A common variant of the pregnancy-associated plasma protein-A (PAPPA) gene encodes a protein with reduced proteolytic activity towards IGF-binding proteins

Pregnancy-associated plasma protein-A (PAPP-A) is a key regulator of insulin-like growth factor (IGF) bioactivity, by releasing the IGFs from their corresponding IGF-binding proteins (IGFBPs). The minor allele of the single nucleotide polymorphism (SNP), rs7020782 (serine < tyrosine), in PAPPA has previously been associated with recurrent pregnancy loss as well as with significant reduced levels of PAPP-A protein in human ovarian follicles. The aim of the present study was to reveal a possible functional effect of the rs7020782 SNP in PAPPA by comparing recombinant PAPP-A proteins from transfected human embryonic kidney 293 T cells. The proteolytic cleavage of IGFBP-4 was shown to be affected by the rs7020782 SNP in PAPPA, showing a significantly reduced cleavage rate for the serine variant compared to the tyrosine variant (p-value < 0.001). The serine variant also showed a trend towards reduced cleavage rates, that was not significant, towards IGFBP-2 and IGFBP-5 compared to the tyrosine variant. No differences were found when analysing cell surface binding, complex formation between PAPP-A and STC2 or proMBP, nor when analysing STC1 inhibition of PAPP-A-mediated IGFBP-4 cleavage. Regulation of IGF bioactivity in reproductive tissues is important and the rs7020782 SNP in PAPPA may disturb this regulation by altering the specific activity of PAPP-A.

www.nature.com/scientificreports www.nature.com/scientificreports/ PAPP-A forming noncovalent high-affinity complexes 19,20 . In contrast, STC2 binds covalently to PAPP-A 21 . Thus, the STCs indirectly inhibit IGF signalling by counteracting PAPP-A activity, adding another level of regulation to the IGF system.
Previous studies have identified different essential protein modules in PAPP-A [22][23][24][25][26][27] . The C-terminal part of PAPP-A contains five complement control protein (CCP) modules. The CCP modules mediate adhesion of PAPP-A to the cell surface through glycosaminoglycans, thereby locating the release of the bioactive IGFs close to their corresponding receptors 26,27 .
A recent study found two inactivating mutations in PAPP-A2, a homolog of PAPP-A, in two independent families with growth failure and markedly lower free IGF-I levels 5 . In addition, carriers of a rare missense variant in the STC2 gene (rs148833559) are approximately 2.1 cm taller compared to non-carriers 28 . Functional in vitro studies revealed that the ability of this variant to bind and inhibit PAPP-A-mediated IGFBP-4 cleavage was reduced, suggesting higher levels of bioactive IGFs that promote growth 28 . These earlier studies highlight the tremendous phenotypic effect small genetic variations may have. Previous studies have investigated single nucleotide polymorphisms (SNPs) in the PAPPA gene and one specific SNP in PAPPA; rs7020782, has been associated with increased risk of recurrent pregnancy loss, risk of gestational diabetes mellitus, risk of developing carotid plaques, and risk for ischemic cerebrovascular disease [29][30][31][32] . Additionally, a recent study revealed a significant effect of this SNP on the level of PAPP-A protein in human ovarian follicles and suggested a possible adverse effect on IGF bioactivity 33 . The rs7020782 SNP is located in exon 14 of the PAPPA gene in the CCP-1 module and causes a change in the amino acid sequence (i.e. tyrosine >serine). Furthermore, the glycosylation of these two variants may differ because the site of variation encodes a potential N-glycosylation motif 22,23 and may hereby affect the properties of the protein, such as ability to bind substrates.
The frequency of the homozygous genotype of the minor allele (the CC genotype: serine variant) of the rs7020782 SNP is reported to be approximately 7-12% in Western and Asian populations 34 .
The present study aims to define a possible functional effect of the rs7020782 SNP in PAPPA by comparison of the recombinant protein variants. Throughout this paper we will address the two rs7020782 SNP variants as: the tyrosine variant and the serine variant.

Surface binding of PAPP-A is not affected by the rs7020782 SNP. Cell surface binding of PAPP-A
to human embryonic kidney 293 T (HEK293T) cells was assessed by flow cytometry and the geometric mean of fluorescence intensities was analysed (Fig. 1). No statistically significant (ns) difference was observed between HEK293T cells incubated with supernatant containing the serine variant (rPA_1144(Ser)) and HEK293T cells incubated with supernatant containing the tyrosine variant (rPA_1144(Tyr)) ( Fig. 1). HEK293T cells incubated with supernatant without PAPP-A (MOCK) showed background fluorescence. Experiments were completed in triplicate and tested with two different primary antibodies against PAPP-A.
