Purified LHR29 and LHR74 antibodies were initially provided by Dr Hugues Loosfelt (INSERM, France) and subsequently the antibody producing clones were obtained from ATCC (Clone ID CRL-2685 and CRL-2686). Antibodies produced in mouse ascites fluid were purified by Protein A affinity chromatography. The PG732 is a goat polyclonal antibody raised against a 19 residue LHCGR peptide (LHCGR 209–227; Swissprot: locus LSHR_HUMAN, accession P22888). The same peptide was used for affinity purification of the antibody from ammonium sulphate precipitated PG732 immune goat serum and was also used to produce mouse monoclonal antibody clone 5A10C9. Therefore, 5A10C9 is a monoclonal version of the PG732 goat anti-LHCGR polyclonal antibody; LHR H-50 (rabbit) and LH K-15 (goat) antibodies against the N-terminus and internal region of human LHCGR respectively, were from Santa Cruz Biotech, USA; anti-hCGbeta monoclonal antibody (clone 094–10627) was from Acris, Germany; anti-FLAG peptide monoclonal antibody was from Sigma-Aldrich (USA). A variety of monoclonal antibodies against hCG or hCGbeta from various commercial sources were also tested during the course of assay development. Antibodies were conjugated using a Lightning Link horse-radish peroxidase (HRP) conjugation kit (Innova Biosciences, Cambridge, UK) and stored at 4°C for up to five months.
Expression, cell extraction and affinity purification of recombinant LHCGR peptides in CHO cells, and western blotting of recombinant LHCGR and hCGbeta-LHCGR protein standards for ELISA
Details of the construction of LHCGR cDNA clones, transfection of Chinese hamster ovary (CHO) cells and expression of the recombinant proteins have been recently described . Briefly, three LHCGR recombinant proteins with 3X FLAG epitope tagged at the C-termini and containing 229, 291 and 318 amino acids of the LHCGR ECD were independently expressed in CHO cells in suspension culture. Following 48h of transfection, the recombinant proteins were extracted either with lysis buffer (25 mM Hepes pH 7.5, 150 mM NaCl, 1% igepal CA-630 [Sigma-Aldrich], 10 mM MgCl2, 1 mM EDTA, 25 mM NaF, 1 mM Na3VO4, and EDTA-free protease inhibitor mix [Sigma-Aldrich] or M-PER reagent [Perbio, Helsinborg, Sweden]. For western blot analysis of the extracts, anti-FLAG, LHR29, LHR74 primary antibodies were diluted at a concentration of 1 μg/mL; PG732 and LHR-H50 were diluted 1 in 5000 and 1 in 2,000, respectively. Both Triton and M-PER lysed extracts expressing LHCGR291 were used for the development of the ELISA assays; however, Triton-lysed extracts were consistently better than the M-PER extracts. The recombinant protein (LHCGR291) was affinity purified using anti-FLAG M2 affinity column according to the protocol provided by the vendor (Sigma Aldrich, USA). The preparation of placental extracts from early pregnancy (12 wks) for western blots were as described .
The sLHCGR and yoked hCG-sLHCGR complex protein standards for ELISA
In order to establish the specificity of the sLHCGR ELISA assay, the LHCGR291 recombinant and mock (control) transfected CHO cell extracts were initially employed. The LHCGR291 recombinant protein, which contains the hCG binding site and binds hCG in plate assays, was routinely used to generate standard curves with capture-detection antibodies as described below. However, the quantitative yield from transfected CHO suspension cell culture following anti-FLAG affinity purification was low (<600 μg/L). Therefore, following initial functionality tests in ELISA, LHCGR ECD was subsequently produced in bulk via a bacterial expression system. Other laboratories had shown that the expression of soluble LHCGR (sLHCGR, N-terminal 336 residues of the extracellular domain) as a thioredoxin fusion protein in E. coli carrying mutations in both thioredoxin reductase (TrxB) and glutathione reductase (gor) genes , had a similar specificity and affinity for hCG as the intact native LHCGR . Therefore, this protocol was followed in order to produce the sLHCGR standard calibrator for ELISA assays. The protocol for expression of the sLHCGR fusion protein and affinity purification through Ni-NTA resin column (Qiagen) were exactly as described . The estimated molecular mass of the fused sLHCGR was 57.54 K with pI of 6.15. The affinity purified protein was >90% pure and the yield varied from 6.63 to 7.34 mg/L.
