Single-cell transcriptome profiling and chromatin accessibility reveal an exhausted regulatory CD4+ T cell subset in systemic lupus erythematosus
单细胞转录组分析和染色质可及性显示系统性红斑狼疮中调节性 CD4+ T 细胞亚群耗竭
Keywords 关键字
Research topic(s) 研究课题
Introduction 介绍
Systemic lupus erythematosus (SLE) is a complex systemic autoimmune disease caused by perturbations in self-tolerance, leading to the activation of autoreactive B cell and T cell immune responses in multiple tissues (Tsokos, 2020). Compelling evidence has shown that epigenetic modifications (e.g., gene regulation that involves chromatin modifications or chromatin accessibility) of T cells are involved in SLE pathogenesis (Tsokos et al., 2016). Specifically, increased acetylation of histone H3/H4 was identified in SLE CD4+ T cells (Hu et al., 2008), and decreased methylation of costimulatory molecules (e.g., CD40LG) and interferon (IFN) signature genes (e.g., IFI44L) (Hedrich et al., 2017) in CD4+ T cells was found to correlate with the extent of inflammation and tissue damage in SLE patients (Coit et al., 2016). Therefore, understanding the impacts of global chromatin accessibility in CD4+ T cells and linking differential accessibility to SLE disease activity can deepen our understanding of SLE pathogenesis.
系统性红斑狼疮 (SLE) 是一种复杂的系统性自身免疫性疾病,由自我耐受的扰动引起,导致多个组织中自身反应性 B 细胞和 T 细胞免疫反应的激活 ( Tsokos, 2020 )。令人信服的证据表明,T 细胞的表观遗传修饰(例如,涉及染色质修饰或染色质可及性的基因调控)与 SLE 的发病机制有关 ( Tsokos et al., 2016 )。具体而言,在 SLE CD4+ T 细胞中发现组蛋白 H3/H4 乙酰化增加 ( Hu et al., 2008 ),并且发现 CD4+ T 细胞中共刺激分子(例如 CD40LG)和干扰素 (IFN) 特征基因(例如 IFI44L) ( Hedrich et al., 2017 ) 的甲基化降低与 SLE 患者的炎症和组织损伤程度相关 ( Coit et al., 2016 )。因此,了解 CD4+ T 细胞中整体染色质可及性的影响并将不同的可及性与 SLE 疾病活动联系起来,可以加深我们对 SLE 发病机制的理解。
系统性红斑狼疮 (SLE) 是一种复杂的系统性自身免疫性疾病,由自我耐受的扰动引起,导致多个组织中自身反应性 B 细胞和 T 细胞免疫反应的激活 ( Tsokos, 2020 )。令人信服的证据表明,T 细胞的表观遗传修饰(例如,涉及染色质修饰或染色质可及性的基因调控)与 SLE 的发病机制有关 ( Tsokos et al., 2016 )。具体而言,在 SLE CD4+ T 细胞中发现组蛋白 H3/H4 乙酰化增加 ( Hu et al., 2008 ),并且发现 CD4+ T 细胞中共刺激分子(例如 CD40LG)和干扰素 (IFN) 特征基因(例如 IFI44L) ( Hedrich et al., 2017 ) 的甲基化降低与 SLE 患者的炎症和组织损伤程度相关 ( Coit et al., 2016 )。因此,了解 CD4+ T 细胞中整体染色质可及性的影响并将不同的可及性与 SLE 疾病活动联系起来,可以加深我们对 SLE 发病机制的理解。
Recent studies have employed single-cell transcriptome sequencing (single-cell RNA sequencing [scRNA-seq]) technology to delineate the landscapes of immune cells in the peripheral blood and kidneys of SLE patients (Arazi et al., 2019; Nehar-Belaid et al., 2020). These foundational studies have provided an empirical basis for dissecting the cellular heterogeneity of patients’ immune systems. However, the current lack of data for specific cell populations—for example, CD4+ T cells—has thus far limited our capacity to make data-driven inferences about the specific impacts of distinct immune cell types on this autoimmune disease.
最近的研究采用单细胞转录组测序 (single-cell RNA sequencing [scRNA-seq])技术来描绘 SLE 患者外周血和肾脏中免疫细胞的景观 ( Arazi et al., 2019 ; Nehar-Belaid et al., 2020 )。这些基础研究为剖析患者免疫系统的细胞异质性提供了实证基础。然而,目前缺乏特定细胞群(例如 CD4+ T 细胞)的数据,迄今为止限制了我们对不同免疫细胞类型对这种自身免疫性疾病的特定影响进行数据驱动推断的能力。
最近的研究采用单细胞转录组测序 (single-cell RNA sequencing [scRNA-seq])技术来描绘 SLE 患者外周血和肾脏中免疫细胞的景观 ( Arazi et al., 2019 ; Nehar-Belaid et al., 2020 )。这些基础研究为剖析患者免疫系统的细胞异质性提供了实证基础。然而,目前缺乏特定细胞群(例如 CD4+ T 细胞)的数据,迄今为止限制了我们对不同免疫细胞类型对这种自身免疫性疾病的特定影响进行数据驱动推断的能力。
Regulatory CD4+ T (Treg) cells are a subset of CD4+ T cells known to function in maintaining immune tolerance and are essential for immune system homeostasis (Sakaguchi et al., 2020). Multiple studies on patients with autoimmune diseases have observed attributes of defective Treg cells (Dominguez-Villar and Hafler, 2018; Wing et al., 2019), such as increased proportions of Th1-like Treg cells, which produce the inflammatory cytokine IFN-γ (Sumida et al., 2018). Interleukin (IL)-17-producing Foxp3-expressing Treg cells were also reported by a previous study (Komatsu et al., 2014). However, there is increasing awareness that some trends observed in early studies may be misleading, owing to the introduction of potentially confounding factors related to the identification of Treg cells (based on the expression of surface markers) (Dominguez-Villar and Hafler, 2018). Single-cell profiling technologies provide an alternative method for identifying and investigating Treg cells under pathological conditions.
调节性 CD4+ T (Treg) 细胞是已知在维持免疫耐受方面发挥作用的 CD4+ T 细胞的一个亚群,对免疫系统稳态至关重要 ( Sakaguchi et al., 2020 )。对自身免疫性疾病患者的多项研究观察到有缺陷的 Treg 细胞的属性 ( Dominguez-Villar and Hafler, 2018 ; Wing et al., 2019 ),例如产生炎性细胞因子 IFN-γ 的 Th1 样 Treg 细胞的比例增加 ( Sumida et al., 2018 )。先前的研究也报道了产生白细胞介素 (IL)-17 的表达 Foxp3 的 Treg 细胞 ( Komatsu et al., 2014 )。然而,人们越来越意识到,由于引入了与 Treg 细胞鉴定相关的潜在混杂因素(基于表面标志物的表达),在早期研究中观察到的一些趋势可能具有误导性 ( Dominguez-Villar and Hafler, 2018 )。单细胞分析技术为在病理条件下识别和研究 Treg 细胞提供了一种替代方法。
调节性 CD4+ T (Treg) 细胞是已知在维持免疫耐受方面发挥作用的 CD4+ T 细胞的一个亚群,对免疫系统稳态至关重要 ( Sakaguchi et al., 2020 )。对自身免疫性疾病患者的多项研究观察到有缺陷的 Treg 细胞的属性 ( Dominguez-Villar and Hafler, 2018 ; Wing et al., 2019 ),例如产生炎性细胞因子 IFN-γ 的 Th1 样 Treg 细胞的比例增加 ( Sumida et al., 2018 )。先前的研究也报道了产生白细胞介素 (IL)-17 的表达 Foxp3 的 Treg 细胞 ( Komatsu et al., 2014 )。然而,人们越来越意识到,由于引入了与 Treg 细胞鉴定相关的潜在混杂因素(基于表面标志物的表达),在早期研究中观察到的一些趋势可能具有误导性 ( Dominguez-Villar and Hafler, 2018 )。单细胞分析技术为在病理条件下识别和研究 Treg 细胞提供了一种替代方法。
Type I IFNs were reported to be blood transcriptional signatures in SLE patients (Banchereau et al., 2016). Several studies have suggested that excess levels of type I IFNs may induce Treg dysregulation, with obvious trends to differing extents in SLE patients with different disease severities (Crow et al., 2019); in contrast, other studies have reported positive effects of IFN mediation on Treg expansion (or function) under inflammatory conditions (Lee et al., 2012). It is thus clear that the impact(s) of type I IFNs on Treg suppression remain quite controversial in autoimmune disease studies.
据报道,I 型 IFN 是 SLE 患者的血液转录特征 ( Banchereau et al., 2016 )。几项研究表明,I 型 IFN 水平过高可能会诱发 Treg 失调,在不同疾病严重程度的 SLE 患者中具有不同程度的明显趋势 ( Crow et al., 2019 );相比之下,其他研究报道了 IFN 介导对炎症条件下 Treg 扩张(或功能)的积极影响 ( Lee et al., 2012 )。因此,很明显,I 型 IFN 对 Treg 抑制的影响在自身免疫性疾病研究中仍然存在很大争议。
据报道,I 型 IFN 是 SLE 患者的血液转录特征 ( Banchereau et al., 2016 )。几项研究表明,I 型 IFN 水平过高可能会诱发 Treg 失调,在不同疾病严重程度的 SLE 患者中具有不同程度的明显趋势 ( Crow et al., 2019 );相比之下,其他研究报道了 IFN 介导对炎症条件下 Treg 扩张(或功能)的积极影响 ( Lee et al., 2012 )。因此,很明显,I 型 IFN 对 Treg 抑制的影响在自身免疫性疾病研究中仍然存在很大争议。
Here, we utilized the assay for transposase-accessible chromatin sequencing (ATAC-seq) and scRNA-seq to analyze peripheral CD4+ T cells from SLE patients and healthy controls and generated high-resolution landscapes of the epigenome and single-cell transcriptome of CD4+ T cells in vivo. We found major epigenetic regulatory patterns corresponding to SLE disease activity based on in-depth analyses of our ATAC-seq data. Additionally, single-cell transcriptome analysis enabled the identification of CD4+ T cell subtypes in SLE patients that were distinct from those of healthy controls and enabled an exploration of their potential contributions to SLE pathogenesis. Ultimately, we discovered that CCR7lowCD74hi Treg cells feature type I IFN-associated functional exhaustion in SLE patients as well as consistently dysregulated patterns at the transcriptome and chromatin accessibility. We also detected apparently similar Treg cells in CD4+ T cell datasets from patients with ulcerative colitis and multiple sclerosis.
在这里,我们利用转座酶可及染色质测序 (ATAC-seq) 和 scRNA-seq 的测定法来分析来自 SLE 患者和健康对照者的外周 CD4 + T 细胞,并生成了体内 CD4 + T 细胞 的表观基因组和单细胞转录组的高分辨率景观。根据对 ATAC-seq 数据的深入分析,我们发现了与 SLE 疾病活动相对应的主要表观遗传调控模式。此外,单细胞转录组分析能够识别 SLE 患者中不同于健康对照的 CD4 + T 细胞亚型,并能够探索它们对 SLE 发病机制的潜在贡献。最终,我们发现 CCR7低CD74hi Treg 细胞在 SLE 患者中具有 I 型 IFN 相关的功能耗竭,以及转录组和染色质可及性的持续失调模式。我们还在溃疡性结肠炎和多发性硬化症患者的 CD4+ T 细胞数据集中检测到明显相似的 Treg 细胞。
在这里,我们利用转座酶可及染色质测序 (ATAC-seq) 和 scRNA-seq 的测定法来分析来自 SLE 患者和健康对照者的外周 CD4 + T 细胞,并生成了体内 CD4 + T 细胞 的表观基因组和单细胞转录组的高分辨率景观。根据对 ATAC-seq 数据的深入分析,我们发现了与 SLE 疾病活动相对应的主要表观遗传调控模式。此外,单细胞转录组分析能够识别 SLE 患者中不同于健康对照的 CD4 + T 细胞亚型,并能够探索它们对 SLE 发病机制的潜在贡献。最终,我们发现 CCR7低CD74hi Treg 细胞在 SLE 患者中具有 I 型 IFN 相关的功能耗竭,以及转录组和染色质可及性的持续失调模式。我们还在溃疡性结肠炎和多发性硬化症患者的 CD4+ T 细胞数据集中检测到明显相似的 Treg 细胞。
Results 结果
Chromatin accessibility landscapes of CD4+ T cells from healthy controls and SLE patients
来自健康对照和 SLE 患者的 CD4+ T 细胞的染色质可及性景观
We generated 102 high-resolution ATAC-seq profiles of primary CD4+ T cells and conducted a genome-wide analysis to map the locations and profile the accessibilities of diverse regulatory elements (Figure 1A). The examined cells were freshly isolated from peripheral blood mononuclear cells (PBMCs) using fluorescence-activated cell sorting (FACS) (Figures S1A and S1B). Our dataset included a total of 65 samples from 63 SLE patients (58 of which had clinical information) and 37 samples from 25 healthy controls (Table S1). For each SLE patient, detailed medication history and comorbidities at the time of blood draw were recorded and summarized (Table S1). Specifically, the patient cohort comprised 40 individuals with new-onset disease (24 of whom were drug naive). According to the disease activity index (DAI) determined using the SLEDAI-2K method (Yee et al., 2007), 38 SLE patients were at a severe stage, with a DAI > 11 (24 of whom were freshly diagnosed). The average DAI was ∼15 in this patient cohort for bulk ATAC-seq of primary CD4+ T cells (Table S1).
我们生成了 102 个原代 CD4+ T 细胞的高分辨率 ATAC-seq 图谱,并进行了全基因组分析,以绘制不同调节元件的位置并分析其可及性(图 1A)。使用荧光激活细胞分选 (FACS) 从外周血单核细胞 (PBMC) 中新鲜分离检查的细胞 (图 S1A 和 S1B)。我们的数据集包括来自 63 名 SLE 患者的 65 个样本(其中 58 名有临床信息)和来自 25 名健康对照的 37 个样本(表 S1)。对于每位 SLE 患者,记录并总结了抽血时的详细用药史和合并症 (表 S1)。具体来说,患者队列包括 40 名新发疾病患者 (其中 24 名未接受过药物治疗)。根据使用 SLEDAI-2K 方法测定的疾病活动指数 (DAI) Yee et al., 2007 ( ),38 例 SLE 患者处于严重阶段,DAI > 11 (其中 24 例为新诊断)。在该患者队列中,原代 CD4+ T 细胞的大量 ATAC-seq 的平均 DAI 为 ∼15(表 S1)。
我们生成了 102 个原代 CD4+ T 细胞的高分辨率 ATAC-seq 图谱,并进行了全基因组分析,以绘制不同调节元件的位置并分析其可及性(图 1A)。使用荧光激活细胞分选 (FACS) 从外周血单核细胞 (PBMC) 中新鲜分离检查的细胞 (图 S1A 和 S1B)。我们的数据集包括来自 63 名 SLE 患者的 65 个样本(其中 58 名有临床信息)和来自 25 名健康对照的 37 个样本(表 S1)。对于每位 SLE 患者,记录并总结了抽血时的详细用药史和合并症 (表 S1)。具体来说,患者队列包括 40 名新发疾病患者 (其中 24 名未接受过药物治疗)。根据使用 SLEDAI-2K 方法测定的疾病活动指数 (DAI) Yee et al., 2007 ( ),38 例 SLE 患者处于严重阶段,DAI > 11 (其中 24 例为新诊断)。在该患者队列中,原代 CD4+ T 细胞的大量 ATAC-seq 的平均 DAI 为 ∼15(表 S1)。

Figure 1 Landscape of the heterogeneity of DNA accessibility in SLE peripheral CD4+ T cells
图 1SLE 外周 CD4+ T 细胞中 DNA 可及性异质性的概况
图 1SLE 外周 CD4+ T 细胞中 DNA 可及性异质性的概况
Each ATAC-seq library was sequenced to obtain an average of more than 20 million paired-end reads (Table S1). With this dataset, we identified a total of 103,317 high-quality peaks for DNA accessibility in healthy and SLE CD4+ T cells using the previously reported tool ATAC-pipe (Zuo et al., 2019); these high-quality peaks had strong signal-to-noise ratios (Figures S1C and S1D) and reproducibility (Figures S1E and S1F). Confirming the plausibility of our analysis, our ATAC-seq analysis successfully detected focal enrichments for open chromatin around the CD3D and CD4 loci (Figures S1G and S1H) but not for CD8A (Figure S1I). In addition, the number of accessible peaks approached saturation (∼100,000) when the number of samples from SLE patients reached 60, indicating that the sample size of our cohort was sufficiently large to identify most of the DNA accessible sites in CD4+ T cells from patients (Figure 1B). Thus, our ATAC-seq analysis of primary CD4+ T cells yielded a large-scale and reliable dataset of genome-wide chromatin accessibility profiles for both healthy individuals and those with SLE.
对每个 ATAC-seq 文库进行测序,以获得平均超过 2000 万个双端读数(表 S1)。借助此数据集,我们使用先前报道的工具 ATAC-pipe ( ) 在健康和 SLE CD4 + T 细胞中共鉴定了 103,317 个高质量 DNA 可及性峰 Zuo et al., 2019 ;这些高质量峰具有很强的信噪比(图 S1、C 和 S1D)和重现性(图 S1、E 和 S1F)。证实了我们分析的合理性,我们的 ATAC-seq 分析成功检测到 CD3D 和 CD4 基因座周围开放染色质的局灶性富集(图 S1G 和 S1H),但没有检测到 CD8A 的局灶性富集(图 S1I)。此外,当 SLE 患者的样本数量达到 60 时,可接近峰的数量接近饱和 (∼100,000),这表明我们队列的样本量足够大,可以识别患者 CD4+ T 细胞中的大部分 DNA 可接近位点(图 1B)。因此,我们对原代 CD4+ T 细胞的 ATAC-seq 分析产生了一个大规模且可靠的全基因组染色质可及性数据集,适用于健康个体和 SLE 患者。
对每个 ATAC-seq 文库进行测序,以获得平均超过 2000 万个双端读数(表 S1)。借助此数据集,我们使用先前报道的工具 ATAC-pipe ( ) 在健康和 SLE CD4 + T 细胞中共鉴定了 103,317 个高质量 DNA 可及性峰 Zuo et al., 2019 ;这些高质量峰具有很强的信噪比(图 S1、C 和 S1D)和重现性(图 S1、E 和 S1F)。证实了我们分析的合理性,我们的 ATAC-seq 分析成功检测到 CD3D 和 CD4 基因座周围开放染色质的局灶性富集(图 S1G 和 S1H),但没有检测到 CD8A 的局灶性富集(图 S1I)。此外,当 SLE 患者的样本数量达到 60 时,可接近峰的数量接近饱和 (∼100,000),这表明我们队列的样本量足够大,可以识别患者 CD4+ T 细胞中的大部分 DNA 可接近位点(图 1B)。因此,我们对原代 CD4+ T 细胞的 ATAC-seq 分析产生了一个大规模且可靠的全基因组染色质可及性数据集,适用于健康个体和 SLE 患者。
Chromatin accessibility signatures of SLE CD4+ T cells
SLE CD4+ T 细胞的染色质可及性特征
We next explored the chromatin accessibility signatures related to SLE disease activity. We first performed a principal-component analysis (PCA) of all the ATAC-seq samples; the healthy control samples were closely clustered together, whereas the SLE patient samples were more diffusely distributed (Figure 1C, average Euclidean distance at the PCA space: controls = 0.09 versus patients = 0.17, p < 10−88), indicating a higher extent of heterogeneity for the chromatin accessibility of primary CD4+ T cells from SLE patients (Figures S1J and S1K). We observed no clear pattern that can distinguish between drug-treated SLE patients and drug-naive SLE patients (Figure S1L). This substantial heterogeneity among the SLE patient samples apparently masked obvious differences between the SLE patients and healthy controls per se, as, when we tried to identify uniform chromatin accessibility differences in SLE versus healthy control samples, we found no lupus- or autoimmune disease-related ontologies among the 3,563 sites with differential DNA accessibility (3.4% of total peaks) (|log2 fold change| > 1, false discovery rate [FDR] <0.05, p < 0.05; Figures S1M and S1N).
接下来,我们探索了与 SLE 疾病活动相关的染色质可及性特征。我们首先对所有 ATAC-seq 样本进行了主成分分析 (PCA);健康对照样本紧密聚集在一起,而 SLE 患者样本分布更分散(图 1C,PCA 空间的平均欧几里得距离:对照组 = 0.09 vs 患者 = 0.17,p < 10-88),表明 SLE 患者原代 CD4+ T 细胞的染色质可及性异质性程度更高(图 S1J 和 S1K)。我们没有观察到可以区分药物治疗的 SLE 患者和未接受过药物治疗的 SLE 患者的明确模式(图 S1L)。SLE 患者样本之间的这种巨大异质性显然掩盖了 SLE 患者和健康对照本身之间的明显差异,因为当我们试图确定 SLE 与健康对照样本中均匀的染色质可及性差异时,我们在 3,563 个具有不同 DNA 可及性的位点(占总峰的 3.4%)中没有发现狼疮或自身免疫性疾病相关的本体(|log2 倍变化|> 1, 错误发现率 [FDR] <0.05,p < 0.05;图 S1M 和 S1N)。
接下来,我们探索了与 SLE 疾病活动相关的染色质可及性特征。我们首先对所有 ATAC-seq 样本进行了主成分分析 (PCA);健康对照样本紧密聚集在一起,而 SLE 患者样本分布更分散(图 1C,PCA 空间的平均欧几里得距离:对照组 = 0.09 vs 患者 = 0.17,p < 10-88),表明 SLE 患者原代 CD4+ T 细胞的染色质可及性异质性程度更高(图 S1J 和 S1K)。我们没有观察到可以区分药物治疗的 SLE 患者和未接受过药物治疗的 SLE 患者的明确模式(图 S1L)。SLE 患者样本之间的这种巨大异质性显然掩盖了 SLE 患者和健康对照本身之间的明显差异,因为当我们试图确定 SLE 与健康对照样本中均匀的染色质可及性差异时,我们在 3,563 个具有不同 DNA 可及性的位点(占总峰的 3.4%)中没有发现狼疮或自身免疫性疾病相关的本体(|log2 倍变化|> 1, 错误发现率 [FDR] <0.05,p < 0.05;图 S1M 和 S1N)。
We therefore adopted an alternative approach to help identify informative chromatin accessibility signatures related to SLE disease activity. We extracted the most variable chromatin-accessible sites among all patients (coefficient of variation >0.5) and performed a PCA of the patient samples based on the intensity values for these peaks (STAR methods). The patient samples were clustered into three distinct groups (Figures 1D and 1E). Because these three groupings were supported by the detection of strong correlations with their respective patient disease activity (DA) scores (Figure 1F), we defined these patient groups as DAlow, DAint, and DAhi.
因此,我们采用了另一种方法来帮助识别与 SLE 疾病活动相关的信息性染色质可及性特征。我们提取了所有患者中染色质可及性最可变的位点 (变异系数 >0.5),并根据这些峰的强度值对患者样本进行了 PCA (STAR 方法)。患者样本分为三个不同的组(图 1、D 和 1E)。由于这三个分组得到了与各自患者疾病活动 (DA) 评分的强相关性检测的支持(图 1F),因此我们将这些患者组定义为 DA低、DAint 和 DAhi。
因此,我们采用了另一种方法来帮助识别与 SLE 疾病活动相关的信息性染色质可及性特征。我们提取了所有患者中染色质可及性最可变的位点 (变异系数 >0.5),并根据这些峰的强度值对患者样本进行了 PCA (STAR 方法)。患者样本分为三个不同的组(图 1、D 和 1E)。由于这三个分组得到了与各自患者疾病活动 (DA) 评分的强相关性检测的支持(图 1F),因此我们将这些患者组定义为 DA低、DAint 和 DAhi。
We also assessed potential impacts of a set of comorbidities in healthy controls and in DAlow, DAint, and DAhi patients (Figure S2A). We found that the DAhi SLE patients showed severe disease activity with high proportions for most of the comorbidities, such as hematuria, leukopenia, pericarditis, and pyuria (Figure S2A). The DAint SLE patients showed moderate disease activity, and the DAlow SLE patients had mild disease activity (Figure 1F). These results demonstrated that the three groupings classified based on the divergence of the chromatin accessibility of peripheral CD4+ T cells were closely related to the clinical severity of SLE.
我们还评估了一组合并症对健康对照和 DAlow、DAint 和 DAhi 患者的潜在影响(图 S2A)。我们发现 DAhi SLE 患者表现出严重的疾病活动度,大多数合并症的比例很高,例如血尿、白细胞减少、心包炎和脓尿(图 S2A)。DAint SLE 患者表现出中度疾病活动度,DA低 SLE 患者表现出轻度疾病活动度(图 1F)。这些结果表明,根据外周 CD4+ T 细胞染色质可及性差异分类的 3 组与 SLE 的临床严重程度密切相关。
我们还评估了一组合并症对健康对照和 DAlow、DAint 和 DAhi 患者的潜在影响(图 S2A)。我们发现 DAhi SLE 患者表现出严重的疾病活动度,大多数合并症的比例很高,例如血尿、白细胞减少、心包炎和脓尿(图 S2A)。DAint SLE 患者表现出中度疾病活动度,DA低 SLE 患者表现出轻度疾病活动度(图 1F)。这些结果表明,根据外周 CD4+ T 细胞染色质可及性差异分类的 3 组与 SLE 的临床严重程度密切相关。
Next, we performed a pairwise differential analysis of each patient group and the healthy controls (STAR methods), thereby identifying a total of 11,774 DNA elements across the genome with significant differential chromatin accessibility (Figure 2A). Unsupervised hierarchical clustering grouped these differentially accessible peaks into five distinct clusters of regulatory elements (Figure 2A). Clusters I–III comprised chromatin regions that were more accessible in primary CD4+ T cells from healthy controls, DAhi and DAint SLE patients, and DAint and DAlow SLE patients, respectively. Cluster IV consisted of a small number of peaks that could distinguish the SLE patients from the healthy controls. In cluster V, the CD4+ T cells from DAhi SLE patients contained fewer accessible peaks than those from the other patient groups.
接下来,我们对每个患者组和健康对照(STAR 方法)进行了成对差异分析,从而确定了整个基因组中总共 11,774 个 DNA 元件,这些元件具有显着的差异染色质可及性(图 2A)。无监督分层聚类将这些差异可及的峰分为五个不同的调节元件集群(图 2A)。簇 I-III 分别由健康对照、DAhi 和 DAint SLE 患者以及 DAint 和 DA低 SLE 患者的原代 CD4+ T 细胞中更容易接近的染色质区域组成。聚类 IV 由少量峰组成,可以区分 SLE 患者和健康对照者。在第 V 组,来自 DAhi SLE 患者的 CD4 + T 细胞包含的可接近峰少于来自其他患者组的细胞。
接下来,我们对每个患者组和健康对照(STAR 方法)进行了成对差异分析,从而确定了整个基因组中总共 11,774 个 DNA 元件,这些元件具有显着的差异染色质可及性(图 2A)。无监督分层聚类将这些差异可及的峰分为五个不同的调节元件集群(图 2A)。簇 I-III 分别由健康对照、DAhi 和 DAint SLE 患者以及 DAint 和 DA低 SLE 患者的原代 CD4+ T 细胞中更容易接近的染色质区域组成。聚类 IV 由少量峰组成,可以区分 SLE 患者和健康对照者。在第 V 组,来自 DAhi SLE 患者的 CD4 + T 细胞包含的可接近峰少于来自其他患者组的细胞。
To explore the relationship(s) between the distinct chromatin accessibilities of primary CD4+ T cells and the clinical statuses of the patients, we performed a correlation analysis of the average intensity value of each peak cluster versus the DA of the SLE patients. We observed that the accessible peaks from cluster I showed a trend toward a positive correlation with the SLEDAI score (p = 0.057; Figure 2B). Notably, we found that the chromatin accessibility intensity of peaks in cluster II, but no other clusters, was significantly positively correlated with SLE DA (p = 0.00017, R = 0.47; Figure 2B). In contrast, the chromatin accessibility intensity of peaks in cluster III was significantly negatively correlated with SLE DA (p = 0.0095, R = −0.42; Figure 2B).
