Early Diagnosis and Prognosis of Pre-Eclampsia

Abstract

The invention relates to a process for the in vitro diagnosis or prognosis of pre-eclampsia, comprising determining an increase or decrease in the concentration of one or more markers in a sample of biological fluid.

Claims

1. A process for the in vitro diagnosis or prognosis of pre-eclampsia, comprising: a) measuring, in a sample of biological fluid originating from a pregnant woman, the concentration of one or more markers chosen from LIFR, TTR (transthyretin), ApoA2 and Pzp (A2M), preferably LIFR and/or TTR, alone or in combination with one or more other markers, b) using the result of the measurement of step a) in the diagnosis or prognosis of pre-eclampsia, wherein a positive diagnosis or a positive prognosis is given by an increase or a decrease in the concentration of the marker(s) relative to the normal concentration of the marker(s) obtained in pregnant women.

2. The process as claimed in claim 1, said biological fluid is chosen from blood, plasma, serum, urine, saliva or amniotic fluid, preferably serum.

3. The process as claimed in either one of claims 1 and 2, said pregnant woman being at a gestation period of less than or equal to 20 weeks, preferably less than or equal to 16 weeks, preferably at a gestation period of between 7 weeks and 16 weeks.

4. The process as claimed in any one of claims 1 to 3, said marker being in a circulating form (sLIFR, sTTR, sApoA2 and sPzp).

5. The process as claimed in any one of claims 1 to 4, said one or more other markers being chosen from sENG, sFLT1, PIGF, PAPP-A, fetal hemoglobin and/or hemopexin.

6. The process as claimed in any one of claims 1 to 5, comprising a step a) of calculating one or more ratios of two markers and optionally calculating one or more ratios of two other markers, preferably calculating the LIFR/TTR, sFLT1/PIGF and/or sLIFRsFLT1 ratio.

7. The use of one or more markers in the in vitro evaluation of pre-eclampsia in a sample of biological fluid originating from a pregnant woman, wherein a positive diagnosis or a positive prognosis is given by an increase or a decrease in the concentration of the marker(s) relative to the normal concentration of the marker(s) obtained in pregnant women, said marker(s) being chosen from LIFR, TTR (transthyretin), ApoA2 and Pzp (A2M), preferably LIFR and/or TTR, alone or in combination with one or more other markers.

8. The use of one or more antibodies directed against one or more markers in the in vitro evaluation of pre-eclampsia in a sample of biological fluid originating from a pregnant woman, wherein a positive diagnosis or a positive prognosis is given by an increase or a decrease in the concentration of the marker(s) relative to the normal concentration of the marker(s) obtained in pregnant women, said marker(s) being chosen from LIFR, TTR (transthyretin), ApoA2 and Pzp (A2M), preferably LIFR and/or TTR, alone or in combination with one or more other markers.

9. The use as claimed in either one of claims 7 and 8, said one or more other markers being chosen from sENG, sFLT1, PIGF, PAPP-A, fetal hemoglobin and/or hemopexin.

10. A diagnosis or prognosis kit of use for carrying out the process as claimed in one of claims 1 to 6.

11. A method for treating pre-eclampsia in a pregnant woman, comprising: (i) carrying out the process as claimed in any one of claims 1 to 6, and (ii) when the diagnosis is positive or the prognosis is positive, introducing a suitable treatment for the pregnant woman.

12. The method as claimed in claim 11, said suitable treatment being one or more medicaments chosen from aspirin (acetylsalicylic acid), magnesium sulfate, antihypertensives, corticosteroids, low molecular weight heparins, cholesterol regulators and oxidative stress regulators, preferably aspirin.

Description

FIGURE LEGENDS

[0059] FIG. 1: diagram depicting the results of the iTRAQ experiment showing the increase or the decrease in the concentration of several markers present in the plasma of mice having pre-eclamptic gestation relative to the normal concentration of these markers in the plasma of mice having normal gestation, at 6.5 days of gestation (the total gestation for the mice line studies is 18.5 days). Among the markers studied with a modified concentration in the pre-eclamptic mice which are useable in humans due to gene homology, it may be noted that the concentrations of the proteins LIFR and TTR are highly modified. It should be noted that the family A serpins (Serpina #) are not readily transposable to humans due to significant polygenism of this gene in mice.

