Prokineticin 1 is up‐regulated by insulin in decidualizing human endometrial stromal cells

Abstract Prokineticin 1 (PROK1), a hypoxia‐regulated angiogenic factor, has emerged as a crucial regulator of embryo implantation and placentation. Dysregulation of PROK1 has been linked to recurrent pregnancy loss, pre‐eclampsia, foetal growth restriction and preterm birth. These pregnancy complications are common in women with obesity and polycystic ovary syndrome, i.e. conditions associated with insulin resistance and compensatory hyperinsulinaemia. We investigated the effect of insulin on PROK1 expression during in vitro decidualization. Endometrial stromal cells were isolated from six healthy, regularly menstruating women and decidualized in vitro. Insulin induced a significant dose‐dependent up‐regulation of PROK1 on both mRNA and protein level in decidualizing endometrial stromal cells. This up‐regulation was mediated by hypoxia‐inducible factor 1‐alpha (HIF1α) via the phosphatidylinositol 3‐kinase (PI3K) pathway. Furthermore, we demonstrated that PROK1 did not affect the viability, but significantly inhibited the migration of endometrial stromal cells and the migratory and invasive capacity of trophoblast cell lines. This in vitro study provides new insights into the regulation of PROK1 by insulin in human decidualizing endometrial stromal cells, the action of PROK1 on migration of endometrial stromal cells, as well as migration and invasion of trophoblasts. We speculate that hyperinsulinaemia may be involved in the mechanisms by which PROK1 is linked to placenta‐related pregnancy complications.


Introduction
The tightly coordinated decidual transformation is indispensable for the formation of a functional feto-maternal interface in order to successfully establish and maintain pregnancy. The central roles of decidualization are the protection of feto-maternal interface against oxidative stress, local suppression of maternal immune response and control of extravillous trophoblast migration and invasion [1]. The latter is mediated, at least in part, by a recently discovered hypoxiaregulated angiogenic factor PROK1, or alternatively named endocrine gland-derived vascular endothelial growth factor (EG-VEGF).
PROK1 has been described in the human endometrium throughout the menstrual cycle in the stroma, glandular epithelium and endothelial cells with increasing levels in the secretory phase [2].
Furthermore, PROK1 expression is significantly increased in first trimester decidua, compared with non-pregnant endometrium [3]. PROK1 is expressed in the placenta and is predominantly regulated by oxygen tension. In the hypoxic period of placental development in the first trimester, increased level of placental PROK1 promotes proliferation of anchoring cytotrophoblasts and formation of trophoblastic plugs in the maternal spiral arteries [4,5]. These trophoblastic plugs protect the developing placenta and embryo from highly oxygenated maternal blood and early oxidative damage. Placental PROK1 level declines after week 10-12 of pregnancy as the oxygen tension increases, facilitating deeper invasion of trophoblasts and remodelling of the maternal spiral arteries in order to supply the developing foetus [5,6]. Impaired invasion and thin or fragmented trophoblastic plugs have been associated with early pregnancy loss, while poor invasion of maternal vessels has been correlated to both pre-eclampsia and foetal growth restriction [4,[6][7][8][9].
PROK1 was shown to be up-regulated by the hypoxia-inducible factor 1 (HIF1) in human placenta and human adrenal carcinoma cell lines and by human chorionic gonadotrophin in human placenta [5,8,10]. HIF1 is a basic helix-loop-helix transcription factor composed of two subunits, HIF1a and HIF1b. The latter is constitutively expressed, whereas HIF1a expression is up-regulated predominantly by hypoxia. Under normoxia, HIF1a is rapidly ubiquitinated by the von Hippel-Lindau tumour suppressor E3 ligase complex and subjected to proteasomal degradation [11]. HIF1a was shown to be upregulated in luteal phase endometrium in both the stromal and the epithelial cells compared to the proliferative phase of the menstrual cycle [12]. It was demonstrated previously in a variety of cell types, but not in endometrial stromal cells, that HIF1a can be regulated not only by hypoxia but also by insulin and other growth factors under normoxic conditions, mainly by up-regulating HIF1a protein expression or preventing it from proteasomal degradation [13][14][15][16]. It was also shown in arising retinal pigment epithelia cell line 19 and primary rat hepatocytes, that HIF1a could be induced by insulin via the PI3K pathway [13,17].
We have recently reported data suggesting adverse effect of insulin on endometrial function and in vitro decidualization [32,33]. In this in vitro study, we aimed to investigate the effect of insulin on the regulation of PROK1 in primary decidualizing human endometrial stromal cells, as well as the effect of PROK1 on migration of human endometrial stromal cells and migration and invasion of trophoblast cells.

