Rhinitis, sinusitis, and ocular allergy 鼻炎、鼻窦炎和眼部过敏
Single cell RNA sequencing of human eosinophils from nasal polyps reveals eosinophil heterogeneity in chronic rhinosinusitis tissue 人鼻息肉中人类嗜酸性粒细胞的单细胞 RNA 测序揭示了慢性鼻窦炎组织中嗜酸性粒细胞的异质性
Chronic rhinosinusitis with nasal polyps (CRSwNP) is characterized by type 2 inflammation in the United States, but the actual roles that eosinophils play in CRSwNP remain largely unclear. 慢性鼻窦炎伴 鼻息肉（CRSwNP）在美国以 2 型炎症为特征，但 嗜酸性粒细胞在 CRSwNP 中的实际作用仍然 largely unclear。
Objective 目标
To reveal the roles and heterogeneity of eosinophils in nasal polyp (NP) tissue, we performed single cell RNA sequencing (scRNA-Seq) analysis of NP tissue.
Methods 方法
Sinonasal tissues (NP and control sinus tissue) and patient matched peripheral blood (PB) samples were obtained from 5 control patients and 5 patients with CRSwNP. Eosinophils were enriched before processing for scRNA-Seq. The gene expression profiles in eosinophils were determined by microwell-based scRNA-Seq technology (BD Rhapsody platform). We predicted the overall function of NP eosinophils by Gene Ontology (geneontology.org) enrichment and pathway analyses and confirmed expression of selected genes by flow cytometry.
Results 结果
After filtering out contaminating cells, we detected 5,542 eosinophils from control PB, 3,883 eosinophils from CRSwNP PB, 101 eosinophils from control sinus tissues (not included in further analyses), and 9,727 eosinophils from NPs by scRNA-Seq. We found that 204 genes were downregulated and 354 genes upregulated in NP eosinophils compared to all PB eosinophils (>1.5-fold, Padj < .05). Upregulated genes in NP eosinophils were associated with activation, cytokine-mediated signaling, growth factor activity, NF-κB signaling, and antiapoptotic molecules. NP eosinophils displayed 4 clusters revealing potential heterogeneity of eosinophils in NP tissue.
Conclusions 结论
Elevated eosinophils in NP tissue appear to exist in several subtypes that may play important pathogenic roles in CRSwNP, in part by controlling inflammation and hyperproliferation of other cells. 在 NP 组织中升高的嗜酸性粒细胞似乎存在几种亚型，这些亚型可能在 CRSwNP 中发挥重要的致病作用，部分通过控制炎症和其他细胞的过度增殖。
Chronic rhinosinusitis (CRS) is a common chronic illness affecting tens of millions of Americans and has an outsized impact on quality of life. CRS accounts for more than 250,000 surgeries per year in the United States, resulting in direct and indirect health care utilization costs of $8 billion to $20 billion annually.1, 2, 3 CRS is frequently divided into 2 main phenotypes: CRS with nasal polyps (CRSwNP) and CRS without nasal polyps (NPs). In Western countries, including the United States, 80% to 90% of CRSwNP patients are characterized by the type 2 endotype.3, 4, 5, 6, 7 慢性鼻窦炎（CRS）是一种常见的慢性疾病，影响数千万美国人，并对生活质量产生重大影响。CRS 每年在美国占据超过 250,000 例手术，导致每年直接和间接的医疗费用高达 80 亿到 200 亿美元。1，2，3 CRS 通常分为两种主要表型：伴有鼻息肉的 CRS（CRSwNP）和不伴有鼻息肉的 CRS（NPs）。在包括美国在内的西方国家，80%到 90%的 CRSwNP 患者被特征为 2 型内型。3，4，5，6，7
Eosinophilic inflammation is highly associated with severity of CRS. Surgical failure rates for CRSwNP have been associated with the presence and intensity of eosinophilic inflammation.8,9 Activation of eosinophils is also associated with epithelial barrier dysfunction through release of granule proteins and eosinophil extracellular traps.10 Mepolizumab, a monoclonal antibody against IL-5 (a key cytokine promoting eosinophil differentiation, recruitment, activation, and survival), has been approved for treatment of patients with CRSwNP.11 In contrast, a small clinical trial of dexpramipexole, an oral drug that also depletes eosinophils, showed >95% reduction of peripheral blood (PB) and NP tissue eosinophils unattended by any reduction in polyp size or improvement of patient-reported symptoms.12 This unexpected result has prompted much discussion regarding the role of eosinophils in the pathogenesis of CRSwNP. To date, single cell molecular analysis of eosinophil gene expression in NPs has not been technically possible. The current study overcomes this technical barrier, allowing analysis of eosinophil activation and function in NPs. 嗜酸性炎症与慢性鼻窦炎的严重程度高度相关。CRSwNP 的手术失败率与嗜酸性炎症的存在和强度有关。89 嗜酸细胞的激活也与上皮屏障功能障碍相关，这通过释放颗粒蛋白和嗜酸细胞细胞外陷阱实现。10 美泊利单抗是一种针对 IL-5（促进嗜酸细胞分化、招募、激活和存活的关键细胞因子）的单克隆抗体，已被批准用于治疗 CRSwNP 患者。11 相比之下，一项关于小型临床试验 的dexpramipexole（一种也能消耗嗜酸细胞的口服药物）显示外周血（PB）和鼻息肉组织中的嗜酸细胞减少超过 95%，但并未伴随息肉大小的减少或患者报告症状的改善。12 这一意外结果引发了关于嗜酸性粒细胞在慢性鼻窦炎伴鼻息肉（CRSwNP）发病机制中作用的广泛讨论。迄今为止，嗜酸性粒细胞在鼻息肉中的基因表达的单细胞分子分析在技术上尚不可行。本研究克服了这一技术障碍，使得能够分析嗜酸性粒细胞在鼻息肉中的激活和功能。
Transcriptomics is a useful tool for studying comprehensive and unbiased gene expression profiles, offering insights into cell types and identifying pathogenic factors in tissue,13 but many published studies, including our own, have failed to detect eosinophils in whole NP tissue by single cell RNA sequencing (scRNA-Seq), although eosinophils were highly accumulated in NP tissue.14, 15, 16, 17, 18 Furthermore, several groups also failed to detect human eosinophils by scRNA-Seq in tissues and body fluids of other type 2 (eosinophilic) inflammatory diseases, including eosinophilic esophagitis, allergic asthma, and atopic dermatitis.19, 20, 21 These technical challenges have prevented the investigation of eosinophils in human diseases by scRNA-Seq. In the present study, we successfully combined microwell-based scRNA-Seq technology with a manual enrichment protocol for granulocytes in order to detect eosinophils by scRNA-Seq and assess their potential roles in NPs. 转录组学是研究全面和无偏基因表达谱的有用工具，提供了对细胞类型的洞察，并识别组织中的致病因素，13 但许多已发表的研究，包括我们自己的研究，未能通过单细胞 RNA 测序（scRNA-Seq）在整个 NP 组织中检测到嗜酸性粒细胞，尽管嗜酸性粒细胞在 NP 组织中高度积累。14，15，16，17，18 此外，几个研究小组也未能通过 scRNA-Seq 在其他类型 2（嗜酸性）炎症性疾病的组织和体液中检测到人类嗜酸性粒细胞，包括嗜酸性食管炎、过敏性哮喘和特应性皮炎。19, 20, 21 这些技术挑战阻碍了通过 scRNA-Seq 对人类疾病中的嗜酸性粒细胞进行研究。在本研究中，我们成功地将基于微孔的 scRNA-Seq 技术与手动富集颗粒细胞的协议相结合，以便通过 scRNA-Seq 检测嗜酸性粒细胞并评估它们在 NPs 中的潜在作用。
Methods 方法
Patients and tissue collection 患者和组织收集
Patients with CRSwNP and control patients were recruited from the otolaryngology clinic and the Northwestern Sinus Center of Northwestern Medicine. NP tissues were obtained during routine endoscopic sinus surgery performed on patients with CRS. All CRS patients met the criteria for the International Consensus Statement on Allergy and Rhinology: Rhinosinusitis.1 All CRS patients recruited had disease that had failed to respond to appropriate and maximal medical therapy before surgery. Patients excluded from the study included those with an established immunodeficiency, pregnancy, coagulation disorder, diagnosis of eosinophilic granulomatous polyangiitis, or cystic fibrosis. We also excluded patients who had received biologic therapies such as mepolizumab, benralizumab, dupilumab, and omalizumab. Control tissue was obtained from patients without CRS who were undergoing either removal of a skull-base tumor or septoplasty. For each control patient, we pooled all available sinus tissues (ie, uncinate tissue, inferior turbinate tissue, and ethmoid tissue) for our analysis. Characteristics of the subjects in this study are shown in Table E1 in the Online Repository available at www.jacionline.org. All patients signed informed consent forms, and the protocols and governing procedures for this study were approved by the institutional review board of the Northwestern University Feinberg School of Medicine. 患有 CRSwNP 的患者和对照患者是从耳鼻喉科诊所和西北医学中心的西北鼻窦中心招募的。NP 组织是在对 CRS 患者进行常规内窥镜鼻窦手术时获得的。所有 CRS 患者均符合国际过敏与鼻科学共识声明的标准： 鼻窦炎。1 所有招募的 CRS 患者在手术前均未对适当和最大限度的药物治疗产生反应。被排除在研究之外的患者包括那些有既往免疫缺陷、怀孕、凝血障碍、嗜酸性肉芽肿性血管炎或囊性纤维化诊断的患者。我们还排除了接受过生物治疗的患者，如美泊利单抗、贝那利单抗、杜匹单抗和奥马珠单抗。对照组织来自于没有 CRS 的患者，这些患者正在接受颅底肿瘤切除或鼻中隔成形术。 对于每个对照患者，我们汇总了所有可用的鼻窦组织（即钩突组织、下鼻甲组织和筛骨组织）进行分析。本研究中受试者的特征显示在表 E1中，在线资料库可在www.jacionline.org找到。所有患者均签署了知情同意书，本研究的协议和管理程序已获得西北大学范伯格医学院机构审查委员会的批准。
Granulocyte enrichment 粒细胞富集
