from in-solution digestion coupled to LC-MS/MS (Table 1).However, A1AG1 (1.53fold increase, AUC = 0.75) and A1AT (1.84-fold decline, AUC = 0.67) showed acceptable diagnostic values for discriminating between pSS patients and HCs according to a Western blot analysis (Figure 1C).In this study, significantly higher serum levels of HNE-protein adducts indicated increments in the oxidative stress status of pSS patients (Supplementary Table 1); these results are consistent with those of previous studies [5,6].
A1AG is an acute-phase protein, and its serum levels are elevated in response to a local inflammatory stimulus in several diseases including depression, cancer, and acquired autoimmune deficiency syndrome [25].A1AG may have anti-inflammatory and immunomodulatory properties [26].In the macrophage deactivation process, A1AG1 may act as a signaling molecule in the maintenance of tissue homeostasis and remodeling [27].Rantapaa-Dahlqvist et al. reported that pSS patients with pericarditis had significantly higher levels of A1AG than did pSS patients without pericarditis; however, no information on A1AG1's involvement in the development of pSS has been reported.In this study, serum protein levels of A1AG1 in pSS were significantly higher than those of HCs by 1.53-fold (Figure 1A).Saroha et al. indicated that altered glycosylation and expression of plasma A1AG may play a role in RA progression [28].A1AT is also an acute-phase protein that has anti-inflammatory and tissueprotective properties and is an immune regulator [12,29].Human A1AT protein levels can increase to inhibit elastase and serine-type proteinase during inflammation [13].Serum protein levels of A1AT in patients with pSS were significantly lower than those of HCs by 1.84-fold (Figure 1B); these results are consistent with previous studies [14,15].In this study, patients with pSS showed a feature of low A1AT level (Figure 1B).Thus, low serum level of A1AT is a risk factor for the development of pSS (Table 2).Further, serum levels of anti-A1AT 50-63 IgG and anti-A1AT 50-63 HNE IgA were significantly lower in pSS patients (Figure 3).However, Elshikha et al.
demonstrated that the human A1AT protein has protective effects through inhibition of dendritic cell (DC) activation and function to attenuate autoimmunity in RA mouse models [30].Ciobanu et al. indicated that significantly lower levels of A1AT in rheumatoid synovial fluid can decrease the anti-protease activity in RA [31].
Stefanescu et al. showed that levels of anti-A1AT antibodies were significantly elevated in RA [32].In several previous studies, elevated IgA-A1AT complex levels were reported in RA, SLE, mixed connective tissue disease, and ankylosing spondylitis compared to HCs [33][34][35].Further, Lacki et al. suggested that a high level of the IgA-A1AT complex may cause worsening of bone erosion in RA [35].In this study, levels of IgA-A1AT 50-63 and their HNE-modified peptide complexes did not significantly differ among patients with RA and SLE compared to HCs (Figure 3A, B, right panel).Levels of IgG-A1AT 50-63 and their HNE-modified peptide complexes were significantly higher among patients with RA and SLE compared to HCs (Figure 3A, B, left panel).Further, we observed that low levels of the anti-A1AT 50-63 IgG antibody obviously increased the risk against pSS, but the anti-A1AT 50-63 HNE IgG antibody reduced the risk against pSS (Table 2).The presence of self-reactive IgG autoantibodies in human serum is thought to represent as pathogenic antibodies in patients with pSS [29].Further, the HNE-modified epitope belong to oxidationspecific epitopes (OSEs) [30].OSEs are present on damaged proteins and induce specific autoantibodies formation [11,12].Anti-OSEs autoantibodies have conveyed protection from autoimmune pathogenesis [29,31,32].Importantly, oxidative stress remained in patients with pSS (Supplementary Table 1).League Against Rheumatism classification criteria [33] or the 1987 ACR classification criteria [34].pSS patients were diagnosed according to the American-European Consensus Group (AECG) classification criteria [35].SLE patients fulfilled the 1997 ACR SLE classification criteria [36].Differentially expressed serum proteins were identified through in-solution digestion and nano-LC-MS/MS using pooled depleted-albumin and IgG serum protein samples randomly selected from nine RA patients and nine age-matched HCs.Two differentially expressed proteins, A1AG1 and A1AT, exhibited 2-fold differences in patients with pSS compared to HCs, and these were selected to examine protein levels through Western blotting using individual serum samples randomly selected from another 40 pSS patients and 40 age-matched HCs.HNE modifications of A1AT and A1AG1 were identified by ingel digestion and nano-LC-MS/MS.HNE modifications of proteins were evaluated through immunoprecipitation (IP) and Western blotting using the aforementioned 40  (Bioinformatics Solutions, Waterloo, Canada) [38].Details are provided in "Supplementary information".were run on 10% SDS-PAGE with in-gel digestion according to a previously described method with minor modifications (Supplementary Figure 1A) [39].HNE modifications were identified in triplicate using tryptic peptide mixtures of gel slices by the aforementioned nano-LC-MS/MS (nanoAcquity system and LTQ-Orbitrap XL TM hybrid mass spectrometer).HNE-modified peptide sequences and sites of serum A1AG1 and A1AT were identified using the PeaksPTM module of the PEAKS 7 software (Bioinformatics Solutions).Details are provided in "Supplementary information".

