第一部分：

目的  T淋巴细胞介导的造血干/祖细胞免疫损伤是再生障碍性贫血（AA）最主要的病例生理机制。本文旨在通过基于测序的转录组谱，全面分析mRNA、长非编码RNA（lncRNA）和microRNA（miRNA）在重型再生障碍性贫血（SAA）和健康对照（HC）中的表达，为探索SAA发病机制提供线索。方法  （1）采用淋巴细胞分离液密度梯度离心法分离外周血单个核细胞（PBMCs）；（2）采用免疫磁珠法分选CD4⁺T细胞；（3）采用Trizol提取RNA；（4）对外周血SAA（n = 4）和HC（n = 3）的CD4⁺T细胞进行完整转录组RNA测序（RNA-seq）；（5）进行总体表达模式及差异表达（DE）分析；并利用辅助性T细胞1/2（Th1/Th2） lncRNA集及差异lncRNA确定核心lncRNA；对核心lncRNA进行功能富集分析，转录因子预测以及建立竞争性内源RNA（ceRNA）网络；对ceRNA网络中的核心mRNA进行可变剪切分析并利用GEO数据库进行交叉验证。结果  超过75%的RNA在SAA患者和健康对照共表达，但亦表现出不同的转录表达谱。（1）送检的7个样品中检测到39456个mRNA、10496个已知的lncRNA和658个miRNA。在SAA和HC中，所有三种类型的RNA的表达水平均相当。（2）以差异倍数（FC） ≥ 1.5且P < 0.05为标准，我们检测到78个DE miRNA，847个DE lncRNA，3753个DE mRNA和2670个DE基因。（3）无监督聚类和主成分分析显示lncRNA表达在患者与对照组中差异显著，另发现新lncRNA 1839种。（4）847个DE lncRNA中，其中320个lncRNA在SAA中上调，其余部分则下调。（5）lncRNA靶基因预测显示：80.8％的lncRNA与其对应的基因表达水平成正相关。（6）lncRNA转录因子预测发现：分别有13个lncRNA和8个lncRNA参与调控IFNG-IFNG通路中的关键因子干扰素调节因子1（IRF1）及信号转导转录激活因子3（STAT3）。（7）GO生物学功能富集显示：lncRNA预测的mRNA主要参与蛋白质结合、细胞质基质、细胞核、细胞质、核质、膜组成部分和金属离子结合。（8）KEGG功能富集提示：单纯疱疹病毒感染、代谢通路、肿瘤、PI3K−Akt通路和人类免疫缺陷病毒感染相关信号通路被显著富集。（9）根据lncRNA–miRNA，miRNA–mRNA的靶标预测，我们构建一个ceRNA网络。（10）Th细胞相关ceRNA网络核心mRNA功能预测分析显示：破骨细胞分化，类风湿性关节炎，NF− κ B信号通路，疟疾及军团菌感染，CD8⁺ α β-T细胞活化信号通路均被显著富集。（11）利用GEO数据库进行交叉验证，发现炎症因子ICOS, RELB及TNFSF8在SAA患者中高表达，而免疫负调控因子SOCS1同样高表达，提示SAA患者免疫内环境存在稳态调节。结论  （1）lncRNA转录因子预测及靶基因功能预测发现SAA免疫相关通路异常表达，提示lncRNA在其发病机制中可能具有重要作用；（2）炎症因子及免疫负调控因子的共同高表达提示SAA患者免疫内环境可能存在稳态调节。