We isolated eosinophils and neutrophils (another difficult cell type to detect by scRNA-Seq in NPs) from whole peripheral blood using the EasySep Direct Human Eosinophil Isolation Kit (STEMCELL Technologies, Vancouver, British Columbia, Canada) and the EasySep Direct Human Neutrophil Isolation Kit (STEMCELL Technologies), respectively, according to the manufacturer’s instructions. The purity of peripheral blood eosinophils and neutrophils was 90-99% and 92-97%, respectively (see Fig E1, A, in the Online Repository available at www.jacionline.org). We then pooled peripheral blood (PB) eosinophils and neutrophils in a 1:1 ratio before staining cells with antibodies and a sample tag (Fig E1, B). We used a bead-free CD66b-positive selection to enrich granulocytes from tissue because CD66b is specifically expressed on both eosinophils and neutrophils.22, 23, 24 Isolated NP cells were incubated with 0.5 μg/mL PE anti-CD66b (clone 6/40c). Following this, we used the EasySep Release Human PE Positive Selection Kit (STEMCELL Technologies) to isolate bead-free CD66b+ granulocytes liberated from NP tissues. The purity of enriched NP tissue granulocytes (eosinophils plus neutrophils) was 55-85% (Fig E1, A). Detailed methods are available in the Methods section of the Online Repository. 我们从全外周血中分离了 嗜酸性粒细胞 和 中性粒细胞（在 NPs 中通过 scRNA-Seq 检测的另一种难以识别的细胞类型），使用 EasySep 直接人类 嗜酸性粒细胞分离试剂盒（STEMCELL Technologies，加拿大不列颠哥伦比亚省温哥华）和 EasySep 直接人类 中性粒细胞分离试剂盒（STEMCELL Technologies），按照制造商的说明进行操作。外周血嗜酸性粒细胞和中性粒细胞的纯度分别为 90-99%和 92-97%（见 图 E1, A, 在可用的在线资料库中 www.jacionline.org）。然后，我们将外周血（PB）嗜酸性粒细胞和中性粒细胞以 1:1 的比例混合，然后用抗体和样本标签对细胞进行染色（图 E1, B）。我们使用无珠 CD66b 阳性选择从组织中富集粒细胞，因为 CD66b 在嗜酸性粒细胞和中性粒细胞上均特异性表达。22, 23, 24 隔离的 NP 细胞与 0.5 μg/mL PE 抗 CD66b（克隆 6/40c）孵育。随后，我们使用 EasySep Release Human PE Positive Selection Kit（STEMCELL Technologies）从 NP 组织中分离出无珠 CD66b+ 粒细胞。富集的 NP 组织粒细胞（嗜酸细胞加中性粒细胞）的纯度为 55-85%（图 E1, A）。详细方法可在在线库的“方法”部分找到。
Single cell RNA sequencing 单细胞 RNA 测序
The pooled PB granulocytes were labeled with oligonucleotide-conjugated AbSeq antibodies (BD Biosciences, Franklin Lakes, NJ) against human CD16 (clone 3G8) and CCR3 (clone 5E8), and a sample tag (BD Biosciences) (Fig E1, B). Enriched tissue–derived granulocytes were also labeled with the same AbSeq antibodies but with a different sample tag to distinguish blood and tissue cells during data analysis. Cells were next stained with Calcein AM (Invitrogen; Thermo Fisher Scientific, Waltham, Mass) and DRAQ7 (BD Biosciences), and live cell numbers were determined by the BD Rhapsody Scanner (BD Biosciences). pooled PB 粒细胞使用针对人类 CD16 (克隆 3G8) 和 CCR3 (克隆 5E8) 的寡核苷酸结合的 AbSeq 抗体 (BD Biosciences, Franklin Lakes, NJ) 标记，并使用样本标签 (BD Biosciences) (图 E1, B)。富集的组织来源粒细胞也使用相同的 AbSeq 抗体标记，但使用不同的样本标签以便在数据分析中区分血液和组织细胞。接下来，细胞用 Calcein AM (Invitrogen; Thermo Fisher Scientific, Waltham, Mass) 和 DRAQ7 (BD Biosciences) 染色，活细胞数量通过 BD Rhapsody 扫描仪 (BD Biosciences) 确定。
We performed single cell separation and library preparation using the microwell-based BD Rhapsody scRNA-Seq platform with the BD Rhapsody Whole Transcriptome Analysis (WTA) reagent kit (BD Biosciences), according to the manufacturer’s instructions. Briefly, tissue-derived and PB granulocytes from each donor were mixed, and 12,000 to 20,000 cells were loaded on the BD Rhapsody cartridge for single cell separation (see Table E2 in the Online Repository available at www.jacionline.org). Within 30 minutes of retrieving cell capture beads, we initiated reverse transcription to prepare cDNA. We then prepared libraries for RNA transcripts, AbSeq antibodies, and sample tags according to the manufacturer’s instructions. Libraries were sequenced on the NovaSeq 6000 device (Illumina, San Diego, Calif) to generate 100 or 150 bp paired-end reads. The FASTQ files were processed using the BD Rhapsody WTA analysis pipeline with the GRCh38 human genome as reference (Seven Bridges, Charlestown, Mass), and data were analyzed by SeqGeq 1.8.0 (BD Biosciences). Quality control (QC) measurements for each sample are provided in Table E2. 我们使用基于微孔的 BD Rhapsody scRNA-Seq 平台和 BD Rhapsody 全转录组分析（WTA）试剂盒（BD Biosciences）进行单细胞分离和文库准备，按照制造商的说明进行。简而言之，来自每个供体的组织衍生细胞和 PB 粒细胞被混合，12,000 到 20,000 个细胞被加载到 BD Rhapsody 卡中进行单细胞分离（请参见表 E2，可在www.jacionline.org的在线库中找到）。在提取细胞捕获珠后的 30 分钟内，我们开始反转录以准备 cDNA。然后，我们根据制造商的说明准备 RNA 转录本、AbSeq 抗体和样本标签的文库。文库在 NovaSeq 6000 设备（Illumina，圣地亚哥，加利福尼亚）上进行测序，以生成 100 或 150 bp 的配对末端读取。使用 BD Rhapsody WTA 分析管道处理 FASTQ 文件，以 GRCh38 人类基因组作为参考（Seven Bridges，查尔斯敦，马萨诸塞州），数据通过 SeqGeq 1.8.0（BD Biosciences）进行分析。 每个样本的质量控制（QC）测量结果在表 E2中提供。
We obtained 62,898 putative cell counts from 5 control patients and 5 CRSwNP patients and used the following QC procedure: (1) select sample tag positive cells, (2) exclude doublets, (3) select cells that had more than 200 transcripts per cell, and (4) select cells for which the percentage of mitochondrial genes was less than 15%. After QC, we detected 48,131 cells that were used for analysis. To reduce noise, we selected 3,074 highly dispersed genes that presented a dispersion factor greater than 2.5 and were expressed in more than 10 cells; unsupervised clustering with dimension reduction analysis was performed by the Seurat 4.0.5 plug-in with a clustering resolution of 0.4 at SeqGeq. This analysis included normalization, principal component analysis, clustering, and dimensionality reduction with uniform manifold approximation and projection (UMAP), and differential expression analysis of clusters. 我们从 5 名对照患者和 5 名CRSwNP患者获得了 62,898 个假定细胞计数，并使用以下质量控制程序：(1) 选择样本标签阳性细胞，(2) 排除双重细胞，(3) 选择每个细胞转录本超过 200 的细胞，以及(4) 选择线粒体基因百分比低于 15%的细胞。经过质量控制后，我们检测到 48,131 个细胞用于分析。为了减少噪声，我们选择了 3,074 个高度分散的基因，这些基因的分散因子大于 2.5，并且在超过 10 个细胞中表达；使用 Seurat 4.0.5 插件进行无监督聚类和降维分析，聚类分辨率为 0.4，分析在 SeqGeq 中进行。该分析包括标准化、主成分分析、聚类和使用统一流形近似和投影（UMAP）的降维，以及聚类的差异表达分析。
Flow cytometric analysis 流式细胞仪分析
After isolation of whole blood leukocytes and NP tissue-derived cells, cells were first treated with Aqua LIVE/DEAD fixable dead cell staining reagent (Invitrogen). Cells were then incubated with Fc Block reagent (Miltenyi Biotec, San Diego, Calif), followed by antibodies. The complete list of antibodies is shown in Table E3 in the Online Repository available at www.jacionline.org, and detailed methods are shown in the Methods section of the Online Repository. 在分离全血白细胞和 NP 组织来源细胞后，首先用 Aqua LIVE/DEAD 可固定死细胞染色试剂（Invitrogen）处理细胞。然后用 Fc 阻断试剂（Miltenyi Biotec, San Diego, Calif）孵育细胞，随后添加抗体。抗体的完整列表见于表 E3，可在www.jacionline.org的在线库中找到，详细方法见于在线库的方法部分。
In vitro peripheral blood eosinophil culture 体外外周血嗜酸性粒细胞培养
Human PB eosinophils (1 million cells/mL) were incubated with 100 ng/mL C5a, 10 ng/mL granulocyte-macrophage colony-stimulating factor (GM-CSF), 10 ng/mL TNF, 10 ng/mL IL-1β, 10 ng/mL IL-4, 20 ng/mL IL-5, 10 ng/mL IL-13, 10 ng/mL IL-33, or 10 ng/mL IFN-γ for 24 hours (n = 4-8) and supernatants collected. The protein concentrations in cell-free supernatants were measured by Luminex and ELISA kits. Detailed methods are shown in the Methods section of the Online Repository. 人类 PB 嗜酸性粒细胞（1 百万细胞/mL）与 100 ng/mL C5a、10 ng/mL 粒细胞-巨噬细胞集落刺激因子（GM-CSF）、10 ng/mL TNF、10 ng/mL IL-1β、10 ng/mL IL-4、20 ng/mL IL-5、10 ng/mL IL-13、10 ng/mL IL-33 或 10 ng/mL IFN-γ孵育 24 小时（n = 4-8），并收集上清液。通过 Luminex 和ELISA试剂盒测定细胞无上清液中的蛋白质浓度。详细方法见在线库的“方法”部分。
Statistical analysis 统计分析
All statistical calculations were performed by SeqGeq 1.8.0 (FlowJo; BD), GraphPad Prism 10.0.3 (GraphPad Software, La Jolla, Calif), and Enrichr (maayanlab.cloud/Enrichr/). Differences between groups were analyzed by Bonferroni correction or 1-way ANOVA Kruskal-Wallis Dunn multiple comparison test. Differences between groups in flow cytometry and in vitro eosinophil culture studies were analyzed by paired t test or Wilcoxon matched-pairs signed rank test. Adjusted P value (Padj) or P value (for flow cytometry and eosinophil in vitro study) of less than .05 was considered significant. 所有统计计算均由 SeqGeq 1.8.0（FlowJo；BD）、GraphPad Prism 10.0.3（GraphPad Software, La Jolla, Calif）和 Enrichr（maayanlab.cloud/Enrichr/）进行。组间差异通过 Bonferroni 校正或单因素 ANOVA Kruskal-Wallis Dunn 多重比较检验进行分析。流式细胞术和 体外 嗜酸性粒细胞培养研究中的组间差异通过配对 t 检验或 Wilcoxon 匹配对符号秩检验进行分析。调整后的 P 值（Padj）或 P 值（针对流式细胞术和嗜酸性粒细胞 体外 研究）小于 .05 被视为显著。