Complex formation between PAPP-A and its inhibitors: STC2 and proMBP. Complex formation between PAPP-A and STC2 or proMBP was studied by Western blot analysis. Supernatants from HEK293T cells transfected with the two rs7020782 PAPP-A variants incubated with supernatant containing STC2 (Fig. 2a,b) or proMBP (Fig. 2c,d) were examined. This analysis was performed to examine if the rs7020782 SNP in PAPPA influenced the covalent binding of PAPP-A to STC2 or proMBP.
Both STC2 and proMBP showed complex formation with the two different rs7020782 variants over time. These complexes were identified by a high molecular weight bands on the Western blot above the bands showing PAPP-A dimers (400 kDa). No difference was revealed with this Western blot analysis (Fig. 2).
Supernatants from HEK293T cells co-transfected with PAPP-A and STC2 (Fig. 2a,b: 'co-trans') or PAPP-A and proMBP (Fig. 2c,d: 'co-trans') were used as positive controls showing high molecular weight bands representing complete complex formation between PAPP-A and STC2 or proMBP. Supernatants from HEK293T cells transfected with PAPP-A alone (Serine variant: PA(Ser) or Tyrosine variant: PA(Tyr)) showed bands representing PAPP-A dimers solely at 400 kDa.
When incubating the two rs7020782 PAPP-A variants with supernatant from transfected HEK293T cells with empty vector (MOCK) no complex formation was observed (Supp. Fig. 2.1e). Experiments were completed in duplicate for both STC2 and proMBP.
Cleavage rate of IGFBP-2, -4, and -5 was lower for the serine variant compared to the tyrosine variant of the rs7020782 SNP in PAPPA. Proteolytic activity towards IGFBP-2, IGFBP-4, and IGFBP-5 was assessed in order to examine whether the activity of PAPP-A was different between the two rs7020782 variants (Fig. 3).
Both rs7020782 variants of PAPP-A cleaved the three radiolabelled IGFBPs resulting in co-migrating cleavage products. Four independent proteinase assays revealed that the serine variant (rPA_1144(Ser)) presented a significantly lower cleavage rate of radiolabelled IGFBP-4 compared to the tyrosine variant (rPA_1144(Tyr)) (p-value < 0.001) (Fig. 3, upper panel). Furthermore, the results showed a trend towards reduced cleavage rates of radiolabelled IGFBP-2 and IGFBP-5 (Fig. 3, lower panel) for the serine variant compared to the tyrosine variant, which were derived from three independent proteinase assays.

STC1 binds and inhibits the proteolytic activity of both PAPP-A rs7020782 variants.
Cleavage of radiolabelled IGFBP-4 by the two PAPP-A rs7020782 variants with increasing concentrations of STC1 was studied in order to evaluate a possible effect on STC1 inhibition (Fig. 4). The proteolytic activities of both PAPP-A rs7020782 variants were inhibited by STC1 showing decreasing cleavage rates with increasing concentrations of STC1. No significant difference was detected in this assay.
www.nature.com/scientificreports www.nature.com/scientificreports/ Discussion PAPP-A-induced IGF signalling has shown to be dependent and regulated by various factors, and these should be addressed when studying the functionality of PAPP-A. First, cell surface adherence of PAPP-A through the CCP modules facilitates IGF receptor activation by releasing bioactive IGFs close to their receptors on the target cell. Second, proMBP and STCs ensure strict regulation of PAPP-A by inactivating PAPP-A-mediated proteolysis of IGFBPs and thereby downregulate IGF signalling. Thus, when evaluating PAPP-A functionality it is important to address the ability of PAPP-A to bind cell surfaces and the ability of PAPP-A to form complexes with the inhibitors in combination with studying its proteolytic activity.
This study provides a detailed analysis of potential functional effects of the rs7020782 SNP in the PAPPA gene using recombinant proteins from transfected HEK293T cells. The proteolytic activity of PAPP-A towards IGFBP-4 was significantly reduced for the serine variant of the rs7020782 SNP compared to the tyrosine variant (Fig. 3). Since PAPP-A acts as a crucial component in regulating IGF bioactivity, we suggest that the rs7020782 SNP disturbs this regulation.