Subjects
Endometrial biopsies were collected under local anaesthesia with an endometrial suction curette (Pipet Curet; CooperSurgical, Trumbull, Connecticut, USA) from six regularly cycling, non-smoking healthy volunteers at cycle day 5-9. All participants were between 18 and 35 years with a body mass index ranging 19-28. Exclusion criteria were hormonal medication within 3 months prior to examination, current chronic disease, endocrine disorder or continuous medication. All women gave their written informed consent, and the Regional Ethical Committee in Stockholm approved the study (Dnr 2008/865-32).

Isolation of human endometrial stromal cells
Isolation of endometrial stromal cells was carried out as previously described [33]. Purity of stromal cells was ensured by sequential culturing and assessed by cytokeratin and CD10 staining for epithelial and stromal cells, respectively.

Culture conditions
Endometrial stromal cells were seeded in six-well Costar plates (Sigma-Aldrich, St. Louis, Missouri, USA) and cultured in DMEM/F12-Glutamax (Thermo Fischer Scientific, Waltham, Massachusetts, USA) supplemented with 10% heat-inactivated foetal bovine serum (HI-FBS) (Sigma-Aldrich) and 0.2% penicillin-streptomycin (Sigma-Aldrich) until~80% confluency. In vitro decidualization was performed with a well-established procedure, as described previously [33,34]. Briefly, media were changed to phenol red-free DMEM/F12 (Thermo Fischer Scientific), supplemented with 2% charcoal-stripped foetal bovine serum (Sigma-Aldrich) and 0.2% penicillinstreptomycin. To investigate the kinetics of decidualization, we treated the cells with 1 lM medroxyprogesterone-17-acetate (MPA) (Sigma-Aldrich) and 0.5 mM N 6 , 2'-O-dibutyryladenosine cAMP (db-cAMP) (Sigma-Aldrich) for 3 and 6 days. In order to clarify the contribution of each of MPA and db-cAMP to induce decidualization, endometrial stromal cells were treated with 1 lM MPA, 0.5 mM db-cAMP or combined treatment of MPA and db-cAMP for 6 days. To investigate the effect of insulin on PROK1 during decidualization, cells were treated with 1 lM MPA and 0.5 mM db-cAMP to induce decidualization for 6 days in the presence or absence of 5, 50 or 500 nM insulin (Sigma-Aldrich).
To test whether wortmannin, a PI3K inhibitor, could block the effect of insulin, stromal cells were pre-decidualized for 3 days and then pre-treated according to previous publications [35,36] with 500 nM wortmannin (Sigma-Aldrich) for 1 hr prior to treatment with 100 nM insulin for 2 days in the presence of decidualization agents db-cAMP and MPA. To evaluate the involvement of HIF1a in the regulation of PROK1 by insulin, stromal cells were pre-decidualized for 3 days and then pre-treated with 5 nM of echinomycin (Merck Millipore, Darmstadt, Germany), a HIF1a inhibitor, for 1 hr prior to addition of 100 nM insulin for 2 more days along with the decidualization agents db-cAMP and MPA. All the treatments were carried out under normoxic conditions. The culture media were renewed every 3 days. After treatments conditioned media were collected, and cells were harvested for RNA isolation. All the treatments have been performed at least twice.

RNA isolation, cDNA synthesis and RT-PCR
Total RNA was extracted, and cDNA was synthesized as described previously [32]. 10 ng cDNA was subjected to RT-PCR using TaqMan method. Ribosomal protein L13a (RPL13A) served as an internal control and was used to normalize gene expression levels. The employed TaqMan assays (Thermo Fischer Scientific) were Hs00951617_m1 for PROK1, Hs00168730_m1 for prolactin (PRL), Hs00236877_m1 for insulin-like growth factor binding protein 1 (IGFBP1) and Hs01926559_g1 for RPL13A. The reactions were run in a StepOnePlus TM Real-Time PCR System (Thermo Fischer Scientific), and gene expression levels were determined using the DDC T method. The PCR efficiency of all amplicons was 90-100%, and all determinations were performed in triplicate. PROK1 was measured in cells from six healthy volunteers.