Results 结果
Detection of eosinophils by scRNA-Seq 通过 scRNA-Seq 检测嗜酸性粒细胞
On clustering of the filtered data set that included 48,131 cells after all QC steps, 11 unique clusters were identified (Fig 1, A, and see Fig E2 in the Online Repository available at www.jacionline.org). We found that RNA markers of eosinophils25,26 including Charcot-Leyden crystal protein (CLC, also known as galectin-10), adhesion G protein-coupled receptor E1 (ADGRE1, also known as EMR1), serine protease 33 (PRSS33), sialic acid binding immunoglobulin-like lectin 8 (SIGLEC8), and arachidonate 15-lipoxygenase (ALOX15) were expressed in clusters 2 and 3, although the pattern of anti-CCR3 antibody staining was not clear for these clusters (Fig 1, B and C). Identification of other cell types by scRNA-Seq is discussed in the Results section of the Online Repository. We also found that sample tags from both tissue-derived cells and PB cells were detected in eosinophil clusters (Fig 1, D). Taken together, these results suggest that in contrast to the first-generation approaches that failed to detect eosinophils, our scRNA-Seq technology can detect single cell gene expression profiles from human eosinophils derived from NP tissues and PB. 在经过所有质量控制步骤后，对包含 48,131 个细胞的过滤数据集进行聚类，识别出 11 个独特的簇（图 1，A，并参见图 E2，可在www.jacionline.org在线库中获取）。 我们发现，嗜酸性粒细胞2526的 RNA 标记物，包括查尔科-莱登晶体蛋白（CLC, 也称为 galectin-10）、粘附 G 蛋白偶联受体 E1（ADGRE1, 也称为 EMR1）、丝氨酸蛋白酶 33（PRSS33）、唾液酸结合免疫球蛋白样凝集素 8（SIGLEC8）和花生四烯酸 15-脂氧合酶（ALOX15）在 2 和 3 簇中表达，尽管这些簇的抗 CCR3 抗体染色模式并不清晰（图 1, B 和 C）。通过 scRNA-Seq 识别其他细胞类型的讨论在在线库的结果部分。我们还发现，来自组织衍生细胞和PB 细胞的样本标签在嗜酸性粒细胞簇中被检测到（图 1, D）。 综合来看，这些结果表明，与未能检测嗜酸性粒细胞的第一代方法相比，我们的 scRNA-Seq 技术能够检测来自NP组织和外周血的人的嗜酸性粒细胞的单细胞基因表达谱。
Fig 1. Detection of human eosinophils and neutrophils by scRNA-Seq. We enriched granulocytes from NPs or control tissue and matched blood samples from 5 CRSwNP patients and 5 control patients and performed scRNA-Seq. After performing unsupervised clustering with dimension reduction analysis, we identified 11 unique clusters. (A) UMAP results. (B) Distribution of cell-specific protein (CCR3 and CD16) and RNA markers within UMAP plots by heat map. Assigned cell populations (C) and origin of cells (NP, control tissue [CT], and PB from CRS and control [Cont] patients) (D) are shown. 图 1。通过 scRNA-Seq 检测人类嗜酸细胞和中性粒细胞。我们从 NP 或对照组织中富集粒细胞，并匹配来自 5 名 CRSwNP 患者和 5 名对照患者的血液样本，进行了 scRNA-Seq。在进行无监督聚类和降维分析后，我们识别出 11 个独特的簇。(A) UMAP 结果。(B) UMAP 图中细胞特异性蛋白（CCR3 和 CD16）和 RNA 标记的分布热图。分配的细胞群体(C)和细胞来源（NP、对照组织[CT]以及来自 CRS 和对照[Cont]患者的 PB）(D)如图所示。
Because past studies failed to detect eosinophils by RNA sequencing, we hypothesized 2 reasons: eosinophils are damaged during single cell separation with droplet-based scRNA-Seq, and eosinophils have much less RNA than other cell types. To investigate our second hypothesis, we analyzed RNA counts in eosinophils and contaminating cell populations, specifically T cells and epithelial cells, within our scRNA-Seq result. Because epithelial cells included several subtypes, we performed clustering on the epithelial cell cluster (cluster 9) in Fig 1. We identified 4 main epithelial cell subsets and used 2 major types of epithelial cells (basal cells and apical cells [including 2 types of goblet cells]) for comparison with eosinophils (see Fig E3, A and B, in the Online Repository available at www.jacionline.org). We found that T cells (2,550 transcripts per cell), basal epithelial cells (8,106 transcripts per cell), and apical epithelial cells (16,488 transcripts per cell) had 3.8-, 12.0-, and 24.3-fold higher RNA transcripts per cell compared to eosinophils (678 transcripts per cell), respectively (Fig E3, C). We also analyzed our published results using droplet-based scRNA-Seq (10x Genomics, Pleasanton, Calif) on whole NP tissues15 and found that the mean level of transcripts per cell in T cells, basal epithelial cells, and apical epithelial cells were 1,318, 3,265 and 4,452, respectively (Fig E3, D), and showed an overall 2- to 4-fold lower transcript level per cell in our whole tissue scRNA-Seq (10x Genomics) study compared to our current study. This suggests that RNA transcripts in eosinophils from our published study were 170-339 transcripts per cell, and most of these eosinophils might be excluded during the QC steps even if droplet-based scRNA-Seq was able to detect eosinophils. 由于过去的研究未能通过 RNA 测序检测到嗜酸性粒细胞，我们假设有两个原因：嗜酸性粒细胞在基于液滴的单细胞分离过程中受到损伤，以及嗜酸性粒细胞的 RNA 量远低于其他细胞类型。为了调查我们的第二个假设，我们分析了在我们的 scRNA-Seq 结果中嗜酸性粒细胞和污染细胞群体（特别是 T 细胞和上皮细胞）的 RNA 计数。由于上皮细胞包括几种亚型，我们对上皮细胞簇（簇 9）进行了聚类分析，如图 1所示。我们识别了 4 个主要的上皮细胞亚群，并使用 2 种主要类型的上皮细胞（基底细胞和顶端细胞[包括 2 种杯状细胞]）与嗜酸性粒细胞进行比较（见图 E3，A和B，在www.jacionline.org的在线库中）。我们发现 T 细胞（每个细胞 2,550 个转录本）、基底上皮细胞（每个细胞 8,106 个转录本）和顶端上皮细胞（每个细胞 16,488 个转录本）分别具有 3.8 倍、12.0 倍和 24 倍的差异。每个细胞的 RNA 转录本数量是嗜酸性粒细胞的 3 倍（每个细胞 678 个转录本），分别为（图 E3, C）。我们还分析了使用基于液滴的单细胞 RNA 测序（10x Genomics，加州普莱森顿）对整个 NP 组织的已发表结果15，发现 T 细胞、基底上皮细胞和顶端上皮细胞的每个细胞转录本平均水平分别为 1,318、3,265 和 4,452（图 E3, D），并且在我们的整体组织单细胞 RNA 测序（10x Genomics）研究中，每个细胞的转录本水平整体低了 2 到 4 倍，与我们当前的研究相比。这表明我们已发表研究中嗜酸性粒细胞的 RNA 转录本数量为每个细胞 170-339 个，并且即使基于液滴的单细胞 RNA 测序能够检测到嗜酸性粒细胞，这些嗜酸性粒细胞在质量控制步骤中可能大部分被排除。
We did not include control sinus tissue eosinophils in further analyses because we only detected 101 eosinophils from control sinus tissues. In addition, because we detected minor subsets of neutrophils in the eosinophil clusters (Fig 1), we excluded CXCR2-positive cells (predominantly neutrophils) from the eosinophil cluster before analyzing eosinophils. We also separately excluded FCGR3B (another marker of neutrophils)-positive cells from the eosinophil cluster, which showed similar findings in all the following analyses (data not shown). We next performed a secondary cluster analysis in cells from the eosinophil cluster after excluding CXCR2-positive cells and found two major clusters that generally separated PB eosinophils and NP eosinophils (Fig 2, A). We also found a minor cluster (cluster 3) that contained contaminating neutrophils or granulocyte precursors (Fig 2, A, and data not shown). We excluded cells from this cluster and used control PB (5,542 cells), CRS PB (3,883 cells), and NP (9,727 cells) eosinophils for further analysis. 我们没有将对照鼻窦组织中的嗜酸性粒细胞纳入进一步分析，因为我们仅从对照鼻窦组织中检测到 101 个嗜酸性粒细胞。此外，由于我们在嗜酸性粒细胞簇中检测到少量的中性粒细胞（图 1），我们在分析嗜酸性粒细胞之前排除了CXCR2阳性细胞（主要是中性粒细胞）。我们还单独排除了FCGR3B（另一个中性粒细胞标记物）阳性细胞，这在所有后续分析中显示了类似的结果（数据未显示）。接下来，我们在排除CXCR2阳性细胞后，对嗜酸性粒细胞簇中的细胞进行了二次聚类分析，发现两个主要簇，通常将 PB 嗜酸性粒细胞和 NP 嗜酸性粒细胞分开（图 2, A）。 我们还发现了一个小的簇（簇 3），其中包含污染的中性粒细胞或粒细胞前体（图 2，A，以及未显示的数据）。我们排除了来自该簇的细胞，并使用对照 PB（5,542 个细胞）、CRS PB（3,883 个细胞）和 NP（9,727 个细胞）嗜酸性粒细胞进行进一步分析。
Fig 2. Detection of differential expression of genes in PB and NP eosinophils by scRNA-Seq. We performed secondary clustering analysis on eosinophil cluster in Fig 1 after excluding CXCR2-positive cells and found 3 clusters (A) that generally separated eosinophils from NP (cluster 1), that comprised PB eosinophils (cluster 2), and that contained contaminating cells (cluster 3). After excluding cells from cluster 3, we examined DEG in CRS PB (3,883 cells) compared to control PB (5,542 cells) (B) and in NP eosinophils (9,727 cells) compared to all PB eosinophils (9,425 cells) (C) by Benjamini-Hochberg algorithm. DEG are shown in violin plots; top 9 dysregulated genes exhibiting more than 2-fold change (Padj < .05) are listed in table. 图 2。通过 scRNA-Seq 检测 PB 和 NP 嗜酸细胞中基因的差异表达。我们在排除 CXCR2 阳性细胞后，对 图 1 中的嗜酸细胞簇进行了二次聚类分析，发现了 3 个簇 (A)，这些簇一般将嗜酸细胞与 NP（簇 1）分开，包含 PB 嗜酸细胞（簇 2），以及含有污染细胞（簇 3）。在排除簇 3 的细胞后，我们比较了 CRS PB（3,883 个细胞）与对照 PB（5,542 个细胞）中的 DEG (B)，以及 NP 嗜酸细胞（9,727 个细胞）与所有 PB 嗜酸细胞（9,425 个细胞）中的 DEG (C)，使用 Benjamini-Hochberg 算法。DEG 以小提琴图的形式展示；前 9 个表现出超过 2 倍变化的失调基因（Padj < .05）列在表中。
