Article submitted to Global Updates, November 2018
Laboratory challenges with cell-based NIPT
Ripudaman Singh1, Niels Uldbjerg2, Lotte Hatt1, Palle Schelde1, Ida Vogel3,4
- ARCEDI Biotech ApS
- Department of Obstetrics and Gynecology, Aarhus University Hospital, Denmark
- Department of Clinical Genetics, Aarhus University Hospital, Denmark
- Center for Prenatal Diagnostics, Aarhus University Hospital, Denmark
In accordance with ISPD policy, the authors of this article Lotte Hatt, Palle Schelde Ripudaman Singh disclose that they have a financial interest in ARCEDI Biotech ApS, which may be affected by the research reported in the following article. Ida Vogel and Niels Uldbjerg had no conflict to disclose.
With the advent of efficient technologies for the isolation and analysis of rare cells of any kind, fetal cells circulating in maternal blood are again attracting attention for their potential in noninvasive prenatal testing/diagnosis [1,2,3,4]. Even though the opportunities that this brings to the table in terms of low-risk and high diagnostic potential are immense (Figure 1), the challenges of isolating these rare fetal cells and then using them for prenatal genetic analysis are not trivial. In this article we will share some of our experiences of developing a fetal cell based noninvasive prenatal test (cbNIPT) which was launched in the Central Region of Denmark in May of this year. Under this provision women who are identified as ‘high-risk’ (>1:300) based on a combined first trimester screening (age, PAPP-A and β-hCG and nuchal translucency) now have the opportunity to choose between a cbNIPT along with a cell-free noninvasive prenatal test (cfNIPT), or an invasive testing and chromosomal microarray.
The journey of developing and optimizing the technology in the laboratory to a point where it has now been implemented in the clinical practice has been long. Henceforth we discuss some challenges that we faced along the way, and how those challenges were addressed.
The target cell type and the markers
It’s an old knowledge that some cells from the fetus invade the maternal blood. Attempts have been made over the last half century to categorize these cells and find specific markers which could in turn be used for isolating, enriching and identifying them from the peripheral blood [5]. Few fetal cell types that are known to circulate in maternal blood are lymphocytes, nucleated red blood cells, and trophoblasts. And several investigations over the years, including ours, have shown that trophoblasts might be the most common cells of fetal origin to circulate in the maternal blood [1,6,4]. Some of these placental cells are known to invade the decidua and replace the cells in the lumen of the maternal vessels, modify them, and ensure adequate blood flow to and from the growing fetus [7].
Connected to the knowledge of the cell type is the knowhow of cell specific markers which are both sensitive and specific in isolating these rare cells from maternal blood. When the aim is to develop a technology based on rare cells which will be implemented in the clinical practice, the issues surrounding no-call rate, turnaround time and positive predictive value become equally relevant. The question then is not whether it is possible to enrich rare cells, but whether it is possible to enrich rare cells consistently and identify them without any risk of misclassification.
Blood sampling and processing
In the absence of the knowledge of the half-life of the fetal cells, and the fact that we are actually targeting ‘foreign’ cells that have invaded the maternal circulation, it might be relevant to stabilize the cells in order to minimize their loss. Commercially available blood collection tubes with mild fixative might be the solution. This can ensure that the sample remains stable over many hours from the collection center to the lab where it will be processed.
The first step in most of the blood processing protocols for rare cell isolation involves the removal of erythrocytes, which can be achieved either by lysis or by density gradient techniques. This follows the pelleting of the leucocytes which also contain a tiny fraction of the rare cells of interest. In our experience, because the density of fetal cells is not known, the process involving lysis of erythrocytes gives a better yield of fetal cells as compared to density gradient techniques.
Rare cell isolation and single cell manipulation
Fetal cells circulating in the maternal blood are rare. There are around 1 to 2 fetal cells per ml of maternal blood [8]. Most of the technologies targeting rare cells either aim for recovery or purity. However, for cbNIPT, both recovery and purity are important – more fetal cells extracted per sample, with very few maternal cell as background. This will not only ensure that there are adequate cells per sample for downstream genetic analysis, but at the same time ensure that most of the maternal cells are removed so that they don’t interfere with the analysis and give mosaic results. And, for us, this fine balance of recovery and purity has been one of the biggest challenges to address. We found this balance by optimizing protocols for magnetic bead enrichment followed by staining of the fetal cells, and also by developing algorithms for fluorescence microscopes for the identification of fetal cells in the background of maternal cells [9] (Figure 2).
Next in the process is the leap from rare cell isolation to single cell analysis. In the end it’s not only about getting a pure fraction of fetal cells, but be able to manipulate these cells so that they are available for genetic analysis. And this step is not trivial either. It is dependent on the precision instruments that have the capacity to physically move a fetal cell from the place where it was localized to a tube/microwell plate where it will be further processed. In this, both the speed and precision are important factors. In our experience ‘picking’ the cells that are air-dried on the glass slides is more challenging than when they are loosely bound to the glass slides in a wet preparation. This step of single cell picking can be the most laborious and time consuming step in the whole process, impeding the downstream analysis.
