【作者】 林江；

【导师】 许文荣；

【作者基本信息】 江苏大学 ， 临床检验诊断学， 2012， 博士

【摘要】 1.研究目的(1)应用高分辨熔解曲线分析(HRMA)技术建立髓系肿瘤中相关基因(C-KIT、IDH、IDH2、DNMT3A、U2AF1和JAK2)突变的分子诊断平台。(2)探讨这些基因突变的临床意义。2.研究方法(1)临床病例收集、细胞分离及DNA提取：收集按FAB和WHO标准诊断的初诊急性髓系白血病(AML)221例、骨髓增生异常综合征(MDS)88例和骨髓增殖性肿瘤(MPN)142例骨髓标本,用人淋巴细胞分离液(Ficoll-Hypaque)分离骨髓单个核细胞,DNA抽提试剂盒提取基因组DNA。(2)PCR扩增：针对不同基因(包括C-KIT、IDH1、IDH2、DNMT3A、U2AF1和JAK2)的基因突变位点建立相应的PCR扩增反应体系,在ABI7300 PCR扩增仪上分别对这些基因进行扩增。(3) HRMA检测：将PCR扩增产物在LightScanner分析仪上进行荧光扫描,应用分析软件进行处理,分析HRMA检测不同基因突变的灵敏度,并进一步检测待测标本的基因突变情况。(4)DNA测序：首先评估DNA测序检测不同基因突变的灵敏度,然后对HRMA扫描鉴定为突变的标本进行测序。(5)突变的临床意义分析：用Spearman秩和相关检验进行突变与性别、年龄、血液学参数、染色体分组、FAB分型、WHO分型等的相关性分析；应用卡方检验或Fisher确切检验对突变阳性组和阴性组之间的分类变量(如性别、染色体分组、FAB分型和WHO分型等)进行差异性比较,应用Mann-WhitneyU检验对突变阳性组和阴性组之间的连续变量(如年龄和血液学参数等)进行差异性比较。应用Kaplan-Meier和Cox分析分别对生存时间进行单因素和多因素分析。3.研究结果(1) C-KIT突变：①HRMA检测C-KIT突变的灵敏度达5%,高于DNA测序的灵敏度(10%)；②对160例AML患者进行检测发现9例(6%)存在C-KIT基因突变,进一步测序证实4例为D816V突变、4例为N822K突变、1例为D816H和N822K双位点突变；9例C-KIT突变主要发生于FAB-M2亚型,其中7例患者伴有t(8;21)易位,2例患者为正常核型。在25例t(8；21)易位AML患者中,C-KIT突变发生率为28%,突变者与非突变者比较,两组在性别、年龄、血象方面并无差异。(2) IDH1/IDH2突变：①HRMA检测IDH1 R132H、IDH2 R140Q和IDH2 R172K突变的灵敏度分别为5%、2%和5%,均高于DNA测序的灵敏度(10%)；②对198例AML患者的进一步检测发现IDH1和IDH2突变发生率分别为2.0%(4例)和5.0%(10例),均为杂合性突变、且均为单独发生。IDH1突变包括3例R132H和1例R132S突变,IDH2突变包括7例R140Q和3例R172K突变。伴IDH1/IDH2突变的AML患者年龄明显高于无突变者(P=0.002)。两组患者在性别、血细胞计数和FAB分型方面并无统计学差异(P>0.05)。染色体核型正常的AML患者中IDH1/IDH2突变率(12/85,14.1%)明显高于核型异常患者(2/99,2.0%)(P=0.004)。③82例MDS患者分别检出2例(2.4%)和3例(3.6%)存在IDH1(R132H和R132S各1例)和IDH2突变(均为R140Q)。IDH1和IDH2突变均为杂合性、且均为单独发生。伴或不伴IDH1/IDH2突变者在性别、年龄、血红蛋白含量和血小板计数方面并无显著性差异(P>0.05),但伴突变者的白细胞计数明显高于无突变者(P=0.010)虽然全部突变均见于核型良好组,但并无统计学意义(P=0.407)。正常核型患者IDH1/IDH2突变率(5/52,9.6%)高于异常核型患者(0/30,0%)(P=0.153)。此外,根据IPSS分组,所有IDH突变均发生于低危组或中危-1组患者。多变量Cox分析揭示年龄和IPSS分组是MDS患者独立的预后因素,而IDH1/IDH2突变对生存时间并无影响。④85例慢性髓系白血病(CML)和57例Ph-MPN患者中均未检测出IDH1/IDH2突变。(3) DNMT3A突变：①HRMA检测DNMT3A R882H突变的灵敏度为2%,高于DNA测序的灵敏度(10%)；②182例AML患者中12例(6.6%)检出DNMT3A突变,均为杂合性突变,包括8例R882H、3例R882C和1例R882P。R882突变与患者的性别和白细胞计数并无相关性,突变者年龄高于无突变者、且突变者血小板计数明显高于无突变者DNMT3A R882突变更常见于单核类AML (M4和M5,7/52例,13.4%)而非单核类AML (M1、M2、M3和M6,5/130,3.8%)(P=0.041)76例正常核型患者中10例(13.2%)存在R882突变,显著高于核型异常患者(2/92例,2.2%)(P=0.007)。③51例MDS患者检出4例(7.8%)DNMT3A R882突变,均为杂合性,包括3例R882H和1例R882C；伴或不伴DNMT3A突变者在年龄、血细胞计数、WHO分型等方面并无显著性差异；4例突变均发生于正常核型患者及IPSS中危组。伴突变者中位生存时间为9个月,而不伴突变者中位生存时间为25个月(P=0.047)。④79例CML和57例Ph-MPN患者中均未检测出DNMT3A突变。(4)U2AF1突变：①HRMA检测U2AF1 S34和Q157突变的灵敏度均为5%,高于DNA测序的灵敏度(10%)；②221例AML患者中4例(1.8%)检出U2AF1突变,均为杂合性突变,包括2例S34Y和2例Q157(1例Q157P和1例Q157R)突变。突变患者与非突变患者在性别、年龄、血细胞、FAB分型和核型分组之间并无差异。伴U2AF1突变的患者生存时间明显短于无突变者(中位时间分别为1和7个月,P=0.034)③88例MDS患者检出4例(4.5%)S34突变,均为杂合性,包括3例S34F和1例S34Y。伴或不伴U2AF1突变者在性别、血细胞计数、WHO分型和IPSS分组方面并无显著性差异。伴U2AF1突变者年龄较无突变者年轻(P=0.059)。72例患者具有随访结果,突变患者和无突变患者在生存时间上并无显著性差异(P=0.954)。④79例CML和57例Ph-MPN患者中均未检测出U2AF1突变。(5)JAK2突变：①HRMA检测JAK2 V617F突变的灵敏度为5%,高于DNA测序的灵敏度(10%)；②48例Ph-的MPN患者中,JAK2 V617F突变可见于3例(100%,均为杂合子突变)真性红细胞增多症(PV)34例(60%,6例纯合子突变、28例杂合子突变)原发性血小板增多症(ET)和3例(60%,1例纯合子突变、2例杂合子突变)原发性骨髓纤维化(PMF)。而57例CML患者均为阴性。HRMA的检测结果均经直接测序证实。诊断的特异性和灵敏度均为100%。4.结论(1) HRMA是检测基因突变的一个高通量技术,其特异性好、灵敏度高、成本低廉,可有效用于对大批量标本进行基因突变的筛选。(2)中国人群AML和MDS患者中存在IDH1和IDH2基因突变,均为杂合子突变,主要见于正常核型患者;IDH1/IDH2突变并非是MDS患者的不良预后因素。(3) DNMT3A R882突变是中国人群AML和MDS患者的一个常见分子事件,该突变均为杂合子突变,主要发生于正常核型患者；DNMT3AR882突变是MDS患者的一个不良预后因素。(4)中国人群AML和MDS中U2AF1突变发生于密码子S34和Q157,均为杂合子突变,但发生率相对较低；伴U2AF1突变的AML患者生存时间较短。更多还原