Enzyme-linked immunosorbent assay
Secreted PROK1 protein produced by the decidualizing stromal cells in the presence of 0, 5, 50 or 500 nM of insulin was measured from conditioned media using enzyme-linked immunosorbent assay (ELISA) (Abcam, Cambridge, UK) according to the manufacturer's instruction. Briefly, stromal cells were decidualized with 1 lM MPA and 0.5 mM db-cAMP in the presence or absence of insulin for 6 days. Media and treatment were renewed after 3 days, and the conditioned media collected on the 6th day were processed for ELISA. The experiment was performed from treatments using six healthy volunteers in duplicates.

Cell viability analysis
To assess cell viability, a MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide) assay was used. Briefly, 10 4 primary endometrial stromal cells were seeded in 200 ll DMEM/F12-Glutamax supplemented with 10% HI-FBS and 0.2% penicillin-streptomycin in a 96-well plate. After 24 hrs, the cells were washed with PBS, and 100 ll phenol red-free DMEM/F12 media were added with or without 50 ng/ml PROK1 (Sigma-Aldrich), supplemented with 2% charcoalstripped foetal bovine serum and 0.2% penicillin-streptomycin. Cell viability was determined after 24 hrs by labelling the cells with 10 ll MTT reagent for 4 hrs. The resulting formazan crystals, which correspond to the amount of viable cells, were solubilized by the addition of 100 ll 10% sodium dodecyl sulphate solution in tris-buffered saline (TBS). Cell viability was quantified by the measurement of the absorbance in a plate-reader with a 595-nm filter. Absorbance of the untreated cells was set to 100%, and the viability of treated cells was calculated accordingly. The experiments were run using cells from three healthy volunteers in quadruplicates.

Wound healing assay
Endometrial stromal cells were seeded in 12-well Costar plates (Sigma-Aldrich) and cultured in DMEM/F12-Glutamax (Thermo Fischer Scientific) supplemented with 10% HI-FBS and 0.2% penicillin-streptomycin until confluency. The cells were scratched with a 200-ll pipette tip, and the media were removed. The cells were washed twice with PBS to remove cell debris and 1 ml phenol red-free DMEM/F12 media was added with or without 50 ng/ml PROK1, supplemented with 2% charcoal-stripped foetal bovine serum and 0.2% penicillin-streptomycin. Microphotographs were taken with 409 magnification at 0 and 24 hrs with a Leica DFC420 C digital camera on (Leica Microsystems, Wetzlar, Germany) a Nikon Eclipse TS 100 inverted microscope (McCrone, Westmont, Illinois, USA). Photographs were analysed using ImageJ software (NIH, Bethesda, Maryland, USA) [37]. The experiments were performed with cells from five healthy volunteers in triplicates.
To track the proliferation and migration of undifferentiated endometrial stromal cells in the presence or absence of 50 ng/ml PROK1, timelapse imaging was applied. Endometrial stromal cells were seeded in 24-well Costar plates (Sigma-Aldrich) and cultured until confluency, scratched with a 200-ll pipette tip, washed twice with PBS and treated as described above. After the addition of the treatment, the plates were relocated to the TIRF microscope incubation chamber, at constant 37°C and under 5% CO 2 . The ZEN2 Software was used to design the experiment and guide the complete microscopic unit automatically. The timelapse imaging was recorded for 24 hrs, with a 15-min. interval per capture. A field of 18 images (3 9 6), which covered a total area of 2506 9 3689.5 lm 2 , was captured using the 109 objective lens. The experiment was performed with cells from three healthy volunteers in triplicates.

