研究背景

传统产前遗传学诊断获取胎儿DNA的侵入性方法存在一定的感染或流产的风险。经宫颈获取的胎盘脱落的滋养层细胞具有完整的胎儿基因组信息，此方法不受孕妇年龄、孕周、产次、体重等因素的影响，从而成为近年来无创产前检测的研究热点。已有研究利用经宫颈获取的细胞(transcervical cell, TCC)样本中分离的滋养层细胞检测常见的胎儿染色体非整倍体、通过测序分析诊断脊髓性肌萎缩和囊性纤维化等单基因病。但TCC样本中存在大量的母源细胞背景，对TCC样本中滋养层细胞进行有效分离是进行下游检测分析的重要前提，但目前仍缺乏快速、有效且重复性高的方法。近年来单基因病无创产前检测技术有了飞速发展，对于TCC样本中分离的滋养层细胞进行单基因遗传病检测的研究尚属于探索阶段，仍有许多难点。

研究目的

本研究的主要目的有以下三个方面：（1）建立快速有效、重复性好的TCC样本滋养层细胞富集方法，为下游检测分析提供足量的细胞；（2）建立对富集分离的脱落滋养层细胞的胎儿源性鉴定方法，为下游检测分析奠定重要基础；（3）以富集分离的脱落滋养层细胞为材料，探索基于胎儿源性细胞的单基因遗传病无创产前检测的可行性。

方法

（1）查阅既往文献报道的女性生殖道细胞表面标记物，通过流式染色鉴定其在宫颈上皮细胞系、JEG3细胞系及TCC样本的表达情况，筛选母源细胞特异标志物。

（2）通过RNA-seq技术对宫颈上皮细胞系、JEG3细胞系进行高通量测序分析，以转录组学测序分析结果为基础，综合差异表达基因组间差异倍数、具有的功能，我们按照细胞膜表达、表达丰度相对高、有商品化抗体的原则，初步筛选有潜力的宫颈上皮细胞标志物，经流式染色筛选特异标志物。

（3）基于筛选的母源细胞特异标志物及绒毛外滋养层细胞(extravillous trophoblast, EVT)特异标志物HLA-G建立基于负筛+正筛的单克隆抗体的磁激活细胞分选法（combining negative and positive selection, CNPS），利用JEG3细胞系+TCT掺入样本初步鉴定CNPS的富集效率。

（4）收集3例正常单胎妊娠男胎孕妇的TCC样本，经CNPS富集后的候选EVT细胞经流式单细胞分选，然后将分选得到的所有单细胞进行全基因组扩增（whole genome amplification, WGA），对WGA产物进行低覆盖度测序分析，根据测序结果鉴定男性胎儿源性细胞。

（5）将低覆盖度测序分析为高质量的胎儿源性EVT细胞及脐血gDNA进行深度全基因组测序（whole genome sequencing, WGS），评估其在全基因组水平上的覆盖度、单核苷酸变异（single nucleotide variant, SNV）、插入缺失(insertion and deletion, InDel)变异的检测能力，探讨其单基因遗传病无创产前检测的可能性。

结果

（1）流式结果显示白细胞表面共同抗原CD45在宫颈上皮细胞系、JEG3细胞系表面的表达为0.01%、0.0%，在TCC样本中的表达为13.1%。而Syndecan-1、ER、PR在HcerEpic宫颈上皮细胞系、JEG3细胞系均有表达。因此纳入CD45作为TCC样本中白细胞阴性筛选的标志物。

（2）基于转录组学结果共筛选到7556 个基因在宫颈上皮细胞系和JEG3细胞系之间呈现差异性表达，其中有3757个基因显著上调表达和3799个基因显著下调表达。按照细胞膜表达、表达丰度相对高、有商品化抗体的原则，综合差异表达基因具有的功能，初步筛选出了CD44、CD227、CD82、CD9、CD49c、CD58、HLA-A共7个有潜力的宫颈上皮细胞标志物。经流式鉴定CD44、CD227在宫颈上皮细胞系及TCC样本中表达明显升高，在JEG3细胞系中无表达或表达很低，符合我们阴性选择母源宫颈上皮细胞的要求。

