Characterization of T-cell repertoire in hairy cell leukemia patients before and after recombinant immuno toxin BL22 therapy
毛细胞白血病患者在接受重组免疫毒素BL22治疗前后的T细胞受体库特征分析
Introduction
引言
Hairy cell leukemia (HCL) is a B-cell lymph o pro lifer ative malignancy representing $2%$ of all leukemias [1]. Hairy cells are mature though not terminally differentiated [2], and display surface CD19, CD20, CD22, immunoglobulin (Ig), CD79A, CD40, CD103, and CD11c [3, 4]. CD25 is expressed in $80%$ of patients and the CD25-negative cases usually belong to a poorprognosis HCL variant (HCLv) group [5–8]. The highly effective purine analogs cladribine (chlo rode oxy a de nosine, or CdA) and pen to statin (de oxy co for my c in or DCF), are not curative, but induce long-term complete remissions in most patients until repeated courses fail to induce responses [9–16]. Purine analogs cause significant decreases in $\mathrm{CD}4+$ and $\mathrm{CD}8+$ lymphocytes [17, 18]; the median time to recovery of $\mathrm{CD}4+$ T-cells is up to 4 years, longer than the time to recovery of $\mathrm{CD}8+$ Tcells. Immuno bio logic targeted treatments include the anti-CD20 monoclonal antibody (MAb) Rituximab [19–22] and the anti-CD22 recombinant immuno toxin BL22 [23, 24], neither of which targets T-cells. Complete and partial remissions have been achieved with both agents in che mo resistant HCL [19–24].
毛细胞白血病 (HCL) 是一种B淋巴细胞增殖性恶性肿瘤，占所有白血病病例的2% [1]。毛细胞虽已成熟但未终末分化 [2]，表面表达CD19、CD20、CD22、免疫球蛋白 (Ig)、CD79A、CD40、CD103和CD11c [3, 4]。80%患者表达CD25，而CD25阴性病例通常属于预后不良的HCL变异型 (HCLv) 组别 [5-8]。高效嘌呤类似物克拉屈滨 (chlorodeoxyadenosine, CdA) 和喷司他丁 (deoxycoformycin, DCF) 虽不能根治疾病，但能使大多数患者获得长期完全缓解，直至重复疗程失效 [9-16]。这类药物会导致CD4+和CD8+淋巴细胞显著减少 [17, 18]；CD4+ T细胞恢复中位时间长达4年，比CD8+ T细胞恢复时间更长。免疫生物靶向治疗包括抗CD20单克隆抗体 (MAb) 利妥昔单抗 [19-22] 和抗CD22重组免疫毒素BL22 [23, 24]，两者均不靶向T细胞。这两种药物在化疗耐药性HCL中均已实现完全和部分缓解 [19-24]。
Large granular lymphocytes (LGL) have been described in HCL as either CD3-negative NK cells or $\mathrm{CD}3+$ cytotoxic T-lymphocytes (CTLs) which are clonal but not malignant [25–28]. Kluin-Nelemans et al. reported T-cell restriction in $25~\mathrm{HCL}$ patients by TCR gamma (TCRG) PCR [29], and in seven patients by RTPCR of each of 22 TCR beta variable (TRBV) families [29, 30]. In one case, expansion of $\mathrm{CD8+/CD57+CTLs}$ was documented [31]. HCL-specific T-cell clones could be isolated from HCL patients [32, 33], but clonally expanded T-cells from two HCL patients were not specific for hairy cells [31]. Thus, it is not clear that T-cell restriction in HCL represents an immune response to the tumor. We have reported that HCL patients after purine analogs frequently have increased percentages of CTLs, defined by ${\mathrm{CD}}8+/{\mathrm{CD}}3+$ , $\mathrm{CD}56+/\mathrm{CD}3+$ , and $\mathrm{CD}57+\mathrm{/CD}3+$ T-cells [34]. In that study, six of six patients analyzed had TCRG restriction by PCR. It has not been possible, due to the limited number of patients previously studied, to determine whether CTL antigens, TCRG PCR, or clinical factors correlate with mono/ oli go cl on al skewing of the $\alpha/\beta$ T-cell repertoire.
大颗粒淋巴细胞 (LGL) 在毛细胞白血病 (HCL) 中被描述为 CD3 阴性 NK 细胞或 $\mathrm{CD}3+$ 细胞毒性 T 淋巴细胞 (CTLs)，这些细胞呈克隆性但非恶性 [25–28]。Kluin-Nelemans 等人通过 TCR γ (TCRG) PCR 在 25 例 HCL 患者中报告了 T 细胞限制性 [29]，并通过 22 个 TCR β 可变 (TRBV) 家族的 RT-PCR 在 7 例患者中证实 [29, 30]。其中一例记录了 $\mathrm{CD8+/CD57+CTLs}$ 的扩增 [31]。虽然可从 HCL 患者中分离出 HCL 特异性 T 细胞克隆 [32, 33]，但两名 HCL 患者的克隆性扩增 T 细胞并非针对毛细胞 [31]。因此，尚不清楚 HCL 中的 T 细胞限制性是否代表对肿瘤的免疫反应。我们曾报道，嘌呤类似物治疗后 HCL 患者常出现 CTLs 比例升高，定义为 ${\mathrm{CD}}8+/{\mathrm{CD}}3+$、$\mathrm{CD}56+/\mathrm{CD}3+$ 和 $\mathrm{CD}57+\mathrm{/CD}3+$ T 细胞 [34]。该研究中，6 例接受 PCR 检测的患者均显示 TCRG 限制性。由于既往研究患者数量有限，尚无法确定 CTL 抗原、TCRG PCR 或临床因素是否与 $\alpha/\beta$ T 细胞库的单/寡克隆偏倚相关。
To explore the relationship between increased percentages of CTL cells and skewing of the T-cell repertoire, we studied 22 purine analog-pretreated HCL patients for TCRG and TRBV restriction, and TRBV expression. We performed flow cytometry to quantify circulating CTLs, NK cells, or $\mathrm{CD}4+$ cells. In 20 of these patients we performed CDR3 spectra typing to determine if each of 22 established TRBV families was expressed in a mono/oligoclonal or polyclonal fashion. To begin to investigate the clinical significance of T-cell restriction in HCL, we studied patients with a wide range of clinical situations with respect to disease burden and prior treatment.
为探究细胞毒性T淋巴细胞(CTL)比例升高与T细胞库偏移的关系，我们研究了22例经嘌呤类似物预处理的毛细胞白血病(HCL)患者的TCRG和TRBV限制性以及TRBV表达情况。通过流式细胞术定量检测循环中的CTL、自然杀伤(NK)细胞或$\mathrm{CD}4+$细胞。其中20例患者进行了CDR3谱型分析，以确定22个已知TRBV家族是否以单克隆/寡克隆或多克隆形式表达。为初步探讨HCL中T细胞限制性的临床意义，我们纳入了疾病负荷和既往治疗情况各异的患者群体。
Material and methods
材料与方法
Patients and controls
患者与对照组
Blood collected in heparin and EDTA tubes was obtained as part of sample acquisition protocols with informed consent approved by the NCI Investigator’s Review Board. All patients had clinical presentation and morphology consistent with HCL and had a complete peripheral blood mono nuclear cell (PBMC) immunophenotypic analysis as part of their standard evaluation. For each patient, dates of blood collection for flow cytometry and PCR results were the same or up to 3 weeks apart. Patients and 18 had hairy cell leukemia variant (HCLv) and the other patients had classic HCL. The two patients with HCLv each had HCL cells negative for CD25 but otherwise consistent with HCL, including positive for CD103 and bright positive for CD22, CD11c and CD20. The diagnosis of HCL by flow cytometry also required the demonstration of monoclonality based on lambda or kappa surface light chains. Peripheral blood specimens from eight normal agematched donors were also collected in EDTA tubes. The clinical status and response to BL22 of most of the patients was recently reported [23].
