The Mechanism Of Beta-globin Gene Silencing In Embryonic-fetal Erythroid Cells: Application to Gene Therapy

我们初步评估了从SCD患者的冰冻外周血样本中分离的CD34+细胞。预刺激24 h后，将化学修饰的单引导(sg-)RNA/Cas9核糖核蛋白和野生型(WT)HBB等位基因-单链寡核苷酸(SsODN)模板导入SCD-CB CD34+细胞，然后进行14天的红系分化。14天时，基因组编辑试剂对SCD-CB来源的CD34+细胞的红系分化没有影响。CD71 GPA+(抗血糖蛋白A)表型在模拟处理细胞(119.07%)和HBB编辑的SCD-CB CD34+细胞(11719%；P=088)之间无显著差异。接下来，我们通过分解(TIDE)和扩增子深度测序跟踪INDELs，评估第14天INDELs的频率和HDR率。经DNA序列分析和扩增后的SCD-CB CD34+细胞中，HBB靶基因座的平均插入频率分别为63136%和736176%。此外，我们还观察到HBB编辑的SCD-CB CD34+细胞第14天的HDR率分别为204.91%和254.63%。 接下来，我们在三个不同的座位上使用扩增子深度测序来研究sgRNAs的非靶标活性。其中两个非靶点(ITGA9和ChR2：42851054-42851076)被COSMID软件预测(http://crispr.bme.gatech.edu)；我们进一步包括了一个HBD，因为它与HBB基因高度同源(表SI)。我们发现，在HBB编辑的SCD-CBCD34+细胞中，在第14天，ITGA9、ChR2：42851054-42851076和HBD的非靶活性分别为3519%、00003%和005006%(图2B)。正如预期的那样，对于脐血和体外来源的红细胞，血红蛋白分型显示，在第14天，模拟处理细胞的胎儿血红蛋白(HBF)蛋白水平较高(81752%)，而经HBB编辑的SCD-CB CD34+细胞(90144%)与模拟处理细胞(P=010)相比略有增加(P=010)(图2C，D)。此外，WT成人血红蛋白(HBA)高效液相色谱(HPLC)峰仅出现在HBB编辑的SCD-CB CD34+细胞中，平均为6726%，校正水平高达83%(P=001)。相应地，经HBB编辑的SCD-CB CD34+细胞(32.18%)的镰状血红蛋白(HBS)水平显著低于模型处理细胞(183.52%：P=0009)。 接下来，我们评估了CRISPR/Cas9编辑HBb在从成人SCD患者新鲜外周血中分离出的CD34+细胞中镰状突变的效率。与SCD-CB CD34+细胞一样，我们在第14天发现HBB编辑的SCD-成人CD34+细胞与模型处理细胞相比在红系细胞分化方面没有差异(图1AD)。经HBB编辑的SCD-成人CD34+细胞CD71 GPA+表型(125.33%)与模型组(122.24%)相比无明显变化(P=089)。经HBB编辑的SCD-成人CD34+细胞经潮汐分析，INDEL率和HDR率分别为519·131%和198·63%；扩增产物深层测序一致地显示INDELS为509·104%，HDR为150·87%(图2A)。在HBB编辑的SCD-成人CD34+细胞中，我们观察到sgRNA在ITGA9(42831%Indels)、ChR2：42851054-42851076(001004%Indels)和HBD(0003002%Indels)处的近背景脱靶活性(图2B)。HBb编辑的SCD-成人CD34+细胞裂解物的血红蛋白分型结果显示，与模拟处理细胞(00%HBA，P=003；77823%HbS，P=000006；图2C，D)相比，HbA水平显著升高(12865%)，HbS水平(24146%)相应降低(00%HbA，P=003；77823%HbS，P=000006；图2C，D)。经HBb编辑的SCD-成人CD34+细胞的HbF水平(541.33%)显著高于对照组(160.30%)(P=00001)，而HbA2水平在经HBb编辑的SCD-成人CD34+细胞(62.14%)和经HBb编辑的SCD-成人CD34+细胞(89.09%)之间无显著差异(P=00001)。 这些结果证明了使用CRISPR/Cas9技术对SCD-CB和SCD-成人CD34+细胞进行有效的体外编辑。校正后的两个血源的SCD CD34+细胞的HbA均显著升高，表明基因校正导致S等位基因向A等位基因的功能性转化，HbS水平相应降低。HBb内的Indels可能导致在HBb编辑的CD34+细胞中观察到的HBF增强。这些INDELs可能参与了表达HBF的细胞在分化过程中的选择性扩增或刺激细胞内HBF的上调。因此，珠蛋白基因座改变导致HbF丰度变化的分子基础需要进一步研究。综上所述，我们的研究建立了冷冻的CB作为SCD CD34+HSPC的替代来源，用于CRISPR/Cas9纠正镰状突变。

We initially evaluated CD34+ cells isolated from frozen CB samples of SCD patients. After 24 h of pre-stimulation, we electroporated chemically modified single guide (sg-)RNA/Cas9 ribonucleoprotein and a wild-type (WT) HBB allele-single-stranded oligonucleotide (ssODN) template into SCD-CB CD34+ cells, followed by 14 days of erythroid differentiation. Treatment with genome editing reagents did not alter erythroid cell differentiation in SCD-CB-derived CD34+ cells at day 14. The CD71GPA+ (anti-glycophorin A) phenotype was not significantly different between mock-treated cells (119 07%) and HBB-edited SCD-CB CD34+ cells (117 19%; P = 088). Next, we evaluated the frequency of indels and HDR rate at day 14 by tracking indels by decomposition (TIDE) and amplicon deep sequencing. Mean indel frequencies for the HBB-targeted loci were 631 36% and 736 176% in HBB-edited SCD-CB CD34+ cells by TIDE and amplicon deep sequencing, respectively. In addition, we observed a 204 91% and 254 63% HDR rate in HBB-edited SCD-CB CD34+ cells at day 14 by TIDE and amplicon deep sequencing, respectively. We next used amplicon deep sequencing at three different loci to investigate the sgRNAs off-target activity. Two of these off-target sites (at ITGA9 and chr2: 42851054-42851076) were