Identification of differentially expressed genes in NP eosinophils 在 NP 嗜酸性粒细胞中鉴定差异表达基因
Before analyzing our eosinophil results, we had to exclude the possibility that eosinophils may be activated during CD66b-positive selection because one study reported activation of eosinophils by anti-CD66b.27 To exclude this possibility, we isolated granulocytes from whole blood leukocytes using the CD66b-positive selection method that we used for enrichment of tissue granulocytes as well as the negative selection kits we used for whole blood granulocytes, and performed scRNA-Seq to compare gene expression profiles in blood eosinophils by the two isolation protocols. We found that there were no significant differences in gene expression profile between eosinophils from the CD66b-positive selection and the eosinophil-negative selection kit (see the Results and Fig E4 in the Online Repository available at www.jacionline.org). 在分析我们的嗜酸性粒细胞结果之前，我们必须排除在 CD66b 阳性选择过程中嗜酸性粒细胞可能被激活的可能性，因为有研究报告了抗 CD66b 对嗜酸性粒细胞的激活。27 为了排除这种可能性，我们使用 CD66b 阳性选择方法从全血白细胞中分离颗粒细胞，这种方法用于富集组织颗粒细胞，以及我们用于全血颗粒细胞的阴性选择试剂盒，并进行了 scRNA-Seq，以比较两种分离方案中血液嗜酸性粒细胞的基因表达谱。我们发现，来自 CD66b 阳性选择和嗜酸性粒细胞阴性选择试剂盒的嗜酸性粒细胞之间的基因表达谱没有显著差异（见结果和图 E4，可在www.jacionline.org的在线库中找到）。
We next compared control PB eosinophils with CRS PB eosinophils. We identified only 9 downregulated and 43 upregulated genes in CRS PB eosinophils (>1.5-fold, Padj < .05), and most of these genes showed a less than 2-fold difference in expression between the two groups (Fig 2, B, and see Table E4 in the Online Repository available at www.jacionline.org). This supports our secondary cluster analysis indicating that eosinophils in both control PB and CRS PB fell into the same cluster and indicates that the gene expression profiles of eosinophils from control PB and CRS PB have only minor differences. We next identified NP eosinophil–specific differentially expressed genes (DEG) by comparing NP eosinophils (9,727 cells) with all PB eosinophils (9,425 cells). We found that 204 genes were downregulated and 354 genes upregulated in NP eosinophils (>1.5-fold, Padj < .05) (Fig 2, C, and see Table E5 in the Online Repository) compared to all PB eosinophils. 我们接下来比较了对照 PB 嗜酸细胞与 CRS PB 嗜酸细胞。我们在 CRS PB 嗜酸细胞中仅识别出 9 个下调基因和 43 个上调基因（>1.5 倍，Padj < .05），这些基因在两组之间的表达差异大多数小于 2 倍（图 2, B, 并参见表 E4，可在www.jacionline.org的在线库中获取）。这支持我们的二次聚类分析，表明对照 PB 和 CRS PB 中的嗜酸细胞属于同一聚类，并表明对照 PB 和 CRS PB 的嗜酸细胞基因表达谱仅有微小差异。接下来，我们通过将 NP 嗜酸细胞（9,727 个细胞）与所有 PB 嗜酸细胞（9,425 个细胞）进行比较，识别了 NP 嗜酸细胞特异性差异表达基因（DEG）。我们发现 NP 嗜酸细胞中有 204 个基因下调，354 个基因上调（>1.5 倍，Padj < .05) (图 2, C, 并参见 表 E5 在在线库) 与所有 PB 嗜酸细胞相比。
Patterns of downregulated genes from NP eosinophils NP 嗜酸性粒细胞中下调基因的模式
We performed gene enrichment analyses of the 204 downregulated genes and found that they were associated with several patterns: the immune system (Reactome [reactome.org/] [R-HSA-168256]), regulation of the actin cytoskeleton (Kyoto Encyclopedia of Genes and Genomes [KEGG; genome.jp/kegg], Gene Ontology [GO; geneontology.org] molecular function [GO:0003779] and GO cellular component [GO:0005856]), and leukocyte transendothelial migration (KEGG and BioPlanet [tripod.nih.gov/bioplanet/]) (Fig 3, A, and see Table E6 in the Online Repository available at www.jacionline.org). Downregulated genes associated with immunity and immune receptors included thioredoxin interacting protein (TXNIP), neutrophil cytosolic factor 2 (NCF2), IL5RA and IL17RA, as well as enzymes such as catalase (CAT), nudix hydrolase 3 (NUDT3), ribonuclease T2 (RNASET2), and dipeptidase 2 (DPEP2) (Fig 3, B,Table I, and Table E6). Identified genes in KEGG regulation of the actin cytoskeleton included integrinsubunit alpha 4 (ITGA4) and CXCR4 (Fig 3, B, and Table E6). Adhesion and migration associated molecules including L-selectin (SELL, also known as CD62L), ITGA4, ITGAL (also known as LFA-1α), and CXCR4, and actin binding proteins including spectrin repeat-containing nuclear envelopeprotein 1 (SYNE1) were also associated with downregulated genes in NP eosinophils (Fig 3, B,Table I, and Table E6). Using flow cytometry, we independently confirmed a few selected genes that SELL (L-selectin), IL5RA (IL-5Rα), and IL17RA (IL-17Rα) were downregulated in NP eosinophils compared to patient-matched PB eosinophils (Fig 3, C). These results suggest that eosinophils in NP tissue have downregulated function of innate immunity and transendothelial migration. 我们对 204 个下调基因进行了基因富集分析，发现它们与几种模式相关：免疫系统（Reactome [reactome.org/] [R-HSA-168256]），肌动蛋白细胞骨架（京都基因和基因组百科全书 [KEGG; genome.jp/kegg]，基因本体 [GO; geneontology.org] 分子功能 [GO:0003779] 和 GO 细胞成分 [GO:0005856]），以及白细胞跨内皮迁移（KEGG 和 BioPlanet [tripod.nih.gov/bioplanet/]）（图 3，A, 并参见表 E6，可在www.jacionline.org）. 与免疫相关的下调基因和免疫受体包括硫氧还原酶相互作用蛋白（TXNIP）、中性粒细胞细胞质因子 2（NCF2）、IL5RA和IL17RA，以及酶类，如过氧化氢酶（CAT）、nudix 水解酶 3（NUDT3）、核糖核酸酶 T2（RNASET2）和二肽酶 2（DPEP2）（图 3，B，表 I，以及表 E6）。 在 KEGG 调控肌动蛋白细胞骨架中识别的基因包括整合素亚基α 4 (ITGA4) 和 CXCR4 (图 3, B, 和 表 E6)。与粘附和迁移相关的分子包括 L-选择素 (SELL, 也称为 CD62L)、ITGA4, ITGAL (也称为 LFA-1α) 和 CXCR4, 以及与下调基因相关的肌动蛋白结合蛋白包括 光谱蛋白 重复含有的核膜蛋白 1 (SYNE1) 也与 NP 嗜酸细胞中的下调基因相关 (图 3, B,表 I, 和 表 E6)。 使用流式细胞术，我们独立确认了一些选定的基因，SELL（L-选择素）、IL5RA（IL-5Rα）和IL17RA（IL-17Rα）在 NP 嗜酸性粒细胞中相较于患者匹配的 PB 嗜酸性粒细胞下调（图 3, C）。这些结果表明，NP 组织中的嗜酸性粒细胞具有下调的先天免疫功能和跨内皮迁移能力。
Fig 3. Gene enrichment analyses of downregulated genes in NP eosinophils. Pathway (Reactome and KEGG) and GO enrichment analyses were performed on >1.5-fold (Padj < .05) downregulated genes in NP eosinophils (204 genes). (A) Top 3 enriched pathways and GO. (B) Expression levels of genes in top associated pathways in each analysis. Expression levels of genes in associated pathways are shown by violin plots with dot plots. ∗∗∗∗Padj < .0001. (C) Representative histograms of flow cytometric plots for L-selectin (SELL), IL-5Rα, and IL-17Rα on eosinophils from blood sample and NP. Levels of cell surface expression of receptors on eosinophils in patient-matched blood and NP samples from CRSwNP patients (n = 5) are shown by mean fluorescence intensity (MFI). ∗∗P < .01, ∗∗∗∗P < .0001. 图 3。NP 嗜酸细胞中下调基因的基因富集分析。对 NP 嗜酸细胞中 >1.5 倍 (Padj < .05) 下调基因（204 个基因）进行了通路（Reactome 和 KEGG）和 GO 富集分析。(A) 前 3 个富集通路和 GO。(B) 每个分析中与顶级相关通路的基因表达水平。与相关通路的基因表达水平通过小提琴图和点图显示。∗∗∗∗Padj < .0001。(C) 来自血液样本和 NP 的嗜酸细胞 L-selectin (SELL)、IL-5Rα 和 IL-17Rα 的流式细胞术图的代表性直方图。来自 CRSwNP 患者（n = 5）的匹配血液和 NP 样本中嗜酸细胞上受体的细胞表面表达水平通过平均荧光强度 (MFI) 显示。∗∗P < .01，∗∗∗∗P < .0001。
Table I. Downregulated genes in NP eosinophils and their associated functions 表 I。NP 嗜酸细胞中下调的基因及其相关功能
Patterns of upregulated expressed genes in NP eosinophils NP 嗜酸性粒细胞中上调表达基因的模式
We next focused on upregulated genes in NP eosinophils. We found that upregulated genes included known eosinophil activation markers such as CD44, CD69, and CXCL8 and orphan nuclear receptorsNR4A1, NR4A2, and NR4A3 (Table E5).28, 29, 30, 31 Gene enrichment analyses of the 354 upregulated genes in NP eosinophils showed associations with the immune system (Reactome [R-HSA-168256]), IL-5 regulation of apoptosis (BioPlanet), negative regulation of apoptotic process (GO biological process [GO:0043066]), nuclear factor kappa–light-chain enhancer of activated B cells (NF-κB) signaling pathway (KEGG), and cytokine-mediated signaling pathway (Reactome [R-HSA-1280215] and GO biological process [GO:0019221]) (Fig 4, A, and see Table E7 in the Online Repository available at www.jacionline.org). Because there were several antiapoptotic genes, including BCL2A1, baculoviral IAP repeat–containing 3 (BIRC3), TNFAIP3, and protein phosphatase 1 regulatory subunit 15A (PPP1R15A) that were elevated in NP eosinophils (Fig 4, B,Table II, and Table E7), NP eosinophils may be more resistant to, or protective from, apoptosis. 我们接下来关注了 NP 嗜酸性粒细胞中上调的基因。我们发现，上调的基因包括已知的嗜酸性粒细胞激活标志物，如CD44，CD69，和CXCL8以及孤儿核受体NR4A1, NR4A2,和NR4A3（表 E5）。28, 29, 30, 31 对 354 个在 NP 嗜酸性粒细胞中上调的基因进行的基因富集分析显示与免疫系统（Reactome [R-HSA-168256]）、IL-5 对凋亡的调控（BioPlanet）、凋亡过程的负调控（GO 生物过程 [GO:0043066]）、活化 B 细胞的核因子κ轻链增强子（NF-κB）信号通路（KEGG）以及细胞因子介导的信号通路（Reactome [R-HSA-1280215] 和 GO 生物过程 [GO:0019221]）(图 4, A, 并参见表 E7，可在www.jacionline.org)。 因为有几个抗凋亡基因，包括BCL2A1, 含有蛹病毒 IAP 重复序列的 3 型(BIRC3), TNFAIP3, 和 蛋白磷酸酶 1调节亚单位 15A (PPP1R15A)在 NP 嗜酸细胞中升高(图 4, B,表 II, 和 表 E7)，NP 嗜酸细胞可能对凋亡更具抵抗力或具有保护作用。