【Abstract】 OBJECTIVE:(1) Establish the molecular diagnosis platform of common gene mutations in myeloid malignancies using high-resolution melting analysis (HRMA).(2) Explore the clinical relevance of these gene muations.METHODES:(1) Collection of clinical cases, cell separation and DNA extraction: Bone marrows were collected from 221 cases with acute myeloid leukemia (AML),88 myelodysplastic syndrome (MDS) and 142 myeloproliferative neoplasm (MPN) which were diagnosed accoriding to FAB and WHO cirtieria. Bone marrow mononuclear cells were separated using Ficoll-Hypaque and genomic DNA was extracted using DNA Purification Kit.(2) PCR amplification:Different PCR systems were established to amplify different mutated genes (C-KIT, IDH1, IDH2, DNMT3A, U2AF1 and JAK2) according to the specific mutated sites of different genes in 7300 Thermo cycler.(3) HRMA detection:PCR products were transferred to the LightscannerTM platform for scanning and were analyzed using software package. The sensitivity of HRMA was evaluated. Gene mutations in the detected samples were further analyzed.(4) DNA sequencing:The sensitivity of direct DNA sequencing was evaluated for different gene mutations. The samples which were identified positive in clinical cases were further sequenced.(5) Analysis of clinical relevance of gene mutations:The correlation between gene mutations and sex, age, hematologic parameters, chromosomal groups, FAB subtypes and WHO subtypes was analyzed with Spearman’s rank correlation. Chi-square analysis or Fisher exact test was carried out to compare the difference of categorical variables between patient groups. Mann-Whitney’s U-test was carried out to compare the difference of continuous variables between patient groups. Survival was analyzed according to the Kaplan-Meier method (univariant analysis) and the Cox regression method (multivariate analysis).RESULTS:(1) C-KIT mutation:①HRMA could distinguish C-KIT mutation with the maximal sensitivity of 5% in a background of wild-type DNA, higher than that of direct DNA sequencing (10%).②In 160 AML patients, 9 (6%) cases were identified with C-KIT mutations, which were further confirmed by DNA sequencing, including 4 D816V mutations,4 N822K mutations, and 1 dual muations of D816H and N822K.8 mutations occurred in the patients with AML-M2 subtype, among which 7 cases had the translocation t(8;21); 2 had normal karyotype. Among 25 cases with t(8;21), the frequency of C-KIT mutation was 28%. There was no difference in sex, age and blood parameters between mutated and wild-type groups.(2) IDH1/IDH2 mutation:①HRMA could distinguish IDH1 R132H, IDH2 R140Q and IDH2 R172K mutations with the maximal sensitivity of 5%,2% and 2%, respectively, higher than that obtained by direct DNA sequencing (10%).