Detection of placental transcription factor mRNA in maternal plasma.

mRNA of placental origin, including chromosome 21encoded mRNA, can be detected reliably in maternal plasma during the first trimester of pregnancy (1, 2). The presence and detectability of placental RNA in maternal plasma permits rapid screening of new markers to test their feasibility for use in noninvasive prenatal diagnostic assays. In contrast to conventional protein-based assays, new markers can include gene products with intracellular localization and noncoding mRNA. We challenged these features by screening maternal plasma for a large number of RNA targets (n 80) known or expected to be present in extraembryonic tissues. This set included genes coding for transcription factors, genes subject to genomic imprinting, genes coding for noncoding RNA, and other genes with restricted or abundant expression in trophoblast cells. Target genes were distributed over all chromosomes except the Y chromosome. Peripheral blood samples were collected from pregnant women attending the Prenatal Diagnostic Centre of the VU University Medical Center. All participants gave informed consent before being included in the study. The study was approved by the VU University Medical Center Ethics Committee. EDTA blood was collected between weeks 9 and 13 of pregnancy. All blood samples were obtained before invasive diagnostic procedures and processed as described previously (2 ). RNA was extracted from 1.6 mL of maternal plasma by silica-based affinity isolation with use of the QIAamp MinElute Virus Vacuum system (Qiagen) with minor modifications (2 ). The amount of carrier in buffer AL was reduced from 28 to 11.4 mg/L. Elution of bound RNA was done with 150 L of AVE buffer instead of MilliQ water. Finally, a 5-min incubation step before elution was introduced in the final step of RNA concentration by Microcon-PCR filters. The two-step, one-tube reverse transcription-PCR (RT-PCR) assay was performed as described previously (2 ) except that for a selected set of genes the number of PCR cycles was increased to 50. The characteristics of the 80 genes selected for expression profiling and the PCR conditions used are listed in Table 1 of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol50/issue8/. Using RT-PCR, we tested this set of 80 genes for their presence in early placental tissue (positive control) and their absence in nonpregnant plasma (negative control) and pregnant plasma. Three patterns were observed. Pattern A consisted of no detectable amounts in pregnant as well as nonpregnant plasma (negative/negative). This was seen for 42 genes (53%). Pattern B consisted of detectable amounts in both pregnant and nonpregnant plasma (positive/positive) and was observed for 30 of 80 genes (37%). Pattern C consisted of detectable amounts in pregnant plasma but no detectable amounts in nonpregnant controls (positive/negative). The latter pattern, the pattern of interest, was observed for eight genes (10%). For five genes, GCM1, ZDHHC1, PAPPA, PSG9, and PLAC1, detection in maternal plasma has never been described. Interestingly, two of these genes (GCM1 and ZDHHC1) code for transcription factors, that is, for gene products not accessible by conventional antibody-based assays. An overview of the RT-PCR results is given in Table 2 of the online Data Supplement. Two major conclusions can be drawn: (a) this expression profiling approach permits rapid screening of a large set of new fetal markers; and (b) the detection of GCM1 is typical of the intrinsic power of the plasma RNA method, i.e., analysis of markers not accessible by conventional antibody-based assays becomes possible. In addition, given the hierarchical importance in genetic control and combinatorial and multiple actions of transcription factors on downstream effector genes, analysis of placental transcription factor RNA in maternal plasma is likely to yield (clinical and basic) information distinct from others. GCM1 mRNA codes for the placenta-specific transcription factor glial cells missing (GCM) (3, 4). We detected GCM1 mRNA in the plasma of all pregnant women (n 6) tested between weeks 9 and 13 of pregnancy (Fig. 1A, lanes 1–6). Negative controls consisting of identically processed plasma from nonpregnant females (n 6) were negative in all cases (Fig. 1A, lanes 7–12). Moreover, after delivery, GCM1 mRNA decreased to undetectable concentrations in maternal plasma after 14 h (Fig. 1B, lanes 3–8). The latter demonstrates the first requirement (no persistence after pregnancy) if this factor is to be used for clinical applications. The GCM1 gene codes for a transcription factor containing the conserved GCM domain (4 ). The GCM domain is a zinc-coordinating, sequence-specific (A/G CCCGCAT) DNA binding domain (5 ). In both mice and humans, the GCM1 gene is expressed in trophoblast cells (3–7). In the mouse placenta, the Gcm-1 protein has been shown to be essential for vascularization of the placenta by fetal vessels (branching morphogenesis) as well in the formation of multinuclear syncytiotrophoblast by fusion of uninuclear cytotrophoblast cells (8 ). The GCM1 gene can be expected to be dysregulated in pregnancies complicated by trisomy 21 or preeclampsia (9, 10). Trisomy 21 placentas have a defect in villus trophoblast fusion as measured and demonstrated in vitro by a reduction in human chorionic gonadotropin -subunit concentrations caused by delayed and reduced trophoblast fusion (9 ). In preeclampsia, decreased placental GCM1 gene expression has been observed (10 ). The approach demonstrated in this report not only permits rapid screening of a large set of potential new markers, it allows the detection of markers not accessible by conventional antibody-based assays. This greatly increases the number of markers that become available for noninvasive prenatal diagnosis. Given the nature of these Technical Briefs