作为样本采集方案的一部分，在NCI审查委员会批准知情同意后获取了肝素管和EDTA抗凝管采集的血液样本。所有患者均符合HCL的临床表现和形态学特征，并作为标准评估的一部分接受了外周血单个核细胞(PBMC)免疫表型分析。每位患者的流式细胞术采血日期与PCR检测结果日期相同或间隔不超过3周。患者#14和18为毛细胞白血病变异型(HCLv)，其余患者均为经典型HCL。两名HCLv患者的肿瘤细胞均不表达CD25，但其他免疫表型(包括CD103阳性、CD22/CD11c/CD20强阳性)符合HCL特征。通过流式细胞术诊断HCL还需证实λ或κ表面轻链的单克隆性。同时采集了8例年龄匹配健康志愿者的EDTA抗凝外周血标本。多数患者的临床状态及对BL22的治疗反应已在近期文献中报道[23]。
Isolation and purification of genomic DNA from whole blood samples
全血样本中基因组DNA的分离与纯化
Whole blood was fraction a ted by cent ri fuga tion in CTP Cell Preparation Tubes (BD, Franklin Lakes, NJ). Nucleated cells were removed and washed with phosphate-buffered saline (without ${\mathrm{Mg}}^{2+}$ or $\mathrm{Ca}^{2+}$ ). The pelleted, washed cells were lysed with DNAzol (Invitrogen, Carlsbad, CA) and genomic DNA was isolated according to the manufacturer’s methods. The concentration of DNA was determined by optical density at $260~\mathrm{nm}$ . One microgram of purified DNA was placed in $0.5~\mathrm{ml}$ PCR tubes, mixed with $20~\mathrm{l}$ of Gene Releaser resin (Bio ventures Inc., Murfreesboro, TN) and layered with mineral oil. The resin/DNA mixtures were pre incubated in a thermo cycle r (Model 9700, Applied Biosystems, Foster City, CA) according to the Gene Releaser protocol, prior to the addition of the specific polymerase chain reaction (PCR) master mixes.
全血通过CTP细胞制备管(BD, Franklin Lakes, NJ)离心分离。去除有核细胞并用磷酸盐缓冲液(不含${\mathrm{Mg}}^{2+}$或$\mathrm{Ca}^{2+}$)洗涤。将沉淀的洗涤细胞用DNAzol(Invitrogen, Carlsbad, CA)裂解，并按制造商方法分离基因组DNA。DNA浓度通过$260~\mathrm{nm}$光密度测定。将1微克纯化DNA置于$0.5~\mathrm{ml}$ PCR管中，与$20~\mathrm{l}$ Gene Releaser树脂(Bio ventures Inc., Murfreesboro, TN)混合并覆盖矿物油。在加入特定聚合酶链反应(PCR)预混液前，树脂/DNA混合物按Gene Releaser方案在热循环仪(Model 9700, Applied Biosystems, Foster City, CA)中进行预孵育。
T-cell receptor gamma chain gene rearrangement studies
T细胞受体γ链基因重排研究
To assess clonality of the TCRG gene, consensus primers directed to conserved sequences in the junctional regions $(\mathbf Ḋ \gamma Ḍ {12}$ , $5^{\prime}$ -CA AGT GT TG TTC C ACT GCC $3^{\prime}$ and Jp12, $5^{\prime}$ -GT TACT AT GAG C $\mathrm{(T/C)}$ TAGTCC $3^{\prime}$ ) and framework determinants of the variable regions $(\mathrm{V}\gamma101$ , $5^{\prime}\cdot$ -CTCACACTC(C/T)CACTTC $3^{\prime}$ and $\mathrm{V}\gamma_{}{11}$ , $5^{\prime}$ -TCTGG(G/A)GTC TAT TACT GT GC-3¢) of the c chain were used to PCR amplify genomic DNA of each sample as previously described [35]. PCR products were separated and analyzed by $16%$ polya cry l amide gel electrophoresis and visualized by staining with ethidium bromide.
为评估TCRG基因的克隆性，我们使用针对连接区保守序列的共有引物$(\mathbf Ḋ \gamma Ḍ {12}$、$5^{\prime}$-CA AGT GT TG TTC C ACT GCC $3^{\prime}$与Jp12、$5^{\prime}$-GT TACT AT GAG C $\mathrm{(T/C)}$ TAGTCC $3^{\prime}$)以及可变区框架决定簇$(\mathrm{V}\gamma101$、$5^{\prime}\cdot$-CTCACACTC(C/T)CACTTC $3^{\prime}$和$\mathrm{V}\gamma_{}{11}$、$5^{\prime}$-TCTGG(G/A)GTC TAT TACT GT GC-3¢)，按照文献[35]所述方法对每个样本的基因组DNA进行PCR扩增。PCR产物通过16%聚丙烯酰胺凝胶电泳分离分析，并经溴化乙锭染色显影。
Purification and separation of peripheral blood $\mathrm{CD}4+$ and $\mathrm{CD}8+$ cells
外周血 $\mathrm{CD}4+$ 和 $\mathrm{CD}8+$ 细胞的纯化与分离
$\mathrm{CD}4+$ and $\mathrm{CD}8+$ T-cells were positively selected from whole peripheral blood by CD4 and CD8 isolation kits using monoclonal antibodies coupled to magnetic beads (Dynal Biotech, Oslo, Norway), according to the manufacturer’s instructions. Separated $\mathrm{CD}4+$ and $\mathrm{CD}8+$ populations were tested by anti-CD4, anti-CD8, and anti-CD19 monoclonal antibodies (BD Bioscience s, San Jose, CA).
$\mathrm{CD}4+$ 和 $\mathrm{CD}8+$ T细胞通过CD4和CD8分离试剂盒(含偶联磁珠的单克隆抗体)从外周血中阳性分选(Dynal Biotech, 挪威奥斯陆)，操作遵循制造商说明。分选后的 $\mathrm{CD}4+$ 和 $\mathrm{CD}8+$ 细胞群采用抗CD4、抗CD8及抗CD19单克隆抗体(BD Bioscience s, 美国圣何塞)进行检测。
Quantitative PCR studies (general)
定量PCR研究（概述）
Reverse transcription was performed followed by quantitative PCR (RT-PCR and RQ-PCR) of total RNA samples from peripheral blood of HCL patients and age-matched control donors. Family-specific forward and reverse primers were used as well as SYBR Green I for reaching high sensitivity. Each family specific reaction was performed in triplicate. A melting curve analysis was done to discriminate between specific and non specific PCR products. Intensity values of TCR beta constant region PCR products from tenfold dilutions were used for a reference standard for relative quant if i cation of TRBV family expression.
对毛细胞白血病(HCL)患者和年龄匹配对照供体的外周血总RNA样本进行逆转录，随后进行定量PCR(RT-PCR和RQ-PCR)。采用家族特异性正向和反向引物及SYBR Green I染料以实现高灵敏度检测。每个家族特异性反应均设置三个复孔。通过熔解曲线分析区分特异性和非特异性PCR产物。以十倍稀释梯度的TCRβ恒定区PCR产物强度值作为参照标准，用于TRBV家族表达的相对定量分析。
Analysis of TCR beta chain variable gene usage
TCR beta链可变区基因使用分析
Quantitative PCR was performed in triplicates in $20~\mathrm{l}$ final volumes containing $10~\mathrm{l}$ QuantiTect SYBR Green PCR Kit (QIAGEN GmbH, Hilden, Germany), and $0.5~\mathrm{l}$ (20 pmole) of each primer in a DNA Engine Opticon 2 Continuous Fluorescence Detector (MJ research, MA). Cycling conditions were $15\mathrm{min}$ at followed by 40 cycles at $10\mathrm{s}$ at $94^{\circ}\mathrm{C}$ , $15~\mathrm{s}$ at $58^{\circ}\mathrm{C}$ , $30~\mathrm{s}$ at $72^{\circ}\mathrm{C}$ , signal reading at the end of this step, then final elongation for $10\mathrm{min}$ at $72^{\circ}\mathrm{C}$ . After the last cycle and final elongation, standard and melting curves were obtained. Amplification efficiencies of different TRBV families were normalized relative to TCR beta chain constant region amplification. For TRBV families [36], specific primers are listed in Table 1 and sequences are presented.