为了探讨原代 CD4+ T 细胞的不同染色质可及性与患者临床状态之间的关系,我们对每个峰簇的平均强度值与 SLE 患者的 DA 进行了相关性分析。我们观察到,来自聚类 I 的可访问峰显示出与 SLEDAI 评分呈正相关的趋势 (p = 0.057;图 2值得注意的是,我们发现簇 II 中峰的染色质可及性强度,但没有其他簇,与 SLE DA 显著呈正相关 (p = 0.00017,R = 0.47; 图 2相比之下,簇 III 中峰的染色质可及性强度与 SLE DA 显著负相关 (p = 0.0095,R = -0.42 ;图 2B).
为了探讨原代 CD4+ T 细胞的不同染色质可及性与患者临床状态之间的关系,我们对每个峰簇的平均强度值与 SLE 患者的 DA 进行了相关性分析。我们观察到,来自聚类 I 的可访问峰显示出与 SLEDAI 评分呈正相关的趋势 (p = 0.057;图 2值得注意的是,我们发现簇 II 中峰的染色质可及性强度,但没有其他簇,与 SLE DA 显著呈正相关 (p = 0.00017,R = 0.47; 图 2相比之下,簇 III 中峰的染色质可及性强度与 SLE DA 显著负相关 (p = 0.0095,R = -0.42 ;图 2B).
We next investigated the potential biological functions of cluster II peaks; as we were still aiming to identify chromatin accessibility signatures related to SLE, we used genomic regions enrichment of annotations tool (GREAT) (McLean et al., 2010) for functional annotation and conducted an enrichment analysis of the peak-associated genes from our analysis against SLE signature gene sets reported from previous SLE studies (Azizi et al., 2018; Hutcheson et al., 2008; Lowther et al., 2016; Smillie et al., 2019; Takeshita et al., 2015) (Table S2; STAR methods). Many genes known to regulate SLE tissue injury (Goropevsek et al., 2017), including STAT1, STAT4, and IFNAR—together with signature genes of Th17 cells, which are known as major contributors to SLE (Goropevsek et al., 2017), such as IL17F, RORA, and IL6R—were associated with cluster II peaks (Figure 2A). We also found that the cluster II peaks associated with the aforementioned genes had elevated chromatin accessibilities in DAhi severe-stage SLE patients compared with DAint and DAlow patients (Figures 2C, 2D, and S2B–S2D).
接下来,我们研究了簇 II 峰的潜在生物学功能;由于我们仍然致力于识别与 SLE 相关的染色质可及性特征,我们使用基因组区域注释富集工具 (GREAT) ( McLean et al., 2010 ) 进行功能注释,并针对先前 SLE 研究中报告的 SLE 特征基因集对我们分析中的峰值相关基因进行了富集分析。 Azizi et al., 2018 Hutcheson et al., 2008 ; Lowther et al., 2016 ; Smillie et al., 2019 ; Takeshita et al., 2015 )(表 S2;STAR 方法)。许多已知调节 SLE 组织损伤的基因 ( Goropevsek et al., 2017 ),包括 STAT1、STAT4 和 IFNAR,以及 Th17 细胞的标志性基因,这些基因被称为 SLE 的主要贡献者 ( Goropevsek et al., 2017 ),如 IL17F、RORA 和 IL6R,与簇 II 峰相关(图 2A)。我们还发现,与 DA int 和 DA低患者相比,与上述基因相关的 cluster II 峰在 DAhi 重度 SLE 患者中的染色质可及性升高(图 2C、2D 和 S2B-S2D)。
接下来,我们研究了簇 II 峰的潜在生物学功能;由于我们仍然致力于识别与 SLE 相关的染色质可及性特征,我们使用基因组区域注释富集工具 (GREAT) ( McLean et al., 2010 ) 进行功能注释,并针对先前 SLE 研究中报告的 SLE 特征基因集对我们分析中的峰值相关基因进行了富集分析。 Azizi et al., 2018 Hutcheson et al., 2008 ; Lowther et al., 2016 ; Smillie et al., 2019 ; Takeshita et al., 2015 )(表 S2;STAR 方法)。许多已知调节 SLE 组织损伤的基因 ( Goropevsek et al., 2017 ),包括 STAT1、STAT4 和 IFNAR,以及 Th17 细胞的标志性基因,这些基因被称为 SLE 的主要贡献者 ( Goropevsek et al., 2017 ),如 IL17F、RORA 和 IL6R,与簇 II 峰相关(图 2A)。我们还发现,与 DA int 和 DA低患者相比,与上述基因相关的 cluster II 峰在 DAhi 重度 SLE 患者中的染色质可及性升高(图 2C、2D 和 S2B-S2D)。
Furthermore, gene set enrichment and disease ontology analyses (STAR methods) showed that genes associated with cluster II peaks had strong enrichment for predicted functions related to the “inflammation” (p < 10−5) and “IFN response” (p < 10−2) terms. There was also enrichment for genes known to contribute to autoimmune diseases, including SLE (p < 10−5), rheumatoid arthritis (p < 10−5), and type 1 diabetes (p < 10−5) (Figures 2E and S2E). Interestingly, genes involved in T cell exhaustion and Treg cell exhaustion (e.g., PDCD1 and TIGIT) (Lowther et al., 2016; Yang et al., 2017) were highly enriched for cluster II peaks (p < 10−2), which were more accessible in DAhi and DAint SLE patients (Figures 2E and S2F). These results suggested that the cluster II peaks represent an SLE disease chromatin signature associated with abnormal inflammation and immune responses in primary CD4+ T cells. Increased accessibility of cluster II peaks may induce the dysfunction of CD4+ T cell subtypes, such as Th17 cells and Treg cells, and promote the aggravation of SLE.
此外,基因集富集和疾病本体分析(STAR 方法)表明,与簇 II 峰相关的基因对与“炎症”(p < 10-5)和“IFN 反应”(p < 10-2)术语相关的预测功能具有很强的富集。已知导致自身免疫性疾病的基因也富集,包括 SLE (p < 10-5)、类风湿性关节炎 (p < 10-5) 和 1 型糖尿病 (p < 10-5)(图 2E 和 S2E)。有趣的是,参与 T 细胞耗竭和 Treg 细胞耗竭的基因 (例如,PDCD1 和 TIGIT) ( Lowther et al., 2016 ; Yang et al., 2017 ) 高度富集了簇 II 峰 (p < 10-2),这在 DAhi 和 DAint SLE 患者中更容易获得(图 2E 和 S2F)。这些结果表明,簇 II 峰代表与原代 CD4+ T 细胞中的异常炎症和免疫反应相关的 SLE 疾病染色质特征。簇 II 峰的可及性增加可能诱导 CD4+ T 细胞亚型(如 Th17 细胞和 Treg 细胞)功能障碍,并促进 SLE 的加重。
此外,基因集富集和疾病本体分析(STAR 方法)表明,与簇 II 峰相关的基因对与“炎症”(p < 10-5)和“IFN 反应”(p < 10-2)术语相关的预测功能具有很强的富集。已知导致自身免疫性疾病的基因也富集,包括 SLE (p < 10-5)、类风湿性关节炎 (p < 10-5) 和 1 型糖尿病 (p < 10-5)(图 2E 和 S2E)。有趣的是,参与 T 细胞耗竭和 Treg 细胞耗竭的基因 (例如,PDCD1 和 TIGIT) ( Lowther et al., 2016 ; Yang et al., 2017 ) 高度富集了簇 II 峰 (p < 10-2),这在 DAhi 和 DAint SLE 患者中更容易获得(图 2E 和 S2F)。这些结果表明,簇 II 峰代表与原代 CD4+ T 细胞中的异常炎症和免疫反应相关的 SLE 疾病染色质特征。簇 II 峰的可及性增加可能诱导 CD4+ T 细胞亚型(如 Th17 细胞和 Treg 细胞)功能障碍,并促进 SLE 的加重。
Single-cell atlas of CD4+ T cells from severe-stage SLE patients
来自重度 SLE 患者的 CD4+ T 细胞单细胞图谱
CD4+ T cells are heterogeneous and comprise multiple subtypes, including naive, memory, effector, and Treg populations. Effector CD4+ T cells can be further divided into cytokine-polarized T helper (Th) 1, Th2, and Th17 cells (Sallusto, 2016). Previous studies on blood from SLE patients have reported a decreased proportion of naive T (Tn) cells and an increased proportion of effector T (Teff) cells (Suarez-Fueyo et al., 2016). The data for altered Treg proportions in blood from SLE patients are inconclusive, and this idea remains controversial (Scheinecker et al., 2020). Chromatin accessibility analysis of our bulk ATAC-seq data indicated that particular CD4+ T cell subtypes may be involved in SLE pathogenesis (Figures 2A and 2B). We therefore used the 10X platform to perform scRNA-seq analysis of the primary CD4+ T cells from six healthy control individuals and 10 severe-stage SLE patients (average DAI of ∼22), a different cohort from that in the bulk ATAC-seq analysis (Figure 1A; Table S1).
CD4+ T 细胞具有异质性,包含多种亚型,包括初始、记忆、效应和 Treg 细胞群。效应 CD4+ T 细胞可进一步分为细胞因子极化辅助性 T 细胞 (Th) 1、Th2 和 Th17 细胞 ( Sallusto, 2016 )。先前对 SLE 患者血液的研究报告称,初始 T (Tn) 细胞的比例降低,效应 T (Teff) 细胞的比例增加 ( Suarez-Fueyo et al., 2016 )。SLE 患者血液中 Treg 比例改变的数据尚无定论,这一想法仍然存在争议 ( Scheinecker et al., 2020 )。我们的大量 ATAC-seq 数据的染色质可及性分析表明,特定的 CD4+ T 细胞亚型可能参与 SLE 发病机制(图 2A 和 2B)。因此,我们使用 10X 平台对来自 6 名健康对照个体和 10 名重度 SLE 患者(平均 DAI 为 ∼22)的原代 CD4+ T 细胞进行 scRNA-seq 分析,与批量 ATAC-seq 分析中的队列不同(图 1A;表 S1)。
CD4+ T 细胞具有异质性,包含多种亚型,包括初始、记忆、效应和 Treg 细胞群。效应 CD4+ T 细胞可进一步分为细胞因子极化辅助性 T 细胞 (Th) 1、Th2 和 Th17 细胞 ( Sallusto, 2016 )。先前对 SLE 患者血液的研究报告称,初始 T (Tn) 细胞的比例降低,效应 T (Teff) 细胞的比例增加 ( Suarez-Fueyo et al., 2016 )。SLE 患者血液中 Treg 比例改变的数据尚无定论,这一想法仍然存在争议 ( Scheinecker et al., 2020 )。我们的大量 ATAC-seq 数据的染色质可及性分析表明,特定的 CD4+ T 细胞亚型可能参与 SLE 发病机制(图 2A 和 2B)。因此,我们使用 10X 平台对来自 6 名健康对照个体和 10 名重度 SLE 患者(平均 DAI 为 ∼22)的原代 CD4+ T 细胞进行 scRNA-seq 分析,与批量 ATAC-seq 分析中的队列不同(图 1A;表 S1)。
After quality control (QC) filtering (Figures S3A–S3D), we obtained a total of 34,176 high-quality single cells, of which 23,464 were from patients and 10,712 were from healthy controls. We then applied Seurat (Stuart et al., 2019) to integrate the cells from SLE patients and healthy controls and identified 16 clusters of CD4+ T cell subtypes (Figures 3A and 3B). We also used Harmony (Korsunsky et al., 2019) to assess the accuracy and robustness of our single-cell clustering results and found strong correlations of the identified CD4+ T cell clusters and gene expression patterns between the two integration methods (Figures S3E–S3I). These results support that we obtained a reliable and high-quality atlas of single-cell transcriptomes of CD4+ T cells from severe-stage SLE patients and healthy controls.
经过质量控制 (QC) 过滤(图 S3A-S3D),我们总共获得了 34,176 个高质量单细胞,其中 23,464 个来自患者,10,712 个来自健康对照。然后,我们应用 Seurat ( Stuart et al., 2019 ) 整合来自 SLE 患者和健康对照者的细胞,并确定了 16 簇 CD4+ T 细胞亚型(图 3A 和 3B)。我们还使用 Harmony ( Korsunsky et al., 2019 ) 来评估单细胞聚类结果的准确性和稳健性,并发现两种整合方法之间鉴定的 CD4 + T 细胞簇和基因表达模式具有很强的相关性(图 S3E-S3I)。这些结果支持我们从重度 SLE 患者和健康对照者那里获得了可靠且高质量的 CD4+ T 细胞单细胞转录组图谱。
经过质量控制 (QC) 过滤(图 S3A-S3D),我们总共获得了 34,176 个高质量单细胞,其中 23,464 个来自患者,10,712 个来自健康对照。然后,我们应用 Seurat ( Stuart et al., 2019 ) 整合来自 SLE 患者和健康对照者的细胞,并确定了 16 簇 CD4+ T 细胞亚型(图 3A 和 3B)。我们还使用 Harmony ( Korsunsky et al., 2019 ) 来评估单细胞聚类结果的准确性和稳健性,并发现两种整合方法之间鉴定的 CD4 + T 细胞簇和基因表达模式具有很强的相关性(图 S3E-S3I)。这些结果支持我们从重度 SLE 患者和健康对照者那里获得了可靠且高质量的 CD4+ T 细胞单细胞转录组图谱。

Figure 3 Single-cell atlas of peripheral CD4+ T cells from normal controls and SLE patients
图 3正常对照和 SLE 患者外周 CD4+ T 细胞的单细胞图谱
图 3正常对照和 SLE 患者外周 CD4+ T 细胞的单细胞图谱
Next, we analyzed known marker genes for particular CD4+ T cell subtypes (STAR methods) and evaluated their expression profiles within each of the 16 cell clusters to identify the cell types in each cluster (Figure 3C). We identified seven major cell subtypes, namely CCR7+ Tn, CXCR5+ T, TBX21+ Th1, GATA3+ Th2, RORC+ Th17, FOXP3+ Treg, and GNLY+ Tct cells (Figure 3D). Although the other clusters each fitted clearly within one of the seven major cell subtypes, the cells of cluster 15 appeared to be in a mixed state among the known CD4+ T cell subtypes and were defined as “Tx”, while the cells of cluster 16 exhibited a small number of genes and were named “undefined” (Figure S3J). Further, we obtained signature gene sets for multiple known CD4+ T cell subtypes from previously reported studies (Azizi et al., 2018; Hutcheson et al., 2008; Lowther et al., 2016; Smillie et al., 2019; Takeshita et al., 2015) and found remarkable enrichment in the corresponding cell subtypes we classified through our scRNA-seq analysis (Figure 3E), confirming our cell subtype identification.
接下来,我们分析了特定 CD4 + T 细胞亚型的已知标记基因(STAR 方法),并评估了它们在 16 个细胞簇中每个细胞簇中的表达谱,以鉴定每个簇中的细胞类型(图 3C)。我们确定了七种主要细胞亚型,即 CCR7 + Tn、CXCR5 + T、TBX21 + Th1、GATA3 + Th2、RORC+ Th17、FOXP3 + Treg 和 GNLY+ Tct 细胞(图 3D)。尽管其他簇都明显符合七种主要细胞亚型之一,但簇 15 的细胞在已知的 CD4+ T 细胞亚型中似乎处于混合状态并被定义为“Tx”,而簇 16 的细胞表现出少量基因并被命名为“未定义”(图 S3J)。此外,我们从先前报道的研究中获得了多种已知 CD4 + T 细胞亚型的特征基因集 ( Azizi et al., 2018 ; Hutcheson et al., 2008 ; Lowther et al., 2016 ; Smillie et al., 2019 ; Takeshita et al., 2015 ),并在我们通过 scRNA-seq 分析分类的相应细胞亚型中发现了显著的富集(图 3E),证实了我们的细胞亚型鉴定。
接下来,我们分析了特定 CD4 + T 细胞亚型的已知标记基因(STAR 方法),并评估了它们在 16 个细胞簇中每个细胞簇中的表达谱,以鉴定每个簇中的细胞类型(图 3C)。我们确定了七种主要细胞亚型,即 CCR7 + Tn、CXCR5 + T、TBX21 + Th1、GATA3 + Th2、RORC+ Th17、FOXP3 + Treg 和 GNLY+ Tct 细胞(图 3D)。尽管其他簇都明显符合七种主要细胞亚型之一,但簇 15 的细胞在已知的 CD4+ T 细胞亚型中似乎处于混合状态并被定义为“Tx”,而簇 16 的细胞表现出少量基因并被命名为“未定义”(图 S3J)。此外,我们从先前报道的研究中获得了多种已知 CD4 + T 细胞亚型的特征基因集 ( Azizi et al., 2018 ; Hutcheson et al., 2008 ; Lowther et al., 2016 ; Smillie et al., 2019 ; Takeshita et al., 2015 ),并在我们通过 scRNA-seq 分析分类的相应细胞亚型中发现了显著的富集(图 3E),证实了我们的细胞亚型鉴定。
We then explored the distributions of the seven major cell subtypes in samples from SLE patients and healthy controls. Multiple studies on PBMCs from SLE patients have reported a decrease in the proportion of naive CD4+ T cells and an increase in effector CD4+ T cells compared with cells from healthy controls (Suarez-Fueyo et al., 2016). We also observed decreases in the proportions of CCR7+ Tn and CXCR5+ T cells in SLE patients and noted increases in the proportions of multiple Teff cell subtypes, including Treg cells and cytotoxic CD4+ T cells (Tct), by comparing the proportion of each cell type directly (Figure 3F).
然后,我们探讨了 SLE 患者和健康对照者样本中 7 种主要细胞亚型的分布。关于 SLE 患者 PBMC 的多项研究报告称,与健康对照细胞相比,初始 CD4+ T 细胞的比例降低,效应 CD4+ T 细胞增加 ( Suarez-Fueyo et al., 2016 )。我们还观察到 SLE 患者 CCR7 + Tn 和 CXCR5 + T 细胞比例的降低,并通过直接比较每种细胞类型的比例,注意到多种 Teff 细胞亚型的比例增加,包括 Treg 细胞和细胞毒性 CD4 + T 细胞 (Tct)(图 3F)。
然后,我们探讨了 SLE 患者和健康对照者样本中 7 种主要细胞亚型的分布。关于 SLE 患者 PBMC 的多项研究报告称,与健康对照细胞相比,初始 CD4+ T 细胞的比例降低,效应 CD4+ T 细胞增加 ( Suarez-Fueyo et al., 2016 )。我们还观察到 SLE 患者 CCR7 + Tn 和 CXCR5 + T 细胞比例的降低,并通过直接比较每种细胞类型的比例,注意到多种 Teff 细胞亚型的比例增加,包括 Treg 细胞和细胞毒性 CD4 + T 细胞 (Tct)(图 3F)。
We next applied Demuxlet (Kang et al., 2018) and Souporcell (Heaton et al., 2020) to deconvolute the cell sources and identified the sources of 15,301 (45%) and 31,752 (93%) sequenced single cells, respectively (STAR methods). We then calculated and compared the proportions of each major cell subtype in SLE patients with those in healthy controls and found that Treg cells were significantly increased in SLE patients (p < 0.05) (Figures 3G and S3K). Pursuing this, we used flow cytometry to measure the distributions of the seven major cell subtypes in CD4+ T cells from the PBMCs of an independent cohort of 27 severe-stage SLE patients and 15 healthy controls (Figure S3L; Table S1) and again noted a remarkable and significant decrease in the proportion of CCR7+ Tn cells in the SLE patients (2.44-fold, p = 0.0001) as well as significant increases in the proportions of Treg cells (1.82-fold, p = 0.005) and Tct cells (1.76-fold, p = 0.02) (Figure 3H).
接下来,我们应用 Demuxlet ( Kang et al., 2018 ) 和 Souporcell ( Heaton et al., 2020 ) 对细胞来源进行去卷积,并分别确定了 15,301 (45%) 和 31,752 (93%) 测序的单细胞的来源(STAR 方法)。然后,我们计算并比较了 SLE 患者与健康对照者中每种主要细胞亚型的比例,发现 SLE 患者的 Treg 细胞显着增加 (p < 0.05)(图 3G 和 S3K)。为此,我们使用流式细胞术测量了 27 名重度 SLE 患者和 15 名健康对照者的独立队列的 PBMC 中 CD4 + T 细胞中 7 种主要细胞亚型的分布(图 S3L;表 S1)并再次注意到 SLE 患者 CCR7 + Tn 细胞比例显着降低 (2.44 倍,p = 0.0001) 以及 Treg 细胞比例 (1.82 倍,p = 0.005) 和 Tct 细胞 (1.76 倍,p = 0.02) 的显着增加(图 3H)。
接下来,我们应用 Demuxlet ( Kang et al., 2018 ) 和 Souporcell ( Heaton et al., 2020 ) 对细胞来源进行去卷积,并分别确定了 15,301 (45%) 和 31,752 (93%) 测序的单细胞的来源(STAR 方法)。然后,我们计算并比较了 SLE 患者与健康对照者中每种主要细胞亚型的比例,发现 SLE 患者的 Treg 细胞显着增加 (p < 0.05)(图 3G 和 S3K)。为此,我们使用流式细胞术测量了 27 名重度 SLE 患者和 15 名健康对照者的独立队列的 PBMC 中 CD4 + T 细胞中 7 种主要细胞亚型的分布(图 S3L;表 S1)并再次注意到 SLE 患者 CCR7 + Tn 细胞比例显着降低 (2.44 倍,p = 0.0001) 以及 Treg 细胞比例 (1.82 倍,p = 0.005) 和 Tct 细胞 (1.76 倍,p = 0.02) 的显着增加(图 3H)。
Divergent changes in the single-cell transcriptomes of CD4+ T cell subtypes from severe-stage SLE patients
重度 SLE 患者 CD4+ T 细胞亚型单细胞转录组的不同变化
To explore the functional divergence of each CD4+ T cell subtype in the disease state, we performed pairwise comparisons of the CD4+ T cell subtypes from SLE patients and the corresponding cell subtypes from healthy controls, which identified a total of 1,331 differentially expressed genes (DEGs) (Figures 4A and 4B; Table S3; STAR methods). Among these DEGs, 197 (28.67%) and 183 (28.42%) genes were up- and downregulated, respectively, in more than five cell subtypes (defined as “shared DEGs”), while 259 (37.70%) and 277 (43.01%) genes were up- and downregulated, respectively, in fewer than three cell subtypes (“specific DEGs”) (Figures 4A, 4B, and S4A–S4C). When accounting for the number of specific DEGs for each of the seven major cell subtypes, there were clearly more specific DEGs that were either up- or downregulated in Treg, Th2, Th17, and Tct cells than in the other cell subtypes (i.e., naive CD4+ T cells, CXCR5+ T cells, Th1 cells) (Figures 4C, S4D, and S4E). Both Euclidean distance and Spearman correlation analyses indicated that the Treg cells showed the largest differences in gene expression between SLE patients and healthy controls (Figure 4D). These results indicate that Treg, Tct, Th17, and Th2 cells dominated the transcriptomic changes in CD4+ T cells from severe-stage SLE patients.
为了探索每种 CD4 + T 细胞亚型在疾病状态下的功能差异,我们对 SLE 患者的 CD4 + T 细胞亚型和健康对照的相应细胞亚型进行了成对比较,共鉴定了 1,331 个差异表达基因 (DEG)(图 4A 和 4B;表 S3;STAR 方法)。在这些 DEGs 中,197 个 (28.67%) 和 183 个 (28.42%) 基因分别在超过 5 个细胞亚型 (定义为“共享 DEGs”) 中上调和下调,而 259 个 (37.70%) 和 277 个 (43.01%) 基因在少于 3 个细胞亚型 (“特异性 DEG”) 中分别上调和下调(图 4A、4B 和 S4A-S4C)。当考虑七种主要细胞亚型中每一种的特异性 DEG 数量时,Treg、Th2、Th17 和 Tct 细胞中明显比其他细胞亚型(即初始 CD4 + T 细胞、CXCR5 + T 细胞、Th1 细胞)中上调或下调的特异性 DEG 更多(图 4C、S4D、 和 S4E)。欧几里得距离和 Spearman 相关分析都表明,Treg 细胞在 SLE 患者和健康对照者之间的基因表达差异最大(图 4D)。这些结果表明,Treg 、 Tct 、 Th17 和 Th2 细胞主导了重度 SLE 患者 CD4+ T 细胞的转录组变化。
为了探索每种 CD4 + T 细胞亚型在疾病状态下的功能差异,我们对 SLE 患者的 CD4 + T 细胞亚型和健康对照的相应细胞亚型进行了成对比较,共鉴定了 1,331 个差异表达基因 (DEG)(图 4A 和 4B;表 S3;STAR 方法)。在这些 DEGs 中,197 个 (28.67%) 和 183 个 (28.42%) 基因分别在超过 5 个细胞亚型 (定义为“共享 DEGs”) 中上调和下调,而 259 个 (37.70%) 和 277 个 (43.01%) 基因在少于 3 个细胞亚型 (“特异性 DEG”) 中分别上调和下调(图 4A、4B 和 S4A-S4C)。当考虑七种主要细胞亚型中每一种的特异性 DEG 数量时,Treg、Th2、Th17 和 Tct 细胞中明显比其他细胞亚型(即初始 CD4 + T 细胞、CXCR5 + T 细胞、Th1 细胞)中上调或下调的特异性 DEG 更多(图 4C、S4D、 和 S4E)。欧几里得距离和 Spearman 相关分析都表明,Treg 细胞在 SLE 患者和健康对照者之间的基因表达差异最大(图 4D)。这些结果表明,Treg 、 Tct 、 Th17 和 Th2 细胞主导了重度 SLE 患者 CD4+ T 细胞的转录组变化。

Figure 4 Compositional and functional differences between normal and SLE CD4+ T subtypes
图 4正常亚型和 SLE CD4+ T 亚型的组成和功能差异
图 4正常亚型和 SLE CD4+ T 亚型的组成和功能差异
Given that nuclear factor κB (NF-κB) signaling pathway genes are known to promote the inflammatory development and progression of SLE (Brightbill et al., 2018), we consistently observed strong expression of these genes, including NFKB1, NFKB2, and REL, in Th17 cells from severe-stage SLE patients (Figure S5A). In addition, strong expression of the effector molecule CCL20 was reported to mediate the Th17-driven inflammatory process and to be associated with SLE pathogenesis (Koga et al., 2016), and this gene was expressed at significantly higher levels in SLE Th17 cells than in healthy Th17 cells (Figure S5B; p = 1.40e−110, Mann-Whitney U test). SLE Tct cells showed obviously increased levels of a series of cytotoxic molecules, including GZMH, NKG7, and GZMB, supporting previous claims about their potential involvement in exacerbating autoimmune diseases (Takeuchi and Saito, 2017) (Figure S5C).
鉴于已知核因子 κB (NF-κB) 信号通路基因可促进 SLE 的炎症发展和进展 ( Brightbill et al., 2018 ),我们始终观察到这些基因(包括 NFKB1、NFKB2 和 REL)在重度 SLE 患者的 Th17 细胞中强烈表达(图 S5A)。此外,据报道,效应分子 CCL20 的强表达介导 Th17 驱动的炎症过程,并与 SLE 发病机制有关 ( Koga et al., 2016 ),并且该基因在 SLE Th17 细胞中的表达水平显著高于健康 Th17 细胞(图 S5B;p = 1.40e-110,Mann-Whitney U 检验)。SLE Tct 细胞显示一系列细胞毒性分子的水平明显升高,包括 GZMH 、 NKG7 和 GZMB,支持先前关于它们可能参与加剧自身免疫性疾病的说法 ( Takeuchi and Saito, 2017 ) (图 S5C)。
鉴于已知核因子 κB (NF-κB) 信号通路基因可促进 SLE 的炎症发展和进展 ( Brightbill et al., 2018 ),我们始终观察到这些基因(包括 NFKB1、NFKB2 和 REL)在重度 SLE 患者的 Th17 细胞中强烈表达(图 S5A)。此外,据报道,效应分子 CCL20 的强表达介导 Th17 驱动的炎症过程,并与 SLE 发病机制有关 ( Koga et al., 2016 ),并且该基因在 SLE Th17 细胞中的表达水平显著高于健康 Th17 细胞(图 S5B;p = 1.40e-110,Mann-Whitney U 检验)。SLE Tct 细胞显示一系列细胞毒性分子的水平明显升高,包括 GZMH 、 NKG7 和 GZMB,支持先前关于它们可能参与加剧自身免疫性疾病的说法 ( Takeuchi and Saito, 2017 ) (图 S5C)。
Interestingly, our single-cell analysis revealed that quite a few T cell exhaustion-related genes, such as TIGIT, PDCD1, and LAG3 (Wherry and Kurachi, 2015), were significantly upregulated in SLE Treg cells (Figures 4A, 4E, and S5D). We purified Treg cells from an additional cohort of healthy controls and severe-stage SLE patients (Table S1) and then performed qPCR and flow cytometry analysis of these genes, and found that each gene was expressed at a significantly higher level in the patients (Figures 4F and S5E; p < 0.05, Student’s t test). We also found that the signature score of Treg exhaustion-like properties was higher in SLE Tregs than in normal Tregs (Figure 4G; p = 2.01 × 10−29). Chromatin accessibility analysis of the isolated Treg cells using ATAC-seq (Table S1) also indicated that these loci were more accessible in SLE patients than in healthy controls (Figures 4H–4J and S5F).