Genetic analysis on fetal cells
Initial analysis on the stained and isolated fetal cells was performed using Fluorescence in situ hybridization (FISH) with an X and an SRY probe on pregnancies with a known male fetus [10,3]. Since then several different DNA-based techniques have been employed: NGS, STR and CMA. For clinical samples the DNA must be amplified to be sufficient for high resolution chromosomal microarray (CMA). Whole genome amplification (WGA) is yet a necessity with 1-4 cells but has the disadvantage that some regions may be amplified more readily than others, as for instance the GC-rich chromosome 19. However, over time even representation may be achieved with a resolution of around 5Mb. Current laboratory efforts are prioritized in setting up STR analyses, Sanger sequencing for monogenetic disorders and NGS - besides continued resolution studies comparing cbNIPT to invasive samples
Many clinical problems, also with cbNIPT, still remain unsolved. Addressing the issue of confined placental mosaicism with cbNIPT is one such challenge (Figure 3). When a pool of 1 to 4 fetal cells are used for prenatal genetic analysis then the risk that mosaicism won’t be detected becomes high. Picking more cells would likely increase the detection rate of mosaicism but would demand a higher number of fetal cells from every single pregnant woman, than what is currently feasible. Repeated sampling of women with low PAPP-A, which is an indication of mosaicism, could be one way to minimize this problem.
Conclusions
The story of challenges with cbNIPT might be from one lab, but these challenges are universal. It’s akin finding new solutions to old problems. For us it has taken 13 years, when we first ventured into the field of fetal cells, to develop a protocol for fetal cell isolation and analysis, which has now been implemented in a clinical setting. While it’s an ongoing process to test the limit of fetal cells for prenatal genetic analysis, one thing that is encouraging is that every step taken from now on is in the forward direction. Currently our focus is on automation and scaling up the complete workflow. We might not yet have a perfect solution for a high volume and complete fetal diagnosis, but what has been achieved to date is an important leap in noninvasive prenatal diagnosis.
References
- Breman AM, Chow JC, U'Ren L et al. Evidence for feasibility of fetal trophoblastic cell-based noninvasive prenatal testing. Prenat. Diagn. (2016).
- Jain CV, Kadam L, van DM et al. Fetal genome profiling at 5 weeks of gestation after noninvasive isolation of trophoblast cells from the endocervical canal. Sci. Transl. Med. 8(363), (2016).
- Kolvraa S, Singh R, Normand EA et al. Genome-wide copy number analysis on DNA from fetal cells isolated from the blood of pregnant women. Prenat. Diagn. (2016).
- Hou S, Chen JF, Song M et al. Imprinted NanoVelcro Microchips for Isolation and Characterization of Circulating Fetal Trophoblasts: Toward Noninvasive Prenatal Diagnostics. ACS Nano. 11(8), 8167-8177 (2017).
- Singh R, Hatt L, Ravn K et al. Fetal cells in maternal blood for prenatal diagnosis: a love story rekindled. Biomark. Med. (2017).
- Hatt L, Brinch M, Singh R et al. Characterization of fetal cells from the maternal circulation by microarray gene expression analysis--could the extravillous trophoblasts be a target for future cell-based non-invasive prenatal diagnosis? Fetal Diagn. Ther. 35(3), 218-227 (2014).
- Moser G, Weiss G, Sundl M et al. Extravillous trophoblasts invade more than uterine arteries: evidence for the invasion of uterine veins. Histochem. Cell Biol. 147(3), 353-366 (2017).
- Hamada H, Arinami T, Sohda S, Hamaguchi H, Kubo T. Mid-trimester fetal sex determination from maternal peripheral blood by fluorescence in situ hybridization without enrichment of fetal cells. Prenat. Diagn. 15(1), 78-81 (1995).
- Kolvraa S, Singh R, Normand EA et al. Genome-wide copy number analysis on DNA from fetal cells isolated from the blood of pregnant women. Prenat. Diagn. (2016).
- Hatt L, Brinch M, Singh R et al. A new marker set that identifies fetal cells in maternal circulation with high specificity. Prenat. Diagn. 34(11), 1066-1072 (2014).
Figures
Figure 1: A schematic representation of the positioning of cbNIPT in the current prenatal diagnosis scenario. Because the test is based on intact fetal cells, not contaminated by maternal genome, it should have the similar strength of analysis as invasive procedure such as chorionic villi sampling. And because cbNIPT is based on a blood draw, the risk to the pregnancy and stress to the pregnant women should be low.
Figure 2: A gallery of fetal cells enriched from maternal blood using ARCEDI technology. The cells are stained with DAPI (Blue) and a pool of fetal cell specific antibodies (Green).
Figure 3: Mosaicism and cbNIPT. With few cells (1-4) used for analysis in cbNIPT, the above matrix shows the probability by which a genetic abnormality (true fetal or confined placental) will be picked or missed depending on the degree of mosaicism.