Dr P. Garred, Tissue Typing Laboratory-7631, Department of Clinical Immunology, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. E-mail: email@example.com
Ficolin-2 (L-ficolin), derived from the FCN2 gene, is an innate immunity pattern recognition molecule found in human serum in which inter-individual variation in serum appears to be under genetic control. To validate and extend this finding, we developed a sandwich ELISA for detection of human Ficolin-2 in serum samples and identified FCN2 genotypes with a Taq Man-based minor groove binder assay and by sequencing. Serum samples were applied to gel-permeation chromatography and fractions were analysed by an ELISA, SDS-PAGE and subsequently Western blotting. In 214 Danish blood donors, the median Ficolin-2 serum concentration was determined to 5.4 μg/ml (range: 1.0–12.2 μg/ml). An ELISA, SDS-PAGE and Western blot analysis of gel-permeation chromatography fractions showed that Ficolin-2 comprises a mixture of covalently and non-covalently linked Ficolin-2 oligomers independent of the individual genotypes. The variation in serum concentration was associated with three polymorphisms in the promoter and one polymorphism in the structural part of the FCN2 gene. Further analysis indicated that two particular alleles on the same haplotype determined a low Ficolin-2 concentration. Our results show that inter-individual variation of Ficolin-2 concentration is associated with polymorphisms in the promoter and the structural part of the FCN2 gene.
Ficolin-2 also named L-ficolin consists of collagen-like triple helices, which are further oligomerized giving an overall bouquet-like structure . The Ficolin-2 polypeptide consists of an N-terminal collagen-like domain, a neck region, and a C-terminal fibrinogen-like (FBG) carbohydrate recognition domain.
Upon binding to distinct pathogen-associated molecular patterns (PAMP), such as carbohydrates (GlcNAc), lipoteichoic acid and acetylated groups, Ficolin-2 may facilitate phagocytosis and activation of complement through the lectin route using the same serine proteases as mannose-binding lectin (MBL) named MASP [2–7]. Collectively, these observations suggest that Ficolin-2 may have a role in innate immunity.
Ficolin-2 is derived from the FCN2 gene located on chromosome 9q34 and is predominantly expressed in the liver  and show sequence homology with Ficolin-1 (M-ficolin) and Ficolin-3 (H-ficolin or Hakata antigen) . It has previously been shown that the Ficolin-2 is present in serum with varying concentrations [10–13]. Ficolin-2 deficiencies have not been reported, but low Ficolin-2 serum levels have been associated with recurrent respiratory infections in children . On the other hand, it has also been reported that Ficolin-2 may contribute to the pathophysiology of IgA-mediated nephropathy and activate complement on necrotic cells [5, 15]. Recently, it has been shown that the FCN2 gene and its promoter are highly polymorphic [12, 16]. Promoter polymorphisms are associated with variation in Ficolin-2 serum levels and polymorphisms in the structural part of the gene influence the affinity of Ficolin-2 to GlcNAc .
Because the knowledge of Ficolin-2 is limited, we wanted to develop an ELISA for quantification and characterization of Ficolin-2 in serum, which may be used both in experimental and clinical situations. Moreover, we wanted to extend the observation that allelic variation in the promoter region influences the Ficolin-2 serum level.
Materials and methods
Peripheral venous blood samples were obtained from 224 unrelated adult Danish blood donors with consent. From 10 individuals serum was not available. The local ethical committee approved the study. Genomic DNA was prepared from each blood sample using the method described by Miller et al. .
Recombinant Ficolin-2 (rFicolin-2) was produced as described . Briefly, the FCN2 gene was amplified from liver cDNA and tagged C-terminally with a penta-His sequence. This construct was cloned into an expression vector and transfected into a Chinese hamster ovary (CHO) cell line, DG44. Subsequently, Ficolin-2 was purified from the cell supernatant on a Ni-NTA column (Qiagen Nordic, West Sussex, UK). Using the same procedure recombinant Ficolin-1 and Ficolin-3 were also produced and used as specificity controls.