[0060] FIG. 2: diagram depicting the concentration of circulating LIFR in the plasma of 30 pre-eclamptic women and 30 control women, taken at the time of labour (A. crude measurement, and B. relative to total proteins). The pre-eclamptic women have a significantly higher concentration of circulating LIFR in their plasma than the control women.

[0061] FIG. 3: diagram depicting the concentration of circulating LIFR and sFLT1 in the plasma of 20 pre-eclamptic patients and 20 control patients, as crude measurement or relative to the total proteins. The pre-eclamptic women have a significantly higher concentration of circulating LIFR in their plasma than the control women. The results show that the LIFR marker appears to be more discriminating than the sFLT1 marker, current reference marker.

[0062] FIG. 4: curve depicting the concentration of circulating LIFR in the plasma of 20 pre-eclamptic patients and 20 control patients, as crude measurement as a function of the gestation date. It is noted that, regardless of the gestation stage, the concentration of circulating LIFR is increased in the plasma of pre-eclamptic patients compared to the concentration of circulating LIFR in control patients. The control patients are defined as pregnant patients in good health, without any known medical complication (no diabetes mellitus, pre-existing hypertension, renal or cardiovascular pathology, no twin pregnancy or other multiple pregnancy).

[0063] FIG. 5: diagram depicting the concentration of PIGF and sENG (markers already known) in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients.

[0064] FIG. 6: diagram depicting the concentration TTR in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients.

[0065] FIG. 7: diagram depicting the concentration sFLT1 (known marker) in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients.

[0066] FIG. 8: diagram depicting the concentration sLIFR in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients.

[0067] FIG. 9: ROC curve depicting the effectiveness of the ratio of sFLT1/PIGF concentration in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients. The closer the area under the curve is to 1, and the more different it is to 0.5, and therefore the greater the area under the curve (ROC area), the more effective is the marker.

[0068] FIG. 10: diagram and curve depicting the ratio of sLIFRsFLT1 concentration in first trimester plasma in 100 patients who later developed pre-eclampsia and 50 control patients. The closer the area under the curve is to 1, and the more different it is to 0.5, and therefore the greater the area under the curve (ROC area), the more effective is the marker.

EXAMPLES

Example 1: Preparation of the Murine Model of Pre-Eclampsia by Overexpression of STOX1

[0069] The cDNA of the complete open reading frame (ORF) of the human STOX1 gene (isoform A) was microinjected into a male pronucleus following fertilization. The pronucleus transfected in this way was implanted into the uterus of a female mouse. After 18.5 days of gestation, transgenic mice were obtained. The number of copies of the transgene STOX1 was evaluated by quantitative polymerase chain reaction (qPCR) using the DNA from the transgenic mouse relative to a single-copy gene.

[0070] Three transgenic mouse lines were obtained and amplified. The analysis of the transmission of the STOX1 transgene through the generations showed that there was a single insertion locus of the transgene STOX1 over the 3 transgenic mouse lines which were amplified. The phenotyping of the transgenic mice was carried out as described in Doridot et aL, Hypertension, 2013, with daily measurement of the blood pressure throughout gestation, and urine sampling for measuring proteinuria. In the article by Doridot et al., it was also shown that markers characteristic of pre-eclampsia (sFLT1 and sENG) are increased in the serum/plasma of transgenic mice carrying fetuses and which are thus pre-eclamptic.