IGFBP-4 exhibits complex interaction with PAPP-A by interacting directly with substrate-binding exosite(s) in the LNR3 module of PAPP-A 35 . The rs7020782 SNP in PAPPA may disturb this proteinase/substrate interaction www.nature.com/scientificreports www.nature.com/scientificreports/ and consequently result in the observed reduced proteolytic cleavage. In contrast, proteolysis of IGFBP-5 and IGFBP-2 by PAPP-A are not known to rely on specific interactions like IGFBP-4 and this may be the reason why the proteolytic cleavage of IGFBP-2 and IGFBP-5 only showed a trend towards reduced cleavage rates (not significant) for the serine variant of the rs7020782.
This study did not find any effect of the rs7020782 SNP on PAPP-A cell surface binding nor on the complex formation with the inhibitors proMBP and STC2. Additionally, the ability of STC1 to inhibit PAPP-A-mediated IGFBP-4 cleavage was not shown to be affected by the rs7020782 SNP.
As aforementioned, the rs7020782 SNP is located in exon 14 of PAPPA in the first CCP module 22,23 . Furthermore, only the serine variant allows for a potential N-glycosylation motif in the protein sequence at residue Asn-1142, however, when the tyrosine variant is present the potential N-glycosylation motif is lost 22,23 . This could potentially exert a significant effect on the protein conformational structure and thereby modulate the proteolytic specificity of PAPP-A. We did not observe an effect of the rs7020782 SNP on the adhesion of PAPP-A to HEK293T cells, thus, the CCP modules do not seem to depend on this site for cell adhesion.
The serine variant of the rs7020782 SNP in PAPPA has previously been associated with various negative physiological outcomes such as recurrent pregnancy loss 29 . Women carrying the rs7020782 SNP in PAPPA encoding the serine variant showed a tendency to have an increased risk of recurrent pregnancy loss, and a significantly increased risk of having at least one pregnancy loss after nine weeks of gestation 29 . Since the current study showed that the serine variant was associated with reduced IGFBP proteolysis, women carrying this variant might present with an altered IGF regulation and thereby a limited amount of free IGF to initiate signalling. Thus, the observed negative outcome on pregnancy is possibly affected by an altered IGF bioactivity through PAPP-A, most likely in combination with other genetic variations and aberrant levels of essential hormones.
In addition, adverse pregnancy outcomes have been associated with low levels of PAPP-A, for example an increased risk of intrauterine growth restriction, extremely premature delivery, preeclampsia, and stillbirth 36 . These studies highlight that the PAPP-A/IGF axis has an important function for pregnancy outcome.
A recent study displays a significant effect of the rs7020782 SNP in vivo 33 . Women carrying the homozygous CC genotype of the rs7020782 SNP (encoding the serine variant) displayed significantly lower concentrations of PAPP-A in follicle fluid (FF) of ovarian antral follicles compared to women carrying the major A allele (encoding the tyrosine variant). Additionally, reduced cleavage of radiolabelled IGFBP-4 was observed when incubated with FF from women with the serine variant, which was suggested to be explained by low PAPP-A levels. A tendency towards reduced estradiol levels and increased androgen levels in FF from women with the serine variant was also found, which was suggested to be influenced by an aberrant IGF regulation. www.nature.com/scientificreports www.nature.com/scientificreports/ Studies of cultured human cumulus cells isolated from women undergoing fertility treatment show importance of FSH together with IGF-1 receptor activity for maximal stimulation of estradiol production, further underlining the significance of IGF signalling for ovarian steroid production 37 .
Overall, the newly found effects of the rs7020782 SNP on PAPP-A proteolytic activity may impact on both human ovarian function and pregnancy outcome.
The observed significant effect of the rs7020782 SNP on PAPP-A-mediated IGFBP-4 cleavage showed an approximately 1.5-fold reduction. Whether this reduction will have biological influence cannot be concluded from the current study, but a recent study showed reduced cleavage of radiolabelled IGFBP-4 when incubated with ovarian FF from women carrying this SNP 33 , which supports this notion. However, the specific activity of PAPP-A was not assessed.
