Aspects of Prenatal Microarray Analysis

Laboratory Aspects of Prenatal Microarray Analysis



Allen N. Lamb, PhD a,b,*allen.n.lamb@aruplab.com,


a Cytogenetics and Genomic Microarray Laboratory, ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT, USA


b Department of Pathology, University of Utah Health Sciences Center, Salt Lake City, UT 84108, USA


* Corresponding author. ARUP Laboratories, MS 115-H01, 500 Chipeta Way, Salt Lake City, UT 84108-1221




Keywords


• Array comparative genomic hybridization • Prenatal microarrays • Maternal cell contamination • Amniocytes • Chorionic villi • Abnormal ultrasound


The recent development and clinical implementation of array comparative genomic hybridization (aCGH) or microarrays has resulted in the most rapid and significant changes in the field of cytogenetics since the development of reliable chromosome banding techniques in the 1970s. aCGH detects gains and losses, which are referred to as copy number variations (CNVs), the size of which is determined by the array design1 (see “Prenatal Array Design” section).


The widespread use of arrays for postnatal cases has revealed many additional abnormalities that were not detectable by G-banded analysis and has altered the view of the extent of the causes of developmental delay, intellectual disability, multiple congenital anomalies, and autism. This has led to the recognition of new microdeletion and microduplication syndromes (see the article by Deak and colleagues elsewhere in this issue), and that the segmental duplication structure of the genome results in nonallelic homologous recombination (NAHR) that causes these genomic disorders.2–5 Analysis of multiple published studies shows that arrays increase the detection rate of clinically significant abnormalities above standard karyotyping by 10% to 15%6 and has led to the recommendation that microarray analysis should be a first-tier test.6,7 Microarrays also have an added benefit of providing a quantitative assessment of the genome, as opposed to the more subjective, qualitative approach with the historical standard of G-banded karyotypes, thus reducing errors or false-negative results. These false-negative results occur more frequently than most clinicians are aware of and are seen in both postnatal and prenatal studies.


The current use of prenatal aCGH has been mostly limited to follow-up of a normal traditional karyotype in cases with an abnormal ultrasound, usually associated with multiple congenital anomalies, and for the further definition of cytogenetic abnormalities, such as unbalanced translocations, marker chromosomes, and checking for losses/gains at the breakpoints of reciprocal translocations. However, the yield for specific ultrasound abnormalities and other indications, such as maternal serum screen positive cases and advanced maternal age (AMA), has not been defined. Therefore, more studies are needed before aCGH becomes a first-tier test for prenatal genetic diagnosis. There is no consensus on what are we trying to achieve in prenatal genetic diagnosis owing to these recent changes in technology. Until recently, the range of genetic abnormalities had been defined by maternal serum screening programs, which includes trisomy 21 and other major aneuploidies. This large-genome-rearrangement view has led some to argue for the use of more rapid aneuploidy screens (fluorescence in situ hybridization [FISH], quantitative fluorescent polymerase chain reaction [QF-PCR], multiplex ligation-dependent probe amplification [MLPA]), and even forgoing traditional karyotype analysis if ultrasound findings are normal, as a cost-effective and adequate testing strategy in the current climate of increasing health care costs.8,9 But this approach does not allow detection of the newer microdeletion/microduplication syndromes that have been discovered recently, and thus will not impact many of the cases of developmental delay, intellectual disability, multiple congenital anomalies, and autism that are seen in the postnatal population.2 There is currently no way to find and concentrate these newer abnormalities in the prenatal population, as maternal serum screening has done for trisomy 21 and other aneuploidies.


One of the major concerns for the use of aCGH in the prenatal genetic diagnosis is the findings of variants of unknown/uncertain/unclear significance10,11 (the term variant of unknown significance [VUS] is used in this article). Many clinicians and genetic counselors would like to keep the VUS to a minimum in prenatal cases. Unlike prenatal cases, postnatal cases with a VUS can wait for the accumulation of evidence that can demonstrate correlation or greatly reduce the likelihood of clinical significance over several months or years as additional studies are completed and published. One of the key steps for determining the clinical significance of a postnatal VUS is not the determination of inheritance, but rather the comparison of the frequencies of a CNV in a patient and normal control populations.3,4 However, with current databases this approach is at present not possible for many CNVs, and it may not be efficient with the limited time available in the context of prenatal diagnosis.


There have been several recent reviews on the role of aCGH in prenatal studies and the reader may wish to consult these for other details not covered in this review.10,12–14



Prenatal Sample Types and Laboratory Processing


Amniotic fluid cells, or amniocytes, obtained by amniocentesis, and chorionic villus cells, obtained by transcervical or transabdominal sampling, are the main source of cells for prenatal chromosome and microarray analysis.15,16 The cells obtained may be used for direct analysis or set up in culture to get dividing cells and to increase the total number of cells available for testing.