②IDH1 mutations were found in 4 of 198 AML patients (2.0%) and IDH2 mutations were found in 10 cases (5.0%). All mutations were heterozygous. IDH1 and IDH2 mutations were mutually exclusive. IDH1 mutations included 3 R132H and 1 R132S. while IDH2 mutations were R140Q (n=7) and R172K (n=3). AML patients with IDH1/IDH2 mutations had significantly older age than those with wild-type IDHs (P=0.002). There were no difference in sex, blood parameters and FAB subtypes between cases with and without mutations (P>0.05). IDH mutations were significantly more frequently observed in cytogenetically normal AML (12/85,14.1%) than in cytogenetically abnormal AML (2/99, 2.0%) (P=0.004).③2 of 82 (2.4%) MDS cases were identified with IDH1 mutations (one R132H and one R132S) and 3 (3.6%) with IDH2 mutations (all R140Q). IDH1 and IDH2 mutations were also heterozygous and mutually exclusive. No significant defference in sex, age, hemoglobin and platelets was observed between MDS patients with and without IDH1/IDH2 mutations (P>0.05). MDS cases with IDH1/IDH2 mutations had higher white blood cell (WBC) counts than those without mutations (P=0.010). All IDH1/IDH2 mutations were observed in cases with favorable risks according to karyotype classification, however, statistical difference was not observed (P=0.407). The frequency of IDH1/IDH2 mutations were higher in MDS patients with normal karyotype (5/52,9.6%) than those with abnormal karyotype (0/30,0%) (P=0.153). In addition, all IDH mutations were identified in patients at Low and Int-1 risks according to IPSS classification. Multivariate analysis showed that age and IPSS grouping were independent prognostic factors. The overall survival (OS) of MDS patients with and without IDH1/IDH2 mutations did not differ.④No IDH1 or IDH2 mutations were detected in 85 chronic myeloid leukemia (CML) and 57 Ph-MPNs.(3) DNMT3A mutation:①HRMA could easily distinguish R882H mutation with the maximal sensitivity of 2%, higher than that obtained by DNA sequencing (10%).②R882 mutations were identified in 12 (6.6%) of 182 AML patients. All mutations were heterozygous, including R882H (n=8), R882C (n=3) and R882P (n=1). No correlation was observed between R882 mutations and gender or WBC counts. AML patients with DNMT3A R882 mutations were more prevalent at older age and present with significantly higher median platelet counts at diagnosis compared to those without mutations. DNMT3A R882 mutations were found more frequently among monoblastic leukemia (M4 and M5,7 of 52,13.4%) compared to non-monoblastic leukemia (M1, M2, M3 and M6,5 of 130, 3.8%) (P=0.041). Among the patients with normal karyotypes,13.2%(10 of 76) cases showed R882 mutation, significantly higher than 2.2%(2 of 92) in those with chromosomal abnormalities (P=0.007).