定量PCR在 $20~\mathrm{l}$ 终体积中进行，一式三份，包含 $10~\mathrm{l}$ QuantiTect SYBR Green PCR试剂盒 (QIAGEN GmbH, Hilden, Germany)、 $0.5~\mathrm{l}$ (20 pmole) 各引物，使用DNA Engine Opticon 2连续荧光检测仪 (MJ research, MA)。循环条件为 下 $15\mathrm{min}$ ，随后进行40个循环： $94^{\circ}\mathrm{C}$ 下 $10\mathrm{s}$ ， $58^{\circ}\mathrm{C}$ 下 $15~\mathrm{s}$ ， $72^{\circ}\mathrm{C}$ 下 $30~\mathrm{s}$ ，在此步骤结束时读取信号，最后在 $72^{\circ}\mathrm{C}$ 下延伸 $10\mathrm{min}$ 。最后一个循环和最终延伸后，获得标准曲线和熔解曲线。不同TRBV家族的扩增效率相对于TCR β链恒定区扩增进行归一化。TRBV家族 [36] 的特异性引物列于表 1 并呈现序列。
Normalization of TRBV specific cDNA concentration
TRBV特异性cDNA浓度的标准化
The TRBV chain-specific cDNA concentration was normalized relative to specific real-time PCR by amplifying part of TCR beta constant genes [37]. Briefly, a $5^{\prime}$ CB-con-F primer ( $5^{\prime}$ -GAG GGT CTC GGC CAC CTT- $^{3^{\prime}}$ and $5^{\prime}$ - GGC AGA CAG GAC CCC TTG- $3^{\prime}$ ) and a $3^{\prime}\mathbf{CB}$ -con-R primer ( $5^{\prime}$ -CC ACT GT GC ACC T CC TTC C-3¢), both specific for TRBC1 and TRBC2 genes, were used in realtime PCR with serial tenfold dilutions of cDNA. The TRBC PCR product intensities were then used by Opticon Monitor 201.10 quant it ation software (MJ Research Inc., Waltham, MA) as standards for the relative determination of TRBV chain-specific concentration.
TRBV链特异性cDNA浓度通过扩增部分TCR beta恒定基因[37]，采用实时定量PCR进行相对标准化。简言之，使用一对特异性针对TRBC1和TRBC2基因的引物：5′ CB-con-F引物(5′-GAG GGT CTC GGC CAC CTT-3′和5′-GGC AGA CAG GAC CCC TTG-3′)与3′ CB-con-R引物(5′-CC ACT GT GC ACC T CC TTC C-3¢)，在连续十倍稀释的cDNA样本中进行实时PCR。随后通过Opticon Monitor 201.10定量分析软件(MJ Research Inc., Waltham, MA)将TRBC PCR产物强度作为标准，用于相对测定TRBV链特异性浓度。
Table 1 Sequences for primers used in TRBV studies
表 1: TRBV研究中使用引物的序列
| TRBV家族 (IMGT分类) | 引物 | 序列 5′→ 3′ |
|---|---|---|
| TRBV2 | BV22 | GAGTTTCTGGTTTCCTTTTATAATAAT TAAACAGGACTCTAAGAAATTT |
| TRBV3-1 | BV9 | CTAAGAAGCCACTGGAGCTC |
| TRBV4-3/2 | BV7 | ATCAAAACGAGAGGACAGCA |
| TRBV5-1 TRBV6 | BV5 BV13a | CACTCAGACCCCAAAATT C |
| TRBV6-4 | BV13b | ATTCAAATACTGCAGGTACC |
| TRBV7 | BV6 | CTCAGG TGTGATCCAATT TC |
| TRBV9 | BV1{9} | CCGCACAACAGTTCCCTGACTTGC |
| TRBV10-3 | BV12a | CAGACTGAGAACCACCGC |
| TRBV10-2 | BV12b | GGAAGGCAGGTGACCTTG |
| TRBV11 | BV21 | AGTAGATGATTCACAGTTGCCTA |
| TRBV12 | BV8 | AGGTGACAGAGATGGGACAA |
| TRBV13 | BV23 | AGAAAAGAGGGAAACAGCC |
| TRBV14 | BV16 | GCAGAACTGGAGGATTCTGG |
| TRBV15 | BV24 | AGCTGCTGTTCCACTACTATGAC |
| TRBV16 | BV25 | TTGATGAAACAGGTA TGCCC |
| TRBV18 | BV18 | AAATTC ATG GTT TAT CTC CAGAAA |
| TRBV19 | BV17 | GAGATT CATCTACTACTCACAGAT |
| TRBV20-1 | BV2 | GCTGATGGCAACTTCCAATG |
| TRBV24-1 | BV15 | CGACAGGCACAGGCTAAA |
| TRBV25-1 | BV11 | ACCTCATCCACTATTCCTATGGA |
| TRBV27 | BV14 | GCAGATCTACTATTCAATGAA |
| TRBV28 | BV3 | TCTCATATGATGTTAAAATGAAAGAA |
| TRBV29-1 | BV4{29} | CGCCCAAACCTAACATTCTC |
| TRBV30 | BV20 | GCTCTTCTACTCCGTTGGT |
Analysis of CDR3 length spectra typing
CDR3长度谱分型分析
CDR3 length polymorphism analysis was performed as described [38]. Briefly, cDNA samples were amplified in 35 cycles of PCR using primers specific for the corresponding TRBV-families as $5^{\prime}$ primers and ${3}^{\prime}\mathbf{CB}.$ - con-R3-HEX as the $3^{\prime}$ primer. The products were purified by QIAquik PCR Purification Kit (QIAGEN, Valencia, CA) and checked on $2%$ agarose gels (Bio-Rad Laboratories, Hercules, CA) stained with ethidium bromide. $10~\mathrm{l}$ containing TRBV chain-specific PCR product and 15:1 deionized formamide (Applied Biosystem, Foster City, CA) GeneScan-500 ROX Size Standard (Applied Biosystem, Warrington, UK) mixture, was loaded onto $12%$ denaturing polyacrylamide sequencing gels and the electrophoresis was run in an ABI Prism 3100 sequencing machine. The data were collected by computer and analyzed by the GeneScan 3.1 software. A dominant CDR3 peak (or peaks) was defined as monoclonal (high-intensity signal) when there was a dramatic reduction of other CDR3 signals within that particular TRBV family, when the area under a single peak was $>50%$ of the total area under the curve area [38, 39].
CDR3长度多态性分析按照文献[38]所述方法进行。简言之，使用相应TRBV家族特异性引物作为5'端引物，3'端引物采用3'CB-con-R3-HEX，对cDNA样本进行35轮PCR扩增。产物经QIAquik PCR纯化试剂盒(QIAGEN, Valencia, CA)纯化后，通过溴化乙锭染色的2%琼脂糖凝胶(Bio-Rad Laboratories, Hercules, CA)进行检测。将10μl含TRBV链特异性PCR产物与15:1去离子甲酰胺(Applied Biosystem, Foster City, CA)、GeneScan-500 ROX分子量标准(Applied Biosystem, Warrington, UK)的混合液上样至12%变性聚丙烯酰胺测序胶，在ABI Prism 3100测序仪上电泳。数据经计算机采集后使用GeneScan 3.1软件分析。当特定TRBV家族内其他CDR3信号显著减弱，且单一峰面积占曲线下总面积的>50%时[38,39]，将该显性CDR3峰(或多个峰)定义为单克隆性(高强度信号)。
Statistical analysis
统计分析
Differences in continuously measured parameter values between two groups of patients were evaluated for statistical significance using the Wilcoxon rank sum test. Differences in binomial (dichotomous) parameters were compared using Fisher’s exact test. Correlations were performed using Spearman (non parametric) rank correlation analysis. Correlations such that $|r|>0.70$ were interpreted as being strong, those such that $0.50<|r|<0.70$ were moderately strong, $0.3<|r|<0.5$ were weak to moderately strong, and those with $|r|<0.30$ were considered weak. The $P$ -value associated with a correlation coefficient is for a test of $r=0$ , and thus is less important in interpretation than the actual magnitude of the coefficient.
两组患者间连续测量的参数值差异采用Wilcoxon秩和检验评估统计学显著性。二项式(二分)参数差异比较采用Fisher精确检验。相关性分析使用Spearman(非参数)秩相关分析法。当$|r|>0.70$时视为强相关，$0.50<|r|<0.70$为中等强度相关，$0.3<|r|<0.5$为弱至中等强度相关，$|r|<0.30$则判定为弱相关。相关系数对应的$P$值用于检验$r=0$的假设，因此在解释时其重要性低于系数的实际量级。
All $P$ -values are two-sided and have not been formally adjusted for multiple comparisons. In this exploratory analysis, when many related comparisons were made, in particular when individual evaluations are made for each TRBV family, only $P$ -values such that $P<0.005$ should be interpreted as being statistically significant, while those such that $0.005<P<0.05$ could be considered trends.
所有$P$值均为双侧检验，且未对多重比较进行正式调整。在此探索性分析中，当进行大量相关比较时（特别是针对每个TRBV家族单独评估时），仅$P<0.005$的$P$值应被视为具有统计学显著性，而$0.005<P<0.05$的数值可视为趋势。
Results
结果
To characterize the previous finding that HCL patients frequently have increased percentages of lymphocytes with CTL phenotype and appear oli go cl on al with respect to TCRG rearrangements [34], we evaluated HCL patients with and without CTL elevations for oli go cl on al restriction in both TCRG and TRBV domains. Blood from patients was used to produce both genomic DNA for TCRG PCR analysis, and cDNA from total RNA to evaluate TRBV usage. In addition to amplification of TRBV chains by PCR for quant it ation, fragments were subjected to CDR3 spectra typing to investigate clonality.