Example 2: Demonstrating the Markers of the Invention in Mice

[0071] Pooled or isolated plasma samples were collected from heparinized blood from the jugular vein of transgenic mice of example 1, carrying fetuses, at 6.5 days of gestation (pre-eclamptic mice) or non-pre-eclamptic mice at 6.5 days of gestation (control group). These plasmas were depleted in the major plasma proteins on MARS-3 columns (Agilent), making it possible to eliminate albumin, IgGs and transferrin. For this purpose the plasmas were diluted 20-fold in the equilibration buffer provided with the MARS-3 column, then filtered by microcentrifugation over 0.22 m filters. Two washes at 100 g for 2.5 min were carried out, and the fraction obtained was concentrated using a 5K MWCO concentrator. The amount of remaining proteins was evaluated by 2D Quant protein assay (GE Healthcare, France). The fraction was subsequently aliquoted and stored at 80 C. Thereafter, samples containing 50 g of proteins were lyophilized, and then resuspended in 25 l of 500 mM TEAB buffer (Sigma) with 1% SDS. The proteins contained in the samples were then alkylated, digested and labelled with the iTRAQ reagents according to the manufacturer's protocols (Applied Biosystems) in order to obtained labelled peptides. The samples containing the labelled peptides are hereinafter referred to as labelled samples. The labelled samples were identified as a function of their status: pre-eclampsia or control. The labelled samples were then combined and lyophilized. In order to limit complexity, the labelled peptides (hereinafter peptides) were washed and fractionated by chromatography on a cation exchange column. 200 g of dry peptides were resuspended in 1 ml of 5 mM KH.sub.2PO.sub.4 and 20% acetonitrile (Carlo Erba), pH 2.8 (buffer A). 80% of the peptides were pre-fractionated with a 5 m PolySULFOETHYL A column (PolyLC, Columbia, Md.) 200 mm long2.1 mm in diameter and with a pore size of 300 , on an HPLC (Waters625 HPLC) with a constant flow of 2 ml/min. The column was first washed with the buffer A from the manufacturer for 30 min, then a gradient of 19 min was applied to increase the concentration of buffer B to 63% then a plateau of 5 min and again 100% of buffer A for 5 min. Fractions were collected every 30 sec, pooled as a function of the UV absorbance at 210 nm (16 g/pool). For the extraction at high salt concentration, desalination was carried out with a C18-Sep-Pak cartridge (Waters). The fractions were subsequently lyophilized and stored at 20 C. The fractions were then resuspended in 0.1% TFA buffer (Fluka) and 10% acetonitrile (Carlo Erba), then 2 g were injected in two goes into a nano-HPLC Ultimate 3000 (Dionex). The peptides were then purified using a C18 PepMap column (Dionex, 0.3 mm I.D.5 mm, pore size 100 , 3 m particles) at 30 l/min in 0.1% TFA (Fluka) and 2% acetonitrile (Carlo Erba). The peptides were then separated with a C18 PepMap 100 column (Dionex, 75 mm I.D.150 mm, pore size 100 , 3 m particles) at 300 nl/min (solution A: 0.1% TFA, 2% acetonitrile, solution B: 20% solution A mixed vol/vol with 80% acetonitrile). After equilibration with 7% of B, a multi-step gradient is started 3 min after injection, with 16% of B 14 min post-injection, 19% at 22 min, 23% at 25 min, 32% at 51 min, 50% at 65 min and a plateau at 95% for 79 min before equilibration for 16 min. Fractionation was then carried out for each 2 g duplicate with a Dionex robot (Probot), 18 min after the start of the injection, and the collection was coupled directly with MALDI plates, every 10 seconds for 384 points per fraction. The eluent and the matrix solution were mixed then placed on the MALDI plates. The matrix of -cyano-4-hydrocinnamic acid (CHCA) was dissolved at 2 mg/ml in a solution of acetonitrile (70%) with 0.1% TFA and 110 M of glu-fibrinopeptide-B for internal calibration (m/z=1570.677).