第二部分：

目的  观察伴Coombs’试验（+）血细胞减少患者的临床特征、疗效及长期生存情况。方法  回顾性分析我科2009年07月01日至2020年06年30日期间伴Coombs’试验（+）的血细胞减少患者，比较获得性骨髓衰竭（BMF组）和继发性血细胞减少（继发组）的临床特征差异，并评估BMF组患者的疗效及预后，采用Kaplan-Meier法分析其5年总体生存率。结果  82例患者符合纳入标准，其中BMF组58例，继发组24例。继发组女性比例明显高于BMF组（83%对41%，P=0.001），该组患者贫血以轻中度为主[血红蛋白中位数78（31~129）g/L]，白细胞水平略低于正常，而血小板显著高于BMF组[71（4~384）´10⁹/L对31.5（4~401）´10⁹/L，P=0.020]。脾脏肿大在继发组更常见，其阳性率高于BMF组（38%对7%，P=0.002）。两组患者红细胞破坏的间接指标如网织红细胞绝对值、总胆红素、间接胆红素及乳酸脱氢酶中位数处于正常范围且差异无统计学意义。BMF组患者淋巴细胞比例明显高于继发组[58.3%（16.4%~94.5%）对34.6%（5.4%~73.7%），P<0.001]，而CD19⁺ B细胞比例低于继发组 [9.3%（0.3%~27.1%）对14.0%（1.2%~35.6%），P=0.010]。继发组患者的风湿免疫相关指标如抗核抗体滴度/阳性率、可提取性核抗原抗体谱阳性率及IgG水平均高于BMF组，但其补体C3、C4水平较BMF组下降（P值均<0.050）。BMF组48例（83%）患者既往有输血史，12例（21%）患者入院前2周有发热病史并予青霉素类或头孢类等抗菌素治疗，4例（7%）患者有丙种球蛋白治疗史（3个月内）。共有17例BMF患者复查Coombs试验，11例（65%）转阴。余6例阳性患者中，1例效价值较前升高，1例持平，4例较前降低。Coombs试验效价转阴或降低的15例患者，其中10例（67%）经免疫抑制及促造血治疗后脱离输血；另5例均为MDS及PNH患者，尚未获得血液学缓解，仍依赖输血。49例BMF组患者纳入随访，其中39例规范服药患者有31例（79.5%）获得血液学缓解，5例死亡；8例未规律服药患者有4例死亡，4例患者带病生存，需依赖血制品输注。本组患者5年总体生存率为（79.8±7.3%），服药患者生存率与未服药组差异无统计学意义。另2例患者行骨髓移植治疗获得血液学缓解。结论  伴Coombs’试验（+）的骨髓衰竭患者应积极给予经典的免疫抑制及促造血治疗，预后良好；而因风湿免疫病继发的伴Coombs’试验（+）血细胞减少患者，临床特征显著，容易鉴别。

第三部分：

目的 本研究旨在检测再生障碍性贫血（AA）患者CD8⁺调节性T细胞（CD8⁺Treg）的数量及生物学功能特征，并探讨其在AA发病机制中的作用。方法 （1）采用淋巴细胞分离液密度梯度离心法分离外周血单个核细胞（PBMCs）；（2）采用流式细胞术（FCM）检测外周血 CD8⁺Treg占CD3⁺淋巴细胞比例；（3）采用FCM检测外周血CD8⁺Treg体外分泌细胞因子的能力；（4）采用CFSE标记法检测外周血CD8⁺Treg体外增殖能力；（5）采用免疫磁珠分选法(MACS)分选CD4⁺T细胞、 CD8⁺细胞毒性T细胞（Tcyt）和CD8⁺Treg；（6）采用Transwell板检测CD8⁺Treg的趋化能力；（7）采用FCM检测外周血CD8⁺Treg趋化因子受体CCR 7、CCR 8的表达水平；（8）采用酶联免疫吸附法（ELISA）检测骨髓（BM）上清中趋化因子配体CCL 21的浓度；（9）采用CFSE标记法检测共培养体系中CD8⁺Treg抑制CD4⁺T/Tcyt 增殖的能力；（10）ELISA 检测共培养上清 IFN-γ的水平；（11）采用FCM检测由外周血单核细胞诱导的成熟树突状细胞（mDC）表面免疫球蛋白样转录物（ILT3）的表达水平；（12）MACS分选脐血来源CD34⁺细胞，通过甲基纤维素半固体培养基（H4435）中加入体外共培养体系获得的上清，比较AA组和健康对照（HC）组CD8⁺Treg改善造血克隆形成的能力。结果 （1）重型再生障碍性贫血[SAA（n = 19）]、非重型再生障碍性贫血[NSAA（n = 7）]患者PBMC中CD8⁺Treg细胞数量较HC组（n = 22）均明显减低（分别为7.61 ± 0.99 % vs 16.08 ± 1.21%，P ＜ 0.0001和10.32 ± 1.38% vs 16.08 ± 1.21%，P = 0.043）；（2）SAA组中CD28^-IFN-γ⁺T细胞占CD3⁺CD8⁺细胞的比例明显低于HC组（8.38 ± 2.25% vs 20.03 ± 2.17%，P = 0.010，n = 4）；与治疗前相比，抗胸腺球蛋白（ATG）治疗后达完全缓解（CR）患者（n = 3）其数量明显恢复（8.38 ± 2.25% vs 28.67 ± 8.22%，P = 0.040）。同时SAA组CD28^- TGF-⁺T细胞也较HC组和CR组降低（分别为0.63 ± 0.14% vs 1.57 ± 0.26%，P = 0.020和0.63 ± 0.13% vs 1.59 ± 0.32%，P = 0.029）；（3）SAA组患者CD8⁺Treg体外增殖能力明显低于HC组（21.68 ± 1.91% vs 53.86 ± 4.94%，P = 0.001，n = 4）；（4）BM上清中CCL 21浓度在SAA组和HC组中无显著差异；（5）SAA组患者CD8⁺Treg表达趋化因子受体CCR 7明显低于HC组（1.54 ± 0.33% vs 4.33 ± 0.39%，P = 0.005，n = 5）；（6）SAA组患者CD8⁺Treg体外趋化指数明显低于HC组（0.47 ± 0.22% vs 1.64 ± 0.62%，P = 0.008，n = 3）；（7）SAA组患者CD8⁺Treg抑制CD4⁺T细胞增殖的能力显著低于HC组；（8）SAA组患者CD8⁺Treg抑制CD4⁺T淋巴细胞产生IFN-γ的能力较HC组明显减弱；（9）SAA组患者CD8⁺Treg上调mDC表达ILT 3的能力较HC组减弱（45.05 ± 2.45% vs 55.9 ± 4.54%，P = 0.028，n = 4）；（10）SAA患者CD4⁺T淋巴细胞体外培养上清可明显抑制造血克隆形成，且SAA患者CD8⁺Treg改善造血克隆形成能力较HC组减弱。结论 （1）AA患者CD8⁺Treg数量减少且增殖能力减弱。（2）AA患者CD8⁺Treg免疫抑制功能存在缺陷。