There are expectations that prenatal aCGH will result in more rapid turnaround time (TAT) by eliminating the need for tissue culture and decreasing the labor involved when compared to traditional cytogenetic analysis. Although TAT will be more rapid for the majority of cases by direct analysis, culturing will be needed on a percentage of cases. Cultures will be needed when direct extractions yield insufficient or poor quality DNA. In addition, some cases may require FISH follow-up to visualize the rearrangement; for example, determining a tandem versus insertional event for a gain, or an unbalanced translocation or derivative chromosome versus a separate gain and loss.


Therefore, laboratories may want to develop protocols to have backup cultures and then discard them if not needed (which incurs added expense in time and materials), or alternatively, to hold material and set up in culture only if needed (which results in longer TAT).


For cases with direct analysis, if results of adequate quality are obtained and FISH is needed, a preliminary report may be issued as it would take additional time before cells can be grown and harvested for FISH analysis to be performed. In the early months and years of a laboratory’s experience with prenatal aCGH, a number of cases and a comfort level may need to be reached before culturing becomes a minor part of the prenatal aCGH process.



Amniocytes


The experience gained over the years for use in traditional chromosome and FISH analyses for uncultured and cultured cells is very useful for aCGH studies.17 Amniocyte quantity is assessed after centrifugation and by gross examination of the cell pellet. This also allows for the assessment of the amount of blood present in the sample, usually of maternal origin. Based on examination of samples used for classic cytogenetics and for rapid aneuploidy assays, the cell pellet size varies from sample to sample and by gestational age (generally increasing with increasing gestational age).


For traditional cytogenetic analysis, cell culture of amniocytes is needed. Only a subset of the total cell population will attach and grow in culture. Amniocyte cultures usually yield a better quality and quantity of DNA for aCGH as compared to extraction of direct amniocytes.


Experience with prenatal aCGH demonstrates that 7 to 10 mL of amniotic fluid should provide enough amniocytes to yield a sufficient quantity and quality of DNA to perform aCGH on the direct extraction for most cases. However, samples and DNA yields from extractions are variable and at times an insufficient quantity or quality of DNA is obtained. This may be particularly challenging for samples less than 16 weeks gestational age. Therefore, for a certain percentage of cases (not yet defined due to the limited number of published cases on directs), amniocytes must be grown in culture to provide adequate material to perform aCGH and provide material for metaphase FISH analysis.



Chorionic Villi


For chorionic villus sampling (CVS), careful dissection of villi from maternal decidua is important. Villi are composed of (1) an outer layer with trophoblasts that are spontaneously dividing and used for direct chromosome analysis and (2) inner core of mesenchymal cells that are enzymatically separated (digested) and grown in culture.


There is an important distinction between a direct study for a traditional chromosome analysis and a direct for a microarray analysis with CVS. For a cytogenetic analysis, the direct uses the spontaneously dividing trophoblasts, and the culture uses the mesenchymal core cells. For aCGH, the direct is a proteinase K digestion of both the trophoblastic and mesenchymal core cells. This difference in a direct for traditional cytogenetic analysis and for an array analysis is important to remember when abnormal results are obtained and explanations and additional testing are being considered.


The trophoblasts are not as closely representative of the fetus as are the mesenchymal core cells. In addition, both layers may also have chromosomal changes that are not seen in the fetus, referred to as confined placental mosaicism (CPM).18 For cytogenetic analysis, if there are discrepancies between the direct and culture results, an amniocentesis is often performed, as amniocytes are usually more representative of the fetus. The algorithm to use for possible mosaicism or CPM with aCGH has not been established because of limited published studies with CVS, but the lessons learned from classic cytogenetics are likely to be the most important guide (eg, see Refs.19,20).



Maternal Cell Contamination


Maternal cell contamination (MCC) has only rarely been a problem in traditional cytogenetic analysis with cultured amniocytes, but there has been an increased concern in cultured CVS. MCC has been of more concern in assays that use uncultured amniocytes and direct CVS. Direct amniotic fluid has a greater chance of MCC because of the presence of blood (usually of maternal origin), and direct CVS may also have some contaminating maternal decidua present.


For amniocytes, the amount of maternal blood is best visualized when cells are centrifuged. Occasionally, maternal fibroblasts may also be present, but cannot be visually detected (usually picked up during the amniocentesis, and if the first part of draw is not omitted, these maternal cells may end up with the fetal cells). Once amniocytes are cultured, the maternal blood does not present problems from an MCC perspective; however, maternal fibroblasts, although not usually presenting problems for in situ culture techniques owing to the low level, may become a problem when expanded with subculturing.


For CVS, samples vary in the ease to clean maternal decidua from the villi, and laboratories vary in the skill of performing this identification and cleaning. In general, with good technique, MCC is rarely an issue during cytogenetic analysis for most laboratories. However, if subculturing is needed, a very low level of maternal cells may be expanded and interfere with obtaining the correct diagnosis.


