Female C57BL/6J and C57BL/6N mice as well as male BALB/c were obtained from Charles River (Sulzfeld, Germany). MC-deficient female C57BL/6J-Kit^W-sh/W-sh (W-sh) mice were bred and maintained at Charité in Berlin, Germany, and then transferred to Magdeburg 2 weeks prior to the experiments. Female IL-15-deficient C57BL/6NTac-IL15^tm1ImxN5 were purchased from Taconic (New York, NY, USA) and bred in our facilities. MC-deficient C57BL/6J-Cpa3^Cre/+ (Cpa3^Cre/+) mice and their controls C57BL/6J-Cpa3^+/+ (Cpa3^+/+) were kindly provided by Prof. H. R. Rodewald (Heidelberg, Germany) and bred in our facilities. Mice were housed in our barrier facility with a 12-h light/dark cycle (7:00 a.m.–7:00 p.m./7:00 p.m.–7:00 a.m.) and received food and water ad libitum. Experiments were performed according to the institutional guidelines upon ministerial approval (42502-2-1296UniMD). 6–8 weeks old females were paired with BALB/c males, checked twice a day for vaginal plugs, and separated from the males if mated. Plug appearance was defined as gestation day (gd) 0.

Female W-sh, Cpa3^Cre/+ and control mice were sacrificed at gd10, and one implantation site from the end of the left uterus horn was removed for the evaluation of uNK numbers. C57BL/6N and IL-15^−/− mice were sacrificed on gd5, gd10, or gd18. For the group sacrificed at gd5, uterus samples were removed for histological, flow cytometry, and RNA analysis of uMCs. On gd10, blood was taken for flow cytometry analysis, and implantation sites were documented and removed for quantification of uNKs and SA analysis. Spleen, mesenteric, and inguinal lymph modes (LNs) were removed for flow cytometry analysis. One decidua piece was snap frozen for later RNA analysis. A further group was employed for blood pressure measurements and sacrificed at gd18; thereafter, fetal and placental weights were measured and the fetus-to-placenta weight ratio index (FPI) was calculated.

Mice were anesthetized, and blood was obtained by retroorbital puncture; afterward, they were sacrificed and the implantation rate was determined by registering the total number of fetuses. Spleen, mesenteric LNs, and inguinal LNs were removed, washed in cold phosphate-buffered saline (PBS), and conserved in cold RPMI until use for flow cytometry analysis. For histological analysis, two implantations per female were removed, fixed in 4% paraformaldehyde (PFA) containing 0.1 M saccharose (pH 7.4) for 6 h at room temperature (RT), and embedded in paraffin after dehydration with ethanol and xylol cycles. One piece of the uterus (gd5) or decidua (gd10) was cut into small pieces and stored in RPMI supplemented with 5% penicillin–streptomycin (P/S) containing 0.005% liberase for flow cytometry. A further piece was fixed in ethanol (96%) at 4°C and later embedded in paraffin after dehydration with ethanol and xylol cycles. Uterus (gd5) or decidua (gd10) tissue was additionally collected for RNA isolation, washed in cold PBS, snap-frozen in liquid nitrogen, and stored at −80°C until use.

Isolation, culture, and transfer of BMMCs were performed as previously described (8). Shortly, wild-type (WT) bone marrow cells from femur and tibia were cultured in Iscove’s modified Dulbecco’s medium supplemented with 10% fetal bovine serum (FBS), 1% P/S, and murine recombinant IL-3 (30 ng/ml, Life technologies, Carlsbad, CA, USA). After confirmation by flow cytometry analysis that >95% of the cells were CD117⁺FcεRIα⁺ BMMCs, they were used for reconstitution. BMMCs were transferred into 10-week-old W-sh females by i. v. injection of 5 × 10⁶ cells in 200 µl of PBS in the morning of two consecutive days (total 1 × 10⁷ cells). 12 weeks after reconstitution, W-sh females were mated with BALB/c males (8).