Part 1: 

Objective T lymphocyte-mediated immune damage of hematopoietic stem/progenitor cells is the main physiological mechanism of aplastic anemia (AA). This article aims to comprehensively analyze the expression of mRNA, long non-coding RNA (lncRNA), and microRNA (miRNA) in severe aplastic anemia (SAA) and healthy controls (HC) through RNA sequencing and bioinformatic analysis for exploring the pathogenesis of SAA. Methods (1) Peripheral blood mononuclear cells (PBMCs) were separated by density gradient centrifugation with lymphocyte separation solution. (2) Magnetic activated cell sorting was used to sort CD4⁺T cells. (3) Total RNA was isolated using the trizol according to the manufacturer's protocol. (4) Deep RNA sequencing with samples from SAA (n = 4) and healthy controls (n = 3) was conducted. (5) Overall expression patterns and differential expression (DE) performed. Hub lncRNA was defined by the intersection between DE lncRNAs and lncRNA of T helper cells. Next, functional enrichment analysis was performed, transcription factors were predicted and Th related competing endogenous RNA (ceRNA) network was constructed. Hub mRNA in the ceRNA were analyzed by Identification of differentially expressed alternative splicing events and cross-validation with the GEO database. Results More than 75% of RNA was co-expressed in SAA patients and healthy controls, but they also showed different transcriptional expression profiles. (1) We detected 39456 mRNAs, 10496 known lncRNAs, and 658 miRNAs in 7 samples of CD4⁺T cells. Overall, the expression levels in all three types of RNA were comparable in patients and healthy controls. (2) Using the criterion of P < 0.05, fold change (FC) > 1.5 in all samples, we detected 78 DE miRNAs, 847 DE lncRNAs, 3753 DE mRNAs, and 2670 DE Gene in SAA patients compared with the healthy controls. (3) Unsupervised clustering and principal component analyses revealed that lncRNA expression profiles distinguish SAA patients from the controls. We detected 10496 known lncRNAs in SAA and found 1839 new lncRNAs. (4) Of the 847 DE lncRNAs, 320 lncRNAs were commonly up-regulated and the others were down-regulated in SAA. (5) The prediction of lncRNA target genes showed that 80.8% of lncRNAs positively correlated with their corresponding gene expression level (either both up or downregulated. (6) The prediction of transcription factors for lncRNA showed the interferon regulatory factor 1 and signal transducer and activator of transcription 3, which were proved to play an essential role in the IFNG-IFNGR pathway, were included in the core TF‑lncRNA network and regulated by 13 lncRNAs and 8 lncRNAs, respectively. (7) Functional annotation revealed that mRNAs predicted by the lncRNAs were mainly involved in eight Gene Ontology terms as protein binding, cytosol, nucleus, cytoplasm, integral component of membrane, nucleoplasm, plasma membrane, and metal ion binding. (8) The enrichment of Kyoto Encyclopedia of Genes and Genomes pathways showed that herpes simplex virus 1 infection, metabolic pathways, pathways in cancer, PI3K-Akt pathway, and human immunodeficiency virus 1 infection-related signaling pathways had significantly enriched. (9) According to the target pairs of lncRNA– miRNA, miRNA– mRNA, we constructed a ceRNA network. (10) Predictive analysis of the hub mRNA function of the ceRNA network related to Th showed that differentiation of osteoclasts, rheumatoid arthritis, NF- κ B signal pathway, disease, and Legionella infection, CD8 ⁺αβ-T cell activation signaling pathway were all enrichment. (11) We found that both the inflammatory factors (ICOS, RELB, and TNFSF8) and the anti-inflammatory factor (SOCS1) were highly expressed in SAA patients, and these genes were cross-validated by the GEO database. It suggested that there may be a homeostatic regulation of the immune environment in SAA patients. Conclusions (1) The prediction of lncRNA transcription factors and target gene function indicates abnormal immune-related pathways in SAA patients meanwhile suggests that lncRNAs may play an important role in the pathogenesis of SAA. (2) The high co-expression of inflammatory factors and anti-inflammatory factors suggests that there may be a homeostatic regulation of the immune internal environment of SAA patients.