[0072] The mass spectra were measured with a 4800 MALDI-TOF-TOF spectrometer (Applied Biosystems), version 3.5.28193, build 1011 with a positive reflection mode at a fixed laser frequency. For each fraction, 50 spectra were collected between 850 and 4000 Da with a laser frequency fixed at 200 Hz. 500 spectra per sample were summed and processed to obtain monoisotopic values with a raw spectrum, the signal/noise ratio of which was greater than 20. Each spectrum made it possible to select the 8 most abundant peaks with a signal/noise ratio >15. 1000 MS/MS spectra were summed for each precursor with the baseline subtracted using the Savitsky-Golay algorithm (4th degree polynomial regression). In order to identify the non-dominant proteins, an exclusion list was produced from a duplicated series.

[0073] The peptides were identified with ProteinPilot version 2.0.1 (Applied Biosystems, MDS-Sciex, Foster City, Calif.). The analysis was made by comparison with the Human Protein Database of the IPI (International protein Index version 3.38). The data was processed taking into account the trypsin cleavage carried out of the carbamidomethylated cysteines, with the search strength set at thorough. A protein is considered to be significantly identified when at least two definitely identified peptides are assigned a confidence score >95% and a degree of difference on the iTRAQ of >1.2, with p<0.05.

[0074] The results are presented in table 1 below.

TABLE-US-00001 TABLE 1 list of markers exhibiting an increase in the concentration or a decrease in the concentration in transgenic mice relative to the normal concentration of the markers obtained in wild-type mice (control group or WT), at 6.5 days of gestation. Name of Number of Tg/WT at marker Full name peptides 6.5 days Gm20547 Protein Gm20547 OS = Mus musculus 3 1.34 GN = Gm20547 PE = 2 SV = 1 LIFR Putative uncharacterized protein OS = Mus musculus 44 1.24 GN = Lifr PE = 2 SV = 1 Serpina1b Serpina1b protein OS = Mus musculus 9 1.24 GN = Serpina1b PE = 2 SV = 1 Hrg Histidine-rich glycoprotein OS = Mus musculus 8 1.10 GN = Hrg PE = 1 SV = 2 Kng1 Kininogen-1 OS = Mus musculus 39 1.08 GN = Kng1 PE = 1 SV = 1 Serpina1a Alpha-1-antitrypsin 1-1 OS = Mus musculus 6 1.07 GN = Serpina1a PE = 1 SV = 4 F2 Coagulation factor II OS = Mus musculus 16 1.05 GN = F2 PE = 2 SV = 1 Serpina6 Serine (Or cysteine) peptidase inhibitor, 28 0.97 clade A, member 6 OS = Mus musculus GN = Serpina6 PE = 2 SV = 1 Cycs Cytochrome c OS = Mus musculus 1 0.97 GN = Cycs PE = 2 SV = 1 Ces1c Carboxylesterase 1C OS = Mus musculus 51 0.88 GN = Ces1c PE = 1 SV = 4 Cp Ceruloplasmin OS = Mus musculus 63 0.88 GN = Cp PE = 2 SV = 1 Pzp(A2M) Alpha-2-macroglobulin OS = Mus musculus 45 0.85 GN = Pzp PE = 2 SV = 1 Kng2 Kng2 protein OS = Mus musculus 5 0.82 GN = Kng2 PE = 2 SV = 1 Apoa2 APOAII OS = Mus musculus 25 0.79 GN = Apoa2 PE = 4 SV = 1 Obp1a MCG117626 OS = Mus musculus 10 0.76 GN = Obp1a PE = 2 SV = 2 TTR Transthyretin OS = Mus musculus 18 0.54 GN = Ttr PE = 2 SV = 1

[0075] The number of peptides column indicates the number of peptides which were clearly identified in the protein during the iTRAQ. The more peptides that are identified, the more is accurate the identification of the corresponding protein. The number of peptides therefore indicates the reliability of the identification of the protein, and hence of the measurement. The correct protein can be considered to have been identified beyond 2 peptides identified in the same protein. The overall results at 6.5 days are presented in FIG. 1.

CONCLUSION

[0076] The concentration of the markers presented in table 1 is either increased or decreased in mice with pre-eclampsia. These markers may therefore be used within the context of the invention.