For the production of the hCG-sLHCGR standard calibrator we employed a different method that aimed to preserve natural eukaryotic modifications of the hCG moiety of the final yoked protein. Previous studies  had shown that a tethered single chain hCG and LHCGR cloned in baculovirus and expressed in insect cells was functional with respect to ligand-receptor interaction. Moreover, the yoked hCG-LHCGR ECD is secreted from insect cells at levels 20-fold higher than conventional eukaryotic expression systems . Our goal was to produce yoked hCG-LHCGR single chain protein containing the epitopes recognized by both hCGbeta and LHCGR antibodies. Therefore, the open reading frame encoding the entire 165 amino acids of hCGbeta was synthesized (ACCESSION AK291552; CGB165). A linker sequence encoding the hCGbeta C-terminal peptide (CTP, constituting amino acids 116–145 of hCGbeta) was ligated at the 3’ end of the above construct (CGB165-CTP). A cDNA clone encoding 115 to 291 amino acid residues of the N-terminal end of LHCGR ECD was produced. The CGB165-CTP was ligated to the 5’-end of the modified LHCGR cloned into p3XFLAG-CMV-14 vector to create 101235–1 clone. The hCGbeta-LHCGR complex was first expressed in transfected CHO cells. The specificity of the yoked hCGbeta-LHCGR protein was established by testing anti-LHCGR, anti-hCGbeta and anti-FLAG monoclonal antibody binding of the recombinant and mock transfected CHO extracts in plate assays as well as by western blotting. Following these initial functionality tests in ELISA, the yoked protein was subsequently produced in recombinant baculovirus transfected insect cells.
For baculovirus expression, the cDNA encoding hCGbeta-CTP-LHCGR (clone 101235–1) with C-terminal 3XFLAG was transferred to a baculovirus vector, DH10Bac strain was used for the recombinant bacmid (rbacmid) generation. The positive rbacmid containing 3xFLAG tagged 101235–1 was confirmed by PCR and the final clone was characterized by DNA sequencing. The rbacmid was transfected into an sf9 insect cell line, using Cellfectin, incubated in SF-900 liquid medium for 5–7 days at 27°C. The supernatant was collected and designated as P1 viral stock. P2 was amplified for later infection. The results of expression evaluation by western blot indicated that the target protein was expressed at the expected relative molecular mass. The recombinant protein was purified from the supernatant by loading on Flag M2 affinity gel, and the beads were eluted with TBS (50 mM Tris–HCl, 150 mM NaCl, pH7.4) containing 200 ng/μl peptide (FLAG: N-Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys-C relative molecular mass, 1013.0). The expression and purification results indicated that the target protein was largely released from the cells into the supernatant. The yield of the yoked hCGbeta-LHCGR protein was 0.42 mg/L with an estimated purity of 70%. The affinity purified protein was stored in 20% glycerol at −20°C.