Transwell migration assay
The motility of AC-1M88 choriocarcinoma-trophoblast hybrid cells (DSMZ, Braunschweig, Germany) was assessed using cell culture inserts with 8-lm pores (Merck Millipore) in 24-well plates. AC-1M88 cells were starved for 24 hrs, and then 10 5 cells were plated to cell culture inserts in 200 ll serum-free F12 media with or without 50 ng/ml PROK1. 750 ll F12 media supplemented with 10% HI-FBS was added to the lower well. After 48 hrs of incubation, inserts were washed twice with PBS, and the cells were fixed in 4% of paraformaldehyde in PBS for 15 min. The cells were stained with 0.5% crystal violet for 20 min., and the non-migrated cells were removed with cotton swab from the upper side of the membrane. Photographs were taken with 409 magnification on a Nikon Eclipse TS 100 inverted microscope using a Leica DFC420 C digital camera. Adobe Photoshop was used to convert photographs to greyscale images, and particles were quantified by ImageJ software. The transwell migration assay was performed in quintuplicates.

Formation of trophoblast spheroids and spheroid invasion assay
Spheroids from HTR-8/SVneo trophoblast cells (a kind gift of Mona Nystad at the University of Tromsø, Norway) were formed on the basis of a procedure described previously [38]. Briefly, for preparation of 100 spheroids, 3 9 10 5 HTR-8/SVneo cells were suspended in 9.11 ml of RPMI 1640 medium and mixed with 890 ll of a 2.8% methylcellulose solution (R&D Systems, Minneapolis, Minnesota, USA). In each well of a non-adherent round-bottom 96well plate, 100 ll of this suspension containing 3000 cells was seeded and incubated at 37°C in 5% CO 2 . Twenty-four hours later, spheroids had formed and were used for the spheroid invasion assay. Spheroids were embedded using 100 ll of 6 mg/ml growth factor reduced Basement Membrane Extract (R&D Systems) diluted with serum-free RPMI 1640 media in the presence or absence of 50 ng/ml PROK1 in round-bottom, low-attachment 96-well plates and centrifuged for 5 min. at 300 g at 4°C. After the gel was solidified at 37°C, 100 ll complete media was added onto the top of the gel, and spheroids were incubated at 37°C in 5% CO 2 . Microphotographs were taken with 409 magnification daily for 5 days with a Leica DFC420 C digital camera on a Nikon Eclipse TS 100 inverted microscope. Photographs were analysed using ImageJ software. The spheroid invasion assay was performed in quintuplicates.

Statistical analysis
Normality of the data was tested with the Kolgomorov-Smirnov test. Statistical analysis was performed by the one-way ANOVA or Friedman test followed by multiple comparisons, based on the distribution of the data, where each treatment group was tested against the untreated group. In the PI3K and HIF1a inhibitory experiments, the insulin-treated groups were compared with their respective controls (decidual, decidual + solvent or decidual + inhibitor). Paired t-test was used to analyse the effect of PROK1 on the viability and migration of endometrial stromal cells. In the cases of transwell migration assay and spheroid invasion assay, unpaired t-test was used to analyse statistical significance between PROK1 treated and untreated samples. P < 0.05 was considered as significant.

PROK1 is up-regulated upon decidualization
To test the effect of decidualization on the expression of PROK1 in human primary endometrial stromal cells, we treated the cells with MPA, db-cAMP or their combination for 6 days. We found that db-cAMP induced morphological transformation of the cells and upregulated PROK1 and the decidualization markers PRL and IGFBP1, whereas MPA did not. The combination of MPA and db-cAMP also induced morphological transformation of the cells and enhanced up-regulation of PROK1 and the decidualization markers PRL and IGFBP1 compared to db-cAMP alone ( Fig. S1A and B). Furthermore, we investigated the kinetics of decidualization in response to the combined use of MPA and db-cAMP and found that decidualization was apparent after 3 days, and even more enhanced after 6 days (Fig. S1C).

Insulin up-regulates PROK1 in decidualizing endometrial stromal cells
Proliferative phase-derived stromal cells were decidualized in the presence or absence of increasing doses of insulin, and relative gene expression level of PROK1 was measured. PROK1 was highly upregulated by insulin in a dose-dependent fashion (ns at 5 nM, P < 0.05 at 50 nM, P < 0.001 at 500 nM) (Fig. 1A).
To verify whether insulin also regulates the protein expression of PROK1 in decidualizing endometrial stromal cells, secreted PROK1 was measured from conditioned media. The protein expression was also highly up-regulated by insulin in a dose-dependent manner (ns at 5 nM, ns at 50 nM, P < 0.01 at 500 nM) (Fig. 1B).

