第一部分

背景

IgA1铰链区O-糖基化异常在IgA肾病(IgA nephropathy, IgAN)发病机制中发挥重要作用。既往蛋白质组学研究发现IgAN患者循环IgA1铰链区O-糖修饰中N-乙酰半乳糖胺(N-acetylgalactosamine, GalNAc)和半乳糖(galactose, Gal)数目显著低于健康人。然而IgAN患者循环IgA1铰链区O-糖型特征对于IgAN的诊断效能尚不清楚。本部分研究目的是利用液相色谱串联质谱(liquid chromatography-tandem mass spectrometry, LC-MS/MS)技术解析IgAN、非IgA其他肾小球疾病对照患者(disease control, DC)和健康对照(healthy control, HC)三组间血浆IgA1铰链区O-糖型差异，并进一步评估循环IgA1铰链区O-糖基化特征的诊断效能。

方法

纳入北京协和医院肾内科住院肾活检明确诊断的原发性IgAN患者62例、其他肾小球疾病30例（10例膜性肾病、10例局灶节段性肾小球硬化和10例微小病变肾病）及30例健康对照。留取上述患者和健康对照血浆分别用KM55试剂盒检测Gd-IgA1水平，同时IgA免疫亲和珠从血浆中提取IgA，经过去N糖、还原烷基化、胰蛋白酶切及去唾液酸化处理，亲水色谱法(hydrophilic interaction liquid chromatography, HILIC)富集IgA1铰链区完整O-糖肽。液相色谱串联质谱(liquid chromatography-tandem mass spectrometry, LC-MS/MS)检测IgA1铰链区O-糖型，用Byonic软件及手工解析血浆IgA1铰链区O-糖型特征。比较三组间血浆IgA1铰链区O-糖型及糖基化特征差异，分析IgAN患者血浆IgA1铰链区O-糖基化修饰特征与临床及病理指标相关性，评估血浆IgA1铰链区O-糖基化特征GalNAc和Gal数目对IgAN的诊断效能。

结果

①122例患者中共鉴定到48种IgA1铰链区完整O-糖肽，IgAN患者血浆IgA1铰链区O-糖型与DC和HC组患者显著不同。

②三组中血浆IgA1铰链区12种主要糖型对应的O糖肽所占比例见表1.2。其中，IgAN组和DC相比，GalNAc3Gal3所占比例显著增高(p＜0.05)，而GalNAc5Gal4和GalNAc6Gal5所占比例显著降低(p＜0.05)。IgAN组和HC相比，GalNAc3Gal3和GalNAc4Gal2所占比例显著升高(p＜0.05)，而GalNAc5Gal5和GalNAc6Gal4所占比例显著降低(p＜0.05)。

③IgAN患者血浆IgA1铰链区GalNAc数目明显少于DC (IgAN vs DC：4.358±0.104 vs 4.471±0.073，p＜0.001)和HC(IgAN vs HC: 4.358±0.104 vs 4.488±0.135, p＜0.001)；IgA1铰链区Gal数目显著低于DC(IgAN vs DC: 3.591±0.126 vs 3.733±0.147)和HC(IgAN vs HC:3.591±0.126 vs 3.712±0.134, p＜0.001)。

④IgA1铰链区Gal数目与血浆Gd-IgA1水平呈负相关（r=-0.283，p=0.026），IgA1铰链区Gal数目与血浆C3水平呈正相关(r=0.335，p=0.008)。

⑤以DC和HC为对照，血浆Gd-IgA1诊断IgAN的AUC分别是0.703 (95% CI: 0.591, 0.815)和0.683 (95% CI: 0.565–0.802)。以DC和HC为对照，血浆IgA1铰链区Gal、GalNAc数目和GalNAc-Gal panel诊断IgAN的AUC显著高于Gd-IgA1水平诊断IgAN的AUC (IgAN vs DC: Gal, GalNAc, GalNAc-Gal panel: 0.793 vs. 0.805 vs. 0.852; IgAN vs HC: Gal, GalNAc, GalNAc-Gal panel: 0.764 vs. 0.809 vs. 0.844)。将血IgA水平加入GalNAc-Gal panel，以DC和HC为对照，GalNAc-Gal-IgA panel诊断IgAN的AUC分别是0.911 (95% CI: 0.854–0.967) 和0.927 (95% CI: 0.866–0.988)。

结论

①与健康对照和其他肾小球疾病相比，IgAN患者血浆IgA1铰链区GalNAc和Gal数目显著少于其他肾小球疾病（MN、MCD及FSGS）及健康人。

②IgAN患者血IgA1铰链区Gal数目与血Gd-IgA1及补体C3水平显著相关。

③血IgA1铰链区GalNAc、Gal数目联合血IgA水平（GalNAc-Gal-IgA panel）用于IgAN的诊断效能，显著优于GalNAc-Gal panel、GalNAc、Gal数目和血Gd-IgA1水平，对IgAN具有良好的诊断效能。