③4 (7.8%) heterozygous DNMT3A R882 mutations were also identified in MDS, including 3 R882H and 1 R882C mutations. The difference in age, hematologic parameters and WHO classifications was not seen between patients with and without mutations. All 4 patients were identified with normal karyotype and were classified in intermediate-risk group according to IPSS classification. The OS of MDS patients with DNMT3A mutation (median 9 months) was shorter than those without mutation (median 25 months) (P=0.047).④No DNMT3A R882 mutations were detected in 79 CML and 57 Ph-MPNs.(4) U2AF1 mutation:①HRMA could easily distinguish U2AF1 S34 and Q157 mutations with the maximal sensitivity of 5%, higher than that obtained by DNA sequencing (10%).②U2AF1 mutations were identified in 4 (1.8%) of 221 AML patients. All mutations were heterozygous, including 2 S34Y and 2 Q157 (1 Q157P and 1 Q157R) mutations. There were no difference in sex, age, blood parameters, FAB subtypes, and karyotype classification between cases with and without mutations (P>0.05). The OS of AML patients with U2AF1 mutation (median 1 months) was shorter than those without mutation (median 7 months) (P=0.034).③4 (4.5%) cases were identified with heterozygous S34 mutations (3 S34F and 1 S34Y). No significance in sex, blood parameters, WHO subtypes, and IPSS classification was observed between MDS patients with and without U2AF1 mutations (P>0.05). The cases with U2AF1 mutations had younger age than those without mutations (.P=0.059). No difference in OS was observed between patients with and without U2AF1 mutations (P=0.954).④No U2AF1 mutations were detected in 79 CML and 57 Ph-MPNs.(5) JAK2 mutation:①HRMA assay was able to distinguish 5% of JAK2 V617F mutant, more sensitive than that obtained by DNA sequencing (10%).②Out of 48 cases with Ph-negative MPN, all 3 (100%) with polycythemia vera (PV) were positive for the presence of JAK V617F heterozygous mutation,34 (70%) with essential thrombocythemia (ET) were positive for JAK V617F mutation (6 homozygous and 28 heterozygous), and 3 (60%) with primary myelofibrosis (PMF) were positive (1 homozygous and 2 heterozygous). JAK2 V617F mutation was not found in all patients with CML. HRMA detection of all cases was fully concordant with the results of direct sequencing. Both the diagnostic specificity and sensitivity were 100%.CONCLUSIONS:(1) HRMA is a high-throughput technique for the detection of gene mutation and can be utilized to screen gene mutation in a large scale samples for good specificity, high sensitivity and low cost.(2) Heterozygous IDH1 and IDH2 mutations are present in Chinese AML and MDS patients and mainly observed in cytogenetically normal cases. IDH1/IDH2 mutations may have no prognostic impact on patients with MDS.(3) DNMT3A R882 mutations are recurrent molecular aberrations in AML and MDS. R882 mutations are all heterozygous and mainly observed in cytogenetically normal patients. DNMT3A R882 muation may be an adverse prognostic event in MDS.(4) U2AF1 mutations, a molecular event of low frequency in Chinese AML and MDS, occur at S34 and Q157 and are all heterozygous. The overall survival of AML patients with U2AF1 mutation was shorter than those without mutation.更多还原