Part 2: 

Objective To assess the clinical characteristics, efficacy, and long-term survival of cytopenia patients with positive Coombs’ test.  Methods Analyze the clinical characteristics of cytopenia patients with positive Coombs’ test in our department from July 1, 2009, to June 30, 2020, retrospectively, compare the differences between acquired bone marrow failure (BMF group) and secondary cytopenia (secondary group) and evaluate the response and outcome of patients in the BMF group. Kaplan-Meier method was used to analyze the 5-year overall survival rate. Results A total of 82 patients met the inclusion criteria, including 58 in the BMF group and 24 in the secondary group. The proportion of women was 83% in secondary group, which was significantly higher than that in the BMF group (41%, P=0.001). Mild to moderate anemia was typical in the secondary group, while white blood cells were slightly lower, and the platelet level in the secondary group [71 (4~384) ´10⁹/L] was significantly higher than that in the BMF group [31.5(4~401) ´10⁹/L, P=0.020]. Splenomegaly was more common in the secondary group, and its positive rate was higher than that of the BMF group (38% vs. 7%, P=0.002). The medians of indirect indicators of red blood cell destruction in the two groups of patients, such as the number of reticulocytes, total bilirubin, indirect bilirubin, and lactate dehydrogenase, were normal while the differences were not statistically significant. The percentage of lymphocytes in the BMF group [58.3% (16.4%~94.5%)] was significantly higher than that of the secondary group [34.6% (5.4%~73.7%), P<0.001], while the proportion of CD19+B cells [9.25% (0.3%~27.1%)] was significantly lower than that of the secondary group[14% (1.2%~35.6%), P=0.010]. Rheumatism-related indicators such as the titer or positive rate of antinuclear antibody, positive rate of extractable nuclear antigen antibody and IgG levels in the secondary group were significantly higher than those in the BMF group, but the levels of complement C3 and C4 in the secondary group were lower than those in the BMF group (P<0.050). In the BMF group, 48 patients (83%) had a history of blood transfusion, 12 patients (21%) had two weeks history of fever before admission treated with antibiotics such as penicillin or cephalosporins, four patients (7%) had gamma globulin Treatment history within three months. A total of 17 BMF patients rechecked the Coombs' test, and 11 cases (65%) became negative. Among the remaining six positive patients, the titer of one was higher, one was the same, and four were lower than before. There were 15 patients whose titers in the Coombs' test turned negative or decreased. 10 (67%) of them didn't need blood transfusions after immunosuppressive and hematopoietic therapy. The others diagnosed as MDS and PNH had not yet achieved hematological remission and still depended on blood transfusion. Forty-nine patients in the BMF group were follow-up. Among them, 31 patients (79.5%) of the 39 patients who took the standard medication achieved hematological remission, and five died; 4 patients died of the eight patients who did not take the medication regularly, and four patients lived with the disease and needed to rely on blood transfusions. The 5-year overall survival rate of the patients in this group was (79.8 ± 7.3%), and there was no statistically significant difference in the survival rate between the patients taking such medicine or not. The other two patients underwent bone marrow transplantation and achieved hematological remission. Conclusions Coombs’ test can be positive in patients with bone marrow failure. The prognosis of those patients is remarkable if they are treated with classic immunosuppressive therapy. Patients with positive Coombs' test and cytopenia secondary to rheumatoid immune disease have distinct clinical features and can be identified easily. 