PROK1 is up-regulated by insulin via the induction of HIF1a
To test whether HIF1a is involved in the insulin-induced up-regulation of PROK1, we treated decidualizing stromal cells with 5 nM echinomycin, a HIF1a inhibitor, in the presence or absence of 100 nM insulin. We found that 5 nM echinomycin completely blocked insulininduced up-regulation of PROK1 mRNA, meaning that insulin exerts its effect on PROK1 via HIF1a (Fig. 2).

Insulin regulates PROK1 through the PI3K pathway
To investigate whether insulin regulates PROK1 through the PI3K pathway, we blocked PI3K signalling with 500 nM wortmannin during the decidualization process in the presence or absence of insulin. PI3K inhibition resulted in an almost complete inhibition of insulin-induced up-regulation of PROK1 (Fig. 3). This suggests that the regulation of insulin on PROK1 is conveyed by the PI3K signalling pathway.

PROK1 does not influence viability of endometrial stromal cells
To test whether PROK1 has an effect on cell viability, i.e. proliferation and cell death, MTT assay was performed on undifferentiated

PROK1 inhibits migration of endometrial stromal cells
To investigate whether PROK1 has an effect on the migration of endometrial stromal cells, we performed a wound healing assay on the undifferentiated endometrial stromal cells. Treatment with 50 ng/ ml PROK1 slightly, but significantly reduced the migration of endometrial stromal cells after 24 hrs (P < 0.05) (Fig. 4B and C and Video S1 and S2).

PROK1 inhibits migration of AC-1M88 trophoblast-choriocarcinoma cells
To demonstrate the effect of PROK1 on the migration of AC-1M88 trophoblast-choriocarcinoma cells, we performed a transwell migration assay in the absence or presence of 50 ng/ml PROK1. We found that PROK1 significantly inhibited the migration of AC-1M88 cells (P < 0.01) ( Fig. 5A and B).

PROK1 inhibits invasion of HTR-8/SVneo trophoblast spheroids
Spheroids were formed using HTR-8/SVneo cells and embedded into basement membrane extract in the presence or absence of 50 ng/ml PROK1. The area covered by individual spheroids was determined daily. We found that invasion of the spheroids was reduced by PROK1, which reached significance in comparison to controls after 3 days (P < 0.05) (Fig. 6A and B).