Immunoprecipitation (IP)
An IP experiment for A1AG1 or A1AT was performed using a mouse anti-A1AG1 monoclonal antibody (sc-69753, Santa Cruz Biotechnology) or a mouse monoclonal antibody (sc-69752, Santa Cruz Biotechnology).HNE modifications of A1AG1 or A1AT were evaluated through a Western blot analysis with a goat polyclonal anti-HNE antibody (MBS536107, MyBioSource, San Diego, CA, USA).
Details are provided in "Supplementary information".

Detection of serum HNE-protein adducts
Levels of HNE-protein adducts were quantified using 168 serum samples for the ELISA protocol of Weber et al. [40].All samples were analyzed in duplicate.
Details are provided in "Supplementary Information".

Statistical analyses
Student's t-test was used to determine the significance of differences in blot densitometry, levels of serum proteins, and levels of HNE-protein adducts, and levels of autoantibody isotypes against A1AT 50-63 and A1AT 50-63 HNE peptides.GraphPad Prism (vers.5.0; Graphpad Software, San Diego, CA, USA) was used to assess differences in Student's t-tests between groups, and a dot plot was drawn.Ages and clinical test results are presented as the mean ± standard deviation (SD).Spectral count data are presented as the mean ± relative SD (RSD).The RSD is a coefficient of variation (CV) and is calculated as a percentage.Multiples of change were defined as (mean of pSS-normalized spectral counts) / (mean of HC-normalized spectral counts).The threshold for up-or downregulated proteins was a 1.0-fold change in expression.
Comparisons of pSS vs. HC serum samples were performed.Proteins that had a 2-fold difference were selected for validation by a Western blot analysis.Univariate and multiple logistic regression models were further used to estimate the adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) for the pSS risk.Power estimations were determined using SAS (vers.9.3; SAS Institute, Cary, NC, USA).
Receiver operating characteristic (ROC) curves were generated to evaluate the diagnostic performance of differentially expressed proteins using MedCalc Statistical Software (vers.15.4; MedCalc Software, Ostend, Belgium).The area under the ROC curve (AUC), sensitivity, and specificity were estimated at a 95% confidence level.
For all statistical tests, the significance level was set to p < 0.05.

Conclusions
We identified HNE modifications on the human serum A1AT protein in vivo to investigate autoantibody isotypes against A1AT 50-63 and A1AT 50-63 HNE peptides associated with pSS patients.Our results showed that low levels of the anti-A1AT 50-63     IgG antibody had an increased risk in pSS patients.However, this possibility needs to be confirmed in larger studies.

4. 2
Depleted-albumin and IgG serum proteins, in-solution digestion, and protein identification by LC-MS/MS Protein concentrations of serum were determined using a Coomassie Plus (Bradford) Assay Kit according to the manufacturer's protocol.Albumin and IgG of serum samples were removed using an Albumin and IgG Depletion SpinTrap column according to the protocol of Uen et al.[37].Three micrograms of depleted-albumin and IgG serum proteins was used to perform in-solution digestion using an In-Solution Tryptic Digestion and Guanidination Kit according to the manufacturer's instructions.Tryptic peptide mixtures were analyzed in triplicate using NanoLC-nanoESi-MS/MS that was performed on a nanoAcquity system (Waters, Milford, MA, USA) connected to an LTQ-Orbitrap XL TM hybrid mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) equipped with a nanospray interface (Proxeon, Odense, Denmark).Differentially expressed proteins were quantified using label-free peptide quantification by the Peaks Q module of the PEAKS 7 software

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 30 October 2017 doi:10.20944/preprints201710.0180.v1 pairs
of pooled serum samples.Next, autoantibody isotypes against unmodified and their HNE-modified peptides were assessed among 49 pSS, 40 RA, and 30 SLE patients, and 49 HCs.Serum was stored at -20 °C until being analyzed.Clinical and demographic characteristics of pSS, RA, and SLE patients, and HCs are presented in Supplementary Table1.However, the age of patients with SLE was significantly lower compared to those of the pSS, RA, and HC cohorts (Supplementary Table1).