Previous studies have investigated the effect of assisted reproductive technology (ART) on first trimester screening parameters such as PAPP-A 38,39 . Significant lower PAPP-A serum levels have been found in women receiving IVF/ICSI treatment compared to women conceiving naturally. Furthermore, the low levels of PAPP-A were significantly correlated with high estradiol levels at ovulation induction 39 . This suggests a regulative role of the hormones produced in the ovary prior to conception on the PAPP-A/IGF axis. In addition, IVF and ICSI are known to result in a higher risk of low birth weight compared to normal pregnancies 38,40 . This is suggested to be affected by PAPP-A/IGF in a study showing that the combination of slow early foetal growth and low PAPP-A lead to an approximately six-fold increased risk of having a small for gestational age infant 40 . Thus, the lower level of PAPP-A in ART associated with high estradiol levels could potentially impact the growth and further development of the foetus due to an abnormal IGF signalling. www.nature.com/scientificreports www.nature.com/scientificreports/ Furthermore, women receiving IVF treatment and classified as poor responders for recombinant FSH have been shown to have significantly lower ovarian FF levels of IGF-1 together with lower serum and ovarian FF levels of oestradiol and progesterone 41 . Furthermore, this study found a positive correlation between total oocytes retrieved and ovarian FF levels of IGF-1 and suggests that the observed decreased level of IGF-1 in poor responders impairs the ovarian steroid production.
The minor allele of the investigated rs7020782 SNP (serine variant) appears to have potentially adverse effects on reproduction but persists at non-negligible frequencies. A potential explanation is that carrying a single copy of the minor allele may either have no phenotypic effect or, in some cases, even lead to potential beneficial effects, in for instance, breast and ovarian cancer where PAPP-A has shown to have tumour-promoting properties 42 . Alternatively, the minor allele may only show adverse effects in a certain genetic background, as, for instance, in Sub-Saharan Africa where the serine allele of the rs7020782 SNP instead of being the minor allele has been reported to be the major allele, which is exactly opposite to European and Asian populations 34 .
Besides recurrent pregnancy loss the rs7020782 SNP in PAPPA also associates with the risk of gestational diabetes mellitus, risk of developing carotid plaques, and risk for ischemic cerebrovascular disease [30][31][32] . An altered IGF bioactivity may be present in these cases, because of the reduced proteolysis of IGFBPs by PAPP-A, however, none of these studies included examinations of PAPP-A activity.
This study shows, for the first time, a direct link between the rs7020782 SNP in PAPPA and reduced proteolytic function, specifically towards IGFBP-4, which may be explained by reduced proteinase/substrate interaction. This may partly explain some of the mechanisms behind the different diseases associated with the SNP and highlights the importance of the IGF pathways in different biological aspects.
We suggest that the rs7020782 SNP in PAPPA could be used as a biomarker for identifying women with high risk of adverse pregnancy outcomes such as recurrent pregnancy loss and perhaps this could even allow for risk-specific treatment and improve the outcomes for these patients, however, further studies are needed to apply this in clinical settings.

Materials and Methods
Mutagenesis. Plasmids encoding human PAPP-A with the tyrosine variant of the rs7020782 SNP were constructed with the QuickChange II site-directed mutagenesis kit (Stratagene) using the primers: 5′-ctggctgtggagaatgcttatctcaattgctccag-3′ (forward primer) 5′-ctggagcaattgagataagcattctccacagccag-3′ (reverse primer) (the rs7020782 change is underlined in the primer sequences).
Human PAPP-A plasmid (pcDNA3.1-PAPP-A) was used as a template described previously 18,27 . The PAPP-A constructs were verified with sequencing. The rs7020782 PAPP-A variant that encodes the serine variant was present in the pcDNA3.1-PAPP-A plasmid.
cell culture and transfection. HEK293T cells were maintained in high glucose Dulbecco's modified Eagle's medium supplemented with 10% foetal bovine serum, 2 mM glutamine, nonessential amino acids, and gentamicin (Invitrogen). Cells were plated in 6-cm tissue culture dishes and transfected with plasmid constructs containing cDNA encoding: human PAPP-A with the serine variant of the rs7020782 SNP, human PAPP-A with the tyrosine variant of the rs7020782 SNP, human proMBP (Human placental oligo-dT-primed cDNA encoding proMBP was used as a template 43 ) and human STC2 (STC2 cDNA in the pcDNA3.1/Myc-His(−) was used as a template 21 ). HEK293T cells transfected with empty vector were included as controls (MOCK). Transfection of cells was performed by calcium www.nature.com/scientificreports www.nature.com/scientificreports/ phosphate co-precipitation using 10 µg of plasmid DNA as previous described 21 . Media were harvested 24 hours and 48 hours post-transfection, cleared by centrifugation, and stored at -20 °C until further analysis.