Generation of monoclonal antibodies.
Female NMRIxBALB/c F1-mice were immunized with rFicolin-2. Fusion of spleen cells was performed essentially as described by Reading . Cells tested positive by an ELISA was cloned by dilution.
A total of 35 positive clones against rFicolin-2 were analysed with respect to cross-reactivity of the highly homologues Ficolin-1 and Ficolin-3. Microtiter plates (Maxisorb; Nunc, Roskilde, Denmark) were coated with rFicolin-1, -2 or -3 (0.5 μg/ml) in phosphate-buffered saline (PBS) (LAB79360; Bie & Bernsten, Roedovre, Denmark) and incubated at 4 °C overnight. The wells were washed three times with PBS containing 0.05% Tween 20 (PBS-T) (15797-1; VWR International APS, Roedovre, Denmark). After washing, 100 μl of hybridoma cell culture supernatant 1:1, 1:50 or 1:100 in PBS-T were added to the wells and incubated for 4 h at 37 °C and the plates were washed. Horseradish peroxidase (HRPO)-conjugated rabbit anti-mouse antibodies (P0260; DAKO Cytomation, Glostrup, Denmark) 1:5000 in PBS-T was added and the plates incubated for 1 h at 37 °C. After a final wash, the plates were developed with OPD substrate solution (to 12 ml distilled H2O, four OPD tablets (S2045, DAKO Cytomation) and 5 μl of H2O2 were added). After 15 min of incubation, the reaction was stopped by adding 100 μl of H2SO4, and the optical density (OD) was measured at 490 nm. A negative control without coating with antigen was included.
Microtiter plates were coated with rFicolin-2 (0.5 μg/ml) in PBS buffer by incubating at 4 °C overnight. The wells were washed three times with PBS-T. After washing, 100 μl of a mixture of biotinylated protein-A purified anti-Ficolin-2 monoclonal antibody FCN219 (0.5 μg/ml in PBS-T) and hybridoma cell culture supernatant from antibody-producing hybridomas in dilution series (1:1–1:100) was added to the wells and incubated for 4 h at 37 °C. After washing, peroxidase-conjugated streptavidin (RPN1231V; GE Healthcare Bio-Sciences, Hilleroed, Denmark) diluted to 1:5000 in PBS-T was added, and incubated for 1 h at 37 °C. After a final wash, the plate was developed as described above. A negative control without coating with antigen was included.
To assess the hybridoma supernatant concentration of antibodies, supernatant was incubated for 4 h at 37 °C on plates coated with rFicolin-2 and subsequently detected with HRPO-conjugated rabbit anti-mouse antibodies (P0260, DAKO Cytomation) 1:5000 in PBS-T. The plate was developed as described above.
Isotyping, purification and biotinylation of the antibodies.
Hybridoma cell lines were cultured in serum-free media (12045-076, Invitrogen, Taastrup, Denmark). Two monoclonal anti Ficolin-2 antibodies FCN216 and FCN219 were isotyped using the Mouse Monoclonal Antibody Isotyping Kit (ISO-1, Sigma Aldrich, Broendby, Denmark). The antibodies were purified using HiTrapTM rProtein A FF columns (GE Healthcare Bio-Sciences) according to manufacturer's guidelines. Purified antibodies were subsequently biotinylated according to instructions of the manufacturer (BK-101, Sigma-Aldrich).
Digestion of Ficolin-2 with collagenase and analysis by ELISA and Western blot.
rFicolin-2, dialysed into 10 mm CaCl2, 25 mm Tris–HCl, pH 7.4, was digested with collagenase type VII (C0773, Sigma-Aldrich) (enzyme substrate ratio 1:2 w/w) for 24 h at 37 °C. To determine to which domains of Ficolin-2 the epitopes recognized by the selected monoclonal anti Ficolin-2 antibodies, FCN216 and FCN219, are located, microtiter plates were coated with 1 μg per well collagenase-digested rFicolin-2 and detected with 0.2 μg per well biotinylated FCN216 or FCN219 antibodies. Further, collagenase-digested rFicolin-2 was subjected to SDS-PAGE and Western blotting as described later.