有趣的是,我们的单细胞分析显示,相当多的 T 细胞耗竭相关基因,如 TIGIT、PDCD1 和 LAG3 ( Wherry and Kurachi, 2015 ),在 SLE Treg 细胞中显著上调(图 4A、4E 和 S5D)。我们从另一组健康对照和重度 SLE 患者(表 S1)中纯化了 Treg 细胞,然后对这些基因进行了 qPCR 和流式细胞术分析,发现每个基因在患者中的表达水平显着更高(图 4F 和 S5E;p < 0.05,学生 t 检验)。我们还发现,SLE Treg 中 Treg 耗竭样特性的特征评分高于正常 Treg(图 4G;p = 2.01 × 10−29)。使用 ATAC-seq 对分离的 Treg 细胞进行染色质可及性分析(表 S1)还表明,这些基因座在 SLE 患者中比在健康对照中更容易接近(图 4H-4J 和 S5F)。
有趣的是,我们的单细胞分析显示,相当多的 T 细胞耗竭相关基因,如 TIGIT、PDCD1 和 LAG3 ( Wherry and Kurachi, 2015 ),在 SLE Treg 细胞中显著上调(图 4A、4E 和 S5D)。我们从另一组健康对照和重度 SLE 患者(表 S1)中纯化了 Treg 细胞,然后对这些基因进行了 qPCR 和流式细胞术分析,发现每个基因在患者中的表达水平显着更高(图 4F 和 S5E;p < 0.05,学生 t 检验)。我们还发现,SLE Treg 中 Treg 耗竭样特性的特征评分高于正常 Treg(图 4G;p = 2.01 × 10−29)。使用 ATAC-seq 对分离的 Treg 细胞进行染色质可及性分析(表 S1)还表明,这些基因座在 SLE 患者中比在健康对照中更容易接近(图 4H-4J 和 S5F)。
To further explore this apparent dysfunction of SLE Treg cells, we isolated Treg cells from another cohort of four healthy controls and four severe-stage SLE patients using flow cytometry (Table S1) and examined their abilities to inhibit Teff cells by a carboxyfluorescein succinimidyl ester (CFSE)-based T cell suppression assay. We found that Treg cells isolated from SLE patients showed lower suppressive activity than healthy controls (Figures 4K and S5G). Together, these results indicate that SLE Treg cells may undergo functional exhaustion and thus fail to suppress the over-reactive immune system. Moreover, Treg with exhaustion-like properties was also observed in the inflamed intestinal mucosa of patients with ulcerative colitis (Smillie et al., 2019) (Figures S5H–S5K).
为了进一步探索 SLE Treg 细胞的这种明显功能障碍,我们使用流式细胞术从另一组 4 名健康对照和 4 名重度 SLE 患者中分离了 Treg 细胞(表 S1),并通过基于羧基荧光素琥珀酰亚胺酯 (CFSE) 的 T 细胞抑制测定检查了它们抑制 Teff 细胞的能力。我们发现从 SLE 患者中分离的 Treg 细胞显示出比健康对照更低的抑制活性(图 4K 和 S5G)。总之,这些结果表明 SLE Treg 细胞可能会发生功能衰竭,因此无法抑制过度反应的免疫系统。此外,在溃疡性结肠炎患者发炎的肠粘膜中也观察到具有疲惫样特性的 Treg ( Smillie et al., 2019 ) (图 S5H-S5K)。
为了进一步探索 SLE Treg 细胞的这种明显功能障碍,我们使用流式细胞术从另一组 4 名健康对照和 4 名重度 SLE 患者中分离了 Treg 细胞(表 S1),并通过基于羧基荧光素琥珀酰亚胺酯 (CFSE) 的 T 细胞抑制测定检查了它们抑制 Teff 细胞的能力。我们发现从 SLE 患者中分离的 Treg 细胞显示出比健康对照更低的抑制活性(图 4K 和 S5G)。总之,这些结果表明 SLE Treg 细胞可能会发生功能衰竭,因此无法抑制过度反应的免疫系统。此外,在溃疡性结肠炎患者发炎的肠粘膜中也观察到具有疲惫样特性的 Treg ( Smillie et al., 2019 ) (图 S5H-S5K)。
CCR7lowCD74hi Treg cells from SLE patients are functionally exhausted
来自 SLE 患者的 CCR7低CD74hi Treg 细胞功能衰竭
A previous study identified three phenotypically and functionally distinct human Treg cell subpopulations based on the expression of CD45RA and FOXP3: CD45RA+FOXP3low naive Treg (nTreg) cells, CD45RA−FOXP3hi effector Treg (eTreg) cells, and CD45RA−FOXP3low fraction (Fr. III) cells (Miyara et al., 2009). In our single-cell transcriptomes of CD4+ T cells from SLE patients and healthy controls, we identified two subpopulations of Treg cells, which we defined as Treg1 and Treg2, from a total of 3,239 cells (Figure 5A). We first obtained the cell subtype-specific genes of the nTreg, eTreg, and Fr. III cells from published microarray datasets (Cuadrado et al., 2018) (Figure S6A; STAR methods). Correlation analysis of the expression of these specific genes between Treg1 and Treg2 cells with nTreg, eTreg, and Fr. III cells (Figure S6B) indicated that Treg1 cells typically exhibited marker expression profiles similar to that of nTreg cells, whereas the Treg2 cell profiles appeared more similar to eTreg and Fr. III cells (Figure S6B).
先前的一项研究根据 CD45RA 和 FOXP3 的表达确定了三种表型和功能不同的人类 Treg 细胞亚群:CD45RA+FOXP3低初始 Treg (nTreg) 细胞、CD45RA−FOXP3高效应 Treg (eTreg) 细胞和 CD45RA−FOXP3低分数 (Fr. III) 细胞 ( Miyara et al., 2009 )。在来自 SLE 患者和健康对照者的 CD4 + T 细胞的单细胞转录组中,我们从总共 3,239 个细胞中鉴定了两个 Treg 细胞亚群,我们将其定义为 Treg1 和 Treg2(图 5A)。我们首先从已发表的微阵列数据集 ( Cuadrado et al., 2018 ) 中获得了 nTreg、eTreg 和 Fr. III 细胞的细胞亚型特异性基因(图 S6A;STAR 方法)。Treg1 和 Treg2 细胞与 nTreg、eTreg 和 Fr. III 细胞之间这些特定基因表达的相关性分析(图 S6B)表明,Treg1 细胞通常表现出与 nTreg 细胞相似的标志物表达谱,而 Treg2 细胞谱似乎更类似于 eTreg 和 Fr. III 细胞(图 S6B)。
先前的一项研究根据 CD45RA 和 FOXP3 的表达确定了三种表型和功能不同的人类 Treg 细胞亚群:CD45RA+FOXP3低初始 Treg (nTreg) 细胞、CD45RA−FOXP3高效应 Treg (eTreg) 细胞和 CD45RA−FOXP3低分数 (Fr. III) 细胞 ( Miyara et al., 2009 )。在来自 SLE 患者和健康对照者的 CD4 + T 细胞的单细胞转录组中,我们从总共 3,239 个细胞中鉴定了两个 Treg 细胞亚群,我们将其定义为 Treg1 和 Treg2(图 5A)。我们首先从已发表的微阵列数据集 ( Cuadrado et al., 2018 ) 中获得了 nTreg、eTreg 和 Fr. III 细胞的细胞亚型特异性基因(图 S6A;STAR 方法)。Treg1 和 Treg2 细胞与 nTreg、eTreg 和 Fr. III 细胞之间这些特定基因表达的相关性分析(图 S6B)表明,Treg1 细胞通常表现出与 nTreg 细胞相似的标志物表达谱,而 Treg2 细胞谱似乎更类似于 eTreg 和 Fr. III 细胞(图 S6B)。

Figure 5 Treg cell subsets and their functional divergence in SLE patients
图 5SLE 患者的 Treg 细胞亚群及其功能差异
图 5SLE 患者的 Treg 细胞亚群及其功能差异
We then utilized Monocle3 (Cao et al., 2019) to perform a pseudotime analysis of the differentiation potential of Treg cell subsets from healthy controls and SLE patients to investigate the specific trajectories of Treg1 and Treg2 cells. We found that Treg2 trajectory-enriched cells strongly expressed genes related to T cell exhaustion and Treg cell exhaustion (Figure 5B), while Treg1 trajectory-enriched cells exhibited higher levels of lineage-specific genes for Th1, Th2, Th17, and Tct cells (Figures S6C, S6D, and Table S4).
然后,我们利用 Monocle3 ( Cao et al., 2019 ) 对健康对照和 SLE 患者的 Treg 细胞亚群的分化潜力进行伪时间分析,以研究 Treg1 和 Treg2 细胞的特定轨迹。我们发现 Treg2 轨迹富集细胞强烈表达与 T 细胞耗竭和 Treg 细胞耗竭相关的基因(图 5B),而 Treg1 轨迹富集细胞对 Th1、Th2、Th17 和 Tct 细胞表现出更高水平的谱系特异性基因(图 S6C、S6D 和表 S4)。
然后,我们利用 Monocle3 ( Cao et al., 2019 ) 对健康对照和 SLE 患者的 Treg 细胞亚群的分化潜力进行伪时间分析,以研究 Treg1 和 Treg2 细胞的特定轨迹。我们发现 Treg2 轨迹富集细胞强烈表达与 T 细胞耗竭和 Treg 细胞耗竭相关的基因(图 5B),而 Treg1 轨迹富集细胞对 Th1、Th2、Th17 和 Tct 细胞表现出更高水平的谱系特异性基因(图 S6C、S6D 和表 S4)。
Among the 1,064 significant DEGs between Treg1 and Treg2 cells (Table S4), we identified two putative surface marker genes, CCR7 and CD74, that could be used to distinguish the two Treg subpopulations (Figures 5C–5E). Flow cytometry analysis indicated that the proportions of both Treg1 (CCR7hiCD74low) and Treg2 (CCR7lowCD74hi) cells were significantly higher in severe SLE patients than in healthy controls (Figure 5F; Student’s t test, ∗p <0.05). There were 830 and 818 DEGs between healthy controls and SLE patients for Treg1 and Treg2 cells, respectively (Figures 6A and 6B; Table S4). We noted that the SLE Treg2 cells highly expressed signature genes of T cell exhaustion, such as PDCD1 and LAG3 (Figures 6C and S6E–S6G) (Lowther et al., 2016). We validated the expression trends for these genes in Treg cells from healthy controls and SLE patients by qPCR and flow cytometry (Figures 6D and S6H). These results suggested that CCR7lowCD74hi Treg2 cells, which are an effective Treg subpopulation in SLE patients, may undergo functional exhaustion that weakens their immunosuppressive function.
在 Treg1 和 Treg2 细胞之间的 1,064 个重要 DEGs 中(表 S4),我们确定了两个推定的表面标记基因 CCR7 和 CD74,可用于区分两个 Treg 亚群(图 5C-5E)。流式细胞术分析表明,严重 SLE 患者中 Treg1 (CCR7hiCD74低) 和 Treg2 (CCR7低CD74hi) 细胞的比例显著高于健康对照(图 5F;学生 t 检验,∗p <0.05)。健康对照和 SLE 患者之间的 Treg1 和 Treg2 细胞分别有 830 和 818 个 DEG(图 6A 和 6B;表 S4)。我们注意到 SLE Treg2 细胞高度表达 T 细胞耗竭的特征基因,例如 PDCD1 和 LAG3(图 6C 和 S6E-S6G) ( Lowther et al., 2016 )。我们通过 qPCR 和流式细胞术验证了这些基因在健康对照和 SLE 患者的 Treg 细胞中的表达趋势(图 6D 和 S6H)。这些结果表明,CCR7低CD74hi Treg2 细胞是 SLE 患者的有效 Treg 亚群,可能会发生功能衰竭,从而削弱其免疫抑制功能。
在 Treg1 和 Treg2 细胞之间的 1,064 个重要 DEGs 中(表 S4),我们确定了两个推定的表面标记基因 CCR7 和 CD74,可用于区分两个 Treg 亚群(图 5C-5E)。流式细胞术分析表明,严重 SLE 患者中 Treg1 (CCR7hiCD74低) 和 Treg2 (CCR7低CD74hi) 细胞的比例显著高于健康对照(图 5F;学生 t 检验,∗p <0.05)。健康对照和 SLE 患者之间的 Treg1 和 Treg2 细胞分别有 830 和 818 个 DEG(图 6A 和 6B;表 S4)。我们注意到 SLE Treg2 细胞高度表达 T 细胞耗竭的特征基因,例如 PDCD1 和 LAG3(图 6C 和 S6E-S6G) ( Lowther et al., 2016 )。我们通过 qPCR 和流式细胞术验证了这些基因在健康对照和 SLE 患者的 Treg 细胞中的表达趋势(图 6D 和 S6H)。这些结果表明,CCR7低CD74hi Treg2 细胞是 SLE 患者的有效 Treg 亚群,可能会发生功能衰竭,从而削弱其免疫抑制功能。
We then examined peripheral blood samples from SLE patients using flow cytometry to detect the secretion of IFN-γ, IL-2, and tumor necrosis factor alpha (TNFα) in CCR7hiCD74low Treg1 and CCR7lowCD74hi Treg2 cells upon phorbol 12-myristate 13-acetate (PMA)/ionomycin stimulation. In contrast to the characteristics of CD8+ T cell exhaustion, which exhibits an insufficient ability in producing IFN-γ, IL-2, and TNFα, we found significantly higher proportions of IFN-γ (p < 0.05, Student’s t test), IL-2 (p < 0.01, Student’s t test), and TNFα (Figure 6E; p < 0.01, Student’s t test) production by Treg2 cells compared with Treg1 cells in SLE patients. As previous studies have shown that these cytokine-producing Treg cells are phenotypically Th1-like Tregs, which contribute to the observed Treg dysfunction in autoimmune diseases (Dominguez-Villar et al., 2011; Korn et al., 2007; McClymont et al., 2011), this result provides further protein-level support for the functional exhaustion-like properties of the CCR7lowCD74hi Treg2 cells in SLE patients.
然后,我们使用流式细胞术检测 SLE 患者的外周血样本,以检测 CCR7hiCD74低 Treg1 和 CCR7低CD74hi Treg2 细胞中 IFN-γ、IL-2 和肿瘤坏死因子 α (TNFα) 的分泌佛波醇 12-肉豆蔻酸酯 13-乙酸酯 (PMA)/离子霉素刺激。与 CD8+ T 细胞耗竭的特征相比,CD8+ T 细胞耗竭在产生 IFN-γ、IL-2 和 TNFα 的能力不足,我们发现 SLE 患者 Treg2 细胞产生 IFN-γ (p < 0.05,学生 t 检验)、IL-2 (p < 0.01,学生 t 检验) 和 TNFα (图 6E;p < 0.01,学生 t 检验) 的比例显着高于 Treg1 细胞。正如以前的研究表明的那样,这些产生细胞因子的 Treg 细胞是表型的 Th1 样 Treg,这有助于在自身免疫性疾病中观察到的 Treg 功能障碍。 Dominguez-Villar et al., 2011 Korn et al., 2007 ; McClymont et al., 2011 ),这一结果为 SLE 患者 CCR7低CD74hi Treg2 细胞的功能耗竭样特性提供了进一步的蛋白质水平支持。
然后,我们使用流式细胞术检测 SLE 患者的外周血样本,以检测 CCR7hiCD74低 Treg1 和 CCR7低CD74hi Treg2 细胞中 IFN-γ、IL-2 和肿瘤坏死因子 α (TNFα) 的分泌佛波醇 12-肉豆蔻酸酯 13-乙酸酯 (PMA)/离子霉素刺激。与 CD8+ T 细胞耗竭的特征相比,CD8+ T 细胞耗竭在产生 IFN-γ、IL-2 和 TNFα 的能力不足,我们发现 SLE 患者 Treg2 细胞产生 IFN-γ (p < 0.05,学生 t 检验)、IL-2 (p < 0.01,学生 t 检验) 和 TNFα (图 6E;p < 0.01,学生 t 检验) 的比例显着高于 Treg1 细胞。正如以前的研究表明的那样,这些产生细胞因子的 Treg 细胞是表型的 Th1 样 Treg,这有助于在自身免疫性疾病中观察到的 Treg 功能障碍。 Dominguez-Villar et al., 2011 Korn et al., 2007 ; McClymont et al., 2011 ),这一结果为 SLE 患者 CCR7低CD74hi Treg2 细胞的功能耗竭样特性提供了进一步的蛋白质水平支持。
We also assessed peripheral blood samples from SLE patients using flow cytometry to detect the expression of exhaustion-associated transcription factors (Tcf1 and Tox) in CCR7hiCD74low Treg1 and CCR7lowCD74hi Treg2 cells upon in vitro PMA/ionomycin stimulation. We observed that there were no significant differences in the expression of Tcf1 between Treg1 and Treg2 cells (Figure 6F). Notably, we found that Treg2 cells displayed significantly higher expression levels of Tox than Treg1 cells (Figure 6F; p < 0.05, Student’s t test). These data are consistent with the findings from previous reports showing Tox as a promoting factor for T cell exhaustion (Kim et al., 2020; Scott et al., 2019) and provide additional support for the conclusion that CCR7lowCD74hi Treg2 cells are Tregs with exhaustion-like properties.
我们还使用流式细胞术评估了 SLE 患者的外周血样本,以检测体外 PMA/离子霉素刺激后 CCR7hiCD74低 Treg1 和 CCR7低CD74hi Treg2 细胞中耗竭相关转录因子 (Tcf1 和 Tox) 的表达。我们观察到 Treg1 和 Treg2 细胞之间 Tcf1 的表达没有显着差异(图 6F)。值得注意的是,我们发现 Treg2 细胞的 Tox 表达水平明显高于 Treg1 细胞(图 6F;p < 0.05,学生 t 检验)。这些数据与先前报告的结果一致,显示 Tox 是 T 细胞耗竭的促进因子 ( Kim et al., 2020 ; Scott et al., 2019 ),并为 CCR7低CD74hi Treg2 细胞是具有耗竭样特性的 Treg 细胞的结论提供额外的支持。
我们还使用流式细胞术评估了 SLE 患者的外周血样本,以检测体外 PMA/离子霉素刺激后 CCR7hiCD74低 Treg1 和 CCR7低CD74hi Treg2 细胞中耗竭相关转录因子 (Tcf1 和 Tox) 的表达。我们观察到 Treg1 和 Treg2 细胞之间 Tcf1 的表达没有显着差异(图 6F)。值得注意的是,我们发现 Treg2 细胞的 Tox 表达水平明显高于 Treg1 细胞(图 6F;p < 0.05,学生 t 检验)。这些数据与先前报告的结果一致,显示 Tox 是 T 细胞耗竭的促进因子 ( Kim et al., 2020 ; Scott et al., 2019 ),并为 CCR7低CD74hi Treg2 细胞是具有耗竭样特性的 Treg 细胞的结论提供额外的支持。
Excessive type I IFN production promotes Treg cell exhaustion in SLE patients
过量的 I 型 IFN 产生促进 SLE 患者的 Treg 细胞耗竭
A previous study showed that chronic stimulation by persistent antigens or inflammatory signals is sufficient to drive the exhaustion and functional disability of T cells (Rome et al., 2020). Therefore, we sought to investigate whether persistent stimulation of inflammatory signals can induce Treg cell exhaustion. We first defined a Treg cell exhaustion signature score (TES) as the average expression of genes involved in Treg cell exhaustion and then ranked the genes whose expression was significantly correlated with the TES score across all Treg cells (Figure 7A; Pearson’s correlation >0.4, p <0.05). We performed functional annotation analyses of these genes by Metascape (Zhou et al., 2019) and gene set enrichment analysis (GSEA) (Mootha et al., 2003) according to the correlation coefficients between the expression of genes and TES (STAR methods). The genomic features of IFN signaling (Figure 7B) and response to type I IFNs (Figure 7C) were significantly enriched. The average expression levels of genes associated with Treg cell exhaustion and genes involved in the type I IFN signaling pathway were also found to be highly correlated along the Treg cell pseudotime trajectory in SLE patients (Figures 7D–7F and S7A). In addition, the chromatin accessibility level of Treg cells between the exhaustion signature genes and type I IFN genes showed a close correlation (Figure S7B). We also found that the expression of IFN-induced genes was higher in SLE Tregs than in Th1, Th2, Th17, and Tct cells, suggesting that Treg cells from SLE patients are more sensitive to IFN stimulation than Th1, Th2, Th17, and Tct cells (Figures S7C–S7F). We therefore hypothesize that persistent stimulation of type I IFN signaling may promote the Treg cell exhaustion observed among the primary CD4+ T cells of SLE patients.
先前的一项研究表明,持续抗原或炎症信号的慢性刺激足以驱动 T 细胞的耗竭和功能残疾 ( Rome et al., 2020 )。因此,我们试图研究炎症信号的持续刺激是否可以诱导 Treg 细胞耗竭。我们首先将 Treg 细胞耗竭特征评分 (TES) 定义为参与 Treg 细胞耗竭的基因的平均表达,然后对所有 Treg 细胞中表达与 TES 评分显著相关的基因进行排序(图 7A;皮尔逊相关性 >0.4,p <0.05)。我们根据基因表达与 TES 之间的相关系数(STAR 方法),通过 Metascape ( Zhou et al., 2019 ) 和基因集富集分析 (GSEA) ( Mootha et al., 2003 ) 对这些基因进行了功能注释分析。IFN 信号转导的基因组特征 (图 7B) 和对 I 型 IFN 的反应 (图 7C) 显著丰富。在 SLE 患者中,与 Treg 细胞耗竭相关的基因和参与 I 型 IFN 信号通路的基因的平均表达水平沿 Treg 细胞伪时间轨迹高度相关(图 7D-7F 和 S7A)。此外,耗竭特征基因和 I 型 IFN 基因之间的 Treg 细胞的染色质可及性水平显示出密切的相关性(图 S7B)。我们还发现 SLE Tregs 中 IFN 诱导基因的表达高于 Th1、Th2、Th17 和 Tct 细胞,这表明 SLE 患者的 Treg 细胞比 Th1、Th2、Th17 和 Tct 细胞对 IFN 刺激更敏感(图 S7C-S7F)。 因此,我们假设 I 型 IFN 信号的持续刺激可能会促进在 SLE 患者的原代 CD4+ T 细胞中观察到的 Treg 细胞耗竭。
先前的一项研究表明,持续抗原或炎症信号的慢性刺激足以驱动 T 细胞的耗竭和功能残疾 ( Rome et al., 2020 )。因此,我们试图研究炎症信号的持续刺激是否可以诱导 Treg 细胞耗竭。我们首先将 Treg 细胞耗竭特征评分 (TES) 定义为参与 Treg 细胞耗竭的基因的平均表达,然后对所有 Treg 细胞中表达与 TES 评分显著相关的基因进行排序(图 7A;皮尔逊相关性 >0.4,p <0.05)。我们根据基因表达与 TES 之间的相关系数(STAR 方法),通过 Metascape ( Zhou et al., 2019 ) 和基因集富集分析 (GSEA) ( Mootha et al., 2003 ) 对这些基因进行了功能注释分析。IFN 信号转导的基因组特征 (图 7B) 和对 I 型 IFN 的反应 (图 7C) 显著丰富。在 SLE 患者中,与 Treg 细胞耗竭相关的基因和参与 I 型 IFN 信号通路的基因的平均表达水平沿 Treg 细胞伪时间轨迹高度相关(图 7D-7F 和 S7A)。此外,耗竭特征基因和 I 型 IFN 基因之间的 Treg 细胞的染色质可及性水平显示出密切的相关性(图 S7B)。我们还发现 SLE Tregs 中 IFN 诱导基因的表达高于 Th1、Th2、Th17 和 Tct 细胞,这表明 SLE 患者的 Treg 细胞比 Th1、Th2、Th17 和 Tct 细胞对 IFN 刺激更敏感(图 S7C-S7F)。 因此,我们假设 I 型 IFN 信号的持续刺激可能会促进在 SLE 患者的原代 CD4+ T 细胞中观察到的 Treg 细胞耗竭。

Figure 7 Chronic type I IFN stimulation promotes Treg cell exhaustion in SLE patients
图 7慢性 I 型 IFN 刺激促进 SLE 患者的 Treg 细胞耗竭
图 7慢性 I 型 IFN 刺激促进 SLE 患者的 Treg 细胞耗竭
Pursuing this, we purified primary Treg cells from healthy controls and performed transcriptome sequencing analysis of Treg cells that were stimulated with different concentrations of type I IFN via multiple protocols. As expected, we noticed that the expression of genes involved in the type I IFN signaling pathway was successfully induced (Figure 7G). We also found that persistent type I IFN stimulation was sufficient to induce the expression of Treg exhaustion-associated genes and genes upregulated in SLE Treg2 cells (Figure 7G; Table S4; STAR methods). The average expression of Treg exhaustion-associated genes in each sample was also found to be significantly positively correlated with that of genes involved in the type I IFN signaling pathway (p = 0.047, R = 0.5; Figures 7H and S7G), suggesting that type I IFN signaling is responsible for the Treg exhaustion-like properties in SLE patients. In addition, a CFSE-based T cell suppression assay indicated that IFNα-treated Treg cells decreased their capacity to inhibit the proliferation and activation of Teff cells (Figure 7I), further supporting that persistent exposure to activated IFN signaling can induce Treg cell exhaustion. In summary, we found that the Treg cells from SLE patients exhibited exhaustion characteristics that were related to IFN signaling and verified in vitro that chronic IFN stimulation could induce Treg dysfunction.
为此,我们从健康对照中纯化了原代 Treg 细胞,并通过多种方案对用不同浓度的 I 型 IFN 刺激的 Treg 细胞进行了转录组测序分析。正如预期的那样,我们注意到参与 I 型 IFN 信号通路的基因表达被成功诱导(图 7G)。我们还发现,持续的 I 型 IFN 刺激足以诱导 SLE Treg 2 细胞中 Treg 耗竭相关基因和基因上调的表达(图 7G;表 S4;STAR 方法)。还发现每个样品中 Treg 耗竭相关基因的平均表达与参与 I 型 IFN 信号通路的基因的平均表达呈显著正相关 (p = 0.047,R = 0.5; 图 7H 和 S7G),表明 I 型 IFN 信号传导是 SLE 患者 Treg 耗竭样特性的原因。此外,基于 CFSE 的 T 细胞抑制测定表明,IFNα 处理的 Treg 细胞降低了其抑制 Teff 细胞增殖和活化的能力(图 7I),进一步支持持续暴露于活化的 IFN 信号传导可诱导 Treg 细胞耗竭。综上所述,我们发现 SLE 患者的 Treg 细胞表现出与 IFN 信号传导相关的耗竭特性,并在体外验证 了慢性 IFN 刺激可诱导 Treg 功能障碍。
为此,我们从健康对照中纯化了原代 Treg 细胞,并通过多种方案对用不同浓度的 I 型 IFN 刺激的 Treg 细胞进行了转录组测序分析。正如预期的那样,我们注意到参与 I 型 IFN 信号通路的基因表达被成功诱导(图 7G)。我们还发现,持续的 I 型 IFN 刺激足以诱导 SLE Treg 2 细胞中 Treg 耗竭相关基因和基因上调的表达(图 7G;表 S4;STAR 方法)。还发现每个样品中 Treg 耗竭相关基因的平均表达与参与 I 型 IFN 信号通路的基因的平均表达呈显著正相关 (p = 0.047,R = 0.5; 图 7H 和 S7G),表明 I 型 IFN 信号传导是 SLE 患者 Treg 耗竭样特性的原因。此外,基于 CFSE 的 T 细胞抑制测定表明,IFNα 处理的 Treg 细胞降低了其抑制 Teff 细胞增殖和活化的能力(图 7I),进一步支持持续暴露于活化的 IFN 信号传导可诱导 Treg 细胞耗竭。综上所述,我们发现 SLE 患者的 Treg 细胞表现出与 IFN 信号传导相关的耗竭特性,并在体外验证 了慢性 IFN 刺激可诱导 Treg 功能障碍。
Discussion 讨论
SLE is a complex systemic disease characterized by a wide spectrum of clinical manifestations and has substantial interindividual heterogeneity (Lisnevskaia et al., 2014). CD4+ T cells from the peripheral blood are an essential contributor and indicator of immune overactivation (Caielli et al., 2018), and the chromatin accessibility of this cell type is highly informative for their identity, activity state, and regulatory programs (Qu et al., 2017). In the present study, we used ATAC-seq to survey the landscape of active regulatory DNA in a single cell type (CD4+ T) sorted from peripheral blood samples of a large cohort of SLE patients. Further analysis of data from this large cohort identified three groups of patients with divergent transcriptional regulatory patterns, and these same groups were clustered similarly based on their DAI. Notably, the SLE disease signature evident in chromatin accessibility patterns was most obvious in severe-stage patients; perhaps patterns in patients with mild or moderate DA can be identified in studies with larger patient cohorts. Our analysis detected a correlation between the epigenome of CD4+ T cells and SLE clinical states and provided a rich empirical foundation from the perspective of epigenetic regulation for understanding SLE heterogeneity.