[0077] The markers having a large number of peptides and an increase or a decrease in the concentration at least equal to 20% relative to the wild-type mice are TTY (transthyretin), LIFR, ApoA2 and Obp1a. LIFR appears to be a particularly beneficial marker for the in vitro diagnosis or prognosis of pre-eclampsia.

Example 3: Study of the LIFR Marker in Pregnant Women at 25 Weeks of Pregnancy

[0078] The concentration of LIFR in the plasma of pregnant women at 25 weeks of pregnancy was measured for FIG. 2 and FIG. 3 with the BOSTER kit (reference EK1200), precisely following the manufacturer's technical recommendations. The detail may be obtained at www.bosterbio.com. The protocol is standardized to operate in strips constituting a 96-well plate. The kit provides a standard which makes it possible to produce a response range by successive dilutions, making it possible to calibrate the unknown samples. After addition of the final solution, a more or less intense yellow staining was read in absorbance at 450 nm. The results were integrated into the standard range (from 156 pg/ml to 10 ng/ml) by linear regression. More precisely, the diluted LIFR standard was prepared 2 hours before the experiment by consecutive dilutions in a diluent provided in the kit. The range was deposited in duplicates in 16 wells (8 concentrations) of the 96-well plate provided. The 80 other samples were dilutions of human samples from serum from pregnant women, either pre-eclamptic or not. The plate was subsequently placed for 1 h 30 at 37 C. After evacuating the supernatant, 100 l of the anti-LIFR antibody from the BOSTER kit (reference EK1200), diluted in the diluent provided in the kit, were added. The plate was then incubated for 1 h at 37 C. The wells are then rinsed 3 times with 300 l of PBS. 100 l per well of the ABC solution (Avidin-Biotin-Peroxidase Complex), diluted in the dilution buffer provided, were subsequently added. Incubation was carried out at 37 C. for 30 minutes. The wells were rinsed 5 times with PBS, then 90 l of TMB buffer were added (containing the peroxidase substrate). Incubation was carried out at 37 C. for 30 minutes, then the TMB stop solution was added (100 l). The absorbance could subsequently be measured at 450 nm with a spectrophotometer.

[0079] The amount of sFLT1a was measured for FIGS. 3 and 4 with the kit sold by MyBioSource under the catalogue number MBS175839, precisely following the manufacturer's technical recommendations. The detail may be obtained at http://www.mybiosource.com/. The protocol is identical to that presented for the LIFR measurement. The protocol is standardized to operate in strips constituting a 96-well plate. The kit provides a standard which makes it possible to produce a response range by successive dilutions, making it possible to calibrate the unknown samples. After addition of the final solution, a more or less intense yellow staining was read in absorbance at 450 nm. The results were integrated into the standard range by linear regression.

LIFR Marker

[0080] The results are presented in FIGS. 2, 3 and 4.

Conclusion for the First Cohort (FIG. 2):

[0081] The mean concentration of LIFR detected in the plasma of pregnant women was 11.1 ng/ml in the controls (pregnant women without PE), as opposed to 12.9 ng/ml in the PE patients (unilateral T test, 0.028, on paired cases 0.023). The results confirm that the concentration of the LIFR marker increases in the plasma of pregnant women with PE.

Conclusion for the Second Cohort and Comparison with sFLT1 (FIGS. 3 and 4):

[0082] Highly significant differences for the concentration of sLIFR were observed between the PE patients and the controls (mean=5.1 ng/ml in the controls, 10.6 ng/ml in the PE patients (p=1.7 10.sup.5)). The reference marker sFLT1 was also tested on the same plasma samples; in this case, the difference was, respectively, 92 ng/ml and 135 ng/ml in the controls and the PE patients (p=0.056).

[0083] FIG. 3 presents the standardization of the LIFR concentration relative to the total protein amount, calculated by the Bradford colorimetric assay relative to a range of bovine serum albumin. We obtained a significant result (1.7 ng/ml vs. 3.8 ng/ml, p=0.007). On the earliest samples (4 controls and 3 PE of 6-7 months, i.e. 25-30 weeks), the controls had a lower concentration of LIFR than the PE patients.