To compare uNK numbers between the experimental groups, implantation sections were stained with DBA lectin. After dewaxing and rehydration procedure, endogenous peroxidases were blocked by incubation with 3% hydrogen peroxide (H₂O₂) in methanol at RT for 30 min in a humidity chamber. Slides were washed in 50 mM PBS twice, followed by 15 min avidin and biotin blocking each. Prior to DBA lectin staining [dilution 1:150 in 1% bovine serum albumin (BSA) in 100 mM PBS, overnight (ON) at 4°C], proteins were blocked with 1% BSA in 100 mM PBS for 30 min. For negative controls, DBA lection staining solution was replaced by 1% BSA in 100 mM PBS. At the next day, sections were washed twice with PBS and treated with horseradish peroxidase-solution for 30 min. After staining, slides were washed twice and incubated with AEC substrate chromogen (Dako) for 6 min and washed twice again. Sections were counter-stainined with hematoxylin solution for 1–2 min, dipping in warm tap water and mounted with Aqua Tex. uNKs were identified as DBA lectin reactive, brown cells, and their number was calculated per 1 mm² with the help of an eyepiece-micrometer (Zeiss) and a light microscope (Zeiss).

Flow cytometry was used to compare percentages of peripheral (p) NKs or uMCs among the experimental groups. Single cell suspensions from uterus (gd5) or decidua (gd10) samples were obtained by enzymatic digestion with RPMI containing 5% P/S and 0.005% liberase for 90 min in the CO₂ incubator (37°C, 5% CO₂). Afterward, liberase was inactivated with RPMI supplemented with 10% FBS and 1% P/S, the tissue was squashed and filtered (100 µm net) and incubated further 30 min with RPMI + 10% FBS + 1% P/S in the CO₂ incubator. After centrifugation, the pellet was resuspended in 100 µl FACS buffer (10 g BSA, 1 g NaN₃ added to 1 l with PBS). Besides uterus/decidua samples, spleen, blood, mesenteric, and inguinal LNs were isolated, crushed, and filtered (100 µm net). Afterwards, erythrocytes were lysed with lysis buffer [89 g NH₄CL, 10 g KHCO₃, 0.38 g EDTA add to 1 l with aqua dest for preparing lysis buffer (10×)] for 10 min. Washing with RPMI followed centrifugation (1,200 rpm, 10 min, 4°C), than single cells suspensions were resuspended in 100 µl FACS buffer. Suspensions were stained 30 min in the darkness at 4°C with the following antibody solutions (dilution 1:100 with FACS buffer): FITC-conjugated rat anti-mouse CD122 (clone: 5H4), PerCP-conjugated hamster anti-mouse CD3 (clone: 145-2C11), and PE-conjugated mouse anti-mouse NK1.1 (clone: PK136) have been utilized for analysis of pNKs. PE-conjugated rat anti-mouse CD117 (clone: 2B8) and FITC-conjugated hamster anti-mouse FcεRα (clone: MAR-1) for examination of uMCs. FcεRα antibody was purchased from eBioscience (Frankfurt, Germany), all other antibodies were purchased from BD Pharmingen (Heidelberg, Germany). Unstained cells or cells stained with isotype antibodies served as controls. Cells were washed, resuspended with 100 µl FACS buffer, and measurement with FACS Calibur (BD Bioscience) after staining. Data analysis with CellQuest Pro software (BD Bioscience) and statistical analysis with GraphPad Prism 5.0 followed.

Spiral artery wall and lumen perimeter of two to ten SAs per implantation per female (gd10) were measured in hematoxylin-eosin (H/E)-stained implantation sections using a light microscope (magnification ×200) and the AxioVision 4 software (Zeiss, Germany). Calculation of SA wall and lumen diameter and determination of wall-to-lumen ratios followed. Afterward, the mean of wall to lumen ratio of multiple SA per one female was employed.

After standard dewaxing and rehydration procedure, implantation sections on slides were stained for 2 min with hematoxylin, rinsed twice in warm tap water, stained 1 min with eosin, and finally rinsed in distilled water. Dehydration by dipping the slides 10 times in each ethanol concentration (75, 95, 100%) and incubation in xylene for 2 min twice followed. Roti-Histokitt was used for covering the slides.