Part 3:

Objective This study aims to detect the number and biological functions of CD8⁺ regulatory T cells (CD8⁺Tregs) in aplastic anemia (AA) and explore its role in the pathogenesis of AA. Methods (1) Peripheral blood mononuclear cells (PBMCs) were separated by density gradient centrifugation with lymphocyte separation solution; (2) The proportion of peripheral blood CD8⁺Tregs in CD3⁺CD8⁺T cells was measured by flow cytometry（FCM）; (3) FCM was used to detect the ability of CD8⁺Tregs to secrete cytokines in vitro; (4) CFSE was used to detect the proliferation ability of CD8⁺Tregs in vitro; (5) Magnetic activated cell sorting (MACS) was used to sort CD4 ⁺ T cells, CD8⁺ cytotoxic T cells (Tcyt) and CD8⁺Tregs; (6) The migratory ability of CD8⁺Tregs was detected by chemotaxis assay; (7) FCM was used to detect the expression levels of CCR 7 and CCR 8 on CD8⁺Tregs in peripheral blood; (8) Enzyme-linked immunosorbent assay (ELISA) was used to detect the concentration of CCL 21 in the bone marrow supernatant; (9) CFSE was used to detect the inhibitory ability of CD8⁺Tregs on the proliferation of CD4⁺T cells/ Tcyt in a co-culture system; (10) ELISA was used to detect the level of IFN-γ in the co-culture supernatant; (11) FCM was used to detect the expression levels of immunoglobulin-like transcript 3 (ILT 3) on the surface of mature dendritic cells (mDC) induced from mononuclear cells;(12) MACS was used to sort cord blood-derived CD34⁺ cells treated with the supernatant obtained by the co-culture system in the methylcellulose semi-solid medium (H4435) for addressing whether AA-CD8⁺Treg could improve hematopoietic colony formation units (CFU) by suppressing effector T cells or not. Results (1) The number of CD8⁺Tregs in SAA (n = 19) and NSAA (n = 7) were significantly lower than that in the healthy control (HC) group (n = 22) (7.61 ± 0.99% vs 16.08 ± 1.21%, P ＜ 0.0001 and 10.32 ± 1.38% vs 16.08 ± 1.21%, P = 0.043, respectively); (2) The proportion of CD28^-IFN-γ⁺T cells in CD3⁺CD8⁺T cells in the SAA group was significantly lower than that in the HC group (8.38 ± 2.25% vs 20.03 ± 2.17%，P = 0.010，n = 4), and it was improved significantly in complete response (CR) patients after treatment with anti-thymocyte globulin (ATG) (8.38 ± 2.25% vs 28.67 ± 8.22%, P = 0.040). At the same time, the percentage of CD28^-TGF-⁺T cells in the SAA group was also lower than that in the HC or CR group (0.63 ± 0.14% vs 1.57 ± 0.26%, P = 0.020 and 0.63 ± 0.13% vs 1.59 ± 0.32%, P = 0.029, respectively). (3) The proliferation rate of peripheral blood CD8⁺Tregs in the SAA group was significantly lower than that in the HC group (21.68 ± 1.91% vs 53.86 ± 4.94%, P = 0.001, n = 4). (4) There was no significant difference in the concentration of CCL 21 between the SAA group and HC group in bone marrow supernatant. (5) The expression of CD8⁺Tregs chemokine receptor CCR7 of AA patients was significantly lower than that of the HC group (1.54 ± 0.33% vs 4.33 ± 0.39%, P = 0.005, n = 5). (6) The in vitro chemotaxis index of CD8⁺Tregs in peripheral blood of SAA patients was significantly lower than that of HC (0.47 ± 0.22% vs 1.64 ± 0.62%, P = 0.008, n = 3). (7)The inhibitory ability of CD8⁺Tregs from SAA patients on the proliferation of CD4⁺ T cells was significantly decreased than that from HC group; (8) The inhibitory ability of CD8⁺Tregs from SAA patients on IFN-γ production of CD4⁺ T cells was significantly weaker than that from HC group; (9) The ability of CD8⁺Tregs from SAA patients to up-regulate ILT3 of mDC was impaired (45.05 ± 2.45% vs 55.90 ± 4.54%，P = 0.028，n = 4). (10) AA- CD8⁺Tregs had a worse capacity of improving CFU formations than HC-CD8⁺Tregs. Conclusion (1) The frequency of CD8⁺Tregs from AA patients is lower with reduced proliferation capacity. (2) The immunosuppressive function of CD8⁺Tregs from AA patients was defective.