SLE 是一种复杂的全身性疾病,其特征是临床表现广泛,并且具有很大的个体间异质性 ( Lisnevskaia et al., 2014 )。来自外周血的 CD4 + T 细胞是免疫过度激活的重要贡献者和指标 ( Caielli et al., 2018 ),这种细胞类型的染色质可及性对其身份、活性状态和调节程序具有高度信息性 ( Qu et al., 2017 )。在本研究中,我们使用 ATAC-seq 调查了从一大群 SLE 患者的外周血样本中分选的单细胞类型 (CD4+ T) 中活性调节 DNA 的景观。对来自这个大型队列的数据的进一步分析确定了三组具有不同转录调控模式的患者,并且这些相同的组根据他们的 DAI 进行了类似的聚类。值得注意的是,染色质可及性模式中明显的 SLE 疾病特征在重症患者中最为明显;也许可以在具有更大患者队列的研究中确定轻度或中度 DA 患者的模式。我们的分析检测到 CD4+ T 细胞表观基因组与 SLE 临床状态之间的相关性,并从表观遗传调控的角度为理解 SLE 异质性提供了丰富的实证基础。
SLE 是一种复杂的全身性疾病,其特征是临床表现广泛,并且具有很大的个体间异质性 ( Lisnevskaia et al., 2014 )。来自外周血的 CD4 + T 细胞是免疫过度激活的重要贡献者和指标 ( Caielli et al., 2018 ),这种细胞类型的染色质可及性对其身份、活性状态和调节程序具有高度信息性 ( Qu et al., 2017 )。在本研究中,我们使用 ATAC-seq 调查了从一大群 SLE 患者的外周血样本中分选的单细胞类型 (CD4+ T) 中活性调节 DNA 的景观。对来自这个大型队列的数据的进一步分析确定了三组具有不同转录调控模式的患者,并且这些相同的组根据他们的 DAI 进行了类似的聚类。值得注意的是,染色质可及性模式中明显的 SLE 疾病特征在重症患者中最为明显;也许可以在具有更大患者队列的研究中确定轻度或中度 DA 患者的模式。我们的分析检测到 CD4+ T 细胞表观基因组与 SLE 临床状态之间的相关性,并从表观遗传调控的角度为理解 SLE 异质性提供了丰富的实证基础。
Previous studies have reported an expansion of PD-1+CXCR5+CD4+ T (Tfh) cells and PD-1+CXCR5−CD4+ T (Tph) cells in SLE patients (He et al., 2013; Kim et al., 2018; Makiyama et al., 2019). However, researchers have paid minimal attention to changes in the proportion of PD-1−CXCR5+CD4+ T cells in disease states. The flow cytometry analysis of blood samples from lupus patients showed that the patients exhibited higher proportions of Tfh and Tph and a lower proportion of PD-1−CXCR5+ CD4+ T cells than healthy control individuals (Lin et al., 2019). In line with these results are data from peripheral blood of patients with Sjögren’s syndrome showing diminished CXCR5 expression in T cell subsets (Aqrawi et al., 2018). Therefore, the finding of a significant decrease in the fraction of CXCR5+ T cells in SLE patients in our study may be associated with an expansion of Tph cells. In addition, CXCR5 is also expressed on 20%–25% of peripheral blood human central memory CD4+ T cells, and these cells are a heterogeneous pool consisting of functionally distinct Th1-, Th2-, and Th17-like subsets (Chevalier et al., 2011). The reduction of CXCR5+ T cells in SLE patients may also be attributed to the activation of these central memory CD4+ T cells that transformed into effector T cells, supported by the results that SLE patients exhibited higher proportions of Th1, Th2, Th17, and Tct cells compared with healthy control individuals (Figures 3G and 3H). Nevertheless, biological function of the CXCR5+ cells, especially the PD-1−CXCR5+CD4+ T cells, remains to be further investigated.
先前的研究报道了 SLE 患者 PD-1+CXCR5+CD4+ T (Tfh) 细胞和 PD-1+CXCR5−CD4+ T (Tph) 细胞的扩增 ( He et al., 2013 ; Kim et al., 2018 ; Makiyama et al., 2019 )。然而,研究人员对疾病状态下 PD-1-CXCR5+CD4+ T 细胞比例的变化关注甚少。对狼疮患者血液样本的流式细胞术分析显示,与健康对照个体相比,患者表现出更高比例的 Tfh 和 Tph 以及更低比例的 PD-1-CXCR5+ CD4+ T 细胞 ( Lin et al., 2019 )。与这些结果一致的是来自干燥综合征患者外周血的数据,显示 T 细胞亚群中 CXCR5 表达降低 ( Aqrawi et al., 2018 )。因此,在我们的研究中发现 SLE 患者 CXCR5 + T 细胞分数显着降低可能与 Tph 细胞的扩增有关。此外,CXCR5 还在 20%–25% 的外周血人中枢记忆 CD4+ T 细胞上表达,这些细胞是一个异质性池,由功能不同的 Th1-、Th2- 和 Th17 样亚群组成 ( Chevalier et al., 2011 )。SLE 患者 CXCR5+ T 细胞的减少也可能归因于这些转化为效应 T 细胞的中枢记忆 CD4+ T 细胞的激活,结果证明 SLE 患者表现出更高比例的 Th1、Th2、Th17 和 Tct 细胞与健康对照个体相比(图 3G 和 3H)。然而,CXCR5 + 细胞的生物学功能,尤其是 PD-1-CXCR5 + CD4 + T 细胞,仍有待进一步研究。
先前的研究报道了 SLE 患者 PD-1+CXCR5+CD4+ T (Tfh) 细胞和 PD-1+CXCR5−CD4+ T (Tph) 细胞的扩增 ( He et al., 2013 ; Kim et al., 2018 ; Makiyama et al., 2019 )。然而,研究人员对疾病状态下 PD-1-CXCR5+CD4+ T 细胞比例的变化关注甚少。对狼疮患者血液样本的流式细胞术分析显示,与健康对照个体相比,患者表现出更高比例的 Tfh 和 Tph 以及更低比例的 PD-1-CXCR5+ CD4+ T 细胞 ( Lin et al., 2019 )。与这些结果一致的是来自干燥综合征患者外周血的数据,显示 T 细胞亚群中 CXCR5 表达降低 ( Aqrawi et al., 2018 )。因此,在我们的研究中发现 SLE 患者 CXCR5 + T 细胞分数显着降低可能与 Tph 细胞的扩增有关。此外,CXCR5 还在 20%–25% 的外周血人中枢记忆 CD4+ T 细胞上表达,这些细胞是一个异质性池,由功能不同的 Th1-、Th2- 和 Th17 样亚群组成 ( Chevalier et al., 2011 )。SLE 患者 CXCR5+ T 细胞的减少也可能归因于这些转化为效应 T 细胞的中枢记忆 CD4+ T 细胞的激活,结果证明 SLE 患者表现出更高比例的 Th1、Th2、Th17 和 Tct 细胞与健康对照个体相比(图 3G 和 3H)。然而,CXCR5 + 细胞的生物学功能,尤其是 PD-1-CXCR5 + CD4 + T 细胞,仍有待进一步研究。
Since Treg cells function in maintaining immune homeostasis (Dominguez-Villar and Hafler, 2018; Sakaguchi et al., 2020; Wing et al., 2019), it seems counterintuitive that Treg numbers were increased in patients with an autoimmune disease. We explored the heterogeneity and dysfunction of SLE Treg cells in SLE patients and healthy controls, identified two distinct subsets of Treg cells, and confirmed their existence by flow cytometry. In healthy controls, these two subsets have polarized cell activities, separately displaying characteristics of naive and effector Treg cells. While both subsets of Treg cells were significantly increased in SLE patients, it was interesting to note that both SLE Treg1 and Treg2 cells in SLE patients adopted an effector-like state, with SLE Treg2 cells tending to display markers of functional exhaustion. ATAC-seq data also supported that this SLE Treg2 exhaustion phenotype results from epigenetic regulation. In addition, when we expanded the scope of our analysis to other autoimmune diseases by examining scRNA-seq? data for Tregs from patients with ulcerative colitis, we again observed upregulated Treg exhaustion-associated genes (Figures S5J–S5M), indicating that these Treg2-like exhausted cells may be features of multiple autoimmune diseases.
由于 Treg 细胞在维持免疫稳态 ( Dominguez-Villar and Hafler, 2018 ; Sakaguchi et al., 2020 ; Wing et al., 2019 ),这似乎有悖常理,即自身免疫性疾病患者的 Treg 值增加。我们探讨了 SLE 患者和健康对照者 SLE Treg 细胞的异质性和功能障碍,确定了 Treg 细胞的两个不同亚群,并通过流式细胞术证实了它们的存在。在健康对照中,这两个亚群具有极化细胞活性,分别表现出幼稚和效应 Treg 细胞的特征。虽然 SLE 患者的两个 Treg 细胞亚群都显著增加,但有趣的是,SLE 患者的 SLE Treg1 和 Treg2 细胞都采用效应样状态,SLE Treg2 细胞倾向于显示功能衰竭的标志物。ATAC-seq 数据还支持这种 SLE Treg2 耗竭表型是表观遗传调控的结果。此外,当我们通过检查 scRNA-seq?溃疡性结肠炎患者的 Tregs 数据,我们再次观察到 Treg 耗竭相关基因上调(图 S5J-S5M),表明这些 Treg2 样耗竭细胞可能是多种自身免疫性疾病的特征。
由于 Treg 细胞在维持免疫稳态 ( Dominguez-Villar and Hafler, 2018 ; Sakaguchi et al., 2020 ; Wing et al., 2019 ),这似乎有悖常理,即自身免疫性疾病患者的 Treg 值增加。我们探讨了 SLE 患者和健康对照者 SLE Treg 细胞的异质性和功能障碍,确定了 Treg 细胞的两个不同亚群,并通过流式细胞术证实了它们的存在。在健康对照中,这两个亚群具有极化细胞活性,分别表现出幼稚和效应 Treg 细胞的特征。虽然 SLE 患者的两个 Treg 细胞亚群都显著增加,但有趣的是,SLE 患者的 SLE Treg1 和 Treg2 细胞都采用效应样状态,SLE Treg2 细胞倾向于显示功能衰竭的标志物。ATAC-seq 数据还支持这种 SLE Treg2 耗竭表型是表观遗传调控的结果。此外,当我们通过检查 scRNA-seq?溃疡性结肠炎患者的 Tregs 数据,我们再次观察到 Treg 耗竭相关基因上调(图 S5J-S5M),表明这些 Treg2 样耗竭细胞可能是多种自身免疫性疾病的特征。
High levels of PD-1 expression on human Treg cells were described as a dysfunctional Treg exhaustion phenotype (Lowther et al., 2016). These exhausted Tregs exhibited high secretion of IFN-γ and were identified in healthy individuals and were enriched in tumor infiltrates (Lowther et al., 2016). Blockade of PD-1 activity reinvigorated the regulatory ability of dysfunctional Treg cells (Yang et al., 2017). In the present study, we identified CCR7lowCD74hi Treg cells with exhaustion-like properties (e.g., high expression of PD-1, TIM3, TIGIT, IFN-γ, IL-2, TNFα, Tox) that enriched in severe-stage SLE patients, as measured by scRNA-seq and supported by protein-level examination results. Treg exhaustion phenotype was also accompanied by several properties, such as an increase in pFoxo1 (Ser319), shortened telomere length, and decreased telomere-specific demethylated region (TSDR) (Lowther et al., 2016). However, the limited number of CCR7lowCD74hi Treg cells from human peripheral blood hindered the further in-depth investigation of the phenotypic and functional exhaustion of this Treg cell subpopulation.
人 Treg 细胞上高水平的 PD-1 表达被描述为功能失调的 Treg 耗竭表型 ( Lowther et al., 2016 )。这些耗竭的 Treg 表现出 IFN γ的高分泌,在健康个体中被发现,并在肿瘤浸润中富集 ( Lowther et al., 2016 )。阻断 PD-1 活性重新激发了功能失调的 Treg 细胞的调节能力 ( Yang et al., 2017 )。在本研究中,我们鉴定了具有耗竭样特性的 CCR7低CD74hi Treg 细胞(例如,PD-1、TIM3、TIGIT、IFN-γ、IL-2、TNFα、Tox的高表达),这些细胞在重度 SLE 患者中富集,通过 scRNA-seq 测量并得到蛋白质水平检查结果的支持。Treg 耗竭表型还伴有多种特性,例如 pFoxo1 (Ser319) 增加、端粒长度缩短和端粒特异性去甲基化区 (TSDR) 减少 ( Lowther et al., 2016 )。然而,来自人外周血的 CCR7低CD74hi Treg 细胞数量有限,阻碍了对该 Treg 细胞亚群的表型和功能耗竭的进一步深入研究。
人 Treg 细胞上高水平的 PD-1 表达被描述为功能失调的 Treg 耗竭表型 ( Lowther et al., 2016 )。这些耗竭的 Treg 表现出 IFN γ的高分泌,在健康个体中被发现,并在肿瘤浸润中富集 ( Lowther et al., 2016 )。阻断 PD-1 活性重新激发了功能失调的 Treg 细胞的调节能力 ( Yang et al., 2017 )。在本研究中,我们鉴定了具有耗竭样特性的 CCR7低CD74hi Treg 细胞(例如,PD-1、TIM3、TIGIT、IFN-γ、IL-2、TNFα、Tox的高表达),这些细胞在重度 SLE 患者中富集,通过 scRNA-seq 测量并得到蛋白质水平检查结果的支持。Treg 耗竭表型还伴有多种特性,例如 pFoxo1 (Ser319) 增加、端粒长度缩短和端粒特异性去甲基化区 (TSDR) 减少 ( Lowther et al., 2016 )。然而,来自人外周血的 CCR7低CD74hi Treg 细胞数量有限,阻碍了对该 Treg 细胞亚群的表型和功能耗竭的进一步深入研究。
Several studies have characterized the impact of type I IFNs on the function of Treg cells in a variety of contexts, leading to contrasting results (Piconese et al., 2015). It has been shown that type I IFN signaling regulates the dynamic balance between human-activated regulatory and effector T cells and attenuates Treg cell function in viral infection and in the tumor microenvironment (Bacher et al., 2013; Gangaplara et al., 2018; Golding et al., 2010; Hashimoto et al., 2014; Srivastava et al., 2014). In line with these studies, we found that type I IFN signaling is responsible for the Treg exhaustion-like properties and Treg dysfunction in SLE patients, as shown by the fact that type I IFN stimulation was sufficient to induce the expression of Treg exhaustion-associated genes and genes upregulated in SLE Treg2 cells (Figure 7G). IFN-mediated positive effects on Treg function have also been reported under stress conditions (Lee et al., 2012; Metidji et al., 2015). This is likely because chronic or sustained exposure to type I IFN signaling may give rise to an opposite effect on Treg homeostasis and functions (Piconese et al., 2015).
几项研究描述了 I 型 IFN 在各种情况下对 Treg 细胞功能的影响,导致了截然不同的结果 ( Piconese et al., 2015 )。已经表明,I 型 IFN 信号转导调节人激活的调节 T 细胞和效应 T 细胞之间的动态平衡,并减弱病毒感染和肿瘤微环境中的 Treg 细胞功能。 Bacher et al., 2013 Gangaplara et al., 2018 ; Golding et al., 2010 ; Hashimoto et al., 2014 ; Srivastava et al., 2014 )。与这些研究一致,我们发现 I 型 IFN 信号传导是 SLE 患者 Treg 耗竭样特性和 Treg 功能障碍的原因,如 I 型 IFN 刺激足以诱导 Treg 耗竭相关基因和 SLE Treg 细胞中上调的基因的表达这一事实所表明(图 7G)。在应激条件下,IFN 介导的对 Treg 功能的积极影响也有报道 ( Lee et al., 2012 ; Metidji et al., 2015 )。这可能是因为长期或持续暴露于 I 型 IFN 信号传导可能会对 Treg 稳态和功能产生相反的影响 ( Piconese et al., 2015 )。
几项研究描述了 I 型 IFN 在各种情况下对 Treg 细胞功能的影响,导致了截然不同的结果 ( Piconese et al., 2015 )。已经表明,I 型 IFN 信号转导调节人激活的调节 T 细胞和效应 T 细胞之间的动态平衡,并减弱病毒感染和肿瘤微环境中的 Treg 细胞功能。 Bacher et al., 2013 Gangaplara et al., 2018 ; Golding et al., 2010 ; Hashimoto et al., 2014 ; Srivastava et al., 2014 )。与这些研究一致,我们发现 I 型 IFN 信号传导是 SLE 患者 Treg 耗竭样特性和 Treg 功能障碍的原因,如 I 型 IFN 刺激足以诱导 Treg 耗竭相关基因和 SLE Treg 细胞中上调的基因的表达这一事实所表明(图 7G)。在应激条件下,IFN 介导的对 Treg 功能的积极影响也有报道 ( Lee et al., 2012 ; Metidji et al., 2015 )。这可能是因为长期或持续暴露于 I 型 IFN 信号传导可能会对 Treg 稳态和功能产生相反的影响 ( Piconese et al., 2015 )。
The use of glucocorticoids or other immunosuppressive agents is still the mainstay of SLE management (Kiriakidou, 2013). However, these agents have substantial adverse effects and do not typically confer adequate therapeutic efficacy (Lisnevskaia et al., 2014). A recent clinical trial of anifrolumab, a human monoclonal antibody that targets the type I IFN receptor, showed a significant therapeutic effect in treating SLE (Morand et al., 2020). Our results also support type I IFN as an effective drug target for the treatment of severe SLE patients and provide a potential treatment strategy for this disorder, suppression of Treg with exhaustion-like properties. Overall, we delineated the chromatin accessibility and single-cell transcriptome atlases of CD4+ T cells from SLE patients and healthy controls and found that chronic IFN signaling pathway induction may induce peripheral Treg cell exhaustion in SLE patients. Our study provides a rich source of data offering an epigenetic and transcriptomic view of CD4+ T cell-related functions in terms of SLE clinical manifestations and immunopathogenesis, supporting a deeper understanding of this autoimmune disease.
使用糖皮质激素或其他免疫抑制剂仍然是 SLE 管理的主要手段 ( Kiriakidou, 2013 )。然而,这些药物具有实质性的不良反应,通常不会赋予足够的治疗效果 ( Lisnevskaia et al., 2014 )。anifrolumab 是一种靶向 I 型 IFN 受体的人单克隆抗体,最近的一项临床试验显示,在治疗 SLE 方面具有显着的治疗效果 ( Morand et al., 2020 )。我们的结果还支持 I 型 IFN 作为治疗严重 SLE 患者的有效药物靶点,并为这种疾病提供了潜在的治疗策略,即抑制具有疲惫样特性的 Treg。总体而言,我们描绘了 SLE 患者和健康对照者 CD4+ T 细胞的染色质可及性和单细胞转录组图谱,发现慢性 IFN 信号通路诱导可能诱导 SLE 患者的外周 Treg 细胞耗竭。我们的研究提供了丰富的数据来源,提供了 CD4+ T 细胞相关功能的表观遗传学和转录组学观点,包括 SLE 临床表现和免疫发病机制,支持对这种自身免疫性疾病的更深入理解。
使用糖皮质激素或其他免疫抑制剂仍然是 SLE 管理的主要手段 ( Kiriakidou, 2013 )。然而,这些药物具有实质性的不良反应,通常不会赋予足够的治疗效果 ( Lisnevskaia et al., 2014 )。anifrolumab 是一种靶向 I 型 IFN 受体的人单克隆抗体,最近的一项临床试验显示,在治疗 SLE 方面具有显着的治疗效果 ( Morand et al., 2020 )。我们的结果还支持 I 型 IFN 作为治疗严重 SLE 患者的有效药物靶点,并为这种疾病提供了潜在的治疗策略,即抑制具有疲惫样特性的 Treg。总体而言,我们描绘了 SLE 患者和健康对照者 CD4+ T 细胞的染色质可及性和单细胞转录组图谱,发现慢性 IFN 信号通路诱导可能诱导 SLE 患者的外周 Treg 细胞耗竭。我们的研究提供了丰富的数据来源,提供了 CD4+ T 细胞相关功能的表观遗传学和转录组学观点,包括 SLE 临床表现和免疫发病机制,支持对这种自身免疫性疾病的更深入理解。
Limitations of the study 研究的局限性
Although there have been a few papers supporting an exhausted phenotype in human Treg cells (Lowther et al., 2016; Yang et al., 2017), the concept of Treg exhaustion has not yet been widely recognized. A major limitation of this study is the lack of protein-level intervention experiments, such as knockout or CRISPR of exhaustion-related genes such as PD-1, IFNAR1, and LKB1 in primary cells, and test the cellular functions thereafter. Such experiments are technically not implemented due to the limited number of CCR7lowCD74hi Treg cells and the fragile state of human primary cells. Thereby, further studies are necessary to uncover the molecular features, particularly the exhaustion-like properties, of the human CCR7lowCD74hi Treg cells.
尽管有一些论文支持人类 Treg 细胞中的耗竭表型 ( Lowther et al., 2016 ; Yang et al., 2017 ),Treg 耗竭的概念尚未得到广泛认可。本研究的一个主要局限性是缺乏蛋白质水平的干预实验,例如对原代细胞中 PD-1 、 IFNAR1 和 LKB1 等耗竭相关基因的敲除或 CRISPR,以及此后测试细胞功能。由于 CCR7低CD74hi Treg 细胞的数量有限且人类原代细胞的脆弱状态,因此在技术上未实施此类实验。因此,需要进一步的研究来揭示人 CCR7低CD74hi Treg 细胞的分子特征,特别是耗竭样特性。
尽管有一些论文支持人类 Treg 细胞中的耗竭表型 ( Lowther et al., 2016 ; Yang et al., 2017 ),Treg 耗竭的概念尚未得到广泛认可。本研究的一个主要局限性是缺乏蛋白质水平的干预实验,例如对原代细胞中 PD-1 、 IFNAR1 和 LKB1 等耗竭相关基因的敲除或 CRISPR,以及此后测试细胞功能。由于 CCR7低CD74hi Treg 细胞的数量有限且人类原代细胞的脆弱状态,因此在技术上未实施此类实验。因此,需要进一步的研究来揭示人 CCR7低CD74hi Treg 细胞的分子特征,特别是耗竭样特性。
STAR★Methods STAR★方法
Key resources table 关键资源表
REAGENT or RESOURCE 试剂或资源 | SOURCE | IDENTIFIER |
---|---|---|
Antibodies 抗体 | ||
APC anti-human CD196 (CCR6) Antibody APC 抗人 CD196 (CCR6) 抗体 | Biolegend 生物传奇 | cat#353416; RRID:AB_10945155 猫#353416;RRID:AB_10945155 |
APC anti-human CD197 (CCR7) Antibody APC 抗人 CD197 (CCR7) 抗体 | Biolegend 生物传奇 | cat#353214; RRID:AB_10917387 猫#353214;RRID:AB_10917387 |
APC anti-human CD25 Antibody APC 抗人 CD25 抗体 | eBioscience 电子生物科学 | cat#17-0259-42; RRID:AB_1582219 货号 #17-0259-42;RRID:AB_1582219 |
APC/Cyanine7 anti-human CD45 Antibody APC/Cyanine7 抗人 CD45 抗体 | Biolegend 生物传奇 | cat#304014; RRID:AB_314402 猫#304014;RRID:AB_314402 |
Brilliant Violet 421™ Anti-human CD127 Antibody | Biolegend | cat#562437; RRID:AB_11153481 |
Brilliant Violet 421™ anti-human CD197 (CCR7) Antibody | Biolegend | cat#353207; RRID:AB_10915137 |
FITC anti-human CD4 Antibody | Biolegend | cat#300538; RRID:AB_2562052 |
PE anti-human CD127 Antibody | eBioscience | cat#12-1278-41; RRID:AB_10853334 |
PE anti-human CD185 (CXCR5) Antibody | Biolegend | cat#356904; RRID:AB_2561813 |
PE anti-human CD25 Antibody | BD Bioscience | cat#560989; RRID:AB_10563905 |
PE anti-human FOXP3 Antibody | Biolegend | cat#320208; RRID:AB_492982 |
PE/Cyanine7 anti-human CD183 (CXCR3) Antibody | Biolegend | cat#353720; RRID:AB_11219383 |
PE/Cyanine7 anti-human CD74 Antibody | Biolegend | cat#357609; RRID:AB_2721663 |
PE/Cyanine7 Anti-Human CD45RA Antibody | Biolegend | cat#304125; RRID:AB_10709440 |
PerCP/Cyanine5.5 anti-human CD3 Antibody | Biolegend | cat#344808; RRID:AB_10640736 |
Chemicals, peptides, and recombinant proteins | ||
penicillin/streptomycin | Gibco | cat#15140122 |
Glutamine | Gibco | cat#25030081 |
sodium pyruvate | Thermo Fisher Scientific | cat#11360070 |
nonessential amino acids | Solarbio | cat#N1250 |
human recombinant IFN-α | Abcam | cat#ab48750 |
Critical commercial assays | ||
10X Chromium Single Cell 3′ Kit | 10X Genomics | cat#120237 |
CD4+CD25+ Regulatory T Cell Isolation Kit | Miltenyi Biotec | cat#130091301 |
CellTrace™ CFSE Cell Proliferation Kit | Thermo Fisher Scientific | cat#C34554 |
Dynabeads™ Human T-Activator CD3/CD28 | Invitrogen | cat#11161D |
Roswell Park Memorial Institute (RPMI) 1640 medium | HyClone | cat#SH30809.01 |
Maxima H Minus Reverse Transcriptase | Thermo Fisher Scientific | cat#EP0751 |
SYBR Green PCR Master Mix | Applied Biosystems | cat#4344463 |
Tn5 transposome | Vazyme Biotech | cat#TD501 |
Nonidet P40 Substitute | Roche | cat#11332473001 |
MinElute PCR Purification Kit | Qiagen | cat#28006 |
Foxp3/Transcription Factor Staining Buffer | eBioscience | cat#00-5523-00 |
SPRIselect beads | Beckman Coulter | cat#B23318 |
foetal calf serum | Gibco | cat#16170078 |
Deposited data | ||
MSigDB pathways | Liberzon et al., 2015 | http://software.broadinstitute.org/gsea/msigdb/index.jsp |
Microarray data of blood CD4+ T from SLE patients and controls | N/A | Gene Expression Omnibus (GEO) (GSE4588) |
Microarray data of healthy and SLE blood CD4+ T | Sharma et al., 2015 | Gene Expression Omnibus (GEO) (GSE55447) |
Microarray data of healthy and SLE blood CD4+ T | Hutcheson et al., 2008 | Gene Expression Omnibus (GEO) (GSE10325) |
Microarray data of CD4+ T cell subsets | Takeshita et al., 2015 | Gene Expression Omnibus (GEO) (GSE61697) |
RNA-seq data of three Treg groups (nTreg, eTreg, Fr. III) | Cuadrado et al., 2018 | Gene Expression Omnibus (GEO) (GSE90600) |
Single cell RNA-seq data of colon mucosa of UC patients and controls | Smillie et al., 2019 | Single Cell Portal: SCP259 |
Single-cell RNA-seq data of PBMCs from UC patients and controls | Boland et al., 2020 | Gene Expression Omnibus (GEO) (GSE125527) |
Single-cell RNA-seq data of PBMCs from healthy controls | Wilk et al., 2020 | Gene Expression Omnibus (GEO) (GSE150728) |
Single-cell RNA-seq data of PBMCs from SLE patients | Schafflick et al., 2020 | Gene Expression Omnibus (GEO) (GSE137029) |
Single-cell RNA-seq data of PBMCs from multiple sclerosis patients | Schafflick et al., 2020 | Gene Expression Omnibus (GEO) (GSE138266) |
ATAC-seq and scRNA-seq of CD4+ T cells from SLE patients and healthy controls | This paper | Genome Sequence Archive (GSA) for Humans (HRA000676) |
Bulk RNA-seq of Treg cells stimulated with/without IFN-α | This paper | Genome Sequence Archive (GSA) for Humans (HRA000676) |
Software and algorithms | ||
Bowtie2 | Langmead and Salzberg, 2012 | http://bowtie-bio.sourceforge.net/bowtie2/index.shtml |
DESeq | Anders and Huber, 2010 | http://bioconductor.org/packages/release/bioc/html/DESeq.html |
BedTools | Quinlan and Hall, 2010 | http://bedtools.readthedocs.io/en/latest/ |
ATAC-pipe | Zuo et al., 2019 | https://github.com/QuKunLab/ATAC-pipe |
MetaScape | Zhou et al., 2019 | http://metascape.org/gp/index.html#/main/step1 |
GREAT | McLean et al., 2010 | http://bejerano.stanford.edu/great/public/html/ |
GSEA | Mootha et al., 2003 | https://www.gsea-msigdb.org/gsea/index.jsp |
Demuxlet | Kang et al., 2018 | https://github.com/statgen/demuxlet |
Souporcell | Heaton et al., 2020 | https://github.com/wheaton5/souporcell |
CellRanger V2.0 | 10X Genomics | https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/what-is-cell-ranger |
Seurat V3.0 (R package) | Stuart et al., 2019 | https://satijalab.org/seurat/ |
SAVER (R package) | Huang et al., 2018 | https://github.com/mohuangx/SAVER |
Code in this study | This paper | https://github.com/QuKunLab/SLE; https://zenodo.org/record/7118733. (https://doi.org/10.5281/zenodo.7118733) |
Resource availability 资源可用性
Lead contact 牵头联系人
Further information and requests for resources should be directed to and will be fulfilled by lead contact, Kun Qu (qukun@ustc.edu.cn).