第二部分

背景   

IgA肾病（IgA nephropathy, IgAN）发病核心环节是循环中增高的糖基化缺陷IgA1，补体在IgAN的发病机制中起重要作用。以往采用ELISA方法对IgAN中尿液补体蛋白进行了评估，并发现与疾病严重程度具有一定关联。目前缺乏利用靶向蛋白质组学技术探索IgAN不同补体活化通路尿液蛋白表达的研究。本研究利用平行反应监测(parallel reaction monitoring, PRM)技术解析IgAN患者尿液中补体蛋白变化，并分析其与临床及病理指标的相关性。

方法

纳入北京协和医院肾内科住院肾活检明确诊断的原发性IgAN患者64例，留取患者晨尿。丙酮提取尿液蛋白，进一步还原、烷基化、胰蛋白酶切及C18固相萃取除盐处理。选取27个感兴趣的补体蛋白通过Skyline软件在混合尿液样品中进行补体多肽筛选，所有样品经过反相C18毛细管色谱柱分离，利用Orbitrap Fusion Lumos Tribrid质谱仪利用PRM方法对筛选到的补体蛋白在各个尿液样本中进行定量分析。鉴定到补体蛋白分成五类，包括经典途径（classical pathway，CP），凝集素途径（lectin pathway，LP），替代途径（alternative pathway，AP）调节蛋白，膜攻击复合物（membrane attack complex，MAC）及补体C3。将补体蛋白定量结果与IgAN患者临床及病理指标进行Spearman相关性分析。

结果

①入组64例患者，37例男性，27例女性，平均年龄40.37±12.80岁，eGFR为73.72±24.71 ml/kg/1.73m²，24h尿蛋白定量为1.86±1.06g。

②混合样品中共鉴定到30个尿液补体肽段，补体肽段定量中位CV值是0.45，82.8%补体肽段的CV值小于0.6，多数肽段定量稳定。

③所有患者尿液中鉴定到19个补体蛋白，尿液补体C5与eGFR负相关。

④尿液C3与尿蛋白正相关。补体LP中纤维蛋白2、甘露糖结合凝集素(mannose binding lectin, MBL)相关的丝氨酸蛋白酶-1均与尿蛋白负相关，C4a与尿蛋白正相关。AP中衰变加速因子(decay-accelerating factor, DAF)与尿蛋白负相关。

⑤尿液C3与MEST-C评分正相关，DAF与MEST-C评分呈负相关，MAC中C9与MEST-C评分正相关。

结论

① 尿液补体，特别是C3、C4a、MAC及DAF等与疾病严重程度有一定相关性；

② 尿液补体变化提示替代及凝集素途径激活在IgAN发病中起突出作用。

Section 1

Background

Aberrant O-glycosylation of IgA1 plays an important role in IgA nephropathy pathogenesis. Previous proteomic studies analyzed O-glycans of the circulating IgA1 hinge region and found that the N-acetylgalactosamine (GalNAc) and galactose number in the hinge region of IgA1 of patients with IgA nephropathy was lower than that in healthy participants. However, the diagnostic performance of the O-glycosylation traits in the hinge region of plasma IgA1 for IgA nephropathy remains unelucidated. The present study aimed to determine the difference in plasma IgA1 hinge region O-glycoforms among IgA nephropathy, non-IgA nephropathy disease controls(DCs), and healthy controls(HCs) via liquid chromatography tandem mass spectrometry (LC-MS/MS), and to further evaluate the diagnostic performance of plasma IgA1 glycosylation traits.

Methods

Sixty-two patients with biopsy-proven primary IgA nephropathy, 30 age- and sex-matched non-IgA nephropathy disease controls (10 patients with membranous nephropathy, 10 with focal segmental glomerulosclerosis, and 10 with minimal change disease), and 30 healthy participants were prospectively recruited. Plasma galactose deficient-IgA1 levels were measured using a KM55 kit. Plasma IgA was extracted using IgA immunoaffinity beads. After de-N-glycosylation, reduction, alkylation, trypsin digestion, and O-glycopeptide enrichment via HILIC, LC-MS/MS and Byonic software were applied to analyze the IgA1 O-glycosylation patterns and we derived the plasma IgA1 O-glycosylation traits. We conducted quantitative analysis of IgA1glycosylation patterns and O-glycosylation traits among the three groups, and correlation analysis between IgA1 O-glycosylation characteristics and clinicopathological characteristics in IgAN patients. The diagnostic performance of IgA1 O-glycosylation characteristics for IgAN was further evaluated.　　　　　　　　

Results

①We identified 48 intact IgA1 O-glycopeptides from the 122 enrolled participants. Plasma IgA1 O-glycosylation patterns were significantly changed in IgA nephropathy patients compared to those with non-IgA nephropathy disease controls and healthy participants.

②The representative mass spectrum and corresponding relative abundance of main 12 glycopeptides in IgA1 HR were shown in Table 1.2. Compared with DC, the relative abundance of GalNAc3Gal3 in the IgAN group was significantly higher, while the relative abundance of GalNAc5Gal4 and GalNAc6Gal5 was significantly lower. Compared with HC, the relative abundance of GalNAc3Gal3 and GalNAc4Gal2 in the IgAN group was significantly higher, while the relative abundance of GalNAc5Gal5 and GalNAc6Gal4 was significantly lower.