为了验证先前发现的HCL患者常伴有CTL表型淋巴细胞比例升高且TCRG重排呈现寡克隆性[34]，我们对比评估了伴有和不伴CTL升高的HCL患者在TCRG和TRBV区域的寡克隆限制性。采用患者血液样本分别制备基因组DNA（用于TCRG PCR分析）和总RNA逆转录的cDNA（用于TRBV使用评估）。除通过PCR扩增TRBV链进行定量外，还对片段进行CDR3谱型分析以研究克隆性。
Patient characteristics
患者特征
To examine a wide range of variations in T-cell repertoires, we purposefully used patients with a wide variety of ages, disease burden, type of treatment last used, and time since last treatment. Peripheral blood specimens from 22 patients with a confirmed diagnosis of HCL were examined. As shown in Table 2, the age was 35–83 (median 56), and the male–female ratio was 4.5:1. There were 18–105 (median 51) months since the last purine analog treatment. All patients were treated with BL22 except for patient P00, and two patients were assessed prior to beginning BL22. The remaining 19 patients were tested 10–45 (median 23) months after beginning BL22 and had received 1–23 (median 4) cycles. White blood counts (WBC) were 1.6–61.2 (median 4.4) and circulating hairy cell (HCL) counts were 0–56.9 (median 0.002) cells/ $1\times10^{^{-9}}$ . Half of the patients had prior s ple nec to my. Serious infections were uncommon except for Aspergillus pneumonia in P47 just prior to beginning BL22. Patient P14 had P neum o cyst is carinii pneumonia and patient 17 had an unspecified pneumonia 33 and 41 months prior to assessment by PCR. Thus, the HCL patients were a diverse group expected to vary widely in T-cell repertoire.
为考察T细胞受体库的广泛变异情况，我们特意选取了年龄跨度大、疾病负担各异、末次治疗方案不同且距末次治疗时间不等的患者群体。研究检测了22例确诊为毛细胞白血病(HCL)患者的外周血样本。如表2所示，患者年龄范围为35-83岁(中位数56)，男女比例为4.5:1。末次嘌呤类似物治疗距今18-105个月(中位数51个月)。除患者P00外，所有患者均接受过BL22治疗，其中两名患者在开始BL22治疗前接受了评估。其余19例患者在开始BL22治疗后10-45个月(中位数23个月)接受检测，这些患者已完成1-23个疗程(中位数4个疗程)。白细胞计数(WBC)为1.6-61.2(中位数4.4)，循环毛细胞(HCL)计数为0-56.9(中位数0.002)个细胞/$1\times10^{^{-9}}$。半数患者曾接受脾切除术。除患者P47在开始BL22治疗前罹患曲霉肺炎外，严重感染病例较为罕见。患者P14曾患卡氏肺孢子虫肺炎，患者17在PCR评估前33和41个月患有不明类型肺炎。由此可见，这些HCL患者的T细胞受体库预计存在显著差异。
Table 2 HCL study patient characteristics
| Patient | Age | Sex | Months since last CdA/DCF | Months since beginning BL22 | Total BL22 cycles | Months since last BL22 | WBC cells/L x10-9 | HCL cells/L x10-9 | Spleen Present |
| P00 | 83 | M | 30 | Untreated | 0 | N/A | 1.60 | 0.003 | No |
| P09 | 46 | F | 52 | 43 | 4 | 41 | 4.51 | 0.000 | Yes |
| P14 | 51 | M | 52 | 45 | 20 | 1 | 6.39 | 1.297 | No |
| P17 | 66 | M | 105 | 44 | 3 | 41 | 4.61 | 0.000 | Yes |
| P18 | 74 | M | 50 | 37 | 12 | 6 | 61.20 | 56.871 | No |
| P20 | 53 | M | 60 | 35 | 2 | 34 | 3.80 | 0.002 | Yes |
| P25 | 63 | M | 53 | 36 | 7 | 0 | 1.90 | 0.012 | No |
| P27 | 47 | M | 68 | 35 | 5 | 25 | 4.17 | 0.001 | Yes |
| P28 | 61 | F | 83 | 37 | 2 | 36 | 6.28 | 0.000 | No |
| P29 | 40 | M | 37 | 33 | 23 | 0 | 9.53 | 0.320 | Yes |
| P30 | 56 | M | 65 | 20 | 16 | 6 | 5.60 | 0.000 | Yes |
| P33 | 50 | M | 60 | 23 | 5 | 20 | 5.57 | 0.000 | Yes |
| P36 | 48 | F | 28 | 23 | 3 | 21 | 2.21 | 0.086 | No |
| P38 | 51 | M | 68 | 21 | 3 | 19 | 3.35 | 0.134 | No |
| P40 | 35 | M | 35 | 21 | 1 | 21 | 5.09 | 0.000 | Yes |
| P41 | 72 | M | 23 | 13 | 9 | 1 | 4.14 | 0.000 | Yes |
| P43 | 65 | M | 23 | 10 | 4 | 6 | 4.30 | 0.000 | No |
| P44 | 69 | F | 18 | 11 | 8 | 1 | 7.10 | 0.453 | No |
| P45 | 59 45 | M | 47 | 10 | 3 | 7 | 3.92 | 0.000 | Yes |
| P46 | M | 62 | 10 | 3 | 7 | 24.60 | 22.671 | No | |
| P47 | 56 | M | 27 | Pre | 0 | N/A | 3.79 | 3.566 | No |
| P49 | 65 | M | 26 | Pre | 0 | N/A | 1.78 | 0.128 | Yes |
N/A, not applicable. Pre, prior to BL22 treatment HCL Hairy Cell Leukemia
表 2 HCL研究患者特征
| 患者 | 年龄 | 性别 | 末次CdA/DCF治疗(月) | BL22起始治疗(月) | BL22总周期数 | 末次BL22治疗(月) | WBC (×10^9/L) | HCL细胞 (×10^9/L) | 脾脏存在 |
|---|---|---|---|---|---|---|---|---|---|
| P00 | 83 | 男 | 30 | 未治疗 | 0 | N/A | 1.60 | 0.003 | 否 |
| P09 | 46 | 女 | 52 | 43 | 4 | 41 | 4.51 | 0.000 | 是 |
| P14 | 51 | 男 | 52 | 45 | 20 | 1 | 6.39 | 1.297 | 否 |
| P17 | 66 | 男 | 105 | 44 | 3 | 41 | 4.61 | 0.000 | 是 |
| P18 | 74 | 男 | 50 | 37 | 12 | 6 | 61.20 | 56.871 | 否 |
| P20 | 53 | 男 | 60 | 35 | 2 | 34 | 3.80 | 0.002 | 是 |
| P25 | 63 | 男 | 53 | 36 | 7 | 0 | 1.90 | 0.012 | 否 |
| P27 | 47 | 男 | 68 | 35 | 5 | 25 | 4.17 | 0.001 | 是 |
| P28 | 61 | 女 | 83 | 37 | 2 | 36 | 6.28 | 0.000 | 否 |
| P29 | 40 | 男 | 37 | 33 | 23 | 0 | 9.53 | 0.320 | 是 |
| P30 | 56 | 男 | 65 | 20 | 16 | 6 | 5.60 | 0.000 | 是 |
| P33 | 50 | 男 | 60 | 23 | 5 | 20 | 5.57 | 0.000 | 是 |
| P36 | 48 | 女 | 28 | 23 | 3 | 21 | 2.21 | 0.086 | 否 |
| P38 | 51 | 男 | 68 | 21 | 3 | 19 | 3.35 | 0.134 | 否 |
| P40 | 35 | 男 | 35 | 21 | 1 | 21 | 5.09 | 0.000 | 是 |
| P41 | 72 | 男 | 23 | 13 | 9 | 1 | 4.14 | 0.000 | 是 |
| P43 | 65 | 男 | 23 | 10 | 4 | 6 | 4.30 | 0.000 | 否 |
| P44 | 69 | 女 | 18 | 11 | 8 | 1 | 7.10 | 0.453 | 否 |
| P45 | 59 | 男 | 47 | 10 | 3 | 7 | 3.92 | 0.000 | 是 |
| P46 | 45 | 男 | 62 | 10 | 3 | 7 | 24.60 | 22.671 | 否 |
| P47 | 56 | 男 | 27 | 治疗前 | 0 | N/A | 3.79 | 3.566 | 否 |
| P49 | 65 | 男 | 26 | 治疗前 | 0 | N/A | 1.78 | 0.128 | 是 |
N/A表示不适用。Pre表示BL22治疗前。HCL指毛细胞白血病(Hairy Cell Leukemia)。
PBMC analysis by flow cytometry
流式细胞术分析外周血单个核细胞 (PBMC)
As previously reported [34], patients with HCL had an increase in percent lymphocytes phenotypic ally consistent with CTLs. As shown in Table 3, which lists $%$ lymphocytes relative to total T-cells, 50, 50, and $41%$ of patients had high $%$ $\mathrm{CD}56+/\mathrm{CD}3+$ , $\mathrm{CD}57+/\mathrm{CD}3+,$ and ${\mathrm{CD}}8+/{\mathrm{CD}}3+$ cells, respectively. Thus, 15 $(68%)$ of the patients had T-cells with abnormally high percentages of cells phenotypic ally consistent with CTLs.