更多信息和资源请求应直接发送至首席联系人 Kun Qu (qukun@ustc.edu.cn) 并由其提供。
更多信息和资源请求应直接发送至首席联系人 Kun Qu (qukun@ustc.edu.cn) 并由其提供。
Materials availability 材料可用性
All non-commercial reagents used in this paper are available from the lead contact upon request.
本文中使用的所有非商业试剂均可根据要求从主要联系人处获得。
本文中使用的所有非商业试剂均可根据要求从主要联系人处获得。
Experimental model and subject details
实验模型和主题详细信息
Human subjects 人类受试者
Peripheral blood samples were collected from SLE patients and healthy controls at the First Affiliated Hospital of University of Science and Technology of China. Informed consent was obtained from the patients. Study procedures were followed in accordance with protocols approved by the ethics committee of the University of Science and Technology of China. Detailed clinical information for the patients is described in Table S1. The 24 clinical and laboratory parameters for SLE patients were collected according to the internationally certified method SLEDAI-2K (Gladman et al., 2002).
从中国科学技术大学第一附属医院的 SLE 患者和健康对照者收集外周血样本。获得患者的知情同意。根据中国科学技术大学伦理委员会批准的方案遵循研究程序。患者的详细临床信息见表 S1。根据国际认证的方法 SLEDAI-2K 收集 SLE 患者的 24 个临床和实验室参数 ( Gladman et al., 2002 )。
从中国科学技术大学第一附属医院的 SLE 患者和健康对照者收集外周血样本。获得患者的知情同意。根据中国科学技术大学伦理委员会批准的方案遵循研究程序。患者的详细临床信息见表 S1。根据国际认证的方法 SLEDAI-2K 收集 SLE 患者的 24 个临床和实验室参数 ( Gladman et al., 2002 )。
The ATAC-seq data of CD4+ T cells were obtained from a cohort of 56 female and 7 male SLE patients, with a mean age of 35.48; the scRNA-seq data of CD4+ T cells were obtained from a cohort of 10 female SLE patients, with a mean age of 31.33; the FACS experiment of CD4+ T cells was obtained from a cohort of 27 female SLE patients, with a mean age of 39.5; the ATAC-seq data of Treg cells were obtained from a cohort of 5 female SLE patients, with a mean age of 41; the quantitive PCR experiment of Treg cells was obtained from a cohort of 8 female SLE patients, with a mean age of 37.3; the assay of the suppressive function for Treg cells was obtained from a cohort of 4 female SLE patients, with a mean age of 28.75. Detailed information is also available in Table S1.
CD4+ T 细胞的 ATAC-seq 数据来自 56 例女性和 7 例男性 SLE 患者,平均年龄为 35.48 岁;CD4+ T 细胞的 scRNA-seq 数据来自 10 例女性 SLE 患者队列,平均年龄为 31.33;CD4+ T 细胞的 FACS 实验来自 27 例女性 SLE 患者队列,平均年龄为 39.5;Treg 细胞的 ATAC-seq 数据来自一组 5 例女性 SLE 患者,平均年龄为 41 岁;Treg 细胞的定量 PCR 实验来自 8 例女性 SLE 患者队列,平均年龄为 37.3;Treg 细胞抑制功能的测定是从 4 名女性 SLE 患者的队列中获得的,平均年龄为 28.75 岁。表 S1 中也提供了详细信息。
CD4+ T 细胞的 ATAC-seq 数据来自 56 例女性和 7 例男性 SLE 患者,平均年龄为 35.48 岁;CD4+ T 细胞的 scRNA-seq 数据来自 10 例女性 SLE 患者队列,平均年龄为 31.33;CD4+ T 细胞的 FACS 实验来自 27 例女性 SLE 患者队列,平均年龄为 39.5;Treg 细胞的 ATAC-seq 数据来自一组 5 例女性 SLE 患者,平均年龄为 41 岁;Treg 细胞的定量 PCR 实验来自 8 例女性 SLE 患者队列,平均年龄为 37.3;Treg 细胞抑制功能的测定是从 4 名女性 SLE 患者的队列中获得的,平均年龄为 28.75 岁。表 S1 中也提供了详细信息。
Method details 方法详细信息
Cell isolation 细胞分离
Peripheral blood was drawn in a green-top blood collection tube. PBMCs were then prepared by Ficoll-Paque density gradient centrifugation and stained with fluorochrome-labelled anti-human monoclonal antibodies (Biolegend). Bulk CD4+ T cells were sorted with CD45 (clone HI30), CD4 (clone RPA-T4), and CD3 (SK7) using a SH800S flow cytometer (SONY). For CD4+ T helper cell subtypes, cells were identified as previously described (Morita et al., 2011). Briefly, naïve cells were identified as CD4+ CCR7+ CD45RA+, CXCR5+ T cells as CD4+ CD25- CD45RA− CXCR5+, Th1 cells as CD4+ CD25- CD45RA−CXCR3+ CCR6-, Th2 cells as CD4+ CD25- CD45RA− CXCR3- CCR6-, Th17 cells as CD4+ CD25- CD45RA− CXCR3- CCR6+, Treg cells as CD4+ FOXP3+, and Tct cells as CD4+ CD25- CD45RA+ CCR7-. The post-sort purities were confirmed to be >95% prior to ATAC-seq. For single-cell RNA-seq, CD4+ T cells were sorted and cryopreserved according to the 10X Genomics official recommendation.
外周血在绿顶采血管中抽取。然后通过 Ficoll-Paque 密度梯度离心制备 PBMC,并用荧光染料标记的抗人单克隆抗体 (Biolegend) 染色。使用 SH800S 流式细胞仪 (SONY) 对大量 CD4+ T 细胞进行 CD45 (克隆 HI30)、CD4 (克隆 RPA-T4) 和 CD3 (SK7) 分选。对于 CD4 + 辅助性 T 细胞亚型,如前所述鉴定细胞 ( Morita et al., 2011 )。简而言之,初始细胞被鉴定为 CD4+ CCR7+ CD45RA+,CXCR5+ T 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR5+,Th1 细胞被鉴定为 CD4+ CD25- CD45RA−CXCR3+ CCR6-,Th2 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR3- CCR6-,Th17 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR3- CCR6+,Treg 细胞作为 CD4+ FOXP3+,Tct 细胞作为 CD4+ CD25- CD45RA+ CCR7-。在 ATAC-seq 之前,分选后纯度被确认为 >95%。对于单细胞 RNA-seq,根据 10X Genomics 官方建议对 CD4+ T 细胞进行分选和冻存。
外周血在绿顶采血管中抽取。然后通过 Ficoll-Paque 密度梯度离心制备 PBMC,并用荧光染料标记的抗人单克隆抗体 (Biolegend) 染色。使用 SH800S 流式细胞仪 (SONY) 对大量 CD4+ T 细胞进行 CD45 (克隆 HI30)、CD4 (克隆 RPA-T4) 和 CD3 (SK7) 分选。对于 CD4 + 辅助性 T 细胞亚型,如前所述鉴定细胞 ( Morita et al., 2011 )。简而言之,初始细胞被鉴定为 CD4+ CCR7+ CD45RA+,CXCR5+ T 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR5+,Th1 细胞被鉴定为 CD4+ CD25- CD45RA−CXCR3+ CCR6-,Th2 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR3- CCR6-,Th17 细胞被鉴定为 CD4+ CD25- CD45RA− CXCR3- CCR6+,Treg 细胞作为 CD4+ FOXP3+,Tct 细胞作为 CD4+ CD25- CD45RA+ CCR7-。在 ATAC-seq 之前,分选后纯度被确认为 >95%。对于单细胞 RNA-seq,根据 10X Genomics 官方建议对 CD4+ T 细胞进行分选和冻存。
Intracellular staining 细胞内染色
PBMCs were prepared and stained using the indicated human mAbs. Homologous IgGs served as the negative control. FACS surface marker staining was performed according to the Biolegend antibody instructions. For intracellular staining, cells were blocked, stained with Foxp3 (clone 259D) or CD74 (clone pin.1) and then washed with Foxp3/Transcription Factor Staining Buffer (eBioscience, cat no. 00-5523-00) according to the manufacturer’s instructions.
制备 PBMC 并使用指定的人 mAb 染色。同源 IgG 用作阴性对照。根据 Biolegend 抗体说明进行 FACS 表面标志物染色。对于细胞内染色,细胞被封闭,用 Foxp3(克隆 259D)或 CD74(克隆针.1)染色,然后根据制造商的说明用 Foxp3/转录因子染色缓冲液(eBioscience,货号 00-5523-00)洗涤。
制备 PBMC 并使用指定的人 mAb 染色。同源 IgG 用作阴性对照。根据 Biolegend 抗体说明进行 FACS 表面标志物染色。对于细胞内染色,细胞被封闭,用 Foxp3(克隆 259D)或 CD74(克隆针.1)染色,然后根据制造商的说明用 Foxp3/转录因子染色缓冲液(eBioscience,货号 00-5523-00)洗涤。
Single-cell RNA-seq library preparation and sequencing
单细胞 RNA-seq 文库制备和测序
Single-cell suspensions were prepared as described in the 10X Genomics protocol. Briefly, we sorted CD4+ T cells from 10 SLE patients and 6 healthy controls. The cells reached a final viability of 85%. We then resuspended the cells at a concentration of 700 cells/μL and mixed the same sample types immediately according to the 10X Genomics Chromium single-cell protocol for the v2 reagent kit (10X Genomics). Cell suspensions were loaded onto a chromium single-cell chip along with reverse transcription (RT) master mix and 3′ gel beads. After generation of single-cell gel beads in emulsion (GEMs). RT was performed using a C1000 Touch™ Thermal Cycler (Bio-Rad) using the manufacturer’s standard parameters. cDNA was amplified and purified with SPRIselect beads (Beckman Coulter). Single-cell 3′ libraries were then constructed following fragmentation, end repair, polyA tailing, adaptor ligation and size selection. Single-cell sequencing libraries were generated with one sample index for each sample and sequenced on the Illumina HiSeq X-Ten platform.
按照 10X Genomics 实验步骤中的说明制备单细胞悬液。简而言之,我们对 10 名 SLE 患者和 6 名健康对照者的 CD4+ T 细胞进行了分选。细胞的最终活力达到 85%。然后,我们以 700 个细胞/μL 的浓度重悬细胞,并根据 v2 试剂盒 (10X Genomics) 的 10X Genomics Chromium 单细胞实验方案立即混合相同的样品类型。将细胞悬液与逆转录 (RT) 预混液和 3′ 凝胶珠一起加载到铬单细胞芯片上。在乳液 (GEM) 中生成单细胞凝胶珠后。使用 C1000 Touch™ 热循环仪 (Bio-Rad) 使用制造商的标准参数进行 RT。用 SPRIselect 珠子 (Beckman Coulter) 扩增和纯化 cDNA。然后在片段化、末端修复、polyA 拖尾、衔接子连接和大小选择后构建单细胞 3' 文库。生成单细胞测序文库,每个样本有一个样本索引,并在 Illumina HiSeq X-Ten 平台上进行测序。
按照 10X Genomics 实验步骤中的说明制备单细胞悬液。简而言之,我们对 10 名 SLE 患者和 6 名健康对照者的 CD4+ T 细胞进行了分选。细胞的最终活力达到 85%。然后,我们以 700 个细胞/μL 的浓度重悬细胞,并根据 v2 试剂盒 (10X Genomics) 的 10X Genomics Chromium 单细胞实验方案立即混合相同的样品类型。将细胞悬液与逆转录 (RT) 预混液和 3′ 凝胶珠一起加载到铬单细胞芯片上。在乳液 (GEM) 中生成单细胞凝胶珠后。使用 C1000 Touch™ 热循环仪 (Bio-Rad) 使用制造商的标准参数进行 RT。用 SPRIselect 珠子 (Beckman Coulter) 扩增和纯化 cDNA。然后在片段化、末端修复、polyA 拖尾、衔接子连接和大小选择后构建单细胞 3' 文库。生成单细胞测序文库,每个样本有一个样本索引,并在 Illumina HiSeq X-Ten 平台上进行测序。
ATAC-seq library preparation and sequencing
ATAC-seq 文库制备和测序
ATAC-seq of CD4+ T cells was performed as previously described with minor modifications. Briefly, CD4+ T cells were sorted using a SH800S sorter (SONY). Samples were lysed in cold lysis buffer (10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl2 and 0.1% NP-40 (Roche)) for 3 min on ice to prepare the nuclei. Immediately after cell lysis, nuclei were centrifuged at 500 g for 5 min, and the supernatant was discarded carefully. Nuclei extracts were then incubated with Tn5 transposome (Vazyme Biotech, cat no. TD501) at 37°C for 30 min. After DNA purification with the MinElute Kit (Qiagen), PCR was performed to amplify the library for 12–15 cycles based on the quantitative data regarding the optimum number of PCR cycles. The PCR conditions were as follows: 98°C for 30 sec and cycling at 98°C for 10 sec, 63°C for 30 sec and 72°C for 1 min. After PCR amplification, the sample libraries were purified and sequenced on the Illumina HiSeq X-Ten platform with the 150-bp paired-end configuration.
如前所述进行 CD4 + T 细胞的 ATAC-seq,并稍作修改。简而言之,使用 SH800S 分选仪 (SONY) 对 CD4 + T 细胞进行分选。将样品在冷裂解缓冲液(10 mM Tris-HCl (pH 7.4)、10 mM NaCl、3 mM MgCl2 和 0.1% NP-40 (Roche)中在冰上裂解 3 分钟以制备细胞核。细胞裂解后,立即将细胞核以 500 g 离心 5 分钟,小心弃去上清液。然后将细胞核提取物与 Tn5 转座子酶(Vazyme Biotech,货号。TD501) 在 37°C 下放置 30 分钟。使用 MinElute 试剂盒 (Qiagen) 纯化 DNA 后,根据有关最佳 PCR 循环数的定量数据,进行 PCR 以扩增文库 12–15 个循环。PCR 条件如下:98°C 30 秒,在 98°C 下循环 10 秒,在 63°C 下循环 30 秒,在 72°C 下循环 1 分钟。PCR 扩增后,在 Illumina HiSeq X-Ten 平台上以 150 bp 双端配置纯化和测序样本文库。
如前所述进行 CD4 + T 细胞的 ATAC-seq,并稍作修改。简而言之,使用 SH800S 分选仪 (SONY) 对 CD4 + T 细胞进行分选。将样品在冷裂解缓冲液(10 mM Tris-HCl (pH 7.4)、10 mM NaCl、3 mM MgCl2 和 0.1% NP-40 (Roche)中在冰上裂解 3 分钟以制备细胞核。细胞裂解后,立即将细胞核以 500 g 离心 5 分钟,小心弃去上清液。然后将细胞核提取物与 Tn5 转座子酶(Vazyme Biotech,货号。TD501) 在 37°C 下放置 30 分钟。使用 MinElute 试剂盒 (Qiagen) 纯化 DNA 后,根据有关最佳 PCR 循环数的定量数据,进行 PCR 以扩增文库 12–15 个循环。PCR 条件如下:98°C 30 秒,在 98°C 下循环 10 秒,在 63°C 下循环 30 秒,在 72°C 下循环 1 分钟。PCR 扩增后,在 Illumina HiSeq X-Ten 平台上以 150 bp 双端配置纯化和测序样本文库。
Assessment of the inhibitory function of Treg cells
Treg 细胞抑制功能的评估
For functional analysis, CD4+CD25+ T cells and CD4+CD25- T cells were purified from PBMCs using a CD4+CD25+ Regulatory T Cell Isolation Kit (Miltenyi Biotec, cat no. 130091301). Briefly, a non-CD4+ cocktail and anti-biotin beads were used for the isolation of CD4+ T cells. After detachment, the cells were washed, and CD4+CD25+ Treg cells were positively selected using CD25 microbeads. The cells were reanalysed after sorting and routinely showed a purity greater than 95%. The negative fraction of CD4+CD25- T cells from healthy controls served as effector cells and were stained with a CellTrace™ CFSE Cell Proliferation Kit (Thermo Fisher Scientific, cat no. C34554) at 1 μm for 15 min. Then, CD4+CD25- T cells (2∗10∧4) were incubated with Dynabeads™ Human T-Activator CD3/CD28 (Invitrogen, cat no. 11161D) at a 20:1 ratio in Roswell Park Memorial Institute (RPMI) 1640 medium (HyClone, cat no. SH30809.01) supplemented with 10% foetal calf serum (Gibco, cat no. 16170078), 100 U/mL penicillin/streptomycin (Gibco, cat no. 15140122), 2 mM glutamine (Gibco, cat no. 25030081), sodium pyruvate (Thermo Fisher Scientific, cat no. 11360070) and nonessential amino acids (Solarbio, cat no. N1250). Purified CD4+CD25+ Treg cells from healthy controls or SLE patients were added to the culture at a 1:0 or 1:1 ratio. After 4 days, the proliferation of CD4+CD25- T cells was determined by assessing CFSE dilution by flow cytometry. The data are expressed as the percent inhibition of cell proliferation according to the following formula: inhibition% = 100-(cell proliferation ratio at 1:1/cell proliferation ratio at 1:0).
对于功能分析,使用 CD4+CD25+ 调节性 T 细胞分离试剂盒(Miltenyi Biotec,货号 130091301)从 PBMC 中纯化 CD4+CD25+ T 细胞和 CD4+CD25-T 细胞。简而言之,使用非 CD4 + 混合物和抗生物素微珠分离 CD4 + T 细胞。分离后,洗涤细胞,并使用 CD25 微珠阳性选择 CD4 + CD25 + Treg 细胞。分选后对细胞进行重新分析,通常显示纯度大于 95%。来自健康对照的 CD4+CD25-T 细胞的阴性组分用作效应细胞,用 1 μm 的 CellTrace™ CFSE 细胞增殖试剂盒(Thermo Fisher Scientific,货号 C34554)染色 15 分钟。然后,将 CD4+CD25-T 细胞 (2∗10∧4) 与 Dynabeads™ 人 T 活化剂 CD3/CD28(Invitrogen,货号 11161D)以 20:1 的比例在罗斯威尔公园纪念研究所 (RPMI) 1640 培养基(HyClone,货号 11161D)中以 20:1 的比例孵育。SH30809.01) 补充有 10% 胎牛血清(Gibco,货号 16170078)、100 U/mL 青霉素/链霉素(Gibco,货号 15140122)、2 mM 谷氨酰胺(Gibco,货号 25030081)、丙酮酸钠(Thermo Fisher Scientific,货号 11360070)和非必需氨基酸(Solarbio,货号 N1250)。将来自健康对照或 SLE 患者的纯化 CD4 + CD25 + Treg 细胞以 1:0 或 1:1 的比例添加到培养物中。4 天后,通过流式细胞术评估 CFSE 稀释度测定 CD4+CD25-T 细胞的增殖。 根据以下公式,数据表示为细胞增殖的抑制百分比:抑制% = 100 - (1:1时的细胞增殖比/1:0时的细胞增殖比)。
对于功能分析,使用 CD4+CD25+ 调节性 T 细胞分离试剂盒(Miltenyi Biotec,货号 130091301)从 PBMC 中纯化 CD4+CD25+ T 细胞和 CD4+CD25-T 细胞。简而言之,使用非 CD4 + 混合物和抗生物素微珠分离 CD4 + T 细胞。分离后,洗涤细胞,并使用 CD25 微珠阳性选择 CD4 + CD25 + Treg 细胞。分选后对细胞进行重新分析,通常显示纯度大于 95%。来自健康对照的 CD4+CD25-T 细胞的阴性组分用作效应细胞,用 1 μm 的 CellTrace™ CFSE 细胞增殖试剂盒(Thermo Fisher Scientific,货号 C34554)染色 15 分钟。然后,将 CD4+CD25-T 细胞 (2∗10∧4) 与 Dynabeads™ 人 T 活化剂 CD3/CD28(Invitrogen,货号 11161D)以 20:1 的比例在罗斯威尔公园纪念研究所 (RPMI) 1640 培养基(HyClone,货号 11161D)中以 20:1 的比例孵育。SH30809.01) 补充有 10% 胎牛血清(Gibco,货号 16170078)、100 U/mL 青霉素/链霉素(Gibco,货号 15140122)、2 mM 谷氨酰胺(Gibco,货号 25030081)、丙酮酸钠(Thermo Fisher Scientific,货号 11360070)和非必需氨基酸(Solarbio,货号 N1250)。将来自健康对照或 SLE 患者的纯化 CD4 + CD25 + Treg 细胞以 1:0 或 1:1 的比例添加到培养物中。4 天后,通过流式细胞术评估 CFSE 稀释度测定 CD4+CD25-T 细胞的增殖。 根据以下公式,数据表示为细胞增殖的抑制百分比:抑制% = 100 - (1:1时的细胞增殖比/1:0时的细胞增殖比)。
In vitro stimulation of Treg with IFN-α
用 IFN-α 体外刺激 Treg
PBMCs from healthy controls were isolated by Ficoll-Paque gradient centrifugation and then stimulated with human recombinant IFN-α(0 μg/mL, 0.25 μg/mL, 0.5 μg/mL, 1 μg/mL) (Abcam cat no. ab48750). The cells were cultured for 7 days in RPMI 1640 medium supplemented with 10% heat-inactivated human serum or 10% SLE serum. After 7 days, cells were harvested and isolated by flow cytometry using the same sorting strategy as that used for CD4+CD25+CD127lowCCR7low Treg2 cells. Finally, RNA-seq library construction and sequencing were performed using the same strategies described above.
通过 Ficoll-Paque 梯度离心分离来自健康对照的 PBMC,然后用人重组 IFN-α(0 μg/mL、0.25 μg/mL、0.5 μg/mL、1 μg/mL)(Abcam 货号 ab48750)刺激。将细胞在补充有 10% 热灭活人血清或 10% SLE 血清的 RPMI 1640 培养基中培养 7 天。7 天后,收获细胞并通过流式细胞术分离细胞,使用与 CD4+CD25+CD127低CCR7低 Treg2 细胞相同的分选策略。最后,使用上述相同的策略进行 RNA-seq 文库构建和测序。
通过 Ficoll-Paque 梯度离心分离来自健康对照的 PBMC,然后用人重组 IFN-α(0 μg/mL、0.25 μg/mL、0.5 μg/mL、1 μg/mL)(Abcam 货号 ab48750)刺激。将细胞在补充有 10% 热灭活人血清或 10% SLE 血清的 RPMI 1640 培养基中培养 7 天。7 天后,收获细胞并通过流式细胞术分离细胞,使用与 CD4+CD25+CD127低CCR7低 Treg2 细胞相同的分选策略。最后,使用上述相同的策略进行 RNA-seq 文库构建和测序。
Bulk RNA-seq library preparation and sequencing
大量 RNA-seq 文库制备和测序
Treg (DAPI−CD45+CD3+CD4+CD25+CD127low) and CCR7- Treg2 (DAPI−CD45+CD3+CD4+CD25+CD127lowCCR7low) cells were sorted by flow cytometry. For Treg and CCR7- Treg2 cells, up to 1000 cells were collected directly in a 0.2 mL PCR tube (KIRGEN, Cat No. KG2511), and the RNA-seq library was constructed using the Smart-seq2 method (Picelli et al., 2014) and sequenced on the Illumina NovaSeq 6000 system; at least 20 million paired reads were generated per sample.
通过流式细胞术对 Treg (DAPI−CD45+CD3+CD4+CD25+CD127低)和 CCR7-Treg2 (DAPI-CD45+CD3+CD4+CD25+CD127低CCR7低)细胞进行分选。对于 Treg 和 CCR7-Treg2 细胞,将多达 1000 个细胞直接收集在 0.2 mL PCR 管中(KIRGEN,货号。KG2511),使用 Smart-seq2 方法 ( Picelli et al., 2014 ) 构建 RNA-seq 文库,并在 Illumina NovaSeq 6000 系统上测序;每个样本至少生成 2000 万个配对读数。
通过流式细胞术对 Treg (DAPI−CD45+CD3+CD4+CD25+CD127低)和 CCR7-Treg2 (DAPI-CD45+CD3+CD4+CD25+CD127低CCR7低)细胞进行分选。对于 Treg 和 CCR7-Treg2 细胞,将多达 1000 个细胞直接收集在 0.2 mL PCR 管中(KIRGEN,货号。KG2511),使用 Smart-seq2 方法 ( Picelli et al., 2014 ) 构建 RNA-seq 文库,并在 Illumina NovaSeq 6000 系统上测序;每个样本至少生成 2000 万个配对读数。
Real-time quantitative polymerase chain reaction
实时定量聚合酶链反应
The selected candidate genes were validated by qPCR. Briefly, cDNA was synthesized with Maxima H Minus Reverse Transcriptase (Thermo Fisher Scientific, cat no. EP0751) in accordance with the manufacturer’s instructions. Two-step PCR was performed using SYBR Green PCR Master Mix (Applied Biosystems, cat no. 4344463) in accordance with the manufacturer’s instructions on a LightCycler96 fluorescence sequence detection system (Roche). Gene expression was quantified relative to that of the housekeeping gene GAPDH and normalized to the control by the standard 2-ΔΔCT calculation. The primer sequences were as follows: GAPDH, 5′- GGAGCGAGATCCCTCCAAAAT-3′ and 5′- GGCTGTTGTCATACTTCTCATGG-3′; PDCD1, 5′-ACGAGGGACAATAGGAGCCA-3′ and 5′-GGCATACTCCGTCTGCTCAG-3′; LAG3, 5′-GCCTCCGACTGGGTCATTTT-3′ and 5′- CTTTCCGCTAAGTGGTGATGG-3′; CTLA4, 5′-CATGATGGGGAATGAGTTGACC-3′ and 5′- TCAGTCCTTGGATAGTGAGGTTC-3′; TIGIT, 5′-TGGTCGCGTTGACTAGAAAGA-3′ and 5′- GGGCTCCATTCCTCCTGTC-3′.