③Compared with DCs and HCs, the GalNAc number was lowest in IgA nephropathy patients (IgAN vs DC vs HC: 4.358±0.104 vs 4.471±0.073 vs 4.488±0.135, p＜0.001). In addition, a similar result was observed for the galactose number in the IgA1 hinge region(IgAN vs DC vs HC:3.591±0.126 vs 3.733±0.147 vs 3.712±0.134, p＜0.001).

④The analysis of the correlations between plasma Gd-IgA1 levels and the numbers of GalNAc and Gal in IgAN patients showed that the number of Gal was significantly negatively correlated with the plasma Gd-IgA1 level(r=-0.283，p=0.026). We firstly found that the number of Gal in the IgA1 HR was positively correlated with the level of serum C3(r=0.335，p=0.008).

⑤Compared with DC and HC, the AUC values of Gal, GalNAc, and the combination of Gal and GalNAc in diagnosing IgAN were significantly higher than those of plasma Gd-IgA1 levels (Gal, GalNAc, GalNAc-Gal panel for DC: 0.793 vs. 0.805 vs. 0.852; Gal, GalNAc, GalNAc-Gal panel for HC: 0.764 vs. 0.809 vs. 0.844). We further combined plasma IgA level with IgA1 O-glycosylation traits (combined Gal and GalNAc) to form a panel for IgAN diagnosis. The GalNAc-Gal-IgA panel had the strongest diagnostic performance among all the above five markers for distinguishing IgAN from DC and HC participants, with AUC values of 0.911 (95% CI: 0.854-0.967) the DC group and 0.927 (95% CI: 0.866-0.988) when compared with the HC group.

Conclusions

①Compared with HCs and DCs, the GalNAc and Gal number of plasma IgA1 HR in IgAN patients were significantly lower.  

②The Gal number of plasma IgA1 HR in IgAN patients correlated with plasma Gd-IgA1 and C3 level．

③The panel containing GalNAc, Gal, and circulating IgA showed excellent diagnostic performance and is promising for application in clinical practice.

Section 2

Background

The initiating factor of IgA nephropathy (IgAN) is the elevation of circulating galactose-deficient IgA1. The complement proteins play an important role in the pathogenesis of IgAN. In previous studies, urinary complement proteins (UCP) of IgAN patients were detected by ELISA and association was found between urinary UCP and disease severity. However, there was few studies to explore the UCP profiles in different pathways by targeted proteomics in IgAN patients. In this study, we analyzed UCP profiles and explored the correlation between UCP and clinicopathological parameters by parallel reaction monitoring (PRM).

Methods

Sixty-four patients with primary IgAN diagnosed by renal biopsy in our hospital were enrolled. The urine was collected from the patients in biopsy morning. The urinary proteins were extracted with acetone, and followed by reduction, alkylation, trypsin digestion and C18 extraction for desalting. A total of 27 complement proteins of interest were selected and screened by Skyline software in the pooled urine sample. Each urine sample was analyzed with a reverse-phase C18 capillary LC column. The raw MS data was acquired using Orbitrap Fusion Lumos Tribrid to further quantitatively analyze the screened complement proteins in each urine sample. The complement proteins were classified into five groups, including classical pathway (CP), lectin pathway (LP), alternative pathway (AP) complement regulatory proteins, membrane attack complex (MAC) and complement C3. Spearman correlation analysis was conducted between the complement protein quantification and clinicopathological parameters in IgAN patients. 　　　　　　　　

Results

①Sixty-four patients with biopsy-proven primary IgA nephropathy were enrolled, including 37 males and 27 females, with an average age of 40.37±12.80 years, eGFR of 73.72±24.71 mL/kg/1.73m² and 24h urinary protein excretion rate (UPER) of 1.86±1.06 g.  

②A total of 30 urinary complement peptides were identified in the pooled sample. The median CV value of the identified complement peptides was 0.48, and the CV value of the 80% complement peptide was less than 0.6, and most of the complement peptides were stable.

③Nineteen complement proteins were identified in urine samples of all IgAN patients, and urinary C5 was negatively correlated with eGFR.

④Urinary C3 was positively correlated with UPER. Ficolin-2 and mannose binding lectin (MBL) serine protease-1 of LP were negatively correlated with UPER, while C4a was positively correlated with UPER. The decay-accelerating factor (DAF) in AP was negatively correlated with UPER.  

⑤Urinary C3 was positively correlated with MEST-C total score, DAF was negatively correlated with MEST-C total score, and C9 in MAC was positively correlated with MEST-C total score.

Conclusions

① Urinary complement proteins, especially C3, C4a, MAC and DAF, were correlated with disease severity in IgAN patients.

② The changes of urinary complement proteins suggested that AP and LP activation play an important role in the pathogenesis of IgAN.