此前报道[34]显示，HCL患者中具有CTLs表型的淋巴细胞比例有所增加。如表3所示（列出了淋巴细胞占总T细胞的百分比），分别有50%、50%和41%的患者表现出高比例的CD56+/CD3+、CD57+/CD3+和CD8+/CD3+细胞。因此，15例 (68%) 患者的T细胞中具有异常高比例且表型符合CTLs特征的细胞。
TCR gamma rearrangement studies
TCR gamma重排研究
To detect TCRG restriction, which occurs in the chromosomes of / T-cells, PCR was performed on genomic DNA prepared from PBMCs of HCL patients, and results are shown in Table 3. Examples of restricted and polyclonal gamma chain rearrangement cases are shown in Fig. 1. A restricted (oli go cl on al) TCRG pattern was seen in 17 of 22 patients $(77%)$ and the remaining five patients $(23%)$ had a polyclonal pattern. By an exactWilcoxon rank sum test, patients with TCRG restriction had higher percentages of $_{\mathrm{CD3}+\mathrm{/CD57+}}$ ${P=0.005)}$ , $\mathrm{CD}3+\mathrm{/CD}56+$ $\overset{\vartriangle}{\boldsymbol{P}}=0.048)$ ), and a trend toward higher $%$ ${\mathrm{CD}}3+/{\mathrm{CD}}8+$ $P=0.055)$ cells, but lower percentages of $\mathrm{CD}3+/\mathrm{CD}4+$ $\scriptstyle{P=0.004})$ cells. TCRG restriction was associated with a lower number of BL22 cycles ). Thus, TCRG restriction was not associated with heavier BL22 treatment, and was associated with higher percentages of cells phenotypic ally consistent with CTLs.
为检测TCRG限制性（见于γ/δ T细胞染色体），我们对HCL患者外周血单个核细胞(PBMCs)提取的基因组DNA进行PCR，结果如表3所示。图1展示了限制性与多克隆γ链重排的典型病例。22例患者中17例(77%)呈现限制性(寡克隆)TCRG模式，其余5例(23%)为多克隆模式。经Wilcoxon秩和检验精确分析，TCRG限制性患者具有更高比例的CD3+/CD57+ (P=0.005)、CD3+/CD56+ (P=0.048)细胞，且CD3+/CD8+细胞比例呈升高趋势(P=0.055)，但CD3+/CD4+细胞比例较低(P=0.004)。TCRG限制性与更少的BL22疗程相关(P=0.003)。因此，TCRG限制性与更重的BL22治疗无关，而与具有细胞毒性T细胞(CTL)表型的细胞比例升高相关。
Table 3 Patient results
| PT# | % of T-lymphocytes in blood | Number of TRBVfamily | PCR gamma Restrictrion | |||||
| CD56+/CD3+ | CD57+/CD3+ | CD8+/CD3+ | CD4+/CD3+ | Total CDR3 Peaks/Families | Mono/ oligoclonal | Polyclonal | ||
| BL45 | 8 | 2 | 20 | 72 | 3.21 | 15 | 0 | Yes |
| BL27 | 3 | 39 | 65 | 8.05 | 0 | 17 | No | |
| BL14 | 2 | 6 | 19 | 81 | 6.82 | 5 | 5 | No |
| BL25 | 2 | 7 | 40 | 60 | 7.59 | 3 | 7 | No |
| BL29 | 9 | 9 | 40 | 57 | 7.36 | 0 | 10 | No |
| BL20 | 1 | 11 | 52 | 51 | 5.86 | 9 | 4 | Yes |
| BL40 | 14 | 13 | 45 | 48 | 6.95 | 3 | 7 | Yes |
| BL38 | 7 | 16 | 38 | 60 | 6.41 | 9 | 5 | Yes |
| BL30 | 11 | 18 | 34 | 66 | 7.14 | 5 | 7 | No |
| BL33 | 25 | 24 | 46 | 44 | ND | ND | ND | Yes |
| BL46 | 9 | 27 | 36 | 27 | 7.23 | 9 | 7 | Yes |
| BL28 | 14 | 30 | 45 | 60 | 5.84 | 9 | 3 | Yes |
| BL09 | 11 | 31 | 54 | 48 | 6.37 | 3 | 1 | Yes |
| BL36 | 8 | 33 | 36 | 58 | 4.81 | 18 | 0 | Yes |
| BL17 | 14 | 34 | 52 | 45 | ND | ND | ND | Yes |
| BL47 | 7 | 40 | 51 | 44 | 6.19 | 7 | 2 | Yes |
| BL41 | 2 | 44 | 62 | 37 | 6.05 | 13 | 3 | Yes |
| BL00 | 19 | 47 | 57 | 43 | 8.00 | 4 | 9 | Yes |
| BL18 | 17 | 50 | 50 | 33 | 4.14 | 21 | 0 | Yes |
| BL43 | 38 | 57 | 18 | 54 | 6.20 | 7 | 1 | Yes |
| BL49 | 50 | 71 | 61 | 36 | 3.30 | 19 | 1 | Yes |
| BL44 | 23 | 76 | 57 | 46 | 3.77 | 21 | 0 | Yes |
$\mathbf{N}/\mathbf{A}$ , not applicable. Pre, prior to BL22 treatment. Normal ranges, determined by the NIH Clinical Center Flow Cytometry Lab, are $3.2{-}9.7%$ for $\mathrm{CD}56+/\mathrm{CD}3+$ , $4.9{-}28.3%$ for $\mathrm{CD}57+\mathrm{\Omega}$ $\mathrm{CD}3+$ , $37.\bar{9}-47.8%$ for ${\mathrm{CD8}}+/{\mathrm{CD3}}+$ , and variable for $\mathrm{CD}4+/\mathrm{CD}3+$ , expressed as $%$ of T $\left(\mathrm{CD}3+\right)$ lymphocytes. Patients are arranged in ascending order of $%$ ${\mathrm{CD}}57+/{\mathrm{CD}}3+$ . Total CDR3 peaks/family indicate the average number of peaks for each patient by CDR3 spectra typing, considering the 22 TRBV families evaluable
表 3 患者结果
| PT# | CD56+/CD3+ (%) | CD57+/CD3+ (%) | CD8+/CD3+ (%) | CD4+/CD3+ (%) | Total CDR3 Peaks/Families | Mono/oligoclonal | Polyclonal | PCR gamma Restriction |
|---|---|---|---|---|---|---|---|---|
| BL45 | 8 | 2 | 20 | 72 | 3.21 | 15 | 0 | Yes |
| BL27 | 3 | 39 | 65 | 8.05 | 0 | 17 | No | |
| BL14 | 2 | 6 | 19 | 81 | 6.82 | 5 | 5 | No |
| BL25 | 2 | 7 | 40 | 60 | 7.59 | 3 | 7 | No |
| BL29 | 9 | 9 | 40 | 57 | 7.36 | 0 | 10 | No |
| BL20 | 1 | 11 | 52 | 51 | 5.86 | 9 | 4 | Yes |
| BL40 | 14 | 13 | 45 | 48 | 6.95 | 3 | 7 | Yes |
| BL38 | 7 | 16 | 38 | 60 | 6.41 | 9 | 5 | Yes |
| BL30 | 11 | 18 | 34 | 66 | 7.14 | 5 | 7 | No |
| BL33 | 25 | 24 | 46 | 44 | ND | ND | ND | Yes |
| BL46 | 9 | 27 | 36 | 27 | 7.23 | 9 | 7 | Yes |
| BL28 | 14 | 30 | 45 | 60 | 5.84 | 9 | 3 | Yes |
| BL09 | 11 | 31 | 54 | 48 | 6.37 | 3 | 1 | Yes |
| BL36 | 8 | 33 | 36 | 58 | 4.81 | 18 | 0 | Yes |
| BL17 | 14 | 34 | 52 | 45 | ND | ND | ND | Yes |
| BL47 | 7 | 40 | 51 | 44 | 6.19 | 7 | 2 | Yes |
| BL41 | 2 | 44 | 62 | 37 | 6.05 | 13 | 3 | Yes |
| BL00 | 19 | 47 | 57 | 43 | 8.00 | 4 | 9 | Yes |
| BL18 | 17 | 50 | 50 | 33 | 4.14 | 21 | 0 | Yes |
| BL43 | 38 | 57 | 18 | 54 | 6.20 | 7 | 1 | Yes |
| BL49 | 50 | 71 | 61 | 36 | 3.30 | 19 | 1 | Yes |
| BL44 | 23 | 76 | 57 | 46 | 3.77 | 21 | 0 | Yes |