所选候选基因通过 qPCR 验证。简而言之,用 Maxima H Minus 逆转录酶(Thermo Fisher Scientific,货号。EP0751) 的 Alpha S S Package。根据制造商在 LightCycler96 荧光序列检测系统 (Roche) 上的说明,使用 SYBR Green PCR 预混液(Applied Biosystems,货号 4344463)进行两步法 PCR。相对于看家基因 GAPDH 的基因表达进行定量,并通过标准 2-ΔΔCT 计算归一化为对照。引物序列如下:GAPDH、5′-GGAGCGAGATCCCTCCAAAAT-3′ 和 5′-GGCTGTTGTCATACTTCTCATGG-3′;PDCD1、5'-ACGAGGGACAATAGGAGCCA-3' 和 5'-GGCATACTCCGTCTGCTCAG-3';LAG3, 5'-GCCTCCGACTGGGTCATTTT-3' 和 5'- CTTTCCGCTAAGTGGTGATGG-3';CTLA4, 5'-CATGATGGGGAATGAGTTGACC-3' 和 5'-TCAGTCCTTGGATAGTGAGGTTC-3';TIGIT, 5'-TGGTCGCGTTGACTAGAAAGA-3' 和 5'- GGGCTCCATTCCTCCTGTC-3'。
所选候选基因通过 qPCR 验证。简而言之,用 Maxima H Minus 逆转录酶(Thermo Fisher Scientific,货号。EP0751) 的 Alpha S S Package。根据制造商在 LightCycler96 荧光序列检测系统 (Roche) 上的说明,使用 SYBR Green PCR 预混液(Applied Biosystems,货号 4344463)进行两步法 PCR。相对于看家基因 GAPDH 的基因表达进行定量,并通过标准 2-ΔΔCT 计算归一化为对照。引物序列如下:GAPDH、5′-GGAGCGAGATCCCTCCAAAAT-3′ 和 5′-GGCTGTTGTCATACTTCTCATGG-3′;PDCD1、5'-ACGAGGGACAATAGGAGCCA-3' 和 5'-GGCATACTCCGTCTGCTCAG-3';LAG3, 5'-GCCTCCGACTGGGTCATTTT-3' 和 5'- CTTTCCGCTAAGTGGTGATGG-3';CTLA4, 5'-CATGATGGGGAATGAGTTGACC-3' 和 5'-TCAGTCCTTGGATAGTGAGGTTC-3';TIGIT, 5'-TGGTCGCGTTGACTAGAAAGA-3' 和 5'- GGGCTCCATTCCTCCTGTC-3'。
ATAC-seq primary data processing and peak calling
ATAC-seq 原始数据处理和峰值调用
ATAC-seq raw data were processed using the published ATAC-seq pipeline ATAC-pipe (Zuo et al., 2019). Sequencing reads were mapped using the “--MappingQC” module with option “-c 50” in ATAC-pipe. Adapter sequences were trimmed, and reads were mapped to hg19 using Bowtie2 (Langmead and Salzberg, 2012). PCR duplicates were removed as described. Mapped reads were then shifted +4/-5 bp depending on the read strand such that the first base of each mapped read represented the Tn5 cleavage position. All mapped reads were then extended to 50 bp centred by the cleavage position. Reads mapped to repeated regions and chromosome M were removed. We used the “--PeakCalling” module in ATAC-pipe with the options “--p1 3 --q1 5 --f1 1 -w 50” to call the peaks. The peaks were then filtered, and enriched regions were identified as those with a posterior probability >0.99. Samples from the same cell type classified under the same clinical condition (healthy, SLE) were grouped for peak calling, and peaks for all categories were then merged together to generate a peak list. The number of raw read counts mapped to each peak in each sample was quantified by this module in ATAC-pipe. We then obtained an N×M data matrix where N indicated the number of merged peaks, M indicated the number of samples, and the matrix value Di,j represented the raw read counts falling in peak i (i = 1 to N) of sample j (j = 1 to M). After manually removing the peaks mapping to chromosome Y, this data matrix was then normalized by the “normalize.quantiles” function of the “preprocessCore” package in R, and the normalized matrix was then then log2 transformed. To avoid the effect of the difference between males and females, we excluded the peaks that were significantly differential in male and female samples (N = 212, screened by all male samples vs all female samples, |log2 fold change|>1.2, FDR<0.05). This final data matrix was used for downstream analysis.
ATAC-seq 原始数据使用已发布的 ATAC-seq 管道 ATAC-pipe () 进行处理 Zuo et al., 2019 。使用 ATAC-pipe 中的“--MappingQC”模块和选项 “-c 50”对测序读数进行定位。修剪接头序列,并使用 Bowtie2 ( Langmead and Salzberg, 2012 ) 将读数映射到 hg19。如前所述去除 PCR 重复项。然后根据读取链将映射的读数移动 +4/-5 bp,使得每个映射读数的第一个碱基代表 Tn5 切割位置。然后将所有映射的读数延伸到以切割位置为中心的 50 bp。映射到重复区域和染色体 M 的读数被去除。我们使用了 ATAC-pipe 中的 “--PeakCalling” 模块,并带有选项 “--p1 3 --q1 5 --f1 1 -w 50” 来调用峰值。然后过滤峰,并将富集区域确定为后验概率为 >0.99 的区域。将来自同一临床条件(健康、SLE)分类的相同细胞类型的样本分组以进行峰识别,然后将所有类别的峰合并在一起以生成峰列表。映射到每个样品中每个峰的原始读取计数数由 ATAC-pipe 中的该模块量化。然后,我们得到了一个 N×M 数据矩阵,其中 N 表示合并峰的数量,M 表示样品数量,矩阵值 Di,j 表示样品 j (j = 1 到 M) 的峰 i (i = 1 到 N) 的原始读取计数。在手动去除映射到染色体 Y 的峰后,该数据矩阵由 R 中 “preprocessCore” 包的 “normalize.quantiles” 函数进行归一化,然后对归一化矩阵进行对数2 转换。 为避免雄性和雌性之间差异的影响,我们排除了雄性和雌性样本中显著差异的峰(N = 212,由所有雄性样本与所有雌性样本筛选,|log2 倍变化|>1.2,FDR<0.05)。该最终数据矩阵用于下游分析。
ATAC-seq 原始数据使用已发布的 ATAC-seq 管道 ATAC-pipe () 进行处理 Zuo et al., 2019 。使用 ATAC-pipe 中的“--MappingQC”模块和选项 “-c 50”对测序读数进行定位。修剪接头序列,并使用 Bowtie2 ( Langmead and Salzberg, 2012 ) 将读数映射到 hg19。如前所述去除 PCR 重复项。然后根据读取链将映射的读数移动 +4/-5 bp,使得每个映射读数的第一个碱基代表 Tn5 切割位置。然后将所有映射的读数延伸到以切割位置为中心的 50 bp。映射到重复区域和染色体 M 的读数被去除。我们使用了 ATAC-pipe 中的 “--PeakCalling” 模块,并带有选项 “--p1 3 --q1 5 --f1 1 -w 50” 来调用峰值。然后过滤峰,并将富集区域确定为后验概率为 >0.99 的区域。将来自同一临床条件(健康、SLE)分类的相同细胞类型的样本分组以进行峰识别,然后将所有类别的峰合并在一起以生成峰列表。映射到每个样品中每个峰的原始读取计数数由 ATAC-pipe 中的该模块量化。然后,我们得到了一个 N×M 数据矩阵,其中 N 表示合并峰的数量,M 表示样品数量,矩阵值 Di,j 表示样品 j (j = 1 到 M) 的峰 i (i = 1 到 N) 的原始读取计数。在手动去除映射到染色体 Y 的峰后,该数据矩阵由 R 中 “preprocessCore” 包的 “normalize.quantiles” 函数进行归一化,然后对归一化矩阵进行对数2 转换。 为避免雄性和雌性之间差异的影响,我们排除了雄性和雌性样本中显著差异的峰(N = 212,由所有雄性样本与所有雌性样本筛选,|log2 倍变化|>1.2,FDR<0.05)。该最终数据矩阵用于下游分析。
Normalization of ATAC-seq profiles
ATAC-seq 配置文件的标准化
The UCSC Genome Browser provides aligned annotation tracks and can also display ATAC-seq signal peaks. To avoid the impacts of variable sequencing depths and signal-to-noise ratios among different samples, we normalized the “bedGraph” files before uploading. To do this, we used the R package DESeq (Anders and Huber, 2010) to calculate the size factor for every sample via the sample×peak raw count matrix (output file of ATAC-pipe). These size factors can be used to suitably measure the depth of data at chromatin-accessible regions in all samples. To obtain the standardized data, the raw count in the “bedGraph” file for each sample was divided by the size factor. Normalized “bedGraph” files were then converted to the UCSC Genome Browser input format “bigWig” using the UCSC tool “bedGraphToBigWig”.
UCSC Genome Browser 提供比对的注释轨道,还可以显示 ATAC-seq 信号峰。为了避免不同样本之间可变测序深度和信噪比的影响,我们在上传前对 “bedGraph” 文件进行了标准化。为此,我们使用 R 包 DESeq ( Anders and Huber, 2010 ) 通过样品峰原始计数矩阵(ATAC-pipe 的输出文件)计算每个样品×粒径因子。这些大小因子可用于适当测量所有样品中染色质可接近区域的数据深度。为了获得标准化数据,将每个样品的 “bedGraph” 文件中的原始计数除以大小因子。然后使用 UCSC 工具“bedGraphToBigWig”将标准化的“bedGraph”文件转换为 UCSC 基因组浏览器输入格式“bigWig”。
UCSC Genome Browser 提供比对的注释轨道,还可以显示 ATAC-seq 信号峰。为了避免不同样本之间可变测序深度和信噪比的影响,我们在上传前对 “bedGraph” 文件进行了标准化。为此,我们使用 R 包 DESeq ( Anders and Huber, 2010 ) 通过样品峰原始计数矩阵(ATAC-pipe 的输出文件)计算每个样品×粒径因子。这些大小因子可用于适当测量所有样品中染色质可接近区域的数据深度。为了获得标准化数据,将每个样品的 “bedGraph” 文件中的原始计数除以大小因子。然后使用 UCSC 工具“bedGraphToBigWig”将标准化的“bedGraph”文件转换为 UCSC 基因组浏览器输入格式“bigWig”。
Saturation curve of the sample and peak numbers
样品和峰数的饱和曲线
Given the heterogeneity of CD4+ T cell subsets, complexity of SLE and variety between individuals, a sufficient number of samples is necessary to ensure that all chromatin-accessible regions of SLE CD4+ T cells are detectable. To construct the saturation curve of the numbers of SLE samples used and peaks detected, we first randomly selected a certain number of SLE samples. Next, we obtained the peak list by the same peak calling method described above, and the number of peaks called from this random selection was obtained. This random selection was performed on 1 to 63 samples 10 times. The x-axis of the saturation curve depicts the number of samples used, while the y-axis depicts the number of peaks called from 10 random selections.
鉴于 CD4+ T 细胞亚群的异质性、SLE 的复杂性和个体之间的多样性,需要足够数量的样本来确保 SLE CD4+ T 细胞的所有染色质可及区域都是可检测到的。为了构建使用的 SLE 样本数量和检测到的峰的饱和曲线,我们首先随机选择一定数量的 SLE 样本。接下来,我们通过上述相同的峰调用方法获得峰列表,并得到从该随机选择中调用的峰数。对 1 至 63 个样品进行 10 次随机选择。饱和度曲线的 x 轴描述使用的样本数,而 y 轴描述从 10 个随机选择中调用的峰数。
鉴于 CD4+ T 细胞亚群的异质性、SLE 的复杂性和个体之间的多样性,需要足够数量的样本来确保 SLE CD4+ T 细胞的所有染色质可及区域都是可检测到的。为了构建使用的 SLE 样本数量和检测到的峰的饱和曲线,我们首先随机选择一定数量的 SLE 样本。接下来,我们通过上述相同的峰调用方法获得峰列表,并得到从该随机选择中调用的峰数。对 1 至 63 个样品进行 10 次随机选择。饱和度曲线的 x 轴描述使用的样本数,而 y 轴描述从 10 个随机选择中调用的峰数。
Correlation analysis of all SLE patients
所有 SLE 患者的相关性分析
To obtain a representative correlation landscape of all samples from SLE patients, we first filtered variant peaks among all SLE samples based on a coefficient of variation (COV) greater than a certain threshold, and multiple variant peak lists were obtained from a COV threshold ranging from 0.2 to 0.7 (step by 0.02). Then, these sample×variant peak count matrixes were used to calculate the sample×sample Pearson’s correlation matrixes. The average of these correlation matrixes was used for the final data presentation. After unsupervised hierarchical clustering (Seaborn clustermap, with parameter metric = 'Euclidean', method = 'complete'), SLE patients were divided into three distinct groups.
为了获得 SLE 患者所有样本的代表性相关景观,我们首先根据大于某个阈值的变异系数 (COV) 过滤所有 SLE 样本中的变异峰,并从 0.2 到 0.7 的 COV 阈值范围内获得多个变异峰列表(以 0.02 为步长)。然后,使用这些 sample×variant 峰计数矩阵计算样本×样本 Pearson 相关矩阵。这些相关矩阵的平均值用于最终数据表示。在无监督分层聚类 (Seaborn clustermap,参数 metric = 'Euclidean',方法 = 'complete') 后,SLE 患者被分为三个不同的组。
为了获得 SLE 患者所有样本的代表性相关景观,我们首先根据大于某个阈值的变异系数 (COV) 过滤所有 SLE 样本中的变异峰,并从 0.2 到 0.7 的 COV 阈值范围内获得多个变异峰列表(以 0.02 为步长)。然后,使用这些 sample×variant 峰计数矩阵计算样本×样本 Pearson 相关矩阵。这些相关矩阵的平均值用于最终数据表示。在无监督分层聚类 (Seaborn clustermap,参数 metric = 'Euclidean',方法 = 'complete') 后,SLE 患者被分为三个不同的组。
Summary of the clinical information for the three SLE patient groups
3 组 SLE 患者临床信息摘要
The SLE DAI was calculated for each SLE patient with detailed clinical information according to the internationally certified method SLEDAI-2K (Gladman et al., 2002). Nephritis is a serious and common comorbidity in SLE patients. To quantify the severity of nephritis for each patient, we defined the sum of the four nephritis-related indicators—haematuria, proteinuria, pyuria, and tubular urine—in SLEDAI-2K as the nephritis DAI. We compared the SLE and nephritis DAIs among the three patient groups and summarized the ratio of patients with each comorbidity for each patient group.
根据国际认证的方法 SLEDAI-2K ( ) 计算每位 SLE 患者的 SLE DAI,并提供详细的临床信息 Gladman et al., 2002 。肾炎是 SLE 患者的一种严重且常见的合并症。为了量化每位患者肾炎的严重程度,我们将 SLEDAI-2K 中四个肾炎相关指标(血尿、蛋白尿、脓尿和肾小管尿)的总和定义为肾炎 DAI。我们比较了 3 个患者组的 SLE 和肾炎 DAI,并总结了每个患者组患有每种合并症的患者比例。
根据国际认证的方法 SLEDAI-2K ( ) 计算每位 SLE 患者的 SLE DAI,并提供详细的临床信息 Gladman et al., 2002 。肾炎是 SLE 患者的一种严重且常见的合并症。为了量化每位患者肾炎的严重程度,我们将 SLEDAI-2K 中四个肾炎相关指标(血尿、蛋白尿、脓尿和肾小管尿)的总和定义为肾炎 DAI。我们比较了 3 个患者组的 SLE 和肾炎 DAI,并总结了每个患者组患有每种合并症的患者比例。
Acquisition of microarray and RNA-seq datasets
采集微阵列和 RNA-seq 数据集
We downloaded the following bulk datasets for comparison to our single-cell data: (1) the microarray gene expression data of three CD4+ T cell subsets in each developmental stage, Tn/Tem/Tcm, downloaded from the Gene Expression Omnibus: GSE61697 (Takeshita et al., 2015); (2) the microarray gene expression data of healthy and SLE blood CD4+ T cells, downloaded from the Gene Expression Ominibus: GSE4588, GSE10325 and GSE55447 (Hutcheson et al., 2008; Sharma et al., 2015); (3) the bulk RNA-seq of three Treg subtypes, nTreg, effect Treg, and Fr. III Treg, downloaded from the Gene Expression Omnibus: GSE90600 (Cuadrado et al., 2018); (4) the single-cell RNA-seq expression data of colon mucosa immune cells from ulcerative colitis patients and controls from the Single Cell Portal: SCP259 (Smillie et al., 2019); (5) the single-cell RNA-seq expression data of PBMCs from ulcerative colitis patients and controls from the Gene Expression Omnibus: GSE125527 (Boland et al., 2020); (6) the single-cell RNA-seq expression data of PBMCs from healthy controls from the Gene Expression Omnibus: GSE150728 (Wilk et al., 2020); (7) the single-cell RNA-seq expression data of PBMCs from SLE patients from the Gene Expression Omnibus: GSE137029; and (8) the single-cell RNA-seq expression data of PBMCs from multiple sclerosis patients from the Gene Expression Omnibus: GSE138266 (Schafflick et al., 2020).
我们下载了以下批量数据集,以便与我们的单细胞数据进行比较:(1) 每个发育阶段三个 CD4 + T 细胞亚群的微阵列基因表达数据,Tn/Tem/Tcm,从基因表达综合下载:GSE61697 ( Takeshita et al., 2015 );(2) 健康细胞和 SLE 血液 CD4+ T 细胞的微阵列基因表达数据,从 Gene Expression Ominibus 下载:GSE4588,GSE10325 和 GSE55447 ( Hutcheson et al., 2008 ; Sharma et al., 2015 );(3) 三种 Treg 亚型 nTreg、效果 Treg 和 Fr. III Treg 的大量 RNA-seq,从基因表达综合下载:GSE90600 ( Cuadrado et al., 2018 );(4) 来自单细胞门户的溃疡性结肠炎患者和对照的结肠粘膜免疫细胞的单细胞 RNA-seq 表达数据:SCP259 ( Smillie et al., 2019 );(5) 来自溃疡性结肠炎患者和基因表达综合的对照的 PBMC 的单细胞 RNA-seq 表达数据:GSE125527 ( Boland et al., 2020 );(6) 来自基因表达综合的健康对照 PBMC 的单细胞 RNA-seq 表达数据:GSE150728 ( Wilk et al., 2020 );(7) 来自 Gene Expression Omnibus 的 SLE 患者 PBMC 的单细胞 RNA-seq 表达数据:GSE137029;(8) 来自基因表达综合的多发性硬化症患者 PBMC 的单细胞 RNA-seq 表达数据:GSE138266 ( Schafflick et al., 2020 )。
我们下载了以下批量数据集,以便与我们的单细胞数据进行比较:(1) 每个发育阶段三个 CD4 + T 细胞亚群的微阵列基因表达数据,Tn/Tem/Tcm,从基因表达综合下载:GSE61697 ( Takeshita et al., 2015 );(2) 健康细胞和 SLE 血液 CD4+ T 细胞的微阵列基因表达数据,从 Gene Expression Ominibus 下载:GSE4588,GSE10325 和 GSE55447 ( Hutcheson et al., 2008 ; Sharma et al., 2015 );(3) 三种 Treg 亚型 nTreg、效果 Treg 和 Fr. III Treg 的大量 RNA-seq,从基因表达综合下载:GSE90600 ( Cuadrado et al., 2018 );(4) 来自单细胞门户的溃疡性结肠炎患者和对照的结肠粘膜免疫细胞的单细胞 RNA-seq 表达数据:SCP259 ( Smillie et al., 2019 );(5) 来自溃疡性结肠炎患者和基因表达综合的对照的 PBMC 的单细胞 RNA-seq 表达数据:GSE125527 ( Boland et al., 2020 );(6) 来自基因表达综合的健康对照 PBMC 的单细胞 RNA-seq 表达数据:GSE150728 ( Wilk et al., 2020 );(7) 来自 Gene Expression Omnibus 的 SLE 患者 PBMC 的单细胞 RNA-seq 表达数据:GSE137029;(8) 来自基因表达综合的多发性硬化症患者 PBMC 的单细胞 RNA-seq 表达数据:GSE138266 ( Schafflick et al., 2020 )。
Acquisition of gene sets 获取基因集
All gene sets used in this study and their citations are listed in Table S2. Marker gene lists of Th1 cells (CXCR3, TBX21, IFNG, GZMB, TNF), Th2 cells (GATA3, IL4, IL5, IL13), Th17 cells (CCR6, RORC, IL17A, IL17F), Treg cells (FOXP3, RTKN2, IKZF2, IL10, TGFB2, TGFB3, CTLA4, IL2RA), Tct cells (NKG7, GZMH, FGFBP2) and resting T cells (CCR7, SELL, CD27) were obtained from published studies (Luckheeram et al., 2012; Patil et al., 2018). Signature genes of naïve T (Tn), effector memory T (Tem), and central memory T (Tcm) cells were obtained by differential expression analysis of Tn/Tem/Tcm cells with all other cell types (Tn features: Tn vs Tem and Tcm; Tem features: Tem vs Tn and Tcm; Tcm features: Tn and Tem; |log2 fold change|>1, FDR<0.05) (Cuadrado et al., 2018). However, we obtained 0 Tcm cell signature genes using this method, potentially because Tcm cells are intermediate between Tn and Tem cells. Risk genes of multiple autoimmune diseases, such as SLE, rheumatoid arthritis, type 1 diabetes inflammatory bowel disease, and ulcerative colitis, were obtained from their reported genes in the GWAS catalogue (v 1.0.2) (Welter et al., 2014). SLE CD4+ T cell up-/downregulated gene sets were screened from the published dataset GSE10325 (Hutcheson et al., 2008) (p value<0.01, log2 fold change>1). The gene sets of biological processes in MSigDB (Liberzon et al., 2015) were also used in our analysis. In addition, we excluded the gene sets that were irrelevant to the immune functions of CD4+ T cells (such as those related to the nervous system, embryonic development, reproduction and cellular dynamics) and the gene sets with more than 1500 or less than 10 genes. The final MSigDB gene sets used in our downstream analysis are listed in Table S2.
本研究中使用的所有基因集及其引文均列于表 S2 中。Th1 细胞 (CXCR3、TBX21、IFNG、GZMB、TNF)、Th2 细胞 (GATA3、IL4、IL5、IL13)、Th17 细胞 (CCR6、RORC、IL17A、IL17F)、Treg 细胞 (FOXP3、RTKN2、IKZF2、IL10、TGFB2、TGFB3、CTLA4、IL2RA)、Tct 细胞 (NKG7、GZMH、FGFBP2) 和静息 T 细胞 (CCR7、SELL、CD27) 的标记基因列表是从已发表的研究中获得的 ( Luckheeram et al., 2012 ; Patil et al., 2018 )。通过对 Tn/Tem/Tcm 细胞与所有其他细胞类型的差异表达分析获得初始 T (Tn)、效应记忆 T (Tem) 和中枢记忆 T (Tcm) 细胞的特征基因(Tn 特征:Tn vs Tem 和 Tcm;Tem 特点:Tem vs Tn 和 Tcm;Tcm 特征: Tn 和 Tem;|log2 倍变化|>1, FDR<0.05) ( Cuadrado et al., 2018 ).然而,我们使用这种方法获得了 0 个 Tcm 细胞特征基因,这可能是因为 Tcm 细胞是 Tn 和 Tem 细胞之间的中间体。多种自身免疫性疾病的风险基因,如 SLE、类风湿性关节炎、1 型糖尿病、炎症性肠病和溃疡性结肠炎,是从 GWAS 目录 (v 1.0.2) 中报道的基因中获得 Welter et al., 2014 的 ( )。从已发表的数据集中筛选 SLE CD4 + T 细胞上调/下调的基因集 GSE10325 ( Hutcheson et al., 2008 ) (p 值<0.01,对数2 倍变化>1)。我们的分析中使用了 MSigDB ( Liberzon et al., 2015 ) 中生物过程的基因集。此外,我们排除了与 CD4+ T 细胞免疫功能无关的基因集(例如与神经系统、胚胎发育、繁殖和细胞动力学相关的基因集)和基因集超过 1500 个或少于 10 个基因。我们下游分析中使用的最终 MSigDB 基因集列于表 S2 中。
本研究中使用的所有基因集及其引文均列于表 S2 中。Th1 细胞 (CXCR3、TBX21、IFNG、GZMB、TNF)、Th2 细胞 (GATA3、IL4、IL5、IL13)、Th17 细胞 (CCR6、RORC、IL17A、IL17F)、Treg 细胞 (FOXP3、RTKN2、IKZF2、IL10、TGFB2、TGFB3、CTLA4、IL2RA)、Tct 细胞 (NKG7、GZMH、FGFBP2) 和静息 T 细胞 (CCR7、SELL、CD27) 的标记基因列表是从已发表的研究中获得的 ( Luckheeram et al., 2012 ; Patil et al., 2018 )。通过对 Tn/Tem/Tcm 细胞与所有其他细胞类型的差异表达分析获得初始 T (Tn)、效应记忆 T (Tem) 和中枢记忆 T (Tcm) 细胞的特征基因(Tn 特征:Tn vs Tem 和 Tcm;Tem 特点:Tem vs Tn 和 Tcm;Tcm 特征: Tn 和 Tem;|log2 倍变化|>1, FDR<0.05) ( Cuadrado et al., 2018 ).然而,我们使用这种方法获得了 0 个 Tcm 细胞特征基因,这可能是因为 Tcm 细胞是 Tn 和 Tem 细胞之间的中间体。多种自身免疫性疾病的风险基因,如 SLE、类风湿性关节炎、1 型糖尿病、炎症性肠病和溃疡性结肠炎,是从 GWAS 目录 (v 1.0.2) 中报道的基因中获得 Welter et al., 2014 的 ( )。从已发表的数据集中筛选 SLE CD4 + T 细胞上调/下调的基因集 GSE10325 ( Hutcheson et al., 2008 ) (p 值<0.01,对数2 倍变化>1)。我们的分析中使用了 MSigDB ( Liberzon et al., 2015 ) 中生物过程的基因集。此外,我们排除了与 CD4+ T 细胞免疫功能无关的基因集(例如与神经系统、胚胎发育、繁殖和细胞动力学相关的基因集)和基因集超过 1500 个或少于 10 个基因。我们下游分析中使用的最终 MSigDB 基因集列于表 S2 中。
Screening the differential peaks between three SLE patient groups and healthy controls
筛选三个 SLE 患者组与健康对照之间的差异峰
Differential analysis was performed on CD4+ T cell ATAC-seq samples from each SLE patient group with all samples from the healthy controls. Peaks with |log2 fold change|>1.2 and FDR<0.05 were defined as differential peaks. After unsupervised hierarchical clustering of the sample×peak count matrix, 11775 differential peaks were divided into Clusters I-V.
对来自每个 SLE 患者组的 CD4 + T 细胞 ATAC-seq 样本与来自健康对照的所有样本进行差异分析。具有 |log2 倍变化|>1.2 和 FDR<0.05 的峰被定义为差分峰。在对 sample×peak count 矩阵进行无监督分层聚类后,将 11775 个差分峰划分为聚类 I-V。
对来自每个 SLE 患者组的 CD4 + T 细胞 ATAC-seq 样本与来自健康对照的所有样本进行差异分析。具有 |log2 倍变化|>1.2 和 FDR<0.05 的峰被定义为差分峰。在对 sample×peak count 矩阵进行无监督分层聚类后,将 11775 个差分峰划分为聚类 I-V。
Functional annotation of peak clusters
峰簇的功能注释
We first obtained the regulatory regions of all genes using the “basal plus extension” model in GREAT (McLean et al., 2010) (proximal: 5 kb upstream, 1 kb downstream; plus distal: up to 500 kb). Then, we overlaid all peaks with these regulatory regions using “intersectBed” in bedtools (Quinlan and Hall, 2010) and constructed a gene×peak regulatory matrix in which each element was a 0 (peak not in the regulatory region of a gene) or 1 (peak in the regulatory region of a gene). The regulatory elements of a gene set were defined as peaks located in the regulatory regions of all genes in this gene set. For each gene set, we identified all its regulatory elements and then summarized the number of these regulatory elements in each peak cluster and the average number of regulatory elements per gene. We calculated the statistical significance of the enrichment of certain functional gene sets in a certain peak cluster compared with all peaks (background) by Fisher’s exact test.