Fig 4. Gene enrichment analyses of upregulated genes in NP eosinophils. Pathway (Reactome, BioPlanet, and KEGG) and GO enrichment analyses were performed on >1.5 fold (Padj < .05) upregulated genes in NP eosinophils (354 genes). (A) Top 3 enriched pathways and GO. (B) Violin plots showing expression levels of genes in top associated pathways in each analysis (solid line indicates median and dotted line quartiles). Violin plots showing expression levels of (C) growth factor genes and (D) other uniquely elevated genes in NP eosinophils. Genes that were detected in fewer cells are also included in dot plots (C). ∗∗∗∗Padj < .0001. (E) Representative histograms of flow cytometric plots for ICAM-1, CD44, and PLAUR in eosinophils from blood sample and NP. Levels of cell surface expression on eosinophils in blood and NP samples from patient-matched CRSwNP patients (n = 5) are shown by mean fluorescence intensity (MFI). ∗P < .05. 图 4。NP 嗜酸细胞中上调基因的基因富集分析。对 NP 嗜酸细胞中 >1.5 倍 (Padj < .05) 上调的基因（354 个基因）进行了通路（Reactome、BioPlanet 和 KEGG）和 GO 富集分析。(A) 前 3 个富集通路和 GO。(B) 小提琴图显示每个分析中与顶级相关通路的基因表达水平（实线 表示中位数，虚线 表示四分位数）。小提琴图显示 (C) 生长因子基因和 (D) NP 嗜酸细胞中其他独特上调基因的表达水平。检测到的在较少细胞中的基因也包含在点图中 (C). ∗∗∗∗Padj < .0001。(E) 流式细胞术图中 ICAM-1、CD44 和 PLAUR 在血液样本和 NP 中的代表性直方图。来自患者匹配的 CRSwNP 患者（n = 5）的血液和 NP 样本中嗜酸细胞的细胞表面表达水平以平均荧光强度（MFI）表示。 ∗P < .05.
Table II. Upregulated genes in NP eosinophils and their associated functions 表 II。NP 嗜酸细胞中上调的基因及其相关功能
We further confirmed several selected upregulated genes in NP eosinophils at the protein level by flow cytometry and found that cell surface expression of ICAM-1, CD44, and PLAUR were significantly elevated in eosinophils in NPs compared to patient-matched PB samples (Fig 4, E, and data not shown). 我们进一步通过流式细胞术确认了 NP 嗜酸细胞中几种选择性上调基因的蛋白水平，并发现与患者匹配的外周血样本相比，NP 中的嗜酸细胞表面 ICAM-1、CD44 和 PLAUR 的表达显著升高（图 4，E，以及未显示的数据）。
Identification of eosinophil heterogeneity in NPs 在鼻息肉中嗜酸性粒细胞的异质性识别
While heterogeneity and functional subtypes of eosinophils have been discussed previously,32, 33, 34 it is unclear whether the eosinophils present in NP tissue are heterogeneous. We next performed clustering with the UMAP analyses in NP eosinophils and identified 4 major clusters (Fig 5, A). All 4 clusters showed high levels of expression of the eosinophil marker CLC gene (Fig 5, B). Cluster 1 presented high expression of CC chemokines including CCL3, CCL3L1, CCL4, and CCL4L2 (Fig 5, C, and see Table E8 in the Online Repository available at www.jacionline.org). In contrast, upregulated genes in clusters 3 and 4 included CXCL8 and INHBA (inhibin subunit beta A), respectively (Fig 5, C, and Table E8). Although there was no single clearly elevated gene in cluster 2, some genes, including ARL4C (ADP ribosylation factor–like GTPase 4C), were elevated in a subset of these cells (Fig 5, C, and Table E8). In contrast, upregulated genes found in other clusters were significantly downregulated in cluster 2 (Fig 5, C, and Table E8). We next found that there were differences in the proportion of these eosinophil subsets in 5 NP tissues; for example, cells in cluster 2 were the most common type of eosinophils in NPs except for one NP donor, where the cells in cluster 1 were dominant (donor NP-2), and the frequency of cells in clusters 1 and 3 was similar in donors NP-1 and NP-5 (Fig 5, D). 尽管之前讨论了嗜酸性粒细胞的异质性和功能亚型，32，33，34，但尚不清楚存在于鼻息肉组织中的嗜酸性粒细胞是否是异质的。接下来，我们对鼻息肉嗜酸性粒细胞进行了 UMAP 分析聚类，并识别出 4 个主要簇（图 5，A）。所有 4 个簇均显示出嗜酸性粒细胞标记基因CLC的高表达水平（图 5，B）。簇 1 表现出高水平的< a id=19>CC 趋化因子的表达，包括CCL3、CCL3L1、CCL4、和CCL4L2（图 5，C，并参见表 E8，可在www.jacionline.org的在线库中找到）。 相反，簇 3 和簇 4 中上调的基因包括CXCL8和INHBA（抑制素β亚基A），分别（图 5，C，和表 E8）。尽管在簇 2 中没有单个明显上调的基因，但一些基因，包括ARL4C（ADP 核糖基化因子样GTP 酶4C），在这些细胞的一个子集中上调（图 5，C，和表 E8）。相反，在其他簇中发现的上调基因在簇 2 中显著下调（图 5，C，和表 E8）。 我们接下来发现，在 5 个鼻息肉（NP）组织中，这些嗜酸性粒细胞亚群的比例存在差异；例如，除了一个鼻息肉供体（供体 NP-2）外，簇 2 中的细胞是鼻息肉中最常见的嗜酸性粒细胞类型，而簇 1 中的细胞在该供体中占主导地位，供体 NP-1 和 NP-5 中簇 1 和簇 3 中的细胞频率相似（图 5, D）。
Fig 5. Eosinophils display heterogeneity in NPs. (A) Clustering with UMAP analysis of 9,727 eosinophils in NPs showed 4 clusters. (B) Expression pattern of eosinophil marker CLC. (C) Expressions of CCL4, ARL4C, CXCL8, and INHBA were specifically elevated in clusters 1, 2, 3, and 4, respectively. Expression levels are shown by heat map in UMAP plots (blue indicates negative expression and red high expression; C) and by violin plots (solid line indicates median expression and dotted line quartiles; B, C, and F). (D) Proportion of eosinophil subsets in each NP donor. (E) PB eosinophils (n = 4-8) were stimulated with C5a, GM-CSF, TNF, IL-1β, IL-4, IL-5, IL-13, IL-33, and IFN-γ for 24 hours, and protein concentrations in supernatants were measured by Luminex assay. ∗P < .05 compared to medium control. 图 5。嗜酸性粒细胞在 NPs 中表现出异质性。(A) 对 9,727 个嗜酸性粒细胞进行 UMAP 分析的聚类显示出 4 个簇。(B) 嗜酸性粒细胞标记物 CLC 的表达模式。(C) CCL4、ARL4C、CXCL8 和 INHBA 的表达在簇 1、2、3 和 4 中分别显著升高。表达水平在 UMAP 图中的热图中显示（蓝色表示负表达，红色表示高表达；C）以及小提琴图中显示（实线表示中位数表达，虚线表示四分位数；B、C和F）。(D) 每个 NP 供体中嗜酸性粒细胞亚群的比例。(E) PB 嗜酸性粒细胞（n = 4-8）在 C5a、GM-CSF、TNF、IL-1β、IL-4、IL-5、IL-13、IL-33 和 IFN-γ的刺激下培养 24 小时，超 natant 中的蛋白质浓度通过 Luminex 检测测量。∗P < .05 与培养基对照相比。
Because clusters 1, 3, and 4 showed specific induction of chemokines and cytokines, we examined whether eosinophils could be activated in vitro by different stimuli to produce a similar pattern of production of CCL4 (cluster 1), CXCL8 (cluster 3), and activin A (cluster 4). We purified PB eosinophils and incubated cells with various stimuli for 24 hours. We found that levels of CCL4 and CXCL8 protein in culture supernatants were increased by C5a, GM-CSF, TNF, IL-1β, IL-5, and IL-33 (Fig 5, E). Interestingly, IL-1β and IL-33 showed different capacities to induce CCL4 and CXCL8 (eg, IL-33 induced more CCL4). We also examined the production of activin A (a homodimer of INHBA) in PB eosinophils but could not detect it in culture supernatants (data not shown). Because IL-1β and IL-33 showed different capacities to induce the gene markers of clusters 1 and 3, respectively, we also examined mRNA expression of receptors for IL-1β and IL-33 in the scRNA-Seq results. We found that the expression of IL1RL1 (a receptor of IL-33) was significantly elevated in cluster 1, and the corresponding clusters showed a similar pattern of CCL4 expression (Fig 5, C and F). In contrast, expression of IL-1β receptors IL1R1, IL1R2, and IL1RAP was at nearly undetectable levels by scRNA-Seq (data not shown). This result suggests that eosinophils may be activated by different stimuli, and this may induce the heterogeneity found in NP eosinophils. 由于簇 1、3 和 4 显示出特定的趋化因子和细胞因子的诱导，我们检查了嗜酸性粒细胞是否可以被激活体外通过不同的刺激产生类似的 CCL4（簇 1）、CXCL8（簇 3）和激活素 A （簇 4）产生模式。我们纯化了外周血嗜酸性粒细胞，并用各种刺激物孵育细胞 24 小时。我们发现培养上清液中 CCL4 和 CXCL8 蛋白的水平被C5a、GM-CSF、TNF、IL-1β、IL-5 和 IL-33（图 5，E）显著增加。有趣的是，IL-1β和 IL-33 在诱导 CCL4 和 CXCL8 方面表现出不同的能力（例如，IL-33 诱导更多的 CCL4）。我们还检查了外周血嗜酸性粒细胞中激活素 A（INHBA 的同源二聚体）的产生，但在培养上清液中未能检测到（数据未显示）。由于 IL-1β和 IL-33 在诱导簇 1 和簇 3 的基因标记方面表现出不同的能力，我们还检查了 scRNA-Seq 结果中 IL-1β和 IL-33 受体的 mRNA 表达。 我们发现，IL1RL1（IL-33 的受体）的表达在簇 1 中显著升高，相应的簇显示出 CCL4 表达的类似模式（图 5，C和F）。相比之下，IL-1β受体IL1R1, IL1R2,和IL1RAP的表达在 scRNA-Seq 中几乎无法检测到（数据未显示）。这一结果表明，嗜酸性粒细胞可能会受到不同刺激的激活，这可能导致 NP 嗜酸性粒细胞的异质性。
Discussion 讨论