$\mathbf{N}/\mathbf{A}$，不适用。Pre，BL22治疗前。正常范围由NIH临床中心流式细胞实验室确定：$\mathrm{CD}56+/\mathrm{CD}3+$为$3.2{-}9.7%$，$\mathrm{CD}57+\mathrm{\Omega}$$\mathrm{CD}3+$为$4.9{-}28.3%$，${\mathrm{CD8}}+/{\mathrm{CD3}}+$为$37.\bar{9}-47.8%$，$\mathrm{CD}4+/\mathrm{CD}3+$因患者而异，以T淋巴细胞$\left(\mathrm{CD}3+\right)$的百分比表示。患者按${\mathrm{CD}}57+/{\mathrm{CD}}3+$百分比升序排列。Total CDR3 peaks/family表示通过CDR3谱分型评估22个可评估的TRBV家族后，每位患者的平均峰数。
Fig. 1 Results of gamma chain T-cell receptor $(T C R)$ rearrangement studies by polymerase chain reaction $(P C R)$ . a Example of polyclonal rearrangement, P14. b (P41) and c (phase II patient) represent examples of clonal (restricted) rearrangements. There are two samples at two different time points 9 months apart, demonstrating the stability of the clonal bands. The asterisks represent 118 and 72 bp. The 72 bp marker was used to align the gels
图 1: 聚合酶链反应 (PCR) 检测γ链T细胞受体 (TCR) 重排结果。a 多克隆重排示例 (P14)。b (P41) 和 c (II期患者) 代表克隆性 (限制性) 重排示例。两个相隔9个月的不同时间点样本显示克隆条带的稳定性。星号标记118 bp和72 bp位置，72 bp标记物用于凝胶对齐。
TRBV expression
TRBV 表达
Since the vast majority of T-cells in our patients, as in most T-cells in the normal population, expressed $\alpha/\beta$ rather than $\gamma/\delta$ TCRs, we analyzed TRBV gene usage. Relative usage of different TRBV genes in total T-cell populations for $22\mathrm{HCL}$ patients and eight control donors is shown in Fig. 2a. Compared to normal controls, HCL patients had higher usage of TRBV14, TRBV20-1, and TRBV29-1 and lower usage of TRBV10-3, TRBV19, and TRBV28 (all $P<0.01$ ). To determine whether TCRG restriction is associated with over- or under expression in certain TRBV families, TRBV expression was compared for the 17 patients who had TCRG restriction and the five with TCRG poly cl on ali ty (Fig. 2b). Observed trends included higher TRBV18 and
由于我们患者体内绝大多数T细胞（与正常人群中的大多数T细胞一样）表达的是$\alpha/\beta$而非$\gamma/\delta$ TCRs，因此我们分析了TRBV基因的使用情况。图2a展示了22例HCL患者和8例对照供者总T细胞群体中不同TRBV基因的相对使用情况。与正常对照相比，HCL患者的TRBV14、TRBV20-1和TRBV29-1使用率较高，而TRBV10-3、TRBV19和TRBV28使用率较低（均$P<0.01$）。为了确定TCRG限制是否与特定TRBV家族的表达过高或过低相关，我们比较了17例存在TCRG限制的患者与5例存在TCRG多克隆性患者的TRBV表达情况（图2b）。观察到的趋势包括TRBV18和
Fig. 2 T-cell receptor beta variable $(T R B V)$ gene family usage. a Comparison between mean values $\left(\pm\mathrm{SD}\right)$ of all investigated hairy cell leukemia $\left(H C L\right)$ patients and control donors. b Comparison between TRBV usage by HCL patients with restricted $(n=17$ ) and polyclonal $(n=5)$ T-cell receptor gamma $(T C R G)$ . $^*$ , $P<0.05$ , identifying trends or better
图 2: T细胞受体β可变 $(TRBV)$ 基因家族使用情况。a 所有研究的毛细胞白血病 $(HCL)$ 患者与对照供体均值 $\left(\pm\mathrm{SD}\right)$ 比较。b 限制性 $(n=17)$ 与多克隆性 $(n=5)$ T细胞受体γ $(TCRG)$ 的HCL患者TRBV使用情况对比。$^*$, $P<0.05$, 表示趋势或更优
TRBV29-1 and lower TRBV11 with restriction (all $P{=}0.030)$ , but these differences were only trends in the context of an analysis of 21 TRBV family subsets.
TRBV29-1 及限制性较低的 TRBV11 (所有 $P{=}0.030$ )，但这些差异在分析 21 个 TRBV 家族子集时仅呈现趋势性。
Effect of BL22 on T-cells and TRBV gene usage
BL22对T细胞及TRBV基因使用的影响
To determine whether the anti-CD22 recombinant immunotoxin BL22 would decrease circulating T-cells, as was reported for purine analogs [13, 17, 18], the 21 patients treated with BL22 were assessed by flow cytometry before and at 1, 2, 4, 6, 12, 18, and 24 months after beginning BL22. As shown in Fig. 3, median T-cell counts decreased only slightly at 1 month and after 2 months were higher than baseline. A pairwise comparison of individual values for the patients showed no significant difference from baseline for any of the time points (all $P>0.05$ by Wilcoxon signed rank test). Thus, BL22 showed no evidence of T-cell toxicity which is common with purine analogs, which was expected based on its selective targeting of B-cells.
为评估抗CD22重组免疫毒素BL22是否会像嘌呤类似物那样降低循环T细胞水平[13, 17, 18]，研究人员对21名接受BL22治疗的患者进行了流式细胞术检测，时间点为治疗前及开始用药后1、2、4、6、12、18和24个月。如图3所示，中位T细胞计数仅在1个月时轻微下降，2个月后即高于基线水平。通过Wilcoxon符号秩检验对患者个体数值进行配对比较，所有时间点与基线均无显著差异（所有$P>0.05$）。这表明BL22未出现嘌呤类似物常见的T细胞毒性，这与其选择性靶向B细胞的特性相符。
TRBV CDR3 spectra typing in total T-cell population
TRBV CDR3 谱型分析在总 T 细胞群体中的应用
To determine whether expression in each TRBV family was polyclonal or mono/oli go cl on al, the CDR3 length distribution pattern was determined. The tests were performed on the same T-cell samples as in the TRBV gene usage analyses. As shown in Fig. 4a, the CDR3 distribution patterns, normally Gaussian for families with polyclonal T-cell expression, were compared for HCL patients with or without TCRG restriction, and also with normal age-matched controls. Data were classified for each family of each patient with respect to whether the peaks were monoclonal/oli go cl on al, polyclonal, or mixed/indeterminate. The number of families with mono/oli go cl on al expression varied from 0 in patients 27 and 29 to 21 in patients 18 and 44 (Table 3). TCRG restriction was associated with evidence of TRBV restriction, including fewer total TRBV peaks/ family ${\it P}{=}0.005)$ , higher number of mono/oli go cl on al families $\scriptstyle{P=0.004})$ , and lower number of polyclonal families (0.005).