[0084] The data suggests that sLIFR has better performance than sFLT1 in discriminating between a sample originating from a PE patient or a control. In addition, despite the limited number of early samples (around 25 weeks), the differences observed are just as large early in the pregnancy as late in the pregnancy (FIG. 4).

Example 4: Studies of the Markers LIFR, TTR, sFLT1, PAPP-A, PIGF and of the sFLTt1/PIGF Ratio in Pregnant Women Around 13-14 Weeks of Gestation

[0085] Plasma samples taken around 13-14 weeks of gestation from pregnant women who went on to develop PE were recovered. These samples are systematically analyzed by a plate ELISA method for all the markers sought: LIFR (BOSTER kit), TTR (Abcam kit), sFLT1 (R&D kit), PAPP-A (Thermofisher kit) and PLGF (R&D kit). As a reminder, the standard ELISA measurement protocol is as follows. The kit provides a standard which makes it possible to produce a response range by successive dilutions, making it possible to calibrate the unknown samples. After addition of the final solution, a more or less intense yellow staining can be read in absorbance at 450 nm. The results are integrated into the standard range (from 156 pg/ml to 10 ng/ml) by linear regression.

[0086] The marker concentration is deduced by implementing similar protocols, detailed for each of the kits. For example, for the LIFR marker, the diluted LIFR standard is prepared 2 hours before the experiment by consecutive dilutions in a diluent provided in the kit. The range is deposited in duplicates in 16 wells (8 concentrations) of the 96-well plate provided. The 80 other samples are a dilution of the human samples. The plate is subsequently placed for 1 h 30 at 37 C. After evacuating the supernatant, 100 l of the anti-LIFR antibody, diluted in the diluent provided in the kit, are added. The plate is then incubated for 1 h at 37 C. 100 l per well of the ABC solution (Avidin-Biotin-Peroxidase Complex), diluted in the dilution buffer provided, are subsequently added. Incubation is carried out at 37 C. for 30 minutes. The wells are rinsed 5 times with PBS, then 90 l of TMB buffer are added (containing the peroxidase substrate). Incubation is carried out at 37 C. for 30 minutes, then the TMB stop solution is added (100 l). The absorbance can subsequently be measured at 450 nm with a spectrophotometer. The protein concentrations will be deduced by linear regression or power regression relative to the ranges provided.

Example 5: Studies of the Markers LIFR, TTR, sFLT1, sENG, PIGF, of the sFLTt1/PIGF and sLIFRsFLT Ratios in Pregnant Women During the First Trimester of Pregnancy (<12 Weeks)

Materials and Methods

[0087] Human samples: 150 samples of human plasma were obtained in collaboration with Prof. Stefan Hansson's team (Lund, Sweden). The samples were collected during the first trimester of pregnancy from patients after signing of informed consent at the hospital of Lund. After analysis, 50 control samples and 100 samples from patients suffering from pre-eclampsia were included in the study.

Biochemical Analysis

[0088] The plasma samples were diluted in a diluent compatible with the ELISA technique in a 96-well microtitration plate. The sFLT1, sLIFR, sENG, PIGF and TTR ELISA kits were purchased respectively from Origene, BosterBio, NetBiotech, Cohesion Biosciences and NetBiotech. The ELISA assays were carried out according to the manufacturer's general recommendations as detailed below.