enzyme-linked immunosorbent assay (eLiSA). The concentration of recombinant PAPP-A in the
supernatants from the transfected HEK293T cells was determined by ELISA (picoPAPP-A, AL-101-i, Ansh Labs, Texas, USA) according to the manufacturer's instructions (PBS + 1% BSA was used for dilution). The ELISA measures were used to achieve equal concentration of the two PAPP-A rs7020782 variants in the different functional assays. flow cytometry. HEK293T cells were used for analysing cell surface adhesion of the PAPP-A rs7020782 variants 27 . Non-transfected cultured HEK293T cells were incubated on ice with supernatant containing the serine variant of the rs7020782 SNP, the tyrosine variant of the rs7020782 SNP, or supernatant from HEK293T cells transfected with empty vector (MOCK) for 1 hour. A total of 800,000 cells/well were used. After incubation the cells were washed four times with 1% BSA in PBS and subsequently incubated with primary antibody against PAPP-A (two different antibodies were used: mAb 1/41 44  Western blotting. In order to investigate the ability of the two rs7020782 PAPP-A variants to form complexes with STC2 and proMBP Western blotting was performed. Supernatants harvested from the HEK293T cells transfected with either of the two rs7020782 PAPP-A constructs, STC2, proMBP, or with empty vector (MOCK) as a negative control was used for Western blotting as previously described 20 . To remove possible traces of endogenous STC2 present in the media containing recombinant PAPP-A, immunoprecipitation of the culture media was carried out by incubating (16 h at 4 °C) 1 ml culture media with 30 ul protein G-Sepharose 4 Fast Flow beads (GE Healthcare, Denmark) coupled to 2 mg/ml of murine monoclonal antibodies against STC2 (STC221 20 ,) as described previously 20 . Supernatants with either of the PAPP-A variants were incubated with STC2 or proMBP for 0, 1, 2, 4, 8, 16, and 24 hours (h) or with empty vector (MOCK) for 0, 8, and 24 h at 37 °C. STC2, proMBP and PAPP-A were blotted onto PVDF membranes (Millipore) following separation by 3-8% SDS-PAGE. For PAPP-A/STC2 and PAPP-A/proMBP complex detection, the membranes were incubated with rabbit polyclonal anti-PAPP-A at 0.63 μg/ml in TST supplemented with 2% skim milk for 16 h at 20 °C 45 . Membranes were washed with TST and subsequently incubated with polyclonal swine anti-rabbit IgG-HRP (DAKO, P0217) in TST supplemented with 2% skim milk for 1 h at 20 °C. Blots were developed using enhanced chemiluminescence (ECL Prime, GE Healthcare). Images were captured using an ImageQuant LAS 4000 instrument (GE Healthcare) (ECL; Amersham Pharmacia Biotech).
The PageRuler ™ Plus Prestained Protein Ladder, 10 to 250 kDa was used as marker (Catalog number: 26619, Thermo Scientific ™ ). proteinase assay. To measure the proteolytic activity of PAPP-A in the supernatants harvested from the HEK293T cells transfected with either of the two rs7020782 variants, a proteinase assay based on autoradiography was performed as previously described in detail 16 . Proteinase assays were performed by measuring the proteolytic cleavage of radiolabelled IGFBP-2, IGFBP-4, and IGFBP-5. Different PAPP-A concentrations were used in the three assays, which were optimized for the variants to achieve cleavage percentages below 30%, where a linear relationship between time and cleavage can be assumed, and the cleavage rates are hereby determined by the slopes 16 . For the IGFBP-2 cleavage assay the PAPP-A concentration was adjusted to 10 nM, for the IGFBP-4 assay PAPP-A concentration was adjusted to 150 pM and for the IGFBP-5 assay the PAPP-A concentration was adjusted 80 pM. Intact and cleaved IGFBPs were quantified with the ImageQuant TL 8.1 software (GE Healthcare).
Addition of increasing concentrations of STC1 was performed in a separate IGFBP-4 cleavage assay to examine if the inhibition of PAPP-A was different between the two rs7020782 PAPP-A variants. The STC1 concentrations used in the assay were: 0.001 nM, 0.0312 nM, 0.0625 nM, 0.125 nM, 0.25 nM, 0.5 nM, and 1.0 nM 19 .

Statistical analysis.
An unpaired t-test was used to test significant difference in cleavage rate of IGFBP-2, IGFBP-4, and IGFBP-5 between the two rs7020782 PAPP-A variants. The same model was used to test significant differences in STC1 inhibition of PAPP-A activity and the cell surface binding ability. The p-value was set to 0.05 to assess statistical significance.

Data Availability
Data generated during this study are made available in the supplementary files. Additional data generated and analysed during the current study are available from the corresponding author on reasonable request. www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/