Automated assay for fetal DNA analysis in maternal serum.

To the Editor: Since the first demonstration of the presence of cell-free fetal DNA in maternal plasma of pregnant women (1 ), many reports have been published, mainly about fetal sex and RhD determinations (2, 3). Although good results have been obtained with conventional PCR, real-time PCR is now the most widely used amplification system for fetal DNA analysis because it allows high sensitivity with a high degree of protection against contamination (4 ). Conventional PCR may produce falsepositive results from contamination (5 ), whereas real-time PCR, because it is a closed-tube system, reduces the risk of false-positive results from carryover of PCR products (6, 7). During sample preparation, however, contamination can result either from cross-contamination between samples or from the operator, particularly if a very small amount of human target sequence has to be detected, such as cell-free fetal DNA in maternal serum. We evaluated a fully automated sample preparation system (8 ) combined with real-time PCR. We extracted DNA from maternal serum by use of the MagNA Pure LC apparatus (Roche Biochemicals) with the Total Nucleic Acid LV reagent set (Roche) according to the manufacturer’s instructions. The eluted DNA and PCR reagents were automatically dispensed into PCR capillaries by an integrated PCR reaction set-up procedure. The operator only placed the reagents and samples in the apparatus. In the first part of the study, crosscontamination between samples was evaluated. We introduced 1 mL of serum from a male and a female in an alternating pattern (Fig. 1A). DNA was eluted in 50 L of elution buffer. We used 10 L of each of the 32 eluates for real-time PCR amplification of the SRY gene (3 ). Two clearly distinguishable groups of curves were observed (Fig. 1B): as expected, the 16 extracted DNA samples from male serum gave positive results for the SRY gene, whereas all DNA extracts from the female serum gave negative results. Quantitative results for the male serum, expressed as crossing points (cp), defined as the maximum of the second derivative of the fluorescence curves, revealed a CV of 0.4% at a 30.6 cp value (n 16). In the second part of the study, 108 sera from pregnant women (mean gestational age, 11.7 weeks) were analyzed by the above fully automated procedure. Results were compared with those obtained using the conventional manual procedure (3 ). The results of the two methods were completely concordant. All sera from pregnant women carrying a female fetus were negative for the SRY gene (n 62), whereas all sera from pregnant women carrying a male fetus were positive (n 46). With the fully automated procedure, the concentration of male fetal DNA in maternal serum during the first trimester of pregnancy was estimated to range from 5.2 to 64.9 copies/mL (mean, 26.3 copies/mL). This result was similar to those obtained by others (9 ). The CV of the test in this range of concentration was 2% (one sample analyzed 12 times during one analytical run; mean cp, 38.6; SD, 0.8 cp), which reflects the reproducibility of both DNA extraction and PCR assay. Analysis of fetal DNA in maternal serum offers new possibilities for noninvasive prenatal diagnosis. A lack of contamination is crucial for subsequent PCR applications, particularly in the field of prenatal diagnosis. Exclusion of contamination during extraction has become a major challenge since the introduction of real-time PCR. The results described here indicate that nearly complete automation of the DNA extraction, amplification, and detection steps can be achieved. This automated procedure could have implications for systematic analysis, such as RhD fe-