Ficolin-2 immunopurification and depletion from serum.
Ficolin-2 was immunopurified by incubating 100 μl of the PNHS in six polystyrene wells coated with FCN216 overnight at 4 °C. After washing 50 μl of SDS-PAGE, sample buffer (LA0041, Invitrogen) were applied to the first well and after 10 min transferred to a new well and repeated seven times. Subsequently, the purified material was analysed on SDS-PAGE and Western blotting under non-reducing conditions as described below. Using the supernatant of 200 μl of PNHS incubated overnight with 50 μl of GlcNAc beads (A2278, Sigma-Aldrich) at 4 °C, which subsequently was spun at 1590 g for 5 min, Ficolin-2-depleted serum was produced.
SDS-PAGE and Western blot.
Samples were loaded onto a 3–8% Tris-Acetate-gels NuPAGE gradient gel (EA03752BOX, Invitrogen) under non-reducing conditions and then blotted to nitrocellulose (Hybond, ECL, RPN78D, GE Healthcare) using the Xcell II mini-Cell blot apparatus in NuPAGE transfer buffer (NP0006-1, Invitrogen). Biotinylated FCN219 was used as a detection antibody. Subsequent steps included incubation with HRPO-linked streptavidin and development was performed with SuperSignal West Femto Maximum Sensitivity Substrate (34095, Pierce) on autoradiographic films. As a molecular weight standard; the Precision Prestained Protein Standard (161-0373, BioRad) was used.
Microtiter plates (Maxisorb; Nunc) were coated with 100 μl of monoclonal anti-Ficolin-2 antibody (FCN216, 2.5 μg/ml) in sodium bicarbonate buffer 0.1 mol/l, pH 8.2 (LAB96520.1000; Bie & Berntsen) by incubating at 4 °C overnight. The wells were washed three times with PBS-T. Sera were diluted to 1:40 and 1:160 with PBS-T. A standard twofold dilution series (1:10–1:1280) of the pooled normal human serum (PNHS) in PBS-T was added to the plates in duplicates. A positive control with rFicolin-2 (4 μg/ml) and a negative control without rFicolin-2 were included and the plate was incubated for 3 h at 37 °C. After washing 100 μl of biotinylated FCN219 (2.5 μg/ml) in PBS-T was added to the wells and incubated at 4 °C overnight. HRPO-conjugated streptavidin diluted to 1:5000 in PBS-T was added, and incubated for 1 h at 37 °C. After washing, the plate was developed as described above.
Determination of the Ficolin-2 concentration in serum.
The Ficolin-2 content in the PNHS was defined by correlating a triplet of PNHS (1:40) to a standard curve consisting of Ficolin-2-depleted serum (1:40) spiked with rFicolin-2 in a serial dilution. In addition, rFicolin-1 and-3 were used as specificity controls.
Limit of quantification.
Pooled normal human serum was depleted of Ficolin-2 using GlcNAc beads as described above. A serial dilution of rFicolin-2 was added to Ficolin-2-depleted serum (final dilution of 1:10 in PBS-T) in duplicates. The lower limit of quantification was determined as the lowest measurable Ficolin-2 concentration with signal >background + 3 SD.
The intra-assay variation was determined by analysing 15 replicates on the same plate. For assessments of inter-assay variation, duplicates in five separate assays were analysed. The mean values, standard deviation (SD) and coefficient of variation [CV = (SD/mean) × 100%] were calculated.