（3）制备细胞比例为0.1%的JEG3细胞+TCT掺入样本，利用负筛+正筛的单克隆抗体的磁珠分选法可将JEG3细胞的比例提高至8.2%，明显高于单纯用耦联有 HLA-G 抗体阳性选择的磁珠富集效率（0.3%）。

（4）本研究中3例样本均为单胎妊娠男胎，我们利用GM12878女性细胞系与GM50166男性细胞系的WGA产物进行低覆盖度测序，根据低覆盖度测序分析结果设置RsrPKMy的阈值来区分性别差异。当RsrPKMy值＞10且覆盖度＞30%时我们可以将候选EVT细胞鉴定为男性胎儿源性，男胎鉴定符合率达100%。

（5）我们将低覆盖度测序分析unique map ratio ＞30%且覆盖度 >30%的男性胎儿源性细胞定义为高质量的细胞，3例样本分别筛选出7个、7个、6个数据质量较高的细胞用于后续WGS。研究发现EVT细胞WGA-WGS覆盖度最高达为94.15%，3例样本的单细胞在各条染色体上reads的分布情况与对应的脐血gDNA WGS测序结果相比较显示出相似的分布趋势，而且与脐血gDNA 在全基因组水平上对SNVs、InDels的检测能力相似，表明经TCC分离的EVT细胞尽可能的复原了胎儿基因组信息。

结论

我们初步建立的CNPS富集方法可为下游检测提供足量的EVT细胞，基于RsrPKMy的阈值鉴定男性胎儿源性细胞的符合率可达100%。通过对分离的高质量EVT细胞WGA扩增产物进行深度测序分析，其基因组相对完整且覆盖度高，与脐血gDNA 在全基因组水平上对SNVs、InDels的检测能力相似，初步表明TCC样本中分离的EVT细胞对单基因疾病的无创产前检测有潜在的临床应用价值。

关键词：滋养层细胞；单细胞分析；无创产前检测；单基因疾病

Background

The traditional invasive methods used to obtain fetal DNA for prenatal diagnosis carry certain risks of infection or abortion. Therefore, the collection of trophoblast cells from the cervix has been proposed as a noninvasive alternative for prenatal testing. This method is possible because cervical trophoblast cells contain complete fetal genome information, and this method can be performed regardless of maternal age, gestational weeks, weight; hence, the method has attracted research attention in recent years. Studies have used trophoblast cells isolated from transcervical cell (TCC) samples to detect common fetal chromosome aneuploidy and monogenic diseases such as spinal muscular atrophy and cystic fibrosis by sequencing analysis. However, there are excessive maternal cells in TCC samples. The effective isolation of trophoblast cells from TCC samples is important for downstream detection and analysis, but fast, effective, and reproducible methods are lacking. In recent years, technology for noninvasive prenatal detection of monogenic diseases has developed rapidly, but research on the detection of monogenic genetic diseases from trophoblast cells isolated from TCC samples is still in the exploratory stage, and there are still many difficulties with this method.

Objective

The main objectives of this study are as follows: (1) establish a method for rapid, effective, and reproducible enrichment of trophoblast cells from TCC samples in order to provide sufficient cells for downstream detection and analysis; (2) investigate the fetal origin of trophoblast cells, which will lay an important foundation for downstream detection and analysis; and (3) explore the feasibility of noninvasive prenatal testing of single gene genetic diseases based on trophoblast cells.

Methods

(1) The surface markers of female genital tract cells were identified by consulting existing literature, and their expression levels in the cervical epithelial cell line, JEG3 cell line, and TCC samples were identified by flow cytometry so as to screen the specific markers of maternal cells.

(2) The RNA of the cervical epithelial cell line and the JEG3 cell line were sequenced and analyzed by RNA-seq technology. Based on the results of transcriptome sequencing, potential markers of cervical epithelial cells were preliminarily screened according to the principles of cell membrane expression, relatively high expression abundance, and commercial antibodies, and then specific markers were screened by flow cytometry.

