The investigation was carried out in accordance with the latest version of the Declaration of Helsinki and approved by the Regional Ethics Committee of Research Centre For Medical Genetics (Moscow, Russia; approval no. 5). All participants signed an informed written consent to participate after the nature of the procedures had been fully explained.

Analyzed sample consisted of three groups of women of relatively the same age (22–40 years old) living in Moscow in the same social environment. The groups are as follows: Group I, non-pregnant, healthy women (n=40), consisting of students and employees of the medical facility (volunteers); group II, women with normal pregnancy >37 weeks (n=40), that later bore healthy children without signs of hypoxia or hypotrophy; group III, women with pregnancy complicated with IUGR >30 weeks, that later on bore children with IUGR signs (n=40). A total 5 ml of blood was collected under strict aseptic conditions from a peripheral vein of the subject using a syringe flushed with heparin (0.1 ml/5 ml blood).

Fetal development during pregnancy was screened with help of ultrasound anthropometry. Biparietal diameter of the head, chest circumference, abdominal circumference and femur length were measured. When these measurements were behind the population mean, by 2 weeks IUGR of stage 1 was diagnosed; 2 to 4 weeks stage 2 IUGR; more than 4 weeks stage 3 of IUGR. The final IUGR status was diagnosed after birth based on body mass. Normative parameters were between 75 and 25 percentile curves, first stage of IUGR 25 curve 10; second stage of IUGR 10 curve 3; third stage below curve 3. In addition to this, the weight-for-height index was used: Healthy, >60; IUGR stage 1, 55–60; stage 2, 50–55; stage 3, <50.

The functional state of the fetuses during pregnancy and labor was assessed with help of functional diagnostic methods such as cardiotocography and dopplerometry of uterine, umbilical and fetal blood vessels.

Cells were removed from the blood by centrifugation at 460 × g, followed by mixing of 3 ml plasma with 0.3 ml of the solution containing 1% sodium lauroyl sarcosinate, 0.02 M EDTA, and 75 µg/ml RNase A (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), incubation for 45 min, then the treatment with proteinase K (200 µg/ml; Promega Corporation, Madison, WI, USA) for 24 h at 37°C. Following two cycles of the purification with saturated phenolic solution, DNA fragments were precipitated by adding two volumes of ethanol in the presence of 2 M ammonium acetate. The precipitate was then washed with 75% ethanol twice, then dried and dissolved in water. The concentration of DNA (cfDNA index) was determined by measuring fluorescence intensity on ‘LS 55’ (PerkinElmer, Inc., Waltham, MA, USA) spectrometer after DNA staining with the PicoGreen (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA). The relative standard error of the index cfDNA was 10±4%.

Levels of DNase I activity in plasma samples were measured by the single radial enzyme diffusion method, as described previously (24). The assay method can determine pg to fg quantities of DNase I in 1 ml serum samples within 30 min. One unit of enzyme assayed corresponds to 1 ng purified human DNase I. Relative standard error of the index DNase I was 15±5%.

All reported results were reproduced at least three times as independent biological replicates. The significance of the observed differences was analyzed using the non-parametric Mann-Whitney U-tests. P<0.05 was considered to indicate a statistically significant difference. Data were analyzed with StatPlus2007 Professional software (http://www.analystsoft.com/).

ApoModel2 software was written by Roman Veiko (Research Centre For Medical Genetics, Moscow, Russia) in Object Pascal language with the help of an integrated media Turbo Delphi 2006 development. Software designed to simulate changes for the cfDNA under the influence of external and internal factors (e.g., exposure to radiation or pathology associated with the induction of apoptosis in cells in the body). The input data of the model were: The level of damages in the cell; the level of damages in the cell after reaching which it enters apoptosis; an increase in the number of lesions after exposure; the amount of DNA in the cell (in pg); the number of cells; the initial percentage of cells with damage; the number of damaging actions; DNA percentage remaining in the cell after degradation.

Functional input parameters were: The time function of reducing the amount of extracellular DNA as a result of elimination; function for apoptotic cells release of DNA into the extracellular environment. Functions can be set using a special formula designer and generator values. The output parameters of the model were the magazine states of the system, which were created during the simulation. The magazine reflected the number of dead cells, the number of surviving cells and the amount of extracellular and intracellular DNA.

CfDNA was isolated from the samples of blood plasma of 120 women. DNA concentrations were determined using the DNA binding dye PicoGreen. Fig. 1A presents cfDNA concentration in the blood plasma (Table I for descriptive statistics). Group I (n=40), non-pregnant women; group II (n=40), women with normal pregnancy (gestational age 30–41 weeks); and group III (n=40), women with pregnancies complicated by IUGR (gestational age 30–41 weeks).