为确定各TRBV家族中的表达是多克隆性还是单/寡克隆性，需分析CDR3长度分布模式。检测所用T细胞样本与TRBV基因使用分析相同。如图4a所示，比较了伴有或不伴TCRG限制的HCL患者及年龄匹配正常对照组的CDR3分布模式（多克隆T细胞表达家族通常呈高斯分布）。根据峰型将每个患者各家族数据分为单/寡克隆性、多克隆性或混合/不确定三类。具有单/寡克隆性表达的家族数量从患者27和29的0个到患者18和44的21个不等（表3）。TCRG限制与TRBV限制证据相关，包括：总TRBV峰/家族数更少（${\it P}{=}0.005)$、单/寡克隆性家族数量更多（$\scriptstyle{P=0.004})$，以及多克隆性家族数量更少（0.005）。
Fig. 3 Changes in T-cells after treatment with BL22. Median T-cell concentrations $\left(\mathrm{cells}/\mathrm{mm}^{3}\right.$ ) quantified by flow cytometry in the peripheral blood. Mean $\pm$ SD (range) values were $862\pm1{,}236$ $(54-$ 5,328) prior to BL22 $(n=21)$ ), and $287\pm432$ (120–1,897) at $1\mathrm{month}\pm10$ days $(n=20)$ , $259\pm306$ (119–1,064) at 2 month $\mathrm{~s~\pm~}15$ days $(n=20)$ ), $509\pm615$ (19–2,125) at 4 months $\pm\nobreakspace20\nobreakspace$ days $(n=16)$ , $371\pm493$ (116–2,062) at 6 months $\pm\nobreakspace30\nobreakspace$ days $(n=19)$ ), $644\pm1\small{,}040$ (140–4,847) at 1 year $\pm~60$ days $(n=20)$ , $458\pm714$ (265–3,379) at 1.5 years $\pm~75$ days $(n=18)$ , and $388\pm505$ $(29-$ 1,882) at 2 years $\pm~90$ days $(n=15)$ ) after beginning BL22. The horizontal dashed line indicates the lower limits of normal $(650~\mathrm{cells}/\mathrm{mm}^{3})$
图 3: BL22治疗后T细胞的变化。通过流式细胞术定量检测外周血中T细胞浓度的中位数 (cells/mm³)。平均值±标准差(范围)分别为：BL22治疗前 $862\pm1{,}236$ (54-5,328)(n=21)；治疗后1个月±10天 $287\pm432$ (120–1,897)(n=20)；2个月±15天 $259\pm306$ (119–1,064)(n=20)；4个月±20天 $509\pm615$ (19–2,125)(n=16)；6个月±30天 $371\pm493$ (116–2,062)(n=19)；1年±60天 $644\pm1\small{,}040$ (140–4,847)(n=20)；1.5年±75天 $458\pm714$ (265–3,379)(n=18)；2年±90天 $388\pm505$ (29–1,882)(n=15)。水平虚线表示正常值下限 (650 cells/mm³)。
Relationship of T-cell restriction to clinical factors
T细胞限制性与临床因素的关系
Next, the association of TRBV restriction to clinical factors was investigated. By Spearman correlation analysis, it was found that age correlated positively with the number of mono/oli go cl on al TRBV families $(r=0.51$ , $P=0.022)$ . Time since last purine analog therapy correlated with number of polyclonal families $\scriptstyle(r=0.46$ , $P=0.040$ ). In contrast, time since last course of BL22 (all $|r|<0.15$ , $P>0.65)$ and number of cycles of BL22 (all $|r|<0.25$ , $P>0.35)$ ) were not correlated to TRBV PCR results. Thus, purine analog therapy and age but not BL22 therapy were associated with TRBV restriction.
接下来，研究了TRBV限制与临床因素的相关性。通过Spearman相关性分析发现，年龄与单/寡克隆TRBV家族数量呈正相关 $(r=0.51$ , $P=0.022)$ 。末次嘌呤类似物治疗时间与多克隆家族数量相关 $\scriptstyle(r=0.46$ , $P=0.040$ )。相比之下，末次BL22疗程时间（所有 $|r|<0.15$ , $P>0.65)$ 和BL22周期数（所有 $|r|<0.25$ , $P>0.35)$ ）与TRBV PCR结果无相关性。因此，嘌呤类似物治疗和年龄（而非BL22治疗）与TRBV限制相关。
TRBV CDR3 spectra typing in sorted T-cell populations
TRBV CDR3 谱型分析在分选T细胞群体中的应用
To further explore whether the mono/oli go cl on al CDR3 patterns shown in Fig. 4a were originating from cells consistent with CTLs or from a different type of T-cell, PBMCs were sorted by magnetic beads into $\mathrm{CD}4+$ and $\mathrm{CD}8+$ populations before producing material for PCR. As shown in Fig. 4b, in each of three patients tested, peaks from $\mathrm{CD}8+$ T-cells had much more restriction than did peaks from $\mathrm{CD}4+$ T-cells. Since other markers of CTLs (CD56 and CD57) were present on $\mathrm{CD}8+$ cells but not on $\mathrm{CD}4+$ cells in these patients (data not shown), these data confirm that CTLs were the source for T-cell restriction in these patients.
为了进一步探究图 4a 中展示的单/寡克隆 CDR3 模式是源自细胞毒性 T 细胞 (CTL) 还是其他类型的 T 细胞，我们在进行 PCR 前通过磁珠将外周血单个核细胞 (PBMC) 分选为 $\mathrm{CD}4+$ 和 $\mathrm{CD}8+$ 细胞群。如图 4b 所示，在测试的三名患者中，$\mathrm{CD}8+$ T 细胞峰值的限制性远高于 $\mathrm{CD}4+$ T 细胞。由于这些患者的 $\mathrm{CD}8+$ 细胞上存在 CTL 的其他标志物 (CD56 和 CD57) 而 $\mathrm{CD}4+$ 细胞上不存在 (数据未展示)，这些数据证实了 CTL 是这些患者中 T 细胞限制性的来源。
Relationships between restriction and flow cytometry data
限制条件与流式细胞术数据的关系
As shown in Table 3 and Fig. 5, there was a moderately strong correlation between percent $\left(\mathrm{{^{0}/0}}\right)\mathrm{{CD}}3+\mathrm{{/CD}}57+$ and the number of mono/oli go cl on al TRBV families per patient (Fig. 5a, $r=0.53$ , $P=0.016$ ). There was a moderate inverse correlation between percent $(%)$ $\mathrm{CD}3+/$ $\mathrm{CD}57+$ and the number of polyclonal families (Fig. 5b, $r=-0.51$ , $P{=}0.023.$ ). The average number of peaks per family, which has been used previously as an index of mono cl on ali ty [40], varied from 3.21 to 8.05 (Table 3), correlating inversely with percent $(%)$ $_{\mathrm{CD3}+\mathrm{/CD57+}}$ (Fig. 5c, $r=-0.42$ , $P{=}0.067 $ ). Overall, the HCL patients evaluated for TRBV CDR3 $(n=20)$ ) had much lower total expression from polyclonal TRBV families compared to controls $(n=8)$ (Fig. 5d, $P<0.0001$ ). No moderately strong correlations were observed between mono/oli go cl on ali ty and CD56, CD8, or CD4 (all $|r|<0.35$ , $P>0.10,$ ).
如表 3 和图 5 所示，$\left(\mathrm{{^{0}/0}}\right)\mathrm{{CD}}3+\mathrm{{/CD}}57+$ 百分比与每位患者的单/寡克隆 TRBV 家族数量呈中等强度相关性 (图 5a，$r=0.53$，$P=0.016$)。 百分比与多克隆家族数量呈中等负相关 (图 5b，$r=-0.51$，$P=0.023$)。每个家族的平均峰数 (既往作为单克隆性指标 [40]) 介于 3.21 至 8.05 之间 (表 3)，与 百分比呈负相关 (图 5c，$r=-0.42$，$P=0.067$)。总体而言，接受 TRBV CDR3 评估的 HCL 患者 $(n=20)$ 其多克隆 TRBV 家族总表达量显著低于对照组 $(n=8)$ (图 5d，$P<0.0001$)。单/寡克隆性与 CD56、CD8 或 CD4 之间未观察到中等强度相关性 (所有 $|r|<0.35$，$P>0.10$)。
Fig. 4 CDR3 length polymorphisms. a Patients with restricted and polyclonal T-cell receptor gamma (TCRG), and control donors. b CDR3 spectra typing was repeated from T-cells sorted either by CD4 or CD8 positive selection
图 4: CDR3长度多态性分析。(a) 显示T细胞受体γ链(TCRG)呈限制性及多克隆分布的患者与健康对照。(b) 通过CD4或CD8阳性分选获得的T细胞进行CDR3谱型重复检测
Fig. 5 Correlations and comparisons using percents of T-cells positive for CD57. The number of mono/oli go cl on al (filled circle) and polyclonal (open circle) TRBV families/ patient are shown in (a) and (b). c Average number of peaks per TRBV family (TRBV complexity score) (filled triangle). Total TRBV expression (open triangle) in polyclonal CDR3 families is shown in (d), contrasting patients $(n=20$ ) with normal (controls $n{=}8$ ), $P<0.0001$ for the difference. Coefficients and $P$ -values in the text are from Spearman correlations, while the lines in the figures show simple linear regression of the actual data
图 5: 使用CD57阳性T细胞百分比进行的相关性分析与比较。(a) 和 (b) 分别展示了每位患者中单/寡克隆 (实心圆) 和多克隆 (空心圆) TRBV家族的数量。(c) 每个TRBV家族的平均峰数 (TRBV复杂度评分) (实心三角)。(d) 显示了多克隆CDR3家族中总TRBV表达量 (空心三角)，对比患者组 $(n=20$ ) 与正常对照组 $(n{=}8$ )，差异显著性 $P<0.0001$。文中相关系数与 $P$ 值来自Spearman相关性分析，图中线条表示实际数据的简单线性回归。
Discussion
讨论
Previous TRBV studies in two to seven HCL patients reported TRBV skewing [29–31, 41]. In the present manuscript, we studied enough patients to investigate quantitative relationships between flow cytometry data and restriction in the TRBV repertoire. We found that the percent of T-cells positive for CD57 was moderately strongly correlated with both TCRG and TRBV restriction. Sorting experiments confirmed that the restriction was from the $\mathrm{CD}8+$ CTLs and not from $\mathrm{CD}4+$ T-cells. Our data is able to show for the first time that patients with increased CTLs had loss of polyclonality, worse with more recent purine analog therapy, but not BL22 immuno toxin therapy.