我们首先使用 GREAT ( McLean et al., 2010 ) 中的 “基础加延伸” 模型获得所有基因的调控区域 ( ) (近端:上游 5 kb,下游 1 kb;加上远端:高达 500 kb)。然后,我们使用 bedtools ( Quinlan and Hall, 2010 ) 中的 “intersectBed” 将所有峰与这些调控区域叠加,并构建一个基因×调控矩阵,其中每个元素都是 0(峰不在基因的调控区)或 1(峰在基因的调控区)。基因集的调控元件被定义为位于该基因集中所有基因的调控区域的峰。对于每个基因集,我们确定了其所有调控元件,然后总结了每个峰簇中这些调控元件的数量以及每个基因的平均调控元件数量。我们通过 Fisher 精确检验计算了与所有峰 (背景) 相比,某个峰簇中某些功能基因集富集的统计显着性。
我们首先使用 GREAT ( McLean et al., 2010 ) 中的 “基础加延伸” 模型获得所有基因的调控区域 ( ) (近端:上游 5 kb,下游 1 kb;加上远端:高达 500 kb)。然后,我们使用 bedtools ( Quinlan and Hall, 2010 ) 中的 “intersectBed” 将所有峰与这些调控区域叠加,并构建一个基因×调控矩阵,其中每个元素都是 0(峰不在基因的调控区)或 1(峰在基因的调控区)。基因集的调控元件被定义为位于该基因集中所有基因的调控区域的峰。对于每个基因集,我们确定了其所有调控元件,然后总结了每个峰簇中这些调控元件的数量以及每个基因的平均调控元件数量。我们通过 Fisher 精确检验计算了与所有峰 (背景) 相比,某个峰簇中某些功能基因集富集的统计显着性。
scRNA-seq primary data processing and gene expression imputation
scRNA-seq 原始数据处理和基因表达插补
Cell Ranger v2.0 was used to demultiplex the FASTQ reads, align them to the hg19 human transcriptome, and extract their ‘‘cell’’ and ‘‘UMI’’ barcodes. The output of this pipeline is a digital gene expression (DGE) matrix, which records the number of UMIs for each gene that are associated with each cell barcode. Next, we created the Seurat objects for three DGE matrixes of three batches using Seurat3.0, and merged them into a big Seurat object. Cells with fewer than 400 genes, more than 4000 genes and a percentage of mitochondrial genes greater than 5% were removed, and genes expressed in fewer than 10 cells were removed. Gene expression was normalized by “LogNormalize” in Seurat3.0. To further recover the expression of genes, we input a normalized expression matrix into the published imputation tool SAVER (Huang et al., 2018) with the default settings. The output of SAVER was the recovered gene expression matrix, which was used for downstream analysis.
Cell Ranger v2.0 用于对 FASTQ 读数进行多路复用,将它们与 hg19 人类转录组对齐,并提取它们的“细胞”和“UMI”条形码。该管道的输出是一个数字基因表达 (DGE) 矩阵,它记录了与每个细胞条形码相关的每个基因的 UMI 数量。接下来,我们使用 Seurat3.0 为三个批次的三个 DGE 矩阵创建了 Seurat 对象,并将它们合并成一个大的 Seurat 对象。去除少于 400 个基因、超过 4000 个基因和大于 5% 的线粒体基因百分比的细胞,去除少于 10 个细胞中表达的基因。在 Seurat3.0 中通过 “LogNormalize” 对基因表达进行归一化。为了进一步恢复基因的表达,我们将标准化的表达矩阵输入到已发布的插补工具 SAVER () Huang et al., 2018 中,并使用默认设置。SAVER 的输出是回收的基因表达基质,用于下游分析。
Cell Ranger v2.0 用于对 FASTQ 读数进行多路复用,将它们与 hg19 人类转录组对齐,并提取它们的“细胞”和“UMI”条形码。该管道的输出是一个数字基因表达 (DGE) 矩阵,它记录了与每个细胞条形码相关的每个基因的 UMI 数量。接下来,我们使用 Seurat3.0 为三个批次的三个 DGE 矩阵创建了 Seurat 对象,并将它们合并成一个大的 Seurat 对象。去除少于 400 个基因、超过 4000 个基因和大于 5% 的线粒体基因百分比的细胞,去除少于 10 个细胞中表达的基因。在 Seurat3.0 中通过 “LogNormalize” 对基因表达进行归一化。为了进一步恢复基因的表达,我们将标准化的表达矩阵输入到已发布的插补工具 SAVER () Huang et al., 2018 中,并使用默认设置。SAVER 的输出是回收的基因表达基质,用于下游分析。
Cell clustering 细胞聚集
We first created the Seurat objects for all three batches of 10X single-cell RNA-seq data and identified the top 1000 variable features (genes in the Y chromosome were eliminated) individually for each object with default parameters. Next, we identified anchors for these three objects and integrated them with the top 15 canonical correlation analysis (CCA) dimensions. Then, we performed PCA on the integrated object and found cell clusters with the top 25 PCA dimensions (resolution set to 1).
我们首先为所有三批 10X 单细胞 RNA-seq 数据创建了 Seurat 对象,并使用默认参数为每个对象单独确定了前 1000 个可变特征(Y 染色体中的基因被消除)。接下来,我们确定了这三个对象的锚点,并将它们与前 15 个典型相关性分析 (CCA) 维度集成。然后,我们对集成对象进行 PCA 并找到具有前 25 个 PCA 维度(分辨率设置为 1)的细胞簇。
我们首先为所有三批 10X 单细胞 RNA-seq 数据创建了 Seurat 对象,并使用默认参数为每个对象单独确定了前 1000 个可变特征(Y 染色体中的基因被消除)。接下来,我们确定了这三个对象的锚点,并将它们与前 15 个典型相关性分析 (CCA) 维度集成。然后,我们对集成对象进行 PCA 并找到具有前 25 个 PCA 维度(分辨率设置为 1)的细胞簇。
Determination of the sample identity for each cell type
确定每种细胞类型的样品特性
To capture interindividual variability in this population genetics study, we sequenced a large number of cells each from 4–6 individuals. Therefore, the sample (healthy controls and SLE patients) from which each cell type was derived was unknown. The published tool Demuxlet can be used to determine the sample identity of each single cell based on a comparison of single nucleotide polymorphisms (SNPs) between individuals and cells (Kang et al., 2018). To obtain the input file for Demuxlet (VCF file containing the SNPs of all samples), we first aligned the ATAC-seq sequencing reads to the reference genome hg19 using the “--MappingQC” module with option “-c 100” in ATAC-pipe. Then, we integrated these output bam files into a single mpileup file using “samtools mpileup”. Next, SNPs were called by VarScan (Koboldt et al., 2012) with the options “--min-coverage 5 --p-value 0.01 --output-vcf 1”. The output VarScan file was the VCF file that was used as the input file for Demuxlet, and Demuxlet was then run with default settings. Only the cells in the “.best” file (a Demuxlet output file) marked by “SNG” (singlet) could be determined. In total, we identified 47% of all cells. To further confirm the sample identities of more cells, we applied another tool, Souporcell (Heaton et al., 2020), to classify all CD4+ T cells by genotype. A total of 93% of the CD4+ T cells were divided into 10 clusters, and cells in the same cluster derived from the same sample. We identified the samples in each cell cluster by overlapping the results of Demuxlet and Souporcell.
为了捕捉这项群体遗传学研究中的个体间变异性,我们对 4-6 个个体的大量细胞进行了测序。因此,每种细胞类型的样本 (健康对照和 SLE 患者) 是未知的。已发布的工具 Demuxlet 可用于根据个体和细胞之间单核苷酸多态性 (SNP) 的比较来确定每个单细胞的样本身份 ( Kang et al., 2018 )。为了获得 Demuxlet 的输入文件(包含所有样本的 SNP 的 VCF 文件),我们首先使用 ATAC-pipe 中带有选项“-c 100”的“--MappingQC”模块将 ATAC-seq 测序读数与参考基因组 hg19 对齐。然后,我们使用 “samtools mpileup” 将这些输出 bam 文件集成到一个 mpileup 文件中。接下来,VarScan ( Koboldt et al., 2012 ) 使用选项 “--min-coverage 5 --p-value 0.01 --output-vcf 1” 调用 SNP。输出 VarScan 文件是用作 Demuxlet 输入文件的 VCF 文件,然后 Demuxlet 使用默认设置运行。只能确定 “.best” 文件(Demuxlet 输出文件)中标记为 “SNG” (singlet) 的单元格。总的来说,我们鉴定了 47% 的细胞。为了进一步确认更多细胞的样本身份,我们应用了另一种工具 Souporcell ( Heaton et al., 2020 ),按基因型对所有 CD4+ T 细胞进行分类。共有 93% 的 CD4 + T 细胞被分成 10 个簇,同一簇中的细胞来源于同一样本。我们通过重叠 Demuxlet 和 Souporcell 的结果来识别每个细胞簇中的样本。
为了捕捉这项群体遗传学研究中的个体间变异性,我们对 4-6 个个体的大量细胞进行了测序。因此,每种细胞类型的样本 (健康对照和 SLE 患者) 是未知的。已发布的工具 Demuxlet 可用于根据个体和细胞之间单核苷酸多态性 (SNP) 的比较来确定每个单细胞的样本身份 ( Kang et al., 2018 )。为了获得 Demuxlet 的输入文件(包含所有样本的 SNP 的 VCF 文件),我们首先使用 ATAC-pipe 中带有选项“-c 100”的“--MappingQC”模块将 ATAC-seq 测序读数与参考基因组 hg19 对齐。然后,我们使用 “samtools mpileup” 将这些输出 bam 文件集成到一个 mpileup 文件中。接下来,VarScan ( Koboldt et al., 2012 ) 使用选项 “--min-coverage 5 --p-value 0.01 --output-vcf 1” 调用 SNP。输出 VarScan 文件是用作 Demuxlet 输入文件的 VCF 文件,然后 Demuxlet 使用默认设置运行。只能确定 “.best” 文件(Demuxlet 输出文件)中标记为 “SNG” (singlet) 的单元格。总的来说,我们鉴定了 47% 的细胞。为了进一步确认更多细胞的样本身份,我们应用了另一种工具 Souporcell ( Heaton et al., 2020 ),按基因型对所有 CD4+ T 细胞进行分类。共有 93% 的 CD4 + T 细胞被分成 10 个簇,同一簇中的细胞来源于同一样本。我们通过重叠 Demuxlet 和 Souporcell 的结果来识别每个细胞簇中的样本。
Identifying the statistically significant differences in cell proportions
识别细胞比例的统计学显著性差异
To identify changes in the proportion of each cell subset in blood from healthy controls and SLE patients, we used two statistical tests that each captured distinct but complementary types of information: (1) Student’s t-test and (2) Fisher’s exact test. The proportion of each CD4+ T cell subtype was calculated for each sample according to the results of Demuxlet and Souporcell. We compared the proportions of CD4+ T cell subtypes between healthy controls and SLE patients (Student’s t-test). However, a t-test was used to examine each cell subset independently; for cell proportions, an increase in the percentage of one cell subset would necessitate decreases in the percentages of other cell subsets. We therefore performed Fisher’s exact test on the numbers of cells from each subset that were isolated from the blood of healthy individuals and patients with SLE, and the odds ratios and p values were used to measure the enrichment of each cell subset in each clinical state. These two methods reflected almost the same statistically significant differences.
为了确定健康对照和 SLE 患者血液中每个细胞亚群比例的变化,我们使用了两种统计检验,每种检验都捕获了不同但互补的信息类型:(1) 学生 t 检验和 (2) Fisher 精确检验。根据 Demuxlet 和 Souporcell 的结果计算每个样本的每种 CD4 + T 细胞亚型的比例。我们比较了健康对照和 SLE 患者之间 CD4+ T 细胞亚型的比例 (学生 t 检验)。然而,使用 t 检验独立检查每个细胞亚群;对于细胞比例,一个细胞亚群百分比的增加将需要降低其他细胞亚群的百分比。因此,我们对从健康个体和 SLE 患者血液中分离的每个亚群的细胞数量进行了 Fisher 精确检验,并使用比值比和 p 值来测量每个细胞亚群在每种临床状态下的富集程度。这两种方法反映了几乎相同的统计学显著性差异。
为了确定健康对照和 SLE 患者血液中每个细胞亚群比例的变化,我们使用了两种统计检验,每种检验都捕获了不同但互补的信息类型:(1) 学生 t 检验和 (2) Fisher 精确检验。根据 Demuxlet 和 Souporcell 的结果计算每个样本的每种 CD4 + T 细胞亚型的比例。我们比较了健康对照和 SLE 患者之间 CD4+ T 细胞亚型的比例 (学生 t 检验)。然而,使用 t 检验独立检查每个细胞亚群;对于细胞比例,一个细胞亚群百分比的增加将需要降低其他细胞亚群的百分比。因此,我们对从健康个体和 SLE 患者血液中分离的每个亚群的细胞数量进行了 Fisher 精确检验,并使用比值比和 p 值来测量每个细胞亚群在每种临床状态下的富集程度。这两种方法反映了几乎相同的统计学显著性差异。
Scoring gene signatures based on gene expression and chromatin accessibility
根据基因表达和染色质可及性对基因特征进行评分
We defined the signature score to compare the degrees of enrichment of a gene set among cells or samples. For gene expression data, we first calculated the sums of the normalized expression across all genes in a signature gene set and then normalized them to values from 0 to 1 among cells/samples , which was deemed the gene expression-signature score. For chromatin accessibility data (ATAC-seq), we first overlaid all ATAC-seq peaks with the regulatory regions (defined in “functional annotation of peak clusters”) of all genes by “intersectBed” in bedtools and constructed connections between peaks and genes. We then calculated the sums of the normalized peak intensities across all peaks that were connected to genes in a signature gene set and then normalized them to values from 0 to 1 among samples. This was deemed the chromatin accessibility signature score.
我们定义了特征评分,以比较细胞或样本中基因集的富集程度。对于基因表达数据,我们首先计算了特征基因集中所有基因的标准化表达总和,然后将它们标准化为细胞/样本 中 0 到 1 的值,这被认为是基因表达-特征分数。对于染色质可及性数据 (ATAC-seq),我们首先通过 bedtools 中的 “intersectBed” 将所有 ATAC-seq 峰与所有基因的调节区域(在 “峰簇的功能注释 ”中定义“)叠加,并在峰和基因之间构建连接。然后,我们计算了与特征基因集中的基因相关的所有峰的归一化峰强度之和,然后将它们在样本中归一化为从 0 到 1 的值。这被认为是染色质可及性特征评分。
我们定义了特征评分,以比较细胞或样本中基因集的富集程度。对于基因表达数据,我们首先计算了特征基因集中所有基因的标准化表达总和,然后将它们标准化为细胞/样本
Screening the differentially expressed genes between healthy controls and SLE patients for each CD4+ T cell subtype
筛选健康对照和 SLE 患者之间每种 CD4+ T 细胞亚型的差异表达基因
We first obtained the log-normalized gene expression matrix (after SAVER imputation) through the method described above, removed genes with a maximum expression value of less than 1 across all cells, and then converted the expression to a standardized z-score across cells. For each CD4+ T cell subtype, the DEGs were screened by three thresholds: (1) p value <0.01 (Mann-Whitney U test), FDR <0.05 (Bonferroni correction), (2) average z score of cells from healthy controls >0.1 or average z score of cells from SLE patients >0.1, and (3) difference between and >0.4. The DEGs of each cell subset are listed in Table S3. We summarized the number of upregulated and downregulated DEGs for each cell subset and showed them on a t-SNE scatter plot. The DEGs between Treg cells from healthy controls and SLE patients were also screened in this manner and are listed in Table S4.
我们首先通过上述方法获得对数归一化基因表达矩阵 (SAVER 插补后),去除所有细胞中最大表达值小于 1 的基因,然后将表达转换为跨细胞的标准化 z 分数。对于每种 CD4+ T 细胞亚型,通过三个阈值筛选 DEGs:(1) p 值 <0.01(Mann-Whitney U 检验),FDR <0.05(Bonferroni 校正),(2) 健康对照 细胞的平均 z 评分 >0.1 或 SLE 患者 细胞的平均 z 评分 >0.1,以及 (3) 与 >0.4 之间的差异 。表 S3 中列出了每个单元子集的 DEG。我们总结了每个细胞亚群上调和下调的 DEGs 的数量,并在 t-SNE 散点图上显示它们。健康对照和 SLE 患者的 Treg 细胞之间的 DEGs 也以这种方式进行筛选,并列于表 S4 中。
我们首先通过上述方法获得对数归一化基因表达矩阵 (SAVER 插补后),去除所有细胞中最大表达值小于 1 的基因,然后将表达转换为跨细胞的标准化 z 分数。对于每种 CD4+ T 细胞亚型,通过三个阈值筛选 DEGs:(1) p 值 <0.01(Mann-Whitney U 检验),FDR <0.05(Bonferroni 校正),(2) 健康对照
Construction of the functional change profiles in all cell subsets
在所有细胞亚群中构建功能变化概况
To explore the functional differences in each cell subtype between healthy controls and SLE patients, we first calculated the difference in average expression (DAV) between healthy and SLE cells for all cell subtypes and obtained a differential expression matrix (DEM) in which each row was a gene and each column was a cell subtype. To construct a complete profile of functional changes, we used all MSigDB gene sets and created a 0/1 matrix in which each column was a functional gene set, each row was a gene, and each value was 0 (row gene was not in the column gene set) or 1 (row gene existed in the column gene set). The DEM and 0/1 matrix were used as the “input file” and “gene sets file” of Genomica, respectively, and “Module Map” (parameters: exclude gene sets with less than 3 genes or more than 1000 genes; expression level ≥0.3 denoted upregulation, expression level ≤ -0.3 denoted downregulation; value was displayed as the negative log p value of gene hit enrichment) was then run to obtain the enrichment of each gene set in each cell subset. The top enriched gene sets are shown in Table S3.
为了探讨健康对照和 SLE 患者之间每种细胞亚型的功能差异,我们首先计算了所有细胞亚型的健康细胞和 SLE 细胞 之间的平均表达 (DAV) 差异,并获得了差异表达矩阵 (DEM),其中每行是一个基因,每列是一个细胞亚型。为了构建功能变化的完整概况,我们使用了所有 MSigDB 基因集并创建了一个 0/1 矩阵,其中每列是一个功能基因集,每行是一个基因,每个值为 0(行基因不在列基因集中)或 1(行基因存在于列基因集中)。分别以DEM和0/1矩阵作为Genomica的“输入文件”和“基因集文件”,运行“模块图”(参数:排除少于3个或大于1000个基因的基因集;表达水平≥0.3表示上调,表达水平≤-0.3表示下调;值显示为基因命中富集的负log p值)得到每个细胞亚群中每个基因集的富集。最富集的基因集如表 S3 所示。
为了探讨健康对照和 SLE 患者之间每种细胞亚型的功能差异,我们首先计算了所有细胞亚型的健康细胞和 SLE 细胞
Estimating the differences in gene expression patterns between Treg cells from healthy controls and SLE patients
估计健康对照和 SLE 患者 Treg 细胞之间基因表达模式的差异
We first obtained the log-normalized gene expression matrix (after SAVER imputation) and calculated the average gene expression (AGE) for each CD4+ T cell type from healthy controls and SLE patients. For each CD4+ T cell subtype, genes with AGEs greater than 0.2 were used to calculate the Euclidean distance and Spearman correlation between healthy controls and SLE patients.
我们首先获得对数标准化基因表达矩阵 (SAVER 插补后) 并计算健康对照和 SLE 患者每种 CD4+ T 细胞类型的平均基因表达 (AGE)。对于每种 CD4 + T 细胞亚型,使用 AGEs 大于 0.2 的基因计算健康对照与 SLE 患者之间的欧几里得距离和 Spearman 相关性。
我们首先获得对数标准化基因表达矩阵 (SAVER 插补后) 并计算健康对照和 SLE 患者每种 CD4+ T 细胞类型的平均基因表达 (AGE)。对于每种 CD4 + T 细胞亚型,使用 AGEs 大于 0.2
Screening of differentially expressed genes between two Treg clusters (Treg1 and Treg2)
筛选两个 Treg 簇(Treg1 和 Treg2)之间的差异表达基因
We first obtained the log-normalized gene expression matrix (after SAVER imputation) through the method described above, removed genes with maximum expression values of less than 1 across all Treg cells, and then converted expression to a standardized z-score across all Treg cells. The DEGs between Treg1 and Treg2 cells were screened by a p value <0.01 (Mann-Whitney U test), FDR <0.05 (Bonferroni correction), and the difference in the average z score between Treg1 and Treg2 cells was more than 0.4 . The DEGs between Treg1 and Treg2 cells are listed in Table S4.
我们首先通过上述方法获得对数标准化基因表达矩阵 (SAVER 插补后),去除所有 Treg 细胞中最大表达值小于 1 的基因,然后将表达转换为所有 Treg 细胞的标准化 z 分数。通过 p 值 <0.01 (Mann-Whitney U 检验)、FDR <0.05 (Bonferroni 校正) 筛选 Treg1 和 Treg2 细胞之间的 DEGs,Treg1 和 Treg2 细胞之间平均 z 评分的差异大于 0.4 。表 S4 列出了 Treg1 和 Treg2 信元之间的 DEG。
我们首先通过上述方法获得对数标准化基因表达矩阵 (SAVER 插补后),去除所有 Treg 细胞中最大表达值小于 1 的基因,然后将表达转换为所有 Treg 细胞的标准化 z 分数。通过 p 值 <0.01 (Mann-Whitney U 检验)、FDR <0.05 (Bonferroni 校正) 筛选 Treg1 和 Treg2 细胞之间的 DEGs,Treg1 和 Treg2 细胞之间平均 z 评分的差异大于 0.4
Identifying the SLE Treg2 up-regulated DEGs
识别 SLE Treg2 上调的 DEG
We identified the SLE Treg2 up-regulated DEGs by the same thresholds described in “Screening of differentially expressed genes between two Treg clusters (Treg1 and Treg2).”, and the interferon response genes were removed in the analysis of Figure 6A, the SLE Treg2 up-regulated DEGs were listed in Table S4.
我们通过“筛选两个 Treg 簇(Treg1 和 Treg2)之间的差异表达基因”中描述的相同阈值鉴定了 SLE Treg2 上调的 DEGs,并且在图 6A 的分析中去除了干扰素反应基因,SLE Treg2 上调的 DEGs 列在表 S4 中。
我们通过“筛选两个 Treg 簇(Treg1 和 Treg2)之间的差异表达基因”中描述的相同阈值鉴定了 SLE Treg2 上调的 DEGs,并且在图 6A 的分析中去除了干扰素反应基因,SLE Treg2 上调的 DEGs 列在表 S4 中。
Identifying the signature genes for three published Treg groups
鉴定三个已发表的 Treg 组的特征基因
Previous studies defined three fractions of FOXP3+CD4+ T cells: naïve Tregs (nTregs), effector Tregs (eTregs) and nonsuppression Tregs (Fr. III). We used the bulk RNA-seq data to identify their signature genes (Cuadrado et al., 2018). We performed differential expression analysis (log2 fold change>1, p value<0.05 (t-test) and FDR <0.5 (Bonferroni correction)) for each Treg group versus all other Treg groups (eTreg vs nTreg and Fr. III; nTreg vs eTreg and Fr. III; Fr. III vs eTreg and nTreg). The signature genes of these three Treg groups are listed in Table S2. The relative expression of all signature genes in eTreg, nTreg and Fr. III cells are shown in the heatmap.
以前的研究定义了 FOXP3 + CD4 + T 细胞的三个部分:初始 Treg (nTregs)、效应 Treg (eTregs) 和非抑制性 Tregs (Fr. III)。我们使用大量 RNA-seq 数据来识别它们的特征基因 ( Cuadrado et al., 2018 )。我们对每个 Treg 组与所有其他 Treg 组 (eTreg vs nTreg 和 Fr. III;nTreg vs eTreg 和 Fr. III;Fr. III vs eTreg 和 Fr. III;Fr. III vs eTreg 和 nTreg)进行了差异表达分析(对数2 倍变化>1,p 值<0.05 (t 检验)和 FDR <0.5 (Bonferroni 校正))。这三个 Treg 组的特征基因列于表 S2 中。热图显示了 eTreg、nTreg 和 Fr. III 细胞中所有特征基因的相对表达。
以前的研究定义了 FOXP3 + CD4 + T 细胞的三个部分:初始 Treg (nTregs)、效应 Treg (eTregs) 和非抑制性 Tregs (Fr. III)。我们使用大量 RNA-seq 数据来识别它们的特征基因 ( Cuadrado et al., 2018 )。我们对每个 Treg 组与所有其他 Treg 组 (eTreg vs nTreg 和 Fr. III;nTreg vs eTreg 和 Fr. III;Fr. III vs eTreg 和 Fr. III;Fr. III vs eTreg 和 nTreg)进行了差异表达分析(对数2 倍变化>1,p 值<0.05 (t 检验)和 FDR <0.5 (Bonferroni 校正))。这三个 Treg 组的特征基因列于表 S2 中。热图显示了 eTreg、nTreg 和 Fr. III 细胞中所有特征基因的相对表达。
Screening specific genes for Th1, Th2, Th17, and Tct cells
筛选 Th1、Th2、Th17 和 Tct 细胞的特异性基因
We used the single-cell gene expression data of healthy controls to identify the specific genes of Th1, Th2, Th17 and Tct cells. Differential expression analysis (log2 fold change >0.2 and Mann-Whitney U test p value <0.01) was performed for each subtype of Teff cells versus all other subtypes of Teff cells. The specific genes of each Teff subtype are listed in Table S4.
我们使用健康对照的单细胞基因表达数据来鉴定 Th1 、 Th2 、 Th17 和 Tct 细胞的特定基因。对 Teff 细胞的每种亚型 >0.2 和 Mann-Whitney U 检验 p 值 <0.01) 进行差异表达分析,而不是所有其他 Teff 细胞亚型。表 S4 列出了每种 Teff 亚型的特定基因。
我们使用健康对照的单细胞基因表达数据来鉴定 Th1 、 Th2 、 Th17 和 Tct 细胞的特定基因。对 Teff 细胞的每种亚型 >0.2 和 Mann-Whitney U 检验 p 值 <0.01) 进行差异表达分析,而不是所有其他 Teff 细胞亚型。表 S4 列出了每种 Teff 亚型的特定基因。
Identifying the Treg exhaustion-associated functional gene set
鉴定 Treg 耗竭相关功能基因集
We used two different methods that each captured distinct types of information to identify Treg exhaustion-associated functions: (1) Metascape gene list annotation and (2) GSEA. We described each of these below. We first calculated the Pearson’s correlation between normalized gene expression and the signature score of Treg exhaustion-like properties for all detected genes across all Treg cells and then identified the genes whose expression was significantly correlated with the signature score of Treg exhaustion-like properties (correlation >0.4). We uploaded this gene list to the Metascape main page and clicked “express analysis”. The Metascape’s report page showed the functional annotations of Treg exhaustion-associated genes. We showed the top enriched functions via a bar plot. To further verify the correlation between Treg exhaustion-like properties and the type I IFN response, we performed GSEAPreranked analysis of Treg-detected genes (ranked by their Pearson correlation with the signature score of Treg exhaustion-like properties) and a type I IFN response gene list.