There was no statistical difference between the three groups regarding cfDNA index (P>0.05). However, in nine women of groups II and III (22%), the authors observed elevated concentration of сfDNA as compared to the controls (Fig. 1A). In the third group, cfDNA concentrations were somewhat lower than in group II. If the single highest outlier is removed from group III (596 ng/ml), there is a persistent decrease in cfDNA concentration in group III compared to group II (P<0.05; NII=40, NIII=39).

Thus, the expected increase of cfDNA concentration in the blood of pregnant women compared to non-pregnant was not observed. Moreover, in troubled pregnancies, where a significant increase in cfDNA concentration was expected, the authors discovered a decrease in the concentration of cfDNA compared to healthy pregnancy. One of the primary reasons of cfDNA concentration decrease can be activation of the cfDNA elimination system of blood. One of the factors here is the activity of the blood plasma enzyme responsible for the cfDNA fragmentation, DNase I.

DNase I activity in blood plasma of 120 women was determined by the standard radial diffusion method (24). The DNase I indices for the groups are provided in Fig. 1A and in Table I. No significant difference was detected between the groups I and II by this index (P>0.05).

The third group, however, differs from group I (P<0.0000001) and group II (P<0.00001). Thus, plasma of pregnant women with IUGR contains highly active DNase I compared to blood plasma of healthy pregnant or non-pregnant women. A total of 19 out of 40 (47.5%) women from the group III had very high activity of DNase I in their blood, that were not present in the non-pregnant women sample. In group II, only 4 (10%) patients had increased DNase I activity compared to non-pregnant women.

Dependence of cfDNA concentration on DNase I activity is shown of Fig. 1C. Marked area is the field of data for non-pregnant women (group I). Non-pregnant women can be divided into two groups: Ia and Ib. Group Ia consists of 35 women (87.5%) and is characterized by linear dependence of cfDNA concentration on DNase I activity. Ib consists of 5 women who had low cfDNA concentration (11–22 ng/ml) and low DNase I activity (1.4–2.1 U/ml).

Marked area on the Fig. 1C corresponding to non-pregnant women included 25 (subgroup IIa) and 16 (IIIa) women from groups II and III. It is of importance, that the activity of DNase I for the subgroup IIIa in this area is higher than the activity of DNase I for groups I and IIa (P>0.05).

All the other patients from groups II (subgroup IIb) and III (subgroup IIIb) are beyond the marked area that contains all the non-pregnant women. For these women, the authors identified a linear dependence between cfDNA concentration and DNase I activity. Linear regression equations (cfDNA=A-b*DNase I) reflecting the dependence in all three groups are shown below.

Ia) cfDNA=123-13*DNase I (n=35), k=−0.8, P<<0,0001; IIb) cfDNA=275-21*DNase I (n=15), k=−0.7, P<0.01; IIIb) cfDNA=354-30*DNase I (n=24), k=−0.62, P<0.002.

Coefficient A in the equation is the hypothetic average cfDNA concentration in absence of DNase I (DNase I=0). This level reflects the amount of dying cells under the same conditions. Thus, the levels of cell death are: Group I (non-pregnant) < group II (healthy pregnancy) < group III (IUGR).

Coefficient b reflects the rate of cfDNA elimination when the DNase I is activated. This coefficient is increasing in a row: group I (non-pregnant) < group II (healthy pregnancy) < group III (IUGR).

Increase in DNase I activity affects cfDNA concentration in blood plasma of group III patients much stronger than patients of other groups. The authors introduced an index that reflects the amount of cfDNA for one unit of DNase I activity: cfDNA/DNase I (Fig. 1D and Table I for descriptive statistics). The difference between cfDNA/DNase I indices for samples I and II is not statistically significant (P>0.05). Sample III differs from sample I (P<0.001) and sample II (P<0.001).

IUGR can be characterized by objective parameters such as mass and height of the fetus and mass-to-height index. The authors analyzed the dependence of these characteristics and duration of gestation on cfDNA concentration, DNase I activity and the cfDNA/DNase I index with help of linear regression method (Table II). cfDNA concentration and cfDNA/DNase I index correlate negatively with all characteristics of the fetus. The higher the cfDNA concentration and the cfDNA/DNase I index are, the lower are the mass and height of the fetus and the shorter is duration of gestation. DNase I activity does not correlate with these factors (Table II).

Apomodel 2 software allowed the authors to set the level of injury and subsequent apoptosis, to analyze the amount of DNA entering into the extracellular medium from the dead cells and to take into account the processes that lead to the elimination of cfDNA. It is assumed that a damaging agent is acting on the cells. If the damage is above a given level, the cell is destroyed and DNA enters into the extracellular medium. cfDNA is constantly removed from the medium. Fig. 2 provides the results of modeling. The dependence of the amount of cfDNA in the medium from the time of observation. The time interval of observation was 3 months. For the function of the DNA release from the dead cells arbitrarily accepted expression: y=1/x, where x=time after a regular exposure.