Plasma and serum samples from three normal healthy individuals were collected and pooled. Aliquots were incubated for up to 24 h at either 4 °C, room temperature or 37 °C. After incubation, the aliquots were stored at −20 °C until measurement of Ficolin-2. In addition, aliquots were subjected to three freeze–thaw cycles where samples were frozen for at least 20 h and thawed at room temperature for an hour. Each sample was analysed in triplicates.
Detecting anti-mouse antibodies in serum samples.
Serum samples may contain heterophilic anti-animal antibodies or rheumatoid factors reacting with the capture antibody and subsequently with the detection antibody thus giving rise to erroneous high signals . To avoid this, we performed a control assay in which the capture antibody was replaced with a mouse monoclonal antibody of identical subclass and light chain type in the same concentration with no known specificity. Samples found positive in the control assay were retested with addition of 5 μg/ml of a mouse antibody with an irrelevant specificity but identical isotype to remove the unwanted reactivity.
Serum samples (100 μl) from a total of six individuals with low (1.4 and 2.2 μg/ml), medium (4.8 and 5.5 μg/ml) or high (10.4 and 10.6 μg/ml) level of Ficolin-2 were subjected to chromatography on a Superdex 200 HR 10/30 column (Amersham Bioscience) connected to automated liquid chromatography system (ÄKTA purifier 900; Pharmacia Biotech, Uppsala, Sweden). PBS-T was used as eluate and fractions of 0.5 or 1 ml was collected and quantified in the Ficolin-2 ELISA. The Bio-Rad gel filtration standard (Bio-Rad, Herlev, Denmark) was used to estimate the molecular weight proteins in eluted fractions. Gel filtration fractions were investigated by SDS-PAGE and Western blot as described above. However, to increase yield, 3 μl of StrataClean resin (cat.nr. 400714, Stratagene; AH diagnostics, Aarhus, Denmark) was added and the mixture was vortexed for 1 min. The samples were spun for 1 min at 700 g and the pellet was resuspended in SDS-PAGE sample buffer.
Allelic discrimination using minor groove binder probe assay.
Single-nucleotide variants were detected using Taq Man MGB probe assay (TaqMan® MGB assay, Applied Biosystems, Foster City, CA, USA). The assay uses a dual-labelled fluorescent probe where the 5′-end is labelled with a reporter dye, VICTM or FAMTM, and the 3′-end is labelled with a quenching dye (Table 1). During the polymerization step of PCR, the exact matching hybridization probe remains bound for strand displacement, resulting in efficient probe cleavage and release of the reporter dye. Mismatch between probe and target greatly reduces the efficiency of probe hybridization and cleavage due to the MGB protein coupled to the probe. To each allele-specific reaction, 12.5 μl of TaqMan Universal PCR Master Mix, 0.6, 40× Custom Assay Mix and 20 ng of genomic DNA diluted in 11.9 μl of dH2O was added to a Termofast 96-well 0.2-ml PCR plate (ABgene, Albertslund, Denmark) that was sealed using Optical Caps (ABgene). Four controls were added for every 44 reactions: two no-template controls and two allelic controls containing DNA with either A allele or B allele. PCR was performed under the following conditions: polymerase activation for 10 min at 95 °C, 40 two-step cycles consisting of: 15 s of denaturation at 92 °C followed by 1 min of annealing and elongation at 60 °C. Results were analysed on ABI PRISM 7700 Sequence Detection platform using the SDS software v1.9 (Applied Biosystems), and using the allele discrimination plate read function to detect the end-point fluorescence in each well. Genotype results were manually assigned.
Table 1. Taq Man MGB-probes and primers.
*VIC and FAM code for the reporter fluorophores. Each probe carries a 3′-end dark quencher and a MGB.