[0089] For each ELISA assay, 10 l of plasma were diluted in 90 l of diluent, except for the TTR ELISA for which the plasma samples were diluted to 1/10000, then the diluted samples were analyzed on 96-well plates in which antibodies capable of recognizing the protein of interest (sFLT1, sLIFR, sENG, PIGF or TTR) are fixed, for 1 h 30 to 2 h at 37 C. The liquid was subsequently removed from the wells and a secondary antibody capable of recognizing the protein of interest was added (100 l, diluted to 1/100) then incubated for 1 h at 37 C. The samples were then washed 3-4 times with 300 l of 0.01 PBS or a specific washing buffer. After each washing step, the liquid was removed and the plate turned over on a piece of absorbent paper, while avoiding the samples drying out. Then, a second anti-Fc antibody coupled to HRP (horseradish peroxidase) was added into the wells (100 l, diluted to 1/100) then the plate was incubated for 30 min at 37 C. The samples were subsequently washed again with 0.01 PBS or a specific washing buffer 5 times, and on each washing the liquid was removed and the plate turned over on a piece of absorbent paper. Finally, the HRP substrate was added into the wells (90 l), which led to blue staining proportional to the amount of protein of interest present in the sample. In parallel, 28 wells were used to prepare standard samples for each of the proteins of interest, which served as corresponding calibration ranges (from 10 ng to 156 pg of protein of interest, and one blank). Finally, an acid stop solution (50 l or 100 l, depending on the instructions in the user's manual) was added into each of the wells, which led to a color change from blue to yellow of proportional intensity. The plate was subsequently automatically read by a plate reader (Berthold) set to 450 nm. The results were stored in an Excel file.

Mathematical Analysis

[0090] A calibration curve was produced by regression analysis (linear or non-linear to obtain the optimal adjustment evaluated by the coefficient of determination, R.sup.2) from standard samples, then the concentrations of protein of interest of each plasma sample were estimated by applying the regression curve and taking into account the dilution factor. The results of the analysis of the plasma samples were then analyzed by parametric tests (Student's t-tests) and non-parametric tests (Mann-Whitney), using Excel's Statist'XL software. In addition, the grouping of the markers was analyzed using a hierarchical tree classification tool from Statist'XL from the inter-marker correlation matrix. The ROC curves were calculated with the software developed by John Eng on the website www.rad.jhmi.edu/jeng/javarad/roc/JROCFITi html.

Results

[0091] The results are presented in FIGS. 5 to 10.

[0092] FIG. 5: the markers PIGF and sENG (markers of the prior art) were tested on the 150 samples of human plasma. The concentration of the PIGF marker is decreased in the PE (pre-eclampsia) samples relative to the control (CTL) samples. The concentration of the sENG marker is increased in the PE (pre-eclampsia) samples relative to the control (CTL) samples. These results are therefore consistent with that which has already been described in the literature for the markers PIGF and sENG. On the other hand, the markers PIGF and sENG for diagnosing pre-eclampsia did not appear to give statistically significant differences. Nonetheless, one sample appears to be problematic (outlier) in the CTL cohort (control). By removing this sample, the diagnosis of pre-eclampsia with the sENG marker becomes statistically significant.

[0093] FIG. 6: the concentration of the TTR marker is significantly decreased in parametric tests and non-parametric tests in the PE (pre-eclampsia) samples relative to the control (CTL) samples. However, we observed overlaps.

[0094] FIG. 7: the concentration of the sFLT1 marker is significantly increased in parametric tests and non-parametric tests in the PE (pre-eclampsia) samples relative to the control (CTL) samples. This marker gave excellent results in the detection of pre-eclampsia with a mean factor of 6-fold more in the disease cases.

[0095] FIG. 8: the concentration of the sLIFR marker is significantly increased in parametric tests and non-parametric tests in the PE (pre-eclampsia) samples relative to the control (CTL) samples. This marker proved to have extremely good performance in the detection of pre-eclampsia, in particular after elimination of the control sample behaving as an outlier (the same as for sENG and PIGF). The mean difference was of the order of 6-fold.

[0096] FIGS. 9 and 10: the curves combining two markers show an excellent result of the synthetic markers sFLT1/PIGF and sFLT1sLIFR. For example, values of the order of 10-fold greater in the PE group, and a very good predictive power of the occurrence of pre-eclampsia for sFLT1sLIFR. The area under the ROC curve (0.87) was maximum relative to the isolated markers or in other combinations.