我们使用了两种不同的方法,每种方法都捕获了不同类型的信息来识别 Treg 耗竭相关功能:(1) Metascape 基因列表注释和 (2) GSEA。我们在下面逐一介绍了这些。我们首先计算了所有 Treg 细胞中所有检测到的基因的标准化基因表达与 Treg 耗竭样特性特征评分之间的 Pearson 相关性,然后确定了其表达与 Treg 耗竭样特性的特征评分显著相关的基因 (相关性 >0.4)。我们将此基因列表上传到 Metascape 主页并单击 “express analysis”。Metascape 的报告页面显示了 Treg 耗竭相关基因的功能注释。我们通过条形图显示了最丰富的函数。为了进一步验证 Treg 耗竭样特性与 I 型 IFN 反应之间的相关性,我们对 Treg 检测到的基因进行了 GSEAPreranked 分析(按它们与 Treg 耗竭样特性的特征评分的 Pearson 相关性排序)和 I 型 IFN 反应基因列表。
我们使用了两种不同的方法,每种方法都捕获了不同类型的信息来识别 Treg 耗竭相关功能:(1) Metascape 基因列表注释和 (2) GSEA。我们在下面逐一介绍了这些。我们首先计算了所有 Treg 细胞中所有检测到的基因的标准化基因表达与 Treg 耗竭样特性特征评分之间的 Pearson 相关性,然后确定了其表达与 Treg 耗竭样特性的特征评分显著相关的基因 (相关性 >0.4)。我们将此基因列表上传到 Metascape 主页并单击 “express analysis”。Metascape 的报告页面显示了 Treg 耗竭相关基因的功能注释。我们通过条形图显示了最丰富的函数。为了进一步验证 Treg 耗竭样特性与 I 型 IFN 反应之间的相关性,我们对 Treg 检测到的基因进行了 GSEAPreranked 分析(按它们与 Treg 耗竭样特性的特征评分的 Pearson 相关性排序)和 I 型 IFN 反应基因列表。
Quantification and statistical analysis
定量和统计分析
The analysis, software, and quantification methodology that are specific to ATAC-seq, RNA-seq, and scRNA-seq experiments are included under the relevant subsections of the method details section. Information regarding replicate numbers is provided in figure legends. If error bars are used in figures, information about what error bars represent is also provided in the figure legend. If the degree of significance is provided in the figure legend, further details regarding the statistical test used are provided in the relevant subsections of the method details that are specific to the analysis being performed.
特定于 ATAC-seq、RNA-seq 和 scRNA-seq 实验的分析、软件和定量方法包含在方法详细信息部分的相关小节下。有关重复编号的信息在图例中提供。如果在图窗中使用误差线,则图例中还会提供有关误差线所代表的信息。如果在图例中提供了显著性程度,则有关所用统计检验的更多详细信息,请参阅特定于正在执行的分析的方法详细信息的相关小节。
特定于 ATAC-seq、RNA-seq 和 scRNA-seq 实验的分析、软件和定量方法包含在方法详细信息部分的相关小节下。有关重复编号的信息在图例中提供。如果在图窗中使用误差线,则图例中还会提供有关误差线所代表的信息。如果在图例中提供了显著性程度,则有关所用统计检验的更多详细信息,请参阅特定于正在执行的分析的方法详细信息的相关小节。
Flow cytometry data were analysed using FlowJo V.X.0.7 software (Tree Star). Statistical analyses and approximations were performed with GraphPad Prism 7 software (GraphPad).
使用 FlowJo V.X.0.7 软件 (Tree Star) 分析流式细胞术数据。使用 GraphPad Prism 7 软件 (GraphPad) 进行统计分析和近似。
使用 FlowJo V.X.0.7 软件 (Tree Star) 分析流式细胞术数据。使用 GraphPad Prism 7 软件 (GraphPad) 进行统计分析和近似。
Statistical significance was analyzed with two-tailed Student’s-tests, Mann-Whitney U test, or two-way ANOVA. A p-value of less than 0.05 was considered statistically significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Data are presented as the upper, centre, and lower lines indicate the 75% quantile +1.5 ∗ interquartile range (IQR), 50% quantile, and 25% quantile −1.5 ∗ IQR, respectively which are indicated in the figure legends.
使用双尾 Student's 检验、Mann-Whitney U 检验或双向方差分析分析统计显著性。小于 0.05 的 p 值被认为具有统计学意义,∗p < 0.05,∗∗p < 0.01,∗∗∗p < 0.001,∗∗∗∗p < 0.0001。数据表示为上线、中线和下线表示 75% 分位数 +1.5 ∗四分位距 (IQR)、50% 分位数和 25% 分位数 −1.5 ∗ IQR,分别在图例中表示。
使用双尾 Student's 检验、Mann-Whitney U 检验或双向方差分析分析统计显著性。小于 0.05 的 p 值被认为具有统计学意义,∗p < 0.05,∗∗p < 0.01,∗∗∗p < 0.001,∗∗∗∗p < 0.0001。数据表示为上线、中线和下线表示 75% 分位数 +1.5 ∗四分位距 (IQR)、50% 分位数和 25% 分位数 −1.5 ∗ IQR,分别在图例中表示。
Acknowledgments 确认
Funding: this work was supported by the National Key R&D Program of China (2020YFA0112200 to K.Q.), the National Natural Science Foundation of China grants (T2125012, 91940306, 31970858, and 31771428 to K.Q.; 32270978 to C.G.), CAS Project for Young Scientists in Basic Research (YSBR-005 to K.Q.), Anhui Province Science and Technology Key Program (202003a07020021 to K.Q.), and the Fundamental Research Funds for the Central Universities (YD2070002019, WK9110000141, and WK2070000158 to K.Q.). We thank the USTC supercomputing center and the School of Life Science Bioinformatics Center for providing computing resources for this project.
资金:这项工作得到了国家重点研发计划(2020YFA0112200 至 K.Q.)、国家自然科学基金资助(T2125012、91940306、31970858 和 31771428;32270978 至 C.G.)、CAS 基础研究青年科学家项目(YSBR-005 至 K.Q.)、安徽省科技重点计划(202003a07020021 至 K.Q.), 以及中央大学基本研究基金(YD2070002019、WK9110000141 和 WK2070000158 到 K.Q.)。我们感谢 USTC 超级计算中心和生命科学学院生物信息学中心为本项目提供计算资源。
资金:这项工作得到了国家重点研发计划(2020YFA0112200 至 K.Q.)、国家自然科学基金资助(T2125012、91940306、31970858 和 31771428;32270978 至 C.G.)、CAS 基础研究青年科学家项目(YSBR-005 至 K.Q.)、安徽省科技重点计划(202003a07020021 至 K.Q.), 以及中央大学基本研究基金(YD2070002019、WK9110000141 和 WK2070000158 到 K.Q.)。我们感谢 USTC 超级计算中心和生命科学学院生物信息学中心为本项目提供计算资源。
Author contributions 作者贡献
K.Q. conceived and supervised the project. C.G. and D.Z. performed the experiments and conducted all the sample preparation for next-generation sequencing with help from Q.S., L.Z., Y.L., Y.S., and X.G. Q.L. performed the data analysis with help from W.Z., Z.Z., Q.Y., J.F., P.D., and P.C. W.Z. also contributed to the revision of the manuscript. X.L., Q.W., and J.T. provided blood samples from SLE patients and healthy control individuals. K.Q., C.G., and Q.L. wrote the manuscript with the help of all other authors.
K.Q. 构思并监督了这个项目。C.G. 和 D.Z. 在 Q.S.、L.Z.、Y.L.、Y.S. 和 X.G. 的帮助下进行了实验并进行了下一代测序的所有样品制备。Q.L. 在 W.Z.、Z.Z.、Q.Y.、J.F.、P.D. 和 P.C. W.Z. 的帮助下进行了数据分析。W.Z. 也为手稿的修订做出了贡献。X.L.、Q.W. 和 J.T. 提供了 SLE 患者和健康对照个体的血液样本。K.Q.、C.G. 和 Q.L. 在所有其他作者的帮助下撰写了手稿。
K.Q. 构思并监督了这个项目。C.G. 和 D.Z. 在 Q.S.、L.Z.、Y.L.、Y.S. 和 X.G. 的帮助下进行了实验并进行了下一代测序的所有样品制备。Q.L. 在 W.Z.、Z.Z.、Q.Y.、J.F.、P.D. 和 P.C. W.Z. 的帮助下进行了数据分析。W.Z. 也为手稿的修订做出了贡献。X.L.、Q.W. 和 J.T. 提供了 SLE 患者和健康对照个体的血液样本。K.Q.、C.G. 和 Q.L. 在所有其他作者的帮助下撰写了手稿。
Declaration of interests 利益申报
J.F. is the chief executive officer of HanGene Biotech. The other authors declare no competing interests.
J.F. 是 HanGene Biotech 的首席执行官。其他作者声明没有竞争利益。
J.F. 是 HanGene Biotech 的首席执行官。其他作者声明没有竞争利益。
Supplemental information (6)
补充信息 (6)
Document S1. Figures S1–S7
文档 S1.图 S1-S7
文档 S1.图 S1-S7
Table S1. Sample information, related to Figures 1, 3, and 4
表 S1.示例信息,与图 1、图 3 和 4 相关
表 S1.示例信息,与图 1、图 3 和 4 相关
Table S2. Gene sets collected from published studies or database and functional annotations for cluster I–V peaks, related to Figure 2
表 S2.从已发表的研究或簇 I-V 峰的数据库和功能注释中收集的基因集,与图 2 相关
表 S2.从已发表的研究或簇 I-V 峰的数据库和功能注释中收集的基因集,与图 2 相关
Table S3. DEGs and biological functional changes between SLE patients and healthy controls in each CD4+ T cell subtype, related to Figure 4
表 S3.SLE 患者和健康对照者在每种 CD4+ T 细胞亚型中的 DEG 和生物学功能变化,与 图 4 相关
表 S3.SLE 患者和健康对照者在每种 CD4+ T 细胞亚型中的 DEG 和生物学功能变化,与 图 4 相关
Table S4. DEGs and signature gene list, related to Figures 5, 6, and 7
表 S4.DEGs 和特征基因列表,与图 5、6 和 7 相关
表 S4.DEGs 和特征基因列表,与图 5、6 和 7 相关
Document S2. Article plus supplemental information
文档 S2.文章加补充信息
文档 S2.文章加补充信息
References 引用
Anders, S. ∙ Huber, W.
安德斯,S. ∙ 胡伯,W.
安德斯,S. ∙ 胡伯,W.
Differential expression analysis for sequence count data
序列计数数据的差异表达分析
Genome Biol. 2010; 11:R106序列计数数据的差异表达分析
基因组生物学。2010;11:R106
Aqrawi, L.A. ∙ Ivanchenko, M. ∙ Björk, A. ...
Aqrawi, L.A. ∙ Ivanchenko, M. ∙ Björk, A. ...
Aqrawi, L.A. ∙ Ivanchenko, M. ∙ Björk, A. ...
Diminished CXCR5 expression in peripheral blood of patients with Sjogren's syndrome may relate to both genotype and salivary gland homing
干燥综合征患者外周血中 CXCR5 表达降低可能与基因型和唾液腺归巢有关
Clin. Exp. Immunol. 2018; 192:259-270干燥综合征患者外周血中 CXCR5 表达降低可能与基因型和唾液腺归巢有关
临床实验免疫学。2018;192:259-270
Arazi, A. ∙ Rao, D.A. ∙ Berthier, C.C. ...
Arazi, A. ∙ Rao, DA ∙ Berthier, CC ...
Arazi, A. ∙ Rao, DA ∙ Berthier, CC ...
The immune cell landscape in kidneys of patients with lupus nephritis
狼疮性肾炎患者肾脏中的免疫细胞景观
Nat. Immunol. 2019; 20:902-914狼疮性肾炎患者肾脏中的免疫细胞景观
Nat. 免疫学。2019;20:902-914
Azizi, E. ∙ Carr, A.J. ∙ Plitas, G. ...
阿齐兹,E. ∙ 卡尔,AJ ∙ 普利塔斯,G. ...
阿齐兹,E. ∙ 卡尔,AJ ∙ 普利塔斯,G. ...
Single-cell map of diverse immune phenotypes in the breast tumor microenvironment
Cell. 2018; 174:1293-1308.e36细胞。2018;174:1293-1308.e36
Bacher, N. ∙ Raker, V. ∙ Hofmann, C. ...
Interferon-alpha suppresses cAMP to disarm human regulatory T cells
Cancer Res. 2013; 73:5647-5656Banchereau, R. ∙ Hong, S. ∙ Cantarel, B. ...
Personalized immunomonitoring uncovers molecular networks that stratify lupus patients
Cell. 2016; 165:1548-1550Boland, B.S. ∙ He, Z. ∙ Tsai, M.S. ...
Heterogeneity and clonal relationships of adaptive immune cells in ulcerative colitis revealed by single-cell analyses
Sci. Immunol. 2020; 5:eabb4432Brightbill, H.D. ∙ Suto, E. ∙ Blaquiere, N. ...
NF-kappaB inducing kinase is a therapeutic target for systemic lupus erythematosus
Nat. Commun. 2018; 9:179Caielli, S. ∙ Veiga, D.T. ∙ Balasubramanian, P. ...
A CD4(+) T cell population expanded in lupus blood provides B cell help through interleukin-10 and succinate
Nat. Med. 2018; 25:75-81Cao, J. ∙ Spielmann, M. ∙ Qiu, X. ...
The single-cell transcriptional landscape of mammalian organogenesis
Nature. 2019; 566:496-502Chevalier, N. ∙ Jarrossay, D. ∙ Ho, E. ...
CXCR5 expressing human central memory CD4 T cells and their relevance for humoral immune responses
J. Immunol. 2011; 186:5556-5568Coit, P. ∙ Dozmorov, M.G. ∙ Merrill, J.T. ...
Epigenetic reprogramming in naive CD4+T cells favoring T cell activation and non-Th1 effector T cell immune response as an early event in lupus flares
Arthritis Rheumatol. 2016; 68:2200-2209Crow, M.K. ∙ Olferiev, M. ∙ Kirou, K.A.
Type I interferons in autoimmune disease
Annu. Rev. Pathol. 2019; 14:369-393Cuadrado, E. ∙ van den Biggelaar, M. ∙ de Kivit, S. ...
Proteomic analyses of human regulatory T cells reveal adaptations in signaling pathways that protect cellular identity
Immunity. 2018; 48:1046-1059.e6Dominguez-Villar, M. ∙ Baecher-Allan, C.M. ∙ Hafler, D.A.
Identification of T helper type 1-like, Foxp3+ regulatory T cells in human autoimmune disease
Nat. Med. 2011; 17:673-675Dominguez-Villar, M. ∙ Hafler, D.A.
Regulatory T cells in autoimmune disease
Nat. Immunol. 2018; 19:665-673Gangaplara, A. ∙ Martens, C. ∙ Dahlstrom, E. ...
Type I interferon signaling attenuates regulatory T cell function in viral infection and in the tumor microenvironment
PLoS Pathog. 2018; 14:e1006985Gladman, D.D. ∙ Ibañez, D. ∙ Urowitz, M.B.
Systemic lupus erythematosus disease activity index 2000
J. Rheumatol. 2002; 29:288-291Golding, A. ∙ Rosen, A. ∙ Petri, M. ...
Interferon-alpha regulates the dynamic balance between human activated regulatory and effector T cells: implications for antiviral and autoimmune responses
Immunology. 2010; 131:107-117Goropevsek, A. ∙ Holcar, M. ∙ Avčin, T.
The role of STAT signaling pathways in the pathogenesis of systemic lupus erythematosus
Clin. Rev. Allergy Immunol. 2017; 52:164-181Hashimoto, H. ∙ Ueda, R. ∙ Narumi, K. ...
Type I IFN gene delivery suppresses regulatory T cells within tumors
Cancer Gene Ther. 2014; 21:532-541He, J. ∙ Tsai, L.M. ∙ Leong, Y.A. ...
Circulating precursor CCR7(lo)PD-1(hi) CXCR5(+) CD4(+) T cells indicate Tfh cell activity and promote antibody responses upon antigen reexposure
Immunity. 2013; 39:770-781Heaton, H. ∙ Talman, A.M. ∙ Knights, A. ...
Souporcell: robust clustering of single-cell RNA-seq data by genotype without reference genotypes
Nat. Methods. 2020; 17:615-620Hedrich, C.M. ∙ Mäbert, K. ∙ Rauen, T. ...
DNA methylation in systemic lupus erythematosus
Epigenomics. 2017; 9:505-525Hu, N. ∙ Qiu, X. ∙ Luo, Y. ...
Abnormal histone modification patterns in lupus CD4+T cells
J. Rheumatol. 2008; 35:804-810Huang, M. ∙ Wang, J. ∙ Torre, E. ...
SAVER: gene expression recovery for single-cell RNA sequencing
Nat. Methods. 2018; 15:539-542Hutcheson, J. ∙ Scatizzi, J.C. ∙ Siddiqui, A.M. ...
Combined deficiency of proapoptotic regulators Bim and Fas results in the early onset of systemic autoimmunity
Immunity. 2008; 28:206-217Kang, H.M. ∙ Subramaniam, M. ∙ Targ, S. ...
Multiplexed droplet single-cell RNA-sequencing using natural genetic variation
Nat. Biotechnol. 2018; 36:89-94Kim, K. ∙ Park, S. ∙ Park, S.Y. ...
Single-cell transcriptome analysis reveals TOX as a promoting factor for T cell exhaustion and a predictor for anti-PD-1 responses in human cancer
Genome Med. 2020; 12:22Kim, S.J. ∙ Lee, K. ∙ Diamond, B.
Follicular helper T cells in systemic lupus erythematosus
Front. Immunol. 2018; 9:1793Koboldt, D.C. ∙ Zhang, Q. ∙ Larson, D.E. ...
VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing
Genome Res. 2012; 22:568-576Koga, T. ∙ Otomo, K. ∙ Mizui, M. ...
Calcium/calmodulin-dependent kinase IV facilitates the recruitment of interleukin-17-producing cells to target organs through the CCR6/CCL20 Axis in Th17 cell-driven inflammatory diseases
Arthritis Rheumatol. 2016; 68:1981-1988Komatsu, N. ∙ Okamoto, K. ∙ Sawa, S. ...
Pathogenic conversion of Foxp3+ T cells into TH17 cells in autoimmune arthritis
Nat. Med. 2014; 20:62-68Korn, T. ∙ Reddy, J. ∙ Gao, W. ...
Myelin-specific regulatory T cells accumulate in the CNS but fail to control autoimmune inflammation
Nat. Med. 2007; 13:423-431Korsunsky, I. ∙ Millard, N. ∙ Fan, J. ...
Fast, sensitive and accurate integration of single-cell data with Harmony
Nat. Methods. 2019; 16:1289-1296Langmead, B. ∙ Salzberg, S.L.
Fast gapped-read alignment with Bowtie 2
Nat. Methods. 2012; 9:357-359Lee, S.E. ∙ Li, X. ∙ Kim, J.C.K. ...
Type I interferons maintain Foxp3 expression and T-regulatory cell functions under inflammatory conditions in mice
Gastroenterology. 2012; 143:145-154Liberzon, A. ∙ Birger, C. ∙ Thorvaldsdóttir, H. ...
The Molecular Signatures Database (MSigDB) hallmark gene set collection
Cell Syst. 2015; 1:417-425Lin, J. ∙ Yu, Y. ∙ Ma, J. ...
PD-1+CXCR5-CD4+T cells are correlated with the severity of systemic lupus erythematosus
Rheumatology. 2019; 58:2188-2192Lisnevskaia, L. ∙ Murphy, G. ∙ Isenberg, D.
Systemic lupus erythematosus
Lancet. 2014; 384:1878-1888Lowther, D.E. ∙ Goods, B.A. ∙ Lucca, L.E. ...
PD-1 marks dysfunctional regulatory T cells in malignant gliomas
JCI Insight. 2016; 1:e85935Luckheeram, R.V. ∙ Zhou, R. ∙ Verma, A.D. ...
CD4(+)T cells: differentiation and functions
Clin. Dev. Immunol. 2012; 2012:925135Makiyama, A. ∙ Chiba, A. ∙ Noto, D. ...
Expanded circulating peripheral helper T cells in systemic lupus erythematosus: association with disease activity and B cell differentiation
Rheumatology. 2019; 58:1861-1869McClymont, S.A. ∙ Putnam, A.L. ∙ Lee, M.R. ...
Plasticity of human regulatory T cells in healthy subjects and patients with type 1 diabetes
J. Immunol. 2011; 186:3918-3926McLean, C.Y. ∙ Bristor, D. ∙ Hiller, M. ...
GREAT improves functional interpretation of cis-regulatory regions
Nat. Biotechnol. 2010; 28:495-501Metidji, A. ∙ Rieder, S.A. ∙ Glass, D.D. ...
IFN-alpha/beta receptor signaling promotes regulatory T cell development and function under stress conditions
J. Immunol. 2015; 194:4265-4276Miyara, M. ∙ Yoshioka, Y. ∙ Kitoh, A. ...
Functional delineation and differentiation dynamics of human CD4+ T cells expressing the FoxP3 transcription factor
Immunity. 2009; 30:899-911Mootha, V.K. ∙ Lindgren, C.M. ∙ Eriksson, K.F. ...
PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes
Nat. Genet. 2003; 34:267-273Morand, E.F. ∙ Furie, R. ∙ Tanaka, Y. ...
Trial of anifrolumab in active systemic lupus erythematosus
N. Engl. J. Med. 2020; 382:211-221Morita, R. ∙ Schmitt, N. ∙ Bentebibel, S.E. ...
Human blood CXCR5(+)CD4(+) T cells are counterparts of T follicular cells and contain specific subsets that differentially support antibody secretion (vol 34, pg 108, 2011)
Immunity. 2011; 34:135Nehar-Belaid, D. ∙ Hong, S. ∙ Marches, R. ...
Mapping systemic lupus erythematosus heterogeneity at the single-cell level
Nat. Immunol. 2020; 21:1094-1106Patil, V.S. ∙ Madrigal, A. ∙ Schmiedel, B.J. ...
Precursors of human CD4+ cytotoxic T lymphocytes identified by single-cell transcriptome analysis
Sci. Immunol. 2018; 3:eaan8664Picelli, S. ∙ Faridani, O.R. ∙ Björklund, A.K. ...
Full-length RNA-seq from single cells using Smart-seq2
Nat. Protoc. 2014; 9:171-181Piconese, S. ∙ Pacella, I. ∙ Timperi, E. ...
Divergent effects of type-I interferons on regulatory T cells
Cytokine Growth Factor Rev. 2015; 26:133-141Qu, K. ∙ Zaba, L.C. ∙ Satpathy, A.T. ...
Chromatin accessibility landscape of cutaneous T cell lymphoma and dynamic response to HDAC inhibitors
Cancer Cell. 2017; 32:27-41.e4Quinlan, A.R. ∙ Hall, I.M.
BEDTools: a flexible suite of utilities for comparing genomic features
Bioinformatics. 2010; 26:841-842Rome, K.S. ∙ Stein, S.J. ∙ Kurachi, M. ...
Trib1 regulates T cell differentiation during chronic infection by restraining the effector program
J. Exp. Med. 2020; 217:e20190888Sakaguchi, S. ∙ Mikami, N. ∙ Wing, J.B. ...
Regulatory T cells and human disease
Annu. Rev. Immunol. 2020; 38:541-566Sallusto, F.
Heterogeneity of human CD4+T cells against microbes
Annu. Rev. Immunol. 2016; 34:317-334Schafflick, D. ∙ Xu, C.A. ∙ Hartlehnert, M. ...
Integrated single cell analysis of blood and cerebrospinal fluid leukocytes in multiple sclerosis
Nat. Commun. 2020; 11:247Scheinecker, C. ∙ Göschl, L. ∙ Bonelli, M.
Treg cells in health and autoimmune diseases: new insights from single cell analysis
J. Autoimmun. 2020; 110:102376Scott, A.C. ∙ Dündar, F. ∙ Zumbo, P. ...
TOX is a critical regulator of tumour-specific T cell differentiation
Nature. 2019; 571:270-274Sharma, S. ∙ Jin, Z. ∙ Rosenzweig, E. ...
Widely divergent transcriptional patterns between SLE patients of different ancestral backgrounds in sorted immune cell populations
J. Autoimmun. 2015; 60:51-58Smillie, C.S. ∙ Biton, M. ∙ Ordovas-Montanes, J. ...
Intra- and inter-cellular rewiring of the human colon during ulcerative colitis
Cell. 2019; 178:714-730.e22Srivastava, S. ∙ Koch, M.A. ∙ Pepper, M. ...
Type I interferons directly inhibit regulatory T cells to allow optimal antiviral T cell responses during acute LCMV infection
J. Exp. Med. 2014; 211:961-974Stuart, T. ∙ Butler, A. ∙ Hoffman, P. ...
Comprehensive integration of single-cell data
Cell. 2019; 177:1888-1902.e21Suarez-Fueyo, A. ∙ Bradley, S.J. ∙ Tsokos, G.C.
T cells in systemic lupus erythematosus
Curr. Opin. Immunol. 2016; 43:32-38Sumida, T. ∙ Lincoln, M.R. ∙ Ukeje, C.M. ...
Activated beta-catenin in Foxp3(+) regulatory T cells links inflammatory environments to autoimmunity
Nat. Immunol. 2018; 19:1391-1402Takeshita, M. ∙ Suzuki, K. ∙ Kassai, Y. ...
Polarization diversity of human CD4(+) stem cell memory T cells
Clin. Immunol. 2015; 159:107-117Takeuchi, A. ∙ Saito, T.
CD4 CTL, a cytotoxic subset of CD4(+) T cells, their differentiation and function
Front. Immunol. 2017; 8:194Tsokos, G.C.
Autoimmunity and organ damage in systemic lupus erythematosus
Nat. Immunol. 2020; 21:605-614Tsokos, G.C. ∙ Lo, M.S. ∙ Costa Reis, P. ...
New insights into the immunopathogenesis of systemic lupus erythematosus
Nat. Rev. Rheumatol. 2016; 12:716-730Welter, D. ∙ MacArthur, J. ∙ Morales, J. ...
The NHGRI GWAS Catalog, a curated resource of SNP-trait associations
Nucleic Acids Res. 2014; 42:D1001-D1006Wherry, E.J. ∙ Kurachi, M.
Molecular and cellular insights into T cell exhaustion
Nat. Rev. Immunol. 2015; 15:486-499Wilk, A.J. ∙ Rustagi, A. ∙ Zhao, N.Q. ...
A single-cell atlas of the peripheral immune response in patients with severe COVID-19
Nat. Med. 2020; 26:1070-1076Wing, J.B. ∙ Tanaka, A. ∙ Sakaguchi, S.
Human FOXP3(+) regulatory T cell heterogeneity and function in autoimmunity and cancer
Immunity. 2019; 50:302-316Yang, K. ∙ Blanco, D.B. ∙ Neale, G. ...
Homeostatic control of metabolic and functional fitness of Treg cells by LKB1 signalling
Nature. 2017; 548:602-606Yee, C.S. ∙ Farewell, V. ∙ Isenberg, D.A. ...
British Isles Lupus Assessment Group 2004 index is valid for assessment of disease activity in systemic lupus erythematosus
Arthritis Rheum. 2007; 56:4113-4119Zhou, Y. ∙ Zhou, B. ∙ Pache, L. ...
Metascape provides a biologist-oriented resource for the analysis of systems-level datasets
Nat. Commun. 2019; 10:1523Zuo, Z. ∙ Jin, Y. ∙ Zhang, W. ...
ATAC-pipe: general analysis of genome-wide chromatin accessibility
Brief. Bioinform. 2019; 20:1934-1943Figures (7) 图 (7)
Article metrics 文章指标
-
-
22Citations 引文42Captures 捕获1Mentions 提到
-
Supplemental information (6)
补充信息 (6)
PDF (3.53 MB) PDF下载 (3.53 MB)
Document S1. Figures S1–S7
文档 S1.图 S1-S7
文档 S1.图 S1-S7
Spreadsheet (328.42 KB) 电子表格 (328.42 KB)
Table S1. Sample information, related to Figures 1, 3, and 4
表 S1.示例信息,与图 1、图 3 和 4 相关
表 S1.示例信息,与图 1、图 3 和 4 相关
Spreadsheet (236.99 KB) 电子表格 (236.99 KB)
Table S2. Gene sets collected from published studies or database and functional annotations for cluster I–V peaks, related to Figure 2
表 S2.从已发表的研究或簇 I-V 峰的数据库和功能注释中收集的基因集,与图 2 相关
表 S2.从已发表的研究或簇 I-V 峰的数据库和功能注释中收集的基因集,与图 2 相关
Spreadsheet (125.08 KB) 电子表格 (125.08 KB)
Table S3. DEGs and biological functional changes between SLE patients and healthy controls in each CD4+ T cell subtype, related to Figure 4
表 S3.SLE 患者和健康对照者在每种 CD4+ T 细胞亚型中的 DEG 和生物学功能变化,与 图 4 相关
表 S3.SLE 患者和健康对照者在每种 CD4+ T 细胞亚型中的 DEG 和生物学功能变化,与 图 4 相关
Spreadsheet (77.04 KB) 电子表格 (77.04 KB)
Table S4. DEGs and signature gene list, related to Figures 5, 6, and 7
表 S4.DEGs 和特征基因列表,与图 5、6 和 7 相关
表 S4.DEGs 和特征基因列表,与图 5、6 和 7 相关