Each sequence was amplified using a single primer set (Table 2), where the forward primers contained a 5′-T7 sequence tag (5′-taatacgactcactataggg-3′). For universal sequencing, PCR amplifications were carried out in 20-μl volumes containing: 50 ng of genomic DNA, 0.25 μm of each primer, 2.5 mm MgCl2, 0.2 mm dNTP, 50 mm KCl, 10 mm Tris–HCl, pH 8.4, and 0.4 units of Platinum Taq DNA polymerase (Invitrogen). The PCR reactions were performed at the following cycling parameters: 2m 94 °C, 35× (30s 94 °C, 60s 58 °C, 60s 72 °C), 5m 72 °C and were sequenced in forward direction using the ABI BigDye cycle sequencing terminator kit (Applied Biosystems). Sequence reactions were purified using Dynabeads (G 210.010, Dynal, Taastrup, Denmark) and sequence analysis was performed on an ABI Prism 3100 Genetic Analyser (Applied Biosystems). The resulting DNA sequences were aligned using BioEdit software, and DNA polymorphisms were confirmed visually from sequence electropherograms.
The forward primers all contain a 5′-T7 sequence (5′-taatacgactcactataggg-3′).
FCN2+6359C>T, +6424G>T, +6442_6443delCT>A
Hardy–Weinberg equation was analysed using simple gene counting and the chi-squared test with Yates’ correction. Linkage disequilibrium and estimated haplotypes were estimated using the SNPAlyze programme version 4.1 (Dynacom, Yokohama, Japan). Non-parametric Kruskall–Wallis for unpaired group comparison were used to evaluate the promoter SNP on the Ficolin-2 serum concentration.
Specificity of the monoclonal antibodies
After immunization of female mice with rFicolin-2, we obtained six clones of 35 that were highly specific for Ficolin-2 without cross-reactivity with rFicolin-1 and -3 (data not shown). The FCN219 hybridoma cell line was purified and biotinylated for use in the following analyses. To select pairs, antibodies with non-overlapping epitopes in the Ficolin-2 ELISA, an inhibition assay was carried out: rFicolin-2 was adsorbed to polystyrene plates and biotinylated FCN219 competed with hybridoma culture supernatant from the five other hybridomas for antigen binding. FCN219 binding was partly inhibited with several of the supernatants; however, for the FCN216 cell supernatant, no inhibition was detected. In Fig. 1A, it is illustrated using three clones that the binding of biotinylated FCN219 is inhibited by culture supernatants from clones, FCN219 and FCN235, but not with clone FCN216. Therefore, we decided to further characterize the two clones, FCN216 and FCN219, for use in the construction of a quantitative Ficolin-2 sandwich ELISA.
To disclose the epitope region on the Ficolin-2 protein, rFicolin-2 was collagenase digested and subsequently subjected to Western blotting. Further, the digested rFicolin-2 was absorbed onto microtiter wells for detection with FCN216 or FCN219 (Fig. 1B). Even though FCN216 exhibit low affinity for Ficolin-2 in Western blotting, both Western blotting and the ELISA revealed that only FCN219 binds to collagenase-digested Ficolin-2 implying that FCN216 and FCN219 epitopes are located on different regions of Ficolin-2, the collagen-like (FCN216) and the fibrinogen-like domain (FCN219) respectively.
Construction and validation of the Ficolin-2 sandwich ELISA
Different constructs of a sandwich ELISA based on coating with either FCN216 or FCN219 as capture antibody and signal development with either biotinylated FCN216 or FCN219 were assayed. The highest signal was achieved with FCN216 as capture antibody and biotinylated FCN219 as detection antibody and this antibody combination was subsequently used in the quantitative ELISA (data not shown). To determine the content of Ficolin-2 in the human serum pool dilutions of recombinant Ficolin-2 was added to Ficolin-2-depleted serum. The dilution series was added to a microtiter plate and compared with a dilution series of the PNHS. By using this method, the concentration of Ficolin-2 in the PNHS was determined to be 3.0 μg/ml. A standard sigmoid dose–response curve was obtained by adding a dilution series of the PNHS (1:10–1:1280).