Discussion
This study is the first to show that insulin up-regulates gene and protein expression of the pro-angiogenic factor PROK1 in decidualizing human endometrial stromal cells. We furthermore demonstrated that insulin enhances PROK1 expression via HIF1a, a well-known  regulator of PROK1, and that the PI3K pathway is involved in insulininduced up-regulation of PROK1. It was also shown here for the first time that PROK1 does not affect viability, however decreases migration of endometrial stromal cells, a process that is crucial for adequate embryo implantation. In addition, it was demonstrated that PROK1 restricts the migration of AC-1M88 trophoblast-  In this study, we demonstrated that PROK1 gene and protein expression are up-regulated during in vitro decidualization, and this is in agreement with the in vivo findings in human endometrium by others [2]. This up-regulation in decidualizing human endometrial stromal cells was even more enhanced by insulin in a dose-dependent manner. It must be noted that PROK1 gene expression level after in vitro decidualization showed high variety among the volunteers, and two volunteers showed remarkably higher inducibility of PROK1 in response to insulin. These variances might be due to genetic polymorphisms of PROK1 gene, as certain polymorphisms have been associated with different gene expression levels of PROK1 [39].
In the present study, we demonstrated that when decidualizing endometrial stromal cells were treated with a HIF1a inhibitor or a PI3K inhibitor under normoxic conditions, they both prevented insulin-induced up-regulation of PROK1. These findings are in agreement with previous publications on the ability of HIF1a to up-regulate PROK1 in other cell types [5,10]. Our data also support previous findings on the possible effect of insulin on HIF1a and the involvement of the PI3K pathway in this process and suggest that insulin upregulates PROK1 in decidualizing endometrial stromal cells via PI3Kmediated HIF1a induction [13][14][15][16][17].
We also demonstrated that PROK1 reduced the migratory capacity of endometrial stromal cells. It was previously reported that invasion of the embryonic trophoblast required motility of maternal endometrial stromal cells [38,[40][41][42]. Moreover, endogenous PROK1 may act in an autocrine manner in decidualizing endometrial stromal cells via the prokineticin receptor 1 (PROKR1) [43]. We speculate that PROK1 dysregulation may interfere with normal implantation by inhibiting the migration of endometrial stromal cells. We also demonstrated that PROK1 did not affect the viability, i.e. proliferation, cell death and metabolism of endometrial stromal cells significantly. Furthermore, the negligible enhancement by PROK1 on cell viability can In this study, we could show that PROK1 significantly inhibits the migration of AC-1M88 trophoblast-choriocarcinoma cells. As wound healing assay proved to be challenging with AC-1M88 cells, because of the detachment of the whole cell layer when scratching, we performed transwell migration assay. We also demonstrated that invasion of HTR-8/SVneo spheroids is restricted by PROK1. It was previously shown that PROK1 inhibits migration of HTR-8/SVneo cells, invasion of placental villous explants and tube-like organization of HTR-8/SVneo cells [6]. The authors suggested that sustained PROK1 expression beyond 10-12 weeks of gestation might cause an inhibition of deeper invasion of the decidua by differentiating extravillous trophoblasts that might consequently lead to a failure in maternal spiral artery remodelling [6]. Furthermore, it was recently demonstrated that sustained PROK1 levels beyond the first trimester of gestation cause defects in placental organization and function, increase hypoxia and decrease trophoblast invasion in mice, leading to gestational hypertension and dysregulated maternal kidney function, typical symptoms of pre-eclampsia [44]. Our results further support the previous suggestion that PROK1 may restrict the migration and invasion of embryonic trophoblast cells into maternal decidua [6].
Based on our data, we hypothesize that maternal hyperinsulinaemia could increase decidual PROK1 expression that might lead to impaired invasion of extravillous trophoblasts in the first trimester of the pregnancy. Clinical studies have shown that certain polymorphisms of PROK1 and PROKR1 are associated with altered susceptibility for recurrent pregnancy loss [39,45,46]. It was also demonstrated that gene expression level of PROK1 is higher in decidualizing stromal cells from patients with recurrent pregnancy loss compared to controls. This could result in prolonged endometrial receptivity, i.e. a temporally unique sequence of factors that make the endometrium receptive to embryo attachment and implantation, allowing out-ofphase implantation in an unsupportive uterine environment [47][48][49].
A line of evidence suggests that impaired invasion of extravillous trophoblasts might trigger a number of events. The shallow invasion of extravillous trophoblasts in the first trimester of pregnancy was shown to impair remodelling of the maternal spiral arteries that was associated with foetal growth restriction and pre-eclampsia [50,51], resulting in a relatively hypoxic milieu in the placenta [7,52]. The relative placental hypoxia was demonstrated to dysregulate several angiogenic factors including VEGF [53]. It was shown that these factors are dysregulated, and branching of the placental vascularization is enhanced in pre-eclampsia [54,55], suggesting that switch from branching to non-branching angiogenesis is disturbed in this pregnancy complication. PROK1 was shown to enhance tube-like formation of microvascular placental endothelial cells (HPEC), suggesting that PROK1 contributes to branching of the vascular tree [56]. Furthermore, maternal serum level of PROK1 was shown to be significantly higher in pre-eclampsia and foetal growth restriction than in normal pregnancies [4,6].
Taken together, our data show for the first time that PROK1-a crucial factor in pregnancy complications such as recurrent pregnancy loss, pre-eclampsia, foetal growth restriction and spontaneous preterm birth-is up-regulated by insulin in decidualizing endometrial stromal cells. We demonstrate that PROK1 negatively influences three processes, which are suggested to be essential for successful implantation, placenta development and normal maintenance of the pregnancy, i.e. migration of endometrial stromal cells and migration and invasion of trophoblasts. Our results support the previous suggestions that hyperinsulinaemia might play a role in the mechanisms behind hypoxia-related pregnancy complications such as miscarriage and pre-eclampsia.