Although tissue eosinophilia is a key feature of severe CRSwNP patients, the contributions of eosinophils to the pathogenesis of nasal polyposis have not been well studied beyond reports of drugs that deplete them, where conclusions differ. Attempts to detect eosinophils in NPs by scRNA-Seq have not been successful to date.14, 15, 16, 17, 18 We hypothesized this is because of 2 key reasons. First, eosinophils are fragile cells that are damaged during single cell separation with droplet-based scRNA-Seq (included in the 10x Genomics platform) or cell isolation by cell sorter. However, Ordovas-Montanes et al14 used a microwell-based technology (the Seq-Well platform) and still did not detect eosinophils derived from NP tissue. This suggests that the platform for scRNA-Seq alone cannot explain its failure to detect eosinophils. 尽管组织嗜酸性粒细胞增多是重度 CRSwNP 患者的一个关键特征，但嗜酸性粒细胞在鼻息肉发病机制中的贡献尚未得到充分研究，除了关于消耗它们的药物的报告，其中结论各不相同。迄今为止，尝试通过 scRNA-Seq 在 NP 中检测嗜酸性粒细胞并未成功。14，15，16，17，18我们假设这是由于两个关键原因。首先，嗜酸性粒细胞是脆弱的细胞，在使用基于液滴的 scRNA-Seq（包含在 10x Genomics 平台中）进行单细胞分离或通过细胞分选器进行细胞分离时会受到损伤。然而，Ordovas-Montanes 等人14使用了一种微孔技术（Seq-Well 平台），仍然未能检测到来源于 NP 组织的嗜酸性粒细胞。这表明，仅靠 scRNA-Seq 的平台无法解释其未能检测到嗜酸性粒细胞的原因。
Our second hypothesis for the failure to detect eosinophils by traditional scRNA-Seq was that eosinophils have much less RNA than other cell types, and thus the sequence depth in published studies may not be enough to detect these granulocytes in tissue-derived samples. In the current study, we eliminated both possibilities for the previous failure and utilized a combination of manual granulocyte preenrichment from tissues by a gentle bead-free CD66b-positive selection technology and microwell-based scRNA-Seq technology. We report here the successful use of this new approach to allow for the detection of the single cell transcriptome in human eosinophils from diseased tissues as well as from peripheral blood. Although we think the fragile nature of eosinophils and insufficient sequence reads were both problems in the past studies that failed to detect eosinophils by scRNA-Seq, future study is required to examine whether higher sequence depth aids the detection of eosinophils in droplet-based scRNA-Seq or if a combination of microwell-based scRNA-Seq and high sequence depth is necessary to detect eosinophils. 我们第二个假设认为，传统的 scRNA-Seq 未能检测到嗜酸性粒细胞的原因是嗜酸性粒细胞的 RNA 量远低于其他细胞类型，因此已发布研究中的测序深度可能不足以在组织来源的样本中检测到这些颗粒细胞。在当前研究中，我们排除了之前失败的两种可能性，并利用了一种结合了温和的无珠 CD66b 阳性选择技术和微孔基 scRNA-Seq 技术的手动颗粒细胞预富集方法。我们在此报告成功使用这种新方法，允许在来自病变组织和外周血的人类嗜酸性粒细胞中检测单细胞转录组。尽管我们认为嗜酸性粒细胞的脆弱特性和测序读取不足都是过去研究未能通过 scRNA-Seq 检测到嗜酸性粒细胞的问题，但未来的研究仍需检验更高的测序深度是否有助于在基于液滴的 scRNA-Seq 中检测嗜酸性粒细胞，或者是否需要结合微孔基 scRNA-Seq 和高测序深度来检测嗜酸性粒细胞。
Interestingly, we found a downregulation of IL-5Rα and IL-17Rα at both the mRNA and protein levels in NP eosinophils. It is known that activation of eosinophils by IL-3, IL-5 and GM-CSF downregulates expression of IL-5Rα on eosinophils.35,36 Downregulation of cell surface IL-5Rα on eosinophils and increased soluble IL-5Rα have also been reported in bronchoalveolar lavage fluid after segmental allergen challenge of allergic subjects.37 Importantly, that study also showed that bronchoalveolar lavage eosinophils were refractory to IL-5 treatment in ex vivodegranulation assays.37 This suggests that downregulation of IL-5Rα and perhaps IL-17Rα on eosinophils may be a result of a negative feedback pathway. It will be worthwhile to examine whether the downregulation of cell surface expression of IL-5Rα on tissue eosinophils is associated with decreased responsiveness to IL-5 and IL-5Rα targeting biologics in patients. Liu et al36 reported that stimulation with IL-5 and GM-CSF did not affect the expression of GM-CSF receptor on eosinophils and that IL-5–pretreated eosinophils were further activated by GM-CSF but not by IL-5. A future study to examine the roles of GM-CSF and other ligands for which the receptors are not downregulated on NP eosinophils will be important to further clarify the mechanisms of eosinophil activation in NPs. 有趣的是，我们发现 NP 嗜酸细胞中 IL-5Rα和 IL-17Rα在 mRNA 和蛋白质水平上均出现了下调。已知 IL-3、IL-5 和 GM-CSF 激活嗜酸细胞会下调嗜酸细胞上 IL-5Rα的表达。3536 在过敏患者的支气管肺泡灌洗液中，嗜酸细胞表面 IL-5Rα的下调和可溶性 IL-5Rα的增加也已被报道，尤其是在分段过敏原挑战后。37 重要的是，该研究还表明，支气管肺泡灌洗液中的嗜酸细胞对 IL-5 治疗在体外脱颗粒实验中表现出耐药性。37 这表明，嗜酸细胞上 IL-5Rα和可能的 IL-17Rα的下调可能是负反馈通路的结果。 值得研究的是，组织嗜酸性粒细胞表面 IL-5Rα的下调是否与患者对 IL-5 和 IL-5Rα靶向生物制剂的反应性降低相关。Liu 等人36报告称，IL-5 和 GM-CSF 的刺激对嗜酸性粒细胞上 GM-CSF 受体的表达没有影响，并且 IL-5 预处理的嗜酸性粒细胞在 GM-CSF 的作用下进一步被激活，但不受 IL-5 的影响。未来的研究将重要，以检查 GM-CSF 和其他未下调受体的配体在 NP 嗜酸性粒细胞中的作用，从而进一步阐明嗜酸性粒细胞在 NP 中的激活机制。
In GO and pathway analyses of DEG, we found that the NF-κB signaling pathway was strongly associated with upregulated genes in NP eosinophils. Indeed, 15 of 104 genes in the KEGG NF-κB pathway were elevated in NP eosinophils, and the data indicated that both canonical and noncanonical NF-κB pathways were activated in NP eosinophils (Fig E5). Furthermore, AP-1 family transcription factors38 such as FOS, FOSB, FOSL2, JUN, and JUND were also elevated in NP eosinophils (Table E5). It is known that these transcription factors are activated by different simulations. For example, IL-1 family cytokines, including IL-1 and IL-33, activate canonical NF-κB; TNF superfamily members, including CD30L, activate noncanonical NF-κB; and IL-5 and IL-2 family cytokines activate AP-1.38, 39, 40, 41 This suggests that eosinophils may be activated by multiple stimuli in NPs, and future studies will be needed to identify the panoply of factors that activate eosinophils in NPs. 在GO和 DEG 的通路分析中，我们发现 NF-κB 信号通路与 NP 嗜酸细胞中上调的基因密切相关。实际上，KEGG NF-κB 通路中的 104 个基因中有 15 个在 NP 嗜酸细胞中升高，数据表明 NP 嗜酸细胞中经典和非经典 NF-κB 通路均被激活（图 E5）。此外，AP-1 家族转录因子38如FOS, FOSB, FOSL2, JUN,和JUND在 NP 嗜酸细胞中也升高(表 E5)。已知这些转录因子是由不同的刺激激活的。例如，IL-1 家族细胞因子，包括 IL-1 和 IL-33，激活经典 NF-κB；TNF 超家族成员，包括CD30L，激活非经典 NF-κB；而 IL-5 和 IL-2 家族细胞因子则激活 AP-1。38, 39, 40, 41 这表明嗜酸性粒细胞可能会在鼻息肉中受到多种刺激的激活，未来的研究将需要识别激活鼻息肉中嗜酸性粒细胞的各种因素。
We identified several novel NP-associated eosinophil genes including OLR1, PLAUR, and ICAM1 as well as growth factors (AREG, EREG, and VEGFA). Oxidized low-density lipoprotein receptor 1 (OLR1), also known as LOX-1 (lectin-like oxidized low density lipoprotein receptor 1), is a membrane scavenger receptor involved in the internalization of oxidized low-density lipoprotein and activation of cells via NF-κB, mitogen-activated protein kinase, and caspases.42 Although it is a novel finding and potentially important in CRS, we could not confirm upregulation of OLR1 protein on NP eosinophils by flow cytometry compared to PB eosinophils (n = 5, data not shown). In contrast, we were able to confirm upregulation of cell surface expression of ICAM-1 and PLAUR on NP eosinophils. Interestingly, upregulated cell surface expression of ICAM-1 was also reported on sputum eosinophils from asthmatic subjects.43 Although ICAM-1 is well studied in the role of adhesion and transendothelial migration of leukocytes, including eosinophils on endothelial cells and epithelial cells, ICAM-1 can also contribute to immune cell effector function.44 ICAM-1 plays a role as a costimulatory signal in T cells, affects the generation of reactive oxygen species in neutrophils, and has a phagocytic function in macrophages.44 Earlier microarray data suggested that the common eosinophil activators IL-3, IL-5, and GM-CSF induced expression of ICAM-1 via p38 mitogen–activated protein kinase and NF-κB, suggesting that upregulation of ICAM-1 on NP eosinophils was an indicator of activation.45 PLAUR is a glycosyl-phosphatidyl-inositol–anchored membrane protein and cleaves urokinase plasminogen activator to become an active form that can convert plasminogen into plasmin.46 The binding of urokinase plasminogen activator to PLAUR also activates cells to release several cytokines, chemokines, and growth factors.47 PLAUR can interact with various transmembrane receptors, including integrins, epidermal growth factor receptor, platelet-derived growth factor receptors, VEGF receptor 2, and insulin-like growth factor 1 receptor, as well as regulate signal transduction. However, the functional roles of ICAM-1 and PLAUR on NP eosinophils are still not well known and require further investigation. 我们识别了几个新型的 NP 相关嗜酸性粒细胞基因，包括 OLR1, PLAUR, 和 ICAM1 以及生长因子 (AREG, EREG, 和 VEGFA)。氧化低密度脂蛋白受体 1 (OLR1)，也称为 LOX-1（类凝集素 氧化低密度脂蛋白 受体 1），是一个膜 清道夫受体，参与 内吞氧化低密度脂蛋白和通过 NF-κB、丝裂原激活蛋白激酶以及 半胱天冬酶 激活细胞。42 尽管这是一个新发现，并且在慢性鼻窦炎中可能很重要，但我们无法通过流式细胞术确认 NP 嗜酸细胞上 OLR1 蛋白的上调，与 PB 嗜酸细胞相比（n = 5，数据未显示）。相反，我们能够确认 NP 嗜酸细胞上 ICAM-1 和 PLAUR 的细胞表面表达上调。有趣的是，ICAM-1 的细胞表面表达上调也在哮喘患者的痰嗜酸细胞中被报道。43 尽管 ICAM-1 在白细胞（包括嗜酸细胞）在内皮细胞和上皮细胞上的粘附和跨内皮迁移中的作用得到了很好的研究，但 ICAM-1 也可以促进免疫细胞的效应功能。44 ICAM-1 在T 细胞中作为共刺激信号发挥作用，影响中性粒细胞中活性氧种的生成，并在巨噬细胞中具有吞噬功能。44 早期的微阵列数据表明，常见的嗜酸性粒细胞激活因子 IL-3、IL-5 和 GM-CSF 通过 p38 蛋白激酶和 NF-κB 诱导 ICAM-1 的表达，这表明 NP 嗜酸性粒细胞上 ICAM-1 的上调是激活的一个指标。45 PLAUR 是一种糖基磷脂酰肌醇锚定的 膜蛋白，并切割 尿激酶原激活剂 以成为一种可以将 纤溶酶原 转化为 纤溶酶。46 尿激酶原激活剂与 PLAUR 的结合还激活细胞释放多种细胞因子、趋化因子 和生长因子。47 PLAUR 可以与多种跨膜受体相互作用，包括 整合素、表皮生长因子受体、血小板源生长因子受体、VEGF 受体 2 和类胰岛素生长因子 1 受体，并调节信号转导。然而，ICAM-1 和 PLAUR 在 NP 嗜酸性粒细胞中的功能角色仍不清楚，需要进一步研究。