Adding recombinant Ficolin-1 or recombinant Ficolin-3 in different concentrations to the assay did not influence the signal. By spiking Ficolin-2-depleted serum (diluted to 1:10) with rFicolin-2, the lower limit of quantification of Ficolin-2 in serum was estimated to be 0.01 SD ± 0.001 μg/ml. Because the Ficolin-2 serum levels exhibit large inter-individual variation, the serum samples were diluted to 1:40 and 1:160 in PBS-T prior to analysis and added to the plate in duplicates to have a dilution within the assay range. Intra-assay variation was determined by analysing 15 replicates on the same plate and the inter-assay variation was determined by analysis of duplicates in five separate runs. The coefficient of variation (CV%) was calculated as: (SD/mean) × 100%. The intra-assay variation was 6.9% and the inter-assay variation (CV%) was 13.0% respectively.
The stability of Ficolin-2 upon storage of serum and plasma at different temperatures was examined using pools of serum and plasma collected from three healthy individuals, which were frozen three times. Ficolin-2 is stable in EDTA, citrate and heparin plasma as well as in serum for up to 24 h at 4 °C, room temperature and 37 °C; however, when subjected to repeated freeze/thaw cycles, the Ficolin-2 concentration decreased approximately 20% after the second cycle. The assay may be used for determining Ficolin-2 in plasma as measurements of Ficolin-2 in EDTA, citrate and heparin plasma gave similar results as measurements of Ficolin-2 in serum.
Analysis of the oligomeric forms of Ficolin-2 detected in the ELISA by gel permeation chromatography and SDS-PAGE with subsequent Western blotting
Serum from individuals with different FCN2 genotypes and high, medium or low levels of Ficolin-2 was applied to gel permeation chromatography experiments. Ficolin-2 eluted in one bulk (fraction 7–10 ml) corresponding to a molecule weight of about 870 kDa based on the calibrated protein standard peaks (Fig. 2). Further, fractions 6–10 were subjected to SDS-PAGE and Western blotting (Fig. 3). Strong bands were detected in fractions 7 and 8 (lanes 3 and 4) and weaker bands in fractions 9 and 10 (lanes 5 and 6) corresponding to the higher oligomeric forms of Ficolin-2. This indicates that Ficolin-2 in serum primarily consists of higher order oligomers. For comparison, unfractioned serum and Ficolin-2 immunopurified from the PNHS with FCN216 were run on a separate gel. A ladder of bands with the same appearance as the gel filtrated Ficolin-2 was observed.
Using MGB-probe assay, 224 Danish blood donors were genotyped. For validation of the assay, a minimum of 60 samples were sequenced for each SNP. For four SNP, MGB-probe assays were unavailable due to sequence limitations (FCN2-64A>C, FCN2+6359C>T, FCN2+6424G>T, and FCN2+6442_6443delCT>A). For these SNP, genotypes were obtained by sequencing. Examples of results acquired with the MGB-probe assay are shown in Fig. 4. For each individual sample, VIC-fluorescence was plotted against FAM-fluorescence. In FCN2-986 three separate clusters could be distinguished, having only VIC-fluorescence (G/G alleles), only FAM-fluorescence (A/A alleles) or both VIC and FAM (G/A alleles). Three separate clusters were likewise observed in FCN2-602; however, some background was observed as well, despite attempts to optimize the assay. Genotyping results of the FCN2 promoter region (n = 215) and exon 8 (n = 224) are shown in Table 3. The observed frequencies for the SNP in the promoter were in agreement with previous reported findings in Danish White people; however, the exon 8 SNP displayed a higher frequency than previously reported . All SNP adhered to the Hardy–Weinberg expectations (P > 0.05). Haplotype analysis of the FCN2 promoter SNP including the SNP in exon 8 showed five common haplotypes with frequencies ≥5% (Fig. 5). The observed haplotypes accounted for 79.7% of all chromosomes investigated.