先前在2至7例毛细胞白血病(HCL)患者中开展的TRBV研究报告了TRBV偏斜现象[29–31, 41]。本研究通过足够样本量，首次揭示了流式细胞术数据与TRBV谱系限制性之间的量化关系：CD57阳性T细胞比例与TCRG及TRBV限制性均呈中度强相关。分选实验证实这种限制性源自$\mathrm{CD}8+$细胞毒性T淋巴细胞(CTL)，而非$\mathrm{CD}4+$T细胞。我们首次证实CTL增多的患者会出现多克隆性丢失，且嘌呤类似物治疗近期会加剧该现象，但BL22免疫毒素治疗无此效应。
We found that ${\mathrm{CD3}+\mathrm{/CD57+}}$ T-lymphocyte percentage correlated with more restricted and fewer polyclonal TRBV families, and that fewer polyclonal TRBV families correlated with recent purine analog but not with BL22 therapy. Correlations were not as strong for the percent of T-cells positive for CD56, CD8, or CD4 and restriction. Although clonally skewed T-cells in HCL have been observed in some cases to be $\mathrm{CD}4+$ [28, 29], CTLs are generally $\mathrm{CD}8+/\mathrm{CD}4-$ . The lack of correlation with CD56 is probably because a lower percentage of CTLs are $\mathrm{CD}56+$ than $\mathrm{CD}57+$ . Finally, CD8 is a marker expressed by many types of T-cells not classified as CTLs [42]. We used the $%$ of T-cells which were $_{\mathrm{CD3}+\mathrm{/CD57+}}$ rather than the $%$ of total lymphocytes so the results would not be distorted by patients with high numbers of circulating malignant lymphocytes.
我们发现${\mathrm{CD3}+\mathrm{/CD57+}}$ T淋巴细胞百分比与更多限制性和更少多克隆TRBV家族相关，且较少多克隆TRBV家族与近期嘌呤类似物治疗相关，但与BL22疗法无关。CD56、CD8或CD4阳性T细胞百分比与限制性的相关性较弱。尽管在某些病例中观察到HCL患者的克隆性偏移T细胞为$\mathrm{CD}4+$[28, 29]，但CTLs通常为$\mathrm{CD}8+/\mathrm{CD}4-$。与CD56缺乏相关性可能是因为$\mathrm{CD}56+$的CTLs比例低于$\mathrm{CD}57+$。最后，CD8是许多非CTLs类T细胞表达的标记物[42]。我们使用$_{\mathrm{CD3}+\mathrm{/CD57+}}$ T细胞百分比而非总淋巴细胞百分比，以避免循环恶性淋巴细胞数量高的患者对结果造成干扰。
Monoclonal and oli go cl on al expansion have been reported for T-cells in patients with other malignancies [43–45]. In CLL, myeloma, and chronic my e loge nous leukemia, oli go cl on ally expanded CTLs have been reported to be directed against autologous tumor antigens [43, 46, 47]. After allogeneic stem cell transplant, a low number of TRBV CDR3 peaks/family (restriction) correlated with the development of chronic graft vs host disease [40]. We originally speculated that elevations of CTLs in HCL may reflect an immune response to the tumor, since in several cases CTLs increased during follow-up [34]. The more recent study [31] showed that the CTLs in two HCL patients were not auto reactive. Our data show a substantial decrease in poly cl on ali ty with TRBV restriction, which suggests such patients may have more limited cell-mediated immunity.
单克隆和寡克隆扩增已在其他恶性肿瘤患者的T细胞中被报道[43–45]。在慢性淋巴细胞白血病(CLL)、骨髓瘤和慢性髓性白血病中，寡克隆扩增的细胞毒性T淋巴细胞(CTL)被证实针对自体肿瘤抗原[43,46,47]。异基因造血干细胞移植后，较少的TRBV CDR3峰/家族(限制性)与慢性移植物抗宿主病的发生相关[40]。我们最初推测毛细胞白血病(HCL)中CTL的升高可能反映了对肿瘤的免疫应答，因为在多个病例中CTL在随访期间有所增加[34]。但最新研究[31]显示，两名HCL患者的CTL并无自身反应性。我们的数据显示TRBV限制性导致的多克隆性显著降低，这表明此类患者的细胞介导免疫可能更为受限。
Either cladribine or pen to statin can induce profound T-cell toxicity with reductions in $\mathrm{CD}4+$ lymphocytes until a median of 4 years [13, 17, 18]. Reductions in $\mathrm{CD}8+$ lymphocytes are significant but less prolonged. Median nadirs after cladribine of $139/\mathrm{l}$ and $92/\mathrm{l}$ for CD4 and $\mathrm{CD}8+$ lymphocytes have been reported [17]. It is likely that purine analogs, particularly with repeated courses, may eliminate normal T-cell clones and exacerbate restriction in the T-cell repertoire. In the future, it may be advantageous to use biologic therapies like BL22 as front line therapy for HCL due to their lack of T-cell toxicity. However, at the current time, BL22 is being tested only in patients who are resistant to purine analogs. Currently, there are no firm guidelines for when to retreat HCL patients with purine analogs rather than T-cell sparing biologic agents. It is possible that mono/oli go cl on al skewing of the T-cell repertoire, as easily measured by $_{\mathrm{CD3}+\mathrm{/CD57+}}$ percentage, could be useful to assess immune damage due to prior purine analog therapy. Relationships of these measures to infectious morbidity and mortality will require analysis of several-fold higher numbers of patients. Over the next several years, we anticipate that a high percentage of HCL patients pretreated with purine analogs will receive anti-CD22 immuno toxin and/or ritixumab. To determine when these T-cell sparing agents should be used instead of repeat courses of purine analog for relapsed HCL, it may be useful to follow patients carefully for markers of immune function, and our data suggest that the percent of $\mathrm{CD}57+/$ $\mathrm{CD}3+$ T-cells could be a useful marker.
克拉屈滨或喷司他丁均可引发显著的T细胞毒性，导致CD4+淋巴细胞减少，中位持续时间可达4年[13,17,18]。CD8+淋巴细胞减少程度虽显著但持续时间较短。文献报道克拉屈滨治疗后CD4和CD8+淋巴细胞的中位最低值分别为139/μl和92/μl[17]。嘌呤类似物（尤其是重复疗程）可能会清除正常T细胞克隆，加剧T细胞库的限制性。未来由于BL22等生物制剂不具有T细胞毒性，或可作为毛细胞白血病(HCL)的一线治疗方案。但目前BL22仅用于嘌呤类似物耐药患者的试验性治疗。当前尚无明确指南规定何时应对HCL患者再次使用嘌呤类似物而非T细胞保护性生物制剂。通过CD3+/CD57+百分比可简便检测T细胞库的单/寡克隆偏移，这可能有助于评估既往嘌呤类似物治疗造成的免疫损伤。这些指标与感染发病率及死亡率的相关性需通过更大样本量分析验证。未来数年，预计多数经嘌呤类似物预处理的HCL患者将接受抗CD22免疫毒素和/或利妥昔单抗治疗。为确定复发HCL患者何时应改用这些T细胞保护剂而非重复嘌呤类似物疗程，密切监测免疫功能标志物具有临床价值。我们的数据表明CD57+/CD3+T细胞百分比可能成为有效标志物。