Although the growth factors AREG, EREG, and VEGFA have been reported to be increased in NP tissues,48, 49, 50 their expression in NP eosinophils has not been evaluated. A coculture of EoL-1 cells (human eosinophilic leukemia cell line) and NCI-H292 cells (human mucoepidermoid carcinoma cell line) induced AREG and VEGFA.51 Human PB eosinophils produced AREG after activation with GM-CSF and IL-5.52 In contrast, to our knowledge, there is no report showing that eosinophils release EREG. AREG, EREG, and VEGFA are known to induce proliferation of epithelial cells and facilitate epithelial-to-mesenchymal transition.53 This suggests that elevation of eosinophil-derived growth factors may regulate epithelial barrier function, over epithelial proliferation and other epithelial responses in NPs. 尽管在 NP 组织中已报告生长因子 AREG、EREG 和 VEGFA 的增加，48，49，50，但尚未评估它们在 NP 嗜酸性粒细胞中的表达。A 共培养 EoL-1 细胞（人类嗜酸性白血病细胞系）和 NCI-H292 细胞（人类粘液表皮癌细胞系）诱导了 AREG 和 VEGFA。51 人类 PB 嗜酸性粒细胞在 GM-CSF 和 IL-5 激活后产生 AREG。52 相比之下，据我们所知，尚无报告显示嗜酸性粒细胞释放 EREG。已知 AREG、EREG 和 VEGFA 能诱导上皮细胞增殖并促进上皮-间充质转化。53 这表明，嗜酸性粒细胞衍生生长因子的升高可能会调节上皮屏障功能，而不是上皮增殖和其他上皮反应在鼻息肉中。
We also identified potential heterogeneity of eosinophils in NP tissue. However, most of the core genes were shared among all clusters, and there were no clear cell surface molecules that distinguished between clusters at the mRNA level (data not shown). Although this could be a key limitation of scRNA-Seq because of low sequence reads for each type of eosinophils, we also hypothesized that these eosinophil clusters could be mainly controlled by distinct stimuli within NP tissues (eg, by the specific localizations of eosinophils or interaction with other cells in NPs). To test this hypothesis, we stimulated PB eosinophils with various eosinophil activators and found that IL-1β and IL-33 differentiated expression of CXCL8 and CCL4 in eosinophils, respectively (Fig 5, E). Because IL-33 is mainly released from epithelial cells, whereas IL-1β is produced mainly from macrophages and neutrophils, the local distribution of eosinophils in NPs may control the release of specific cytokines and cytokines in NPs. Furthermore, CCL4-positive eosinophils had a high expression of IL-33 receptor (Fig 5, F), indicating a potential role for epithelial IL-33 in NP eosinophils from cluster 1. Future studies using spatial transcriptomics may address this important question. We also performed an ELISA for activin A in supernatants from stimulated PB eosinophils; however, the levels were below the detection limit (data not shown). INHBA is a subunit of activin A, activin AB, and inhibin A.54 One report showed that only the combination of TNF and IL-3 or GM-CSF induced the release of activin A.55 It will require more careful study to examine whether eosinophils produce activin A, activin AB, or inhibin A in NPs and how expression of INHBA is regulated in eosinophils. 我们还识别了 NP 组织中嗜酸性粒细胞的潜在异质性。然而，大多数核心基因在所有簇中都是共享的，并且在 mRNA 水平上没有明显的细胞表面分子区分各个簇（数据未显示）。尽管这可能是 scRNA-Seq 的一个关键限制，因为每种嗜酸性粒细胞的序列读取较低，但我们也假设这些嗜酸性粒细胞簇可能主要受到 NP 组织内不同刺激的控制（例如，由于嗜酸性粒细胞的特定定位或与 NP 中其他细胞的相互作用）。为了验证这一假设，我们用各种嗜酸性粒细胞激活剂刺激 PB 嗜酸性粒细胞，发现 IL-1β和 IL-33 分别在嗜酸性粒细胞中差异表达 CXCL8 和 CCL4（图 5, E）。由于 IL-33 主要由上皮细胞释放，而 IL-1β主要由巨噬细胞和中性粒细胞产生，因此 NP 中嗜酸性粒细胞的局部分布可能控制特定细胞因子和细胞因子的释放。 此外，CCL4 阳性嗜酸细胞高表达 IL-33 受体（图 5，F），这表明上皮 IL-33 在来自簇 1 的 NP 嗜酸细胞中可能发挥作用。未来的研究使用空间转录组学可能会解决这个重要问题。我们还进行了ELISA以检测来自刺激的 PB 嗜酸细胞的激活素 A 的上清液；然而，水平低于检测限（数据未显示）。INHBA是激活素A、激活素AB 和抑制素 A。54 一项报告显示，只有 TNF 与 IL-3 或 GM-CSF 的组合才诱导了激活素 A的释放。55 需要更仔细的研究来检查嗜酸性粒细胞是否在 NPs 中产生激活素 A、激活素 AB 或抑制素 A，以及INHBA在嗜酸性粒细胞中的表达是如何调控的。
A recent small clinical trial showed that treatment of NP patients with dexpramipexole, a drug initially in development as a treatment for amyotrophic lateral sclerosis, depletes eosinophils by 97% without causing any reduction of NP score.12 Taken alone, this finding completely calls into question the role of eosinophils in the pathogenesis of CRSwNP, but this conclusion ignores other studies showing efficacy of mepolizumab and benralizumab in CRSwNP.11,56 It is still possible that eosinophils play a role in the initiation and growth of NPs but not their maintenance because patients with dexpramipexole-treated disease did not exhibit further growth of NP size during the 6-month treatment.12 It is also possible that IL-5 may play a pathogenic role on other immune cells rather than eosinophils in NPs because IL-5Rα was expressed on a subset of antibody-secreting cells in NPs, and the frequency of IL-5Rα–positive antibody-secreting cells in NPs was further elevated in severe CRS patients.57 Further study is needed to examine the role of IL-5 and eosinophil-derived mediators, including growth factors, on the pathogenesis of CRSwNP. 最近的一项 小型临床试验 显示，使用 dexpramipexole 这种最初开发用于治疗 肌萎缩侧索硬化症 的药物治疗 NP 患者，可以使嗜酸性粒细胞减少 97%，而不会导致 NP 评分的降低。12 单独来看，这一发现完全质疑了嗜酸性粒细胞在 CRSwNP 发病机制中的作用，但这一结论忽略了其他研究显示 mepolizumab 和 benralizumab 在 CRSwNP 中的疗效。1156 仍然有可能嗜酸性粒细胞在 NP 的发生和生长中发挥作用，但不参与其维持，因为接受 dexpramipexole 治疗的患者在 6 个月的治疗期间并未表现出 NP 大小的进一步增长。12 IL-5 也可能在鼻息肉中对其他免疫细胞而非嗜酸性粒细胞发挥致病作用，因为在鼻息肉中，IL-5Rα 在一部分抗体分泌细胞上表达，并且在重症慢性鼻窦炎患者中，IL-5Rα 阳性抗体分泌细胞的频率进一步升高。57 需要进一步研究 IL-5 和嗜酸性粒细胞衍生介质（包括生长因子）在 CRSwNP 发病机制中的作用。
Although to our knowledge this is the first study to successfully define the single cell transcriptome in NP eosinophils and provides new insights into the potential roles of eosinophils in CRSwNP, our present study has several limitations. Because eosinophils have less RNA than other cell types, we used cells with 200 transcripts per cell as the threshold to select cells for analysis during the QC process; this threshold is lower than most other scRNA-Seq studies. Furthermore, we were not able to detect enough eosinophils in control sinus tissues, so we were unable to determine genes that were changed simply as a result of their migration into tissues. There was an imbalance in the female-to-male ratio between control and CRSwNP in our scRNA-Seq study based on our blinded patient selection. Finally, the small sample size is a limitation, and a confirmation study with a larger sample size will be required. 尽管据我们所知，这是第一项成功定义 NP 嗜酸细胞单细胞转录组的研究，并为嗜酸细胞在 CRSwNP 中的潜在作用提供了新的见解，但我们目前的研究存在几个局限性。由于嗜酸细胞的 RNA 含量低于其他细胞类型，我们在质量控制过程中使用每个细胞 200 个转录本作为选择分析细胞的阈值；这个阈值低于大多数其他单细胞 RNA 测序研究。此外，我们未能在对照鼻窦组织中检测到足够的嗜酸细胞，因此无法确定因其迁移到组织中而改变的基因。根据我们盲法选择患者的结果，在我们的单细胞 RNA 测序研究中，对照组和 CRSwNP 组之间的雌雄比例存在不平衡。最后，较小的样本量是一个限制，需要进行更大样本量的确认研究。
In conclusion, we successfully captured human eosinophils by scRNA-Seq using 2 alternative protocols, preenrichment of granulocytes and use of a microwell-based scRNA-Seq platform, compared to the standard droplet-based scRNA-Seq. This new scRNA-Seq approach identifies dysregulated genes in NP eosinophils compared to their PB counterparts and suggests that elevated eosinophils in NP tissue may play important pathogenic roles in CRSwNP, in part by controlling inflammation and hyperproliferation of other cells and the reduced function of innate immune responses. Furthermore, our scRNA-Seq protocol will be useful for the study of other human eosinophilic diseases and may expand our knowledge of eosinophil biology as well as the pathophysiology of eosinophilic disorders.