Table 3. FCN2 SNP frequencies.
Amino acid change
*The numbering indicates the nucleotide position relative to the ATG start site.
**Reference (rs) ID from NCBI.
N, total number of individuals; AA, major type homozygote; AB, major/minor heterozygote; BB, minor homozygote; p, allele frequencies; del, deletion mutation.
Based on the newly developed Ficolin-2 ELISA, Ficolin-2 serum concentration was determined. The serum levels ranged from 1.0 to 12.2 μg/ml, with a median and mean value of 5.4 μg/ml. The Ficolin-2 concentration was compared with age and gender, but no correlation was observed. However, as previously reported with another set of antibodies in other Danish samples , the variation in the Ficolin-2 serum concentration was significantly associated with SNP in the promoter region of FCN2 (Fig. 6), at positions −986 (P < 0.0001), −602 (P < 0.0001) and −4 (P = 0.0162), while a borderline significance was observed for the SNP at position −557 (P = 0.0535) and no association with the SNP at position −64 (P = 0.3136). No significant association was seen with the amino acid substituting SNP at position +6359 (P = 0.3188). However, the Ficolin-2 serum level was significantly associated for the amino acid substituting SNP at position +6424 (P = 0.0003), which has not been observed previously (Fig. 6). For +6442 variant only one allele was identified; hence, no statistical analysis could be performed. The genotypes that were significantly associated with variation in Ficolin-2 serum concentration revealed a phenotype that was gene-dose dependent, i.e. homozygotes did either have the highest or lowest concentration, while heterozygotes had intermediate concentrations. Based on the Ficolin-2 serum variation, each SNP seemed to explain a twofold variation in the serum concentration.
The four SNP that correlated with the Ficolin-2 serum level (−986A>G, −602G>A, −4A>G and +6424G>T) were used for haplotype analysis excluding those at positions−557G>A, −64A>C and +6359C>T. Only individuals homozygous for all the four SNP were used in the analysis. In Fig. 7, the serum levels are displayed according to the haplotype. The presence of the −986A allele gives rise to higher Ficolin-2 levels than when the G allele is present and when both the −986G allele and the +6424T allele are present on the same haplotype, the serum level is low. Linkage analysis shows that the minor allele at position +6424 (T) segregates with the major alleles at positions −602 and −4 (G and A respectively) but does not in particular segregates with any of the −986 alleles. The degree of linkage disequilibrium between the different SNP is shown in Table 4.
Table 4. Pairwise linkage disequilibrium (expressed as D′).
Pairwise linkage disequilibrium (LD) coefficients for seven FCN2 SNP. LD ranging from −1.0 to +1.0 is expressed as D′ and is calculated with genotype data from 214 individuals. Positive values indicate that the minor alleles at each locus segregate together. Negative values indicate the minor allele at one locus segregate with the major allele at the other locus. *P < 0.05, **P < 0.001.
The developed Ficolin-2 ELISA was highly sensitive and specific. Gel permeation chromatography of serum followed by analysis of eluted fractions with the Ficolin-2 ELISA showed that Ficolin-2 eluted at a position corresponding to a molecular weight of 870 kDa suggesting that Ficolin-2 may partly circulate as a dimer of its basic 12-mer structure as has been shown using recombinant pig Ficolin-α  and recombinant human Ficolin-2 . However, the presence of lower oligomeric forms and even single polypeptide chains analysed from high molecular weight gel filtration fractions suggest that highly oligomerized Ficolin-2 in serum partly consists of polypeptides that are not covalently stabilized in the N-terminus of the protein. Thus, incomplete disulphide bonding may be a characteristic of the mature protein.