predicted by COSMID software (http://crispr.bme.gatech.edu); we further included one for HBD because it has high homology to the HBB gene (Table SI). We found that the off-target activity at ITGA9, chr2: 42851054-42851076, and HBD was 351 19%, 00 003%, and 005 006% indels in HBB-edited SCD-CB CD34+ cells at day 14, respectively (Fig 2B). As expected for CB- and in vitro-derived erythroblasts, hemoglobin typing demonstrated that background fetal hemoglobin (HbF) protein levels were high at day 14 (817 52%) in mock-treated cells and were slightly increased in HBB-edited SCD-CB CD34+ cells (901 44%) compared with mock-treated cells (P = 010) (Fig 2C, D). In addition, a WT adult hemoglobin (HbA) high-performance liquid chromatography (HPLC) peak occurred only in HBB-edited SCD-CB CD34+ cells and averaged 67 26%, with correction levels up to 83% observed (P = 001). Correspondingly, sickle hemoglobin (HbS) levels were significantly decreased in HBB-edited SCD-CB CD34+ cells (32 18%) compared with mock-treated cells (183 52%: P = 0009). Next, we assessed the efficiency of CRISPR/Cas9 editing of HBB at the sickle mutation in CD34+ cells isolated from fresh peripheral blood collected from adult SCD patients. Like SCD-CB CD34+ cells, we found no differences in erythroid cell differentiation in HBB-edited SCD-adult CD34+ cells at day 14 compared with mock-treated cells (Fig 1AD). No significant change was detected in the CD71GPA+ phenotype in HBB-edited SCD-adult CD34+ cells (125 33%) compared with mock-treated cells (122 24%) (P = 089). Indel frequency and HDR rate were 519 131% and 198 63%, respectively, in HBB-edited SCD-adult CD34+ cells by TIDE analysis; consistently, amplicon deep sequencing showed 509 104% indels and 150 87% HDR (Fig 2A). We observed near-background off-target activity of the sgRNA at ITGA9 (428 31% indels), chr2: 42851054-42851076 (001 004% indels), and HBD (0003 002% indels) in HBB-edited SCD-adult CD34+ cells (Fig 2B). Haemoglobin typing of HBB-edited SCD-adult CD34+ cell lysates showed significantly substantial levels of HbA (128 65%) and a concomitant decrease in HbS (241 46%) compared with mock-treated cells (00 00% HbA, P = 003; 778 23% HbS, P = 000006; Fig 2C, D). HbF levels were also significantly increased in HBB-edited SCD-adult CD34+ cells (541 33%) compared with mock-treated cells (160 30%) (P = 00001), while HbA2 levels were not significant difference between mock-treated cells (62 14%) and HBB-edited SCD-adult CD34+ cells (89 09%) (P = 006). These results demonstrate efficient in vitro editing of SCD-CB and SCD-adult CD34+ cells using the CRISPR/Cas9 technology. Corrected SCD CD34+ cells from both blood sources showed a significant increase in HbA, indicating that the gene correction led to the functional conversion of the S allele to A allele and a corresponding decrease in HbS levels. Indels within HBB may cause the HbF enhancement observed in HBB-edited CD34+ cells. These indels may be involved in the selective expansion of HbF-expressing cells during differentiation or stimulate upregulation of HbF within cells. Therefore, the molecular basis by which alterations at the -globin locus cause changes in HbF abundance requires further investigation. In conclusion, our studies establish frozen CB as an alternative source of SCD CD34+ HSPCs for CRISPR/Cas9 correction of the sickle mutation.