In contrast to classic cytogenetic analysis, DNA-based testing for MCC (usually STR-based markers) is routinely performed in many laboratories for molecular mutation detection of single genes in prenatal diagnosis. The risk for MCC is especially of concern when PCR-based tests are involved, owing to the risk of preferential amplification and over-representation of maternal DNA.21,22


As many view aCGH as a cytogenetic test, there has been a tendency to rarely request MCC on most prenatal aCGH tests, especially those involving amniocytes. ACGH is currently performed after traditional cytogenetic analysis and requires the additional expansion of cultures to obtain sufficient DNA, and the length of time in culture and the amount of subculturing can significantly expand an initial low level of maternal cells. Table 1 presents some data from our laboratory (Signature Genomics) over a period of several months and shows that the rate of MCC was approximately 1% for amniocyte cultures and approximately 4% for CVS cultures. As expected, there is a higher level of MCC in CVS cultures than in amniocyte cultures.


Table 1 Detection of MCC in 466 prenatal samples















  Total Cases MCC Detected (%)
Cultured amniocytes 327 3 (0.9)
Cultured CVS 139 6 (4.3)

Does the presence of any level of MCC potentially invalidate the aCGH result or can copy number gains or losses still be detected accurately in the presence of various levels of MCC? There are no published studies that systematically examine the effects of different levels of MCC on different sized CNVs in aCGH. A potential source of information is studies of mosaicism, and these suggest that levels of a second cell line as low as 10% may be occasionally be detected; however, more consistent detection is in the 20% to 30% range.23–25 This information implies that if an abnormality is present, it may still be detected even in the presence of significant MCC. But a more important question may be at what level of MCC is an abnormality covered up and not detected? Does this differ for deletions versus duplications and for CNVs of different sizes? For example, is a smaller deletion covered up by a lower percentage of MCC than a larger deletion?


Preliminary unpublished studies performed in our laboratory (Signature Genomics) suggest that a large deletion of 2.5 Mb (eg, the DiGeorge syndrome [DGS]/velocardiofacial syndrome [VCFS] region) can be detected at a higher level of MCC than the same size duplication. This is expected owing to the differences in the log 2 ratio shifts for a deletion (2:1) compared to a duplication (2:3); therefore, a duplication is covered up by a lower level of MCC.


Small deletions in the range of 75 to 100 kb are covered up by lower levels of MCC than are larger deletions. Although the lower limits of size and level of MCC have not been systematically determined for duplications, preliminary data suggest this is likely to be in the 15% to 20% range. The lower limits may also be somewhat variable depending on the quality of the DNA and the experimental noise, so once the MCC level is above 10% there is always the possibility that a small, clinically significant duplication within a gene could be missed. Levels of MCC less than 10% appear not to interfere with detection of small abnormalities, but as platforms become more gene- and exon-centric, this low level may be of concern with prenatal use.


An MCC assay that is quantitative, as some laboratories have validated for opposite sex bone marrow transplants, is more useful as opposed to a simple presence/absence approach. If there is less than 10% MCC present, there is probably little concern, as seen in our preliminary unpublished studies. There is more concern when the amount approaches the 15% to 20% range, and the aCGH results should probably be considered invalidated.


Although most clinicians acknowledge the need for MCC studies for female fetuses, they often do not order the testing for a male fetus, as it is assumed that aCGH testing will readily detect the presence of any maternal DNA. However, preliminary data suggest that we may not be able to detect 10% to 15% MCC in a male fetus in some experiments, seen as a shift in the X and Y chromosome plots (potentially a low-level mosaic gain of X and loss of Y probes). This is the range of MCC that could possibly cover up the presence of a small duplication, which is especially concerning if completely within a gene. Therefore, although this risk may be low, it appears that something clinically significant could potentially be missed in a male fetus and we would not be aware of the contaminating MCC.


The preceding discussion is based on the assumption that a DNA amplification step is not involved. However, SNP-based arrays that use an amplification step will be more sensitive to small amounts of contaminating maternal DNA. This may limit the testing on direct samples with a SNP-based platform, as the presence of any contaminating maternal DNA would invalidate the array results. It could also present problems for testing for some CVS cultures that have been repeatedly expanded.


MCC should be performed on the same DNA extraction as used for the array; if the array needs to be repeated on a new DNA extraction, then the MCC analysis should also be performed on the new extraction as maternal cells may be expanded with additional subculturing.


Samples from pregnancies with oligohydramnios appear to have a higher risk for MCC, where approximately 10% of cases obtained by amniocentesis may have overwhelming MCC.26 Such samples usually do not show evidence of blood contamination and the origin and type of maternal cell is not clear.

Only gold members can continue reading. Log In or Register to continue

Related posts:

Default ThumbnailCytogenetics Evolving Picture of Microdeletion/Microduplication Syndromes in the Age of Microarray Analysis: Variable Expressivity and Genomic Complexity In Situ Hybridization Lymphoblastic Leukemia Syndromes Myeloid Leukemia: Conventional Cytogenetics, FISH, and Moleculocentric Methodologies

Stay updated, free articles. Join our Telegram channel
Tags:
Jun 28, 2017 | Posted by in HEMATOLOGY | Comments Off on Aspects of Prenatal Microarray Analysis

Full access? Get Clinical Tree
Get Clinical Tree app for offline access