After fixation in 4% PFA and ethanol, one implantation per female was embedded in paraffin and cut in 5 µm sections. Immunofluorescence staining was performed after standard dewaxing and rehydration procedure. Staining required antigen retrieval treatment that was carried out in 0.1 M citrate buffer (pH 6.0) for 10 min in the microwave. Afterward, samples were washed and incubated at 4°C ON with anti-smooth muscle (SM) actin (clone: 1A4, Dako, dilution 1:50 in 10% BSA in TBS), followed by washing and incubation with the Alexa Fluor 555-conjugated secondary antibody goat anti-mouse (Life technologies, dilution 1:500 in 10% BSA in TBS) for 2 h at RT. For nuclear counterstaining VECTASHIELD^® mounting medium with DAPI (VECTOR laboratories) was used.

To investigate if insufficient SA remodeling influences blood pressure in pregnant mice, systolic and diastolic blood pressure was determined in female mice by non-invasive tail cuff method using the TSE Blood Pressure Monitoring Systems (TSE Systems) and calculating the mean of 10 single values. Before starting the measurements, mice were located under an infrared lamp for 10 min. A training period of 2 weeks preceded the experiments. Blood pressure was monitored every second day from gd0 to gd18.

Individual fetal and placental weights were measured at gd18 using a micro scale (Kern & Sohn GmbH), with FPI (fetoplacental index) calculated from these values.

In order to visualize uMCs as purple cells with granula, uterine sections from gd5 were stained with toluidine blue. After standard dewaxing and rehydration procedures, slides were incubated for 45 s in 0.1% toluidine blue solution and rinsed briefly in distilled water. Slides were dehydrated by dipping them 10 times in ethanol 75, 95, and 100% and incubation in xylene twice for 2 min each afterward. Slides were covered with Roti-Histokitt.

Total RNA was extracted from frozen uterus tissue (gd5) and frozen decidual tissue (gd10) using TRIzol^® (Life technologies) reagent. Tissue was homogenized using an Ultra Turrax T25 homogenizer (NeoLab) while adding TRIzol^®. RNA extraction with chloroform, precipitation with isopropanol, and washing with ethanol followed. RNA was diluted in RNA-free water, quantification was performed by ultraviolet absorbance at 260 nm, quality check by measuring absorbance at 260/280 nm (Synergy HT, Bio Tek Instruments). RNA storage occurred at −80°C. The RNA was reversed transcribed as followed: 2 µg of total RNA were incubated with oligo dTs and RNA-free water 10 min at 75°C and rested 2 min on ice. Incubation for 30 min at 37°C with dNTPs (2.5 mmol/l), DNase I (2 U/μl), and RNase inhibitor (40 U/μl) in a reaction buffer followed. Afterward, DNAse inactivation for 5 min at 75°C took place. After resting on ice for 2 min, reverse transcriptase (200 U/μl) was added together with RNase inhibitor (40 U/μl) and cDNA-synthesis occurred at 42°C for 60 min. Samples were stored at −20°C subsequent to inactivation of reverse transcriptase for 5 min at 94°C. Real-time polymerase chain reaction amplifications were performed with SYBR Green (Applied Biosystems) and an iQ5 Multicolor Real-Time PCR Detection System (BioRad). Each amplification reactions consisted of 1 µl cDNA, 6.5 µl SYBR Green PCR mastermix, 2 µl RNAse free water, 3 µl primermix, and 0.5 µl fluoresceine (50 nM). No template controls contained water instead of cDNA. The following primer were used: Mcpt1 (fwd: TGG CAC TTC TCT TGC CTT CT; rev: GCT CAC ATC ATG AGC TCC AA), Mcpt6 (fwd: GAC ATC GTG CTG GGT GAC; rev: ATA AGG AGG TGG GAG AGG C), and normalized to the housekeeping β-Actin (fwd: TGC GTC TGG ACC TGG CTG G; rev: ATC CTG TCA GCA ATG CCT GGG). Reactions were performed in duplicates as follows: 10 min 95°C, followed by 40 cycles of 30 s 95°C, 45 s 60°C for Mcpt1 or 45 s 55°C for Mcpt6, 30 s 72°C, and finally once 1 min 95°C.