Ninety-six-well plates (C-well binding capacity 500ng/well, 8-strip, polystyrene; Greiner Bio-one, Germany) were coated with 100 μl per well of LHCGR transfected CHO extracts (for initial ELISA functionality tests) or with 3 μg/mL affinity purified LHR29 (sLHCGR assay) or 5A10C9 (hCG-sLHCGRassay) diluted in 50 mM Carbonate/Bicarbonate buffer pH 9.4 (Thermoscientific/Pierce, UK) at room temperature (RT) overnight. Following removal of the antibody, plates were over-coated with 300 μl per well of 10mM sodium/potassium phosphate buffer at pH 7.6 containing 5% sucrose (Fluka, UK) and 0.5% bovine serum albumin (Sigma-Aldrich, UK) for two hours at RT, before removal and drying overnight at RT. The pre-coated ELISA plates were wrapped and stored at RT and were stable for six months of use. On the day of use, plates were blocked for one hour at RT with 100 μl per well of phosphate buffered saline pH 7.2 (Thermoscientific/Pierce) containing 1% (v/v) casein concentrate (sdt reagents, Germany) prior to adding serum, antigen or standards diluted in 25 mM Bicine (Fluka), 50 mM Tris pH 7.8, 170 mM NaCl and incubation at RT for 2 h. Following binding, plates were washed six times with 300 μl per well 2 mM Tris Cl pH 7.8, 150 mM NaCl, 0.05% Tween 20 prior to incubation with HRP-conjugated antibodies (LHR29, 5A10C9 or anti-hCGbeta) diluted in Immunoshot 2 reagent (Cosmo Bio Co. Ltd, Japan) for 1h. The antigen-binding was detected by adding TMB (3,3′,5,5′-tetramethylbenzidine) substrate (Thermoscientific/Pierce, UK) and the color reaction was stopped by adding an equal volume of 1N HCl. Plates were read at 450-620 nm in a standard plate reader. Data were transferred to Microsoft Excel prior to analysis as described below.
Patient serum samples
The major aim of this study was a prospective examination of the association of serum sLHCGR and hCG-sLHCGR concentrations at early human pregnancy (first trimester) with adverse pregnancy outcome. The study was approved by REC West Midlands, as part of National Research Ethics Services of NHS. Patient information and a patient consent form were given to each patient. As a standard Down’s syndrome screening requirement, pregnancies at 10–14 wks of gestation underwent ultrasound examination as well as nuchal scanning where indicated. A combined trisomy screen comprising biochemical analysis of free hCGbeta and PAPP-A was performed on all serum samples. As part of this standard screening, an aliquot of each serum sample from consenting patients, was stored at −20°C for further analysis of sLHCGR and hCG-LHCGR concentrations. A portion of the study was retrospective with regards to Down’s syndrome; 30 known T21 and 130 control samples (collected from 2006 to 2009) were obtained from the Fetal Medicine Foundation (FMF), UK. Unlike PS, the samples for the retrospective study (RS) were collected in a referral hospital designated for screening high risk fetal aneuploidy. Therefore, the samples for retrospective study obtained from FMF belonged to a high risk pregnancy group.
The Analysis ToolPak (ATP) software was used to compute means, standard deviation (SD), variance (anova) and coefficient of variation (CV) for all data sets. For each ELISA assay, standard curve was generated following examination of the natural log plot of the optical density (OD) at 450-620 nm for all the standard points, following removal of the background diluent signal, to ensure that the points form a straight line. Using the scatter plot function, a standard curve was created using picomoles per mL of the standard as Y-value and OD (450-620 nm) on X-axis. By adding a line of best fit through zero, the equation and regression of a straight line (y = mx), representing the exponential, logarithmic portion of a standard curve, was generated. Generally, the regression, R2 >0.98, was considered valid. This standard curve and the dilution factors were used to measure the analyte concentration in a sample. Correlation testing was performed using the Pearson product moment method (standard R package). General file manipulations and data cleaning were implemented using the Awk programming language or custom programs written in Python. The graphical display of the distributions of data from control and pathological pregnancies were carried out using the ggplot2 package (R statistical software environment). The detection rates for pathological pregnancies were calculated as the proportion of pathological data points found to lie within the critical regions defined by the cut-off values set for two analytes. The false positive rate was calculated as the proportion of all control data points which were found within the critical regions defined by the cut-off values. The statistical significance of the difference in median values for hCGbeta and hCG-sLHCGR between control and pathological pregnancies were calculated using a Wilcoxon signed rank test in the R statistical software package since the data showed a strongly non-gaussian distribution.