Key messages
•
Microwell-based scRNA-Seq technology can detect single cell transcriptomes in human eosinophils.
•
Eosinophils are highly activated in NPs and show downregulation of genes involved in innate immunity and transendothelial migration as well as upregulation of antiapoptotic, proinflammatory, and growth factor genes.
•
NP eosinophils reveal heterogeneity of eosinophils based on the distinct expression of specific genes in each subset.
Supported in part by Regeneron Pharmaceuticals, the National Institutes of Health (grants P01AI145818, R01AI137174, and U19AI136443), and a grant from the Ernest S. Bazley Foundation. 部分由再生元制药、国家卫生研究院（拨款 P01AI145818、R01AI137174 和 U19AI136443）以及厄尼斯特·S·巴兹利基金会的拨款支持。
Disclosure of potential conflict of interest: W. W. Stevens has served on advisory boards for GlaxoSmithKline, Regeneron, and Melinta Therapeutics. A. T. Petershas served on advisory boards for Sanofi-Genzyme/Regeneron, Optinose, AstraZeneca, Novartis, and GSK; and has received research support from Optinose and Sanofi/Regeneron. L. C. Grammer reports personal fees from Astellas Pharmaceuticals. K. C. Welch reports consultant fees from Baxter, OptiNose, and Acclarent. D. B. Conley reports consulting fees from Medtronic and Sanofi/Regeneron. R. P. Schleimer reports personal fees from Intersect ENT, Merck, GlaxoSmithKline, Sanofi, AstraZeneca/Medimmune, Genentech, Actobio Therapeutics, Lyra Therapeutics, Astellas Pharma, and Otsuka; and has royalty rights to Siglec-8 and Siglec-8 ligand–related patents licensed by Johns Hopkins to Allakos. R. C. Kern reports consulting fees from Lyra Therapeutics, Medtronic, GSK, Genentech and Sanofi/Regeneron. B. S. Bochner has received remuneration for serving on scientific advisory board of Allakos and owns stock in Allakos; has served as consultant for GSK, Third Harmonic Bio, Lupagen, Acelyrin, and Sanofi/Regeneron; receives publication-related royalty payments from Elsevier and UpToDate; is coinventor of existing Siglec-8–related patents and thus may be entitled to a share of royalties received by Johns Hopkins University during development and potential sales of such products; and is a cofounder of Allakos, which makes him subject to certain restrictions under university policy (terms of this arrangement are being managed by Johns Hopkins University and Northwestern University in accordance with their conflict-of-interest policies). B. K. Tan reports personal fees from Sanofi/Regeneron/Genzyme and GSK. A. Kato has served on advisory board for AstraZeneca; reports gift for his research from Lyra Therapeutics; and has received research grants from Regeneron and AstraZeneca. The rest of the authors declare that they have no relevant conflicts of interest. 潜在利益冲突的披露：W. W. Stevens 曾担任葛兰素史克、Regeneron 和 Melinta Therapeutics 的顾问委员会成员。A. T. Peters 曾担任赛诺菲-Genzyme/Regeneron、Optinose、AstraZeneca、诺华和 GSK 的顾问委员会成员；并且从 Optinose 和赛诺菲/Regeneron 获得了研究支持。L. C. Grammer 报告称从 Astellas Pharmaceuticals 获得个人费用。K. C. Welch 报告称从 Baxter、OptiNose 和 Acclarent 获得顾问费用。D. B. Conley 报告称从美敦力和赛诺菲/Regeneron获得咨询费用。R. P. Schleimer 报告称从 Intersect ENT、默克、葛兰素史克、赛诺菲、AstraZeneca/Medimmune、基因泰克、Actobio Therapeutics、Lyra Therapeutics、Astellas Pharma 和大冢获得个人费用；并且拥有由约翰霍普金斯授权给 Allakos 的 Siglec-8 和 Siglec-8 配体相关专利的版税权。R. C. Kern 报告称从 Lyra Therapeutics、美敦力、GSK、基因泰克和赛诺菲/Regeneron获得咨询费用。B. S. 博赫纳因担任 Allakos 的科学顾问委员会成员而获得报酬，并持有 Allakos 的股票；曾担任 GSK、Third Harmonic Bio、Lupagen、Acelyrin 和 Sanofi/Regeneron 的顾问；从 Elsevier 和 UpToDate 获得与出版相关的版税；是现有 Siglec-8 相关专利的共同发明人，因此可能有权获得约翰霍普金斯大学在开发和潜在销售此类产品期间收到的版税的一部分；并且是 Allakos 的共同创始人，这使他受到大学政策下某些限制（该安排的条款由约翰霍普金斯大学和西北大学根据其利益冲突政策进行管理）。B. K. Tan 报告从Sanofi/Regeneron/Genzyme和 GSK 获得个人费用。A. Kato 曾担任阿斯利康的顾问委员会成员；报告从 Lyra Therapeutics 获得研究赠款；并从Regeneron和AstraZeneca获得研究资助。其余作者声明他们没有相关的利益冲突。
We gratefully acknowledge James Norton, Roderick Carter, Caroline P. E. Price, and Julia H. Huang (Northwestern University Feinberg School of Medicine) for their skillful technical assistance. We gratefully acknowledge Suchitra Swaminathan and the Flow Cytometry Core Facility, supported by National Cancer Institute CCSG P30 CA060553, awarded to the Robert H. Lurie Comprehensive Cancer Center at Northwestern University for their technical assistance with BD Rhapsody scRNA-Seq. We also gratefully acknowledge the Northwestern University NUSeq Core Facility and the UCLA Technology Center for Genomics & Bioinformatics for their sequencing support of our BD Rhapsody libraries. 我们衷心感谢詹姆斯·诺顿、罗德里克·卡特、卡罗琳·P·E·普赖斯和朱莉亚·H·黄（西北大学范伯格医学院）提供的高超技术支持。我们衷心感谢苏奇特拉·斯瓦米纳坦和流式细胞术核心设施，得到国家癌症研究所CCSG P30 CA060553 的支持，该项目授予西北大学罗伯特·H·卢里综合癌症中心，感谢他们在 BD Rhapsody 单细胞 RNA 测序方面的技术支持。我们还衷心感谢西北大学NUSeq 核心设施和加州大学洛杉矶分校基因组与生物信息学技术中心对我们 BD Rhapsody 文库的测序支持。
C. Bangert, S. Villazala-Merino, M. Fahrenberger, T. Krausgruber, W.M. Bauer, V. Stanek, et al.
Comprehensive analysis of nasal polyps reveals a more pronounced type 2 transcriptomic profile of epithelial cells and mast cells in aspirin-exacerbated respiratory disease
I. Torsteinsdottir, N.G. Arvidson, R. Hallgren, L. Hakansson
Enhanced expression of integrins and CD66b on peripheral blood neutrophils and eosinophils in patients with rheumatoid arthritis, and the effect of glucocorticoids
L. Zhao, S. Xu, G. Fjaertoft, K. Pauksen, L. Hakansson, P. Venge
An enzyme-linked immunosorbent assay for human carcinoembryonic antigen–related cell adhesion molecule 8, a biological marker of granulocyte activities in vivo
B. Gregory, A. Kirchem, S. Phipps, P. Gevaert, C. Pridgeon, S.M. Rankin, et al.
Differential regulation of human eosinophil IL-3, IL-5, and GM-CSF receptor alpha-chain expression by cytokines: IL-3, IL-5, and GM-CSF down-regulate IL-5 receptor alpha expression with loss of IL-5 responsiveness, but up-regulate IL-3 receptor alpha expression
L.Y. Liu, J.B. Sedgwick, M.E. Bates, R.F. Vrtis, J.E. Gern, H. Kita, et al.
Decreased expression of membrane IL-5 receptor alpha on human eosinophils: I. Loss of membrane IL-5 receptor alpha on airway eosinophils and increased soluble IL-5 receptor alpha in the airway after allergen challenge
J. Barreto, S.K. Karathanasis, A. Remaley, A.C. Sposito
Role of LOX-1 (lectin-like oxidized low-density lipoprotein receptor 1) as a cardiovascular risk predictor: mechanistic insight and potential clinical use
Arterioscler Thromb Vasc Biol, 41 (2021), pp. 153-166
Interleukin-3, -5, and granulocyte macrophage colony-stimulating factor–induced adhesion molecule expression on eosinophils by p38 mitogen-activated protein kinase and nuclear factor-κB
T. Homma, A. Kato, M. Sakashita, T. Takabayashi, J.E. Norton, L.A. Suh, et al.
Potential involvement of the epidermal growth factor receptor ligand epiregulin and matrix metalloproteinase-1 in pathogenesis of chronic rhinosinusitis
T.W. Harrison, P. Chanez, F. Menzella, G.W. Canonica, R. Louis, B.G. Cosio, et al.
Onset of effect and impact on health-related quality of life, exacerbation rate, lung function, and nasal polyposis symptoms for patients with severe eosinophilic asthma treated with benralizumab (ANDHI): a randomised, controlled, phase 3b trial