Experimental results were analyzed with GraphPadPrism5 (GraphPad software). Numbers of animals/samples, statistical test, and P-values are indicated in each figure legend. P-values of <0.05 were considered to be statistically significant. Normality of the data was first assessed using the D’Agostino Pearson-Omnibus test before analyzing the data using parametric or non-parametric tests. Depending on their distribution, data are presented as medians or means ± SEM. Implantation rates and data obtained by flow cytometry or real-time PCR were analyzed with the Mann–Whitney U-test. For blood pressure values, histological measurements, fetal weights, placental weights, and FPI unpaired t-test were used for analysis of differences between the animal groups.

Numbers of uNKs were markedly, albeit not significantly, increased in MC-deficient W-sh mice compared to C57BL/6J mice (P = 0.0656) as assessed by staining of implantation cross sections obtained at gestation day (gd) 10 stained for DBA lectin (Figure 1A). Reconstitution of W-sh mice with BMMCs resulted in a significant reduction of uNKs and normalized uNK numbers (Figure 1A). uNK numbers were also significantly increased in MC-deficient Cpa3^Cre/+ mice compared to WT Cpa3^+/+ mice (Figure 1B).

When we used IL-15^−/− mice to assess the effects of NK cells on MCs, we first confirmed that these mice are NK deficient by flow cytometry and immunohistochemistry. Indeed, IL-15^−/− females exhibited dramatically decreased numbers of CD122⁺CD3^-NK1.1⁺ NKs at gd10 in the spleen, blood, as well as mesenteric and inguinal lymph nodes (Figures 2A–D). Their uNK count was near 0 (Figures 2E,F). We next studied whether NK absence influences the number of implantations or SA remodeling. Pregnant IL-15^−/− and C57BL/6N control mice showed comparable numbers of implantations (data not shown). As indicated by a statistically significant increased wall-to-lumen ratio (Figure 3A) at gd10 (P < 0.01), SA remodeling was impaired in IL-15^−/− females. Animals lacking NKs showed SA with a striking small lumen diameter and thicker arterial wall (Figure 3B iv) in contrast to WTs (Figure 3B i). Additionally, there was an increased expression of SM actin in vascular smooth muscle cells of IL-15^−/− mice (Figure 3B v, vi) in contrast to WT animals (Figure 3B ii, iii). Staining for SM actin helps visualizing the thickness of SM layers that should have been removed during an effective SA remodeling process, with a thicker SM layer indicating insufficient SA remodeling. Defects in SA remodeling are often causative of maternal hypertension in humans (2, 11, 12). We, therefore, investigated next whether the impaired SA remodeling in IL-15^−/− mice is associated with maternal hypertension. Using a tail-cuff system, we monitored systolic (Figure 3C) and diastolic (Figure 3D) maternal blood pressure throughout pregnancy and observed a statistically significant increase in systolic blood pressure at gd8, 12, 14, and 16 in IL-15-deficient mice in contrast to C57BL/6N mice. Nevertheless, none of the animals developed hypertension (>140 mmHg) at any pregnancy time point. To evaluate a possible influence of IL-15 deficiency on fetal development, weight of fetuses and placentas was measured and FPI was determined. At gd18, there were no differences in fetal weight (Figure 3E) between the groups. Interestingly, there was a statistically significant increase in placental weight (Figure 3F) in IL-15^−/− mothers in contrast to C57BL/6N controls (P < 0.001). Based on this, a statistically significant lower FPI (Figure 3G), indicative of an insufficient placental development, resulted in IL-15^−/− animals in contrast to the WT group (P < 0.05).

NK-deficient IL-15^−/− mice exhibited significantly higher numbers of uMCs compared to WT as assessed by toluidine blue staining of gd5 uterus cross sections (Figures 4A,B, P < 0.05). Flow cytometry analyses of uterine tissue from animals at gd5 as well as of decidual tissue from animals at gd10 confirmed this and showed markedly increased rates of CD117⁺FcεRα⁺ MCs in uNK-deficient mice as compared to WT (Figure 4C, P < 0.01). Furthermore, real-time PCR analyses demonstrated that gd10 decidua samples of uNK-deficient females have a statistically significantly higher expression of the MC mediators Mcpt1 and Mcpt6 as compared to WT (Figures 4D,E, P < 0.05). In contrast, uterus samples from gd5 showed no differences in the expression of Mcpt1 and Mcpt6 (Figures 4D,E). Collectively, these findings indicate that uMCs and MC-related mediators are upregulated in NK-deficient mice.