(3) Based on the screened maternal cell specific markers and the extravillous trophoblast (EVT) specific marker HLA-G, a new strategy combining negative and positive selection (CNPS) was designed to isolate cervical trophoblast cells. The enrichment efficiency of CNPS was preliminarily identified by JEG3 cell line + TCT incorporation samples.

(4) TCC samples were collected from three women, all of whom had a normal singleton pregnancy with a male fetus. The candidate EVT cells enriched by CNPS were sorted by flow cytometry, and then all the sorted single cells were subjected to whole genome amplification (WGA). The WGA products were sequenced and analyzed with low coverage, and the male fetal cells were identified according to the sequencing results.

(5) Low coverage sequencing of EVT cells was performed to confirm their identity as high-quality EVT cells, and then the high-quality EVT cells and cord blood gDNA were subjected to whole genome sequencing (WGS) to evaluate their coverage rate, single nucleotide variant (SNV) and insertion and deletion (Indel) variation at the whole genome level, and to explore the feasibility of noninvasive prenatal testing for single gene genetic diseases.

Results

(1) The results of flow cytometry showed that the expression of leukocyte surface common antigen CD45 on the surface of the cervical epithelial cell line and the JEG3 cell line was only 0.01% and 0.0%, respectively, whereas it was 13.1% in TCC samples. Therefore, CD45 was included as a marker for leukocyte negative screening in TCC samples.

(2) Based on the results of transcriptomics, 7556 genes were differentially expressed between cervical epithelial cell lines and JEG3 cell lines, among which 3757 genes were significantly upregulated and 3799 genes were significantly downregulated. According to the principle of cell membrane expression, relatively high expression abundance, and commercial antibodies, and the functions of differentially expressed genes, seven potential cervical epithelial cell markers were preliminarily screened: CD44, CD227, CD82, CD9, CD49c, CD58, and HLA-A. According to flow cytometry, the expression levels of CD44 and CD227 in cervical epithelial cell lines and TCC samples increased significantly, but there was either no expression or very low expression in the JEG3 cell lines, which met the requirements of negative selection of maternal cervical epithelial cells.

(3) The proportion of JEG3 cells + TCT mixed with 0.1% cells can be increased to 8.2% by using CNPS, which is significantly higher than the enrichment efficiency of magnetic beads coupled with HLA-G antibody positive selection alone (0.3%).

(4) As previously mentioned, in this study, three TCC samples were collected from women with singleton pregnancies with male fetuses. The WGA products of the GM12878 female cell line and GM50166 male cell line were used for low coverage sequencing, and the threshold of RsrPKMy was set according to the low coverage sequencing analysis results to distinguish gender differences. When RsrPKMy value > 10 and coverage > 30%, the candidate EVT cells are of male fetal origin, and the coincidence rate of male fetal identification is 100%.

(5) The male fetal derived cells with low coverage sequencing analysis unique map ratio > 30% and coverage > 30% were defined as high-quality cells. Seven, seven, and six cells with high data quality were selected from the three samples for subsequent WGS. The WGA-WGS coverage of EVT cells was found to be as high as 94.15%. The distribution of reads on each chromosome of single cells in three samples showed a similar distribution trend compared with the corresponding cord blood gDNA WGS sequencing results. The isolated EVT cells have detection ability for SNVs and Indels at the whole genome level that is similar to those of cord blood gDNA. The results showed that EVT cells isolated from TCC samples recovered fetal genome as much as possible.

Conclusions

This study preliminarily established a CNPS enrichment method, which can provide sufficient EVT cells for downstream detection. The coincidence rate of identifying male fetal cells based on the threshold of RsrPKMy can reach 100%. Through deep depth whole genome sequencing of the WGA amplification products of isolated high-quality EVT cells, the fetal genomic profiling is relatively complete and has a high coverage. The isolated EVT cells have detection ability for SNVs and InDels at the whole genome level that is similar to those of cord blood gDNA. The preliminary results show that EVT cells isolated from TCC samples have potential clinical value for noninvasive prenatal testing of monogenic diseases.

Key words：Trophoblast; single-cell analysis；non-invasive prenatal testing; monogenic disease