The new assay seemed to be very suitable to evaluate the content of Ficolin-2 in human serum. When we determined Ficolin-2 serum concentration from a total of 214 Danish blood donors, we found that it varied between 1.0 and 12.2 μg/ml with a median concentration of 5.4 μg/ml. This result is compatible with previous Ficolin-2 findings with an already established Ficolin-2 ELISA based on another set of monoclonal antibodies . The new assay may be applied to serum samples as well as to plasma samples (EDTA, citrate and heparin). Moreover, it is possible to wait several hours before samples are frozen. Thus, it seems reliable to be used in a routine setting and for epidemiological research. However, repeated freezing and thawing cycles indicated Ficolin-2 may be prone to some degree of in vitro degradation, and results obtained from samples from different freezing and thawing cycles should be interpreted with caution.
It has recently been shown that the FCN2 gene is polymorphic [12, 16] and that polymorphisms in the FCN2 promoter region are associated with variation in Ficolin-2 serum concentrations . To confirm and extend this finding we genotyped the investigated blood donors for eight polymorphisms located in the promoter and within the coding region of the FCN2 using sequencing and a TaqMan-based MGB-probe assay. Even though the conversion rate of the MGB assay was low (four of eight polymorphisms), and the assays proved difficult to optimize, the successful MGB assays were specific and reproducible. We found, in agreement with previous results, that three of the promoter SNP at positions −986, −602 and −4 all significantly correlated with either a twofold decrease or increase in the Ficolin-2 concentration in serum. However, the concentration in samples with different genotypes was overlapping suggesting that other factors contributes to the serum variation. As previously observed, no significant association with Ficolin-2 variation was seen for the SNP at position −64 and for two of the SNP causing amino acid substitutions in exon 8 of the FCN2 gene, +6359 and +6442 respectively . A borderline significance was observed for the SNP at position −557. However, this SNP was in linkage disequilibrium with the SNP at position −602, which probably explain the weak association with the variation in the Ficolin-2 serum concentration. However, we found a highly significant correlation between the serum level and the +6424 SNP (G to T), which has not been observed previously . The allele frequency of +6424 in this study (0.118) is equal to the frequency observed in Dutch Caucasians  but higher than we previously reported . The reason why we did not observe the effect of this SNP on the Ficolin-2 serum concentration in our primary study was probably due to the small number of individuals carrying the allele leading to a lack of power in the statistical analysis . It is of interest that this allele causing an alanine that is substituted with a serine at amino acid position 258 appears to increase the avidity of Ficolin-2 towards GlcNAc , while at the same time it is associated with a low Ficolin-2 serum concentration. Because Ficolin-2 may be important in endogenous waste disposal [3, 5], it may be speculated that increased avidity towards ligands could lead to a lowering of the Ficolin-2 concentration due to consumption. Other possibilities could be linkage disequlibrium to other FCN2 alleles or that the base substitution by itself affects the half-life of mRNA.
Trends of correlation of Ficolin-2 levels to haplotypes were observed. When the A allele at position −986 was present, higher levels of Ficolin-2 was observed on all haplotypes. Further, when the G allele at −986 was present, it is notable that in conjunction with the T allele at +6424 low serum levels were observed. Thus, a contributing effect from both alleles on the serum concentration on this haplotype cannot be excluded. However, an analysis of a much larger sample size is necessary to make firm conclusions on this issue.
In the present study, we have extended the observation that the inter-individual serum concentration of Ficolin-2 may have a genetic origin. Using a new highly specific Ficolin-2 assay revealed that the variation in Ficolin-2 serum concentrations was associated with three SNP in the promoter region and with one amino acid substitution SNP in exon 8 of the FCN2 gene. Thus, inter-individual variation in Ficolin-2 serum concentration is genetically determined. Further studies will reveal the clinical impact of these findings.
Excellent technical assistance was provided by Ms Vibeke Weirup and Vibeke Witved. Grant support was obtained from the Danish Medical Research Council, Danish Rheumatism Association, The Novo Nordisk Foundation, The Benzon Foundation, Copenhagen Hospital Corporation Research Foundation and Rigshospitalet. T.H. is a Copenhagen University research fellow.