Human decidua and invasive trophoblasts are rich sources of nearly all human matrix metalloproteinases

Abstract

Trophoblast cell (CTB) invasion into the maternal endometrium plays a crucial role during human embryo implantation and placentation. As for all invasive cell types, the ability of CTB to infiltrate the uterine wall is facilitated by the activity of matrix metalloproteinases (MMPs), which is regulated by tissue inhibitors of MMPs (TIMPs). There is evidence for the expression of several MMPs and TIMPs in decidua. However, published data are limited. Therefore, to set a foundation for future research, we screened a panel of healthy human deciduas obtained during first, second and third trimester of pregnancy in addition to isolated decidual cell populations for the expression of all known human MMPs and TIMPs by RT–PCR, western blot and immunohistochemistry. In the decidual samples, we detected almost all MMPs and all four TIMPs at mRNA level. While the expression of proMMP-3 and active MMP-13 and -23 was down-regulated in the course of pregnancy, the pro forms of MMP-8, -19 and -23, active MMP-9, -10, -12, -15, -16, -26 and -28, and pro- and active MMP-14 increased towards the end of gestation. All MMPs and TIMPs were expressed in uterine natural killer cells, decidual fibroblasts and/or trophoblasts, with the exception of MMP-20 and -25. In summary, a remarkably broad spectrum of MMPs was expressed at the human feto-maternal interface, reflecting the highly invasive and remodelling effect on placenta formation. It can be speculated that expression of MMPs correlates with the invasive potential of CTBs together with a crucial role in activation of labour at term.

decidua, invasion, matrix metalloproteinases, pregnancy, trophoblasts

Introduction

In contrast to most non-primate pregnancies, human trophoblast cells (CTBs) are characterized by their strong invasiveness which ensures an adequate contact to the maternal blood system. This invasion process seems to be tightly regulated, since CTB invasion is limited to the decidualized endometrium (decidua) and the proximal third of the myometrium during normal pregnancy (Cartwright et al., 2010). For invasion CTBs initially attach to the uterine epithelium, degrade its basement membrane and extracellular matrix (ECM) and subsequently migrate into the decidual stroma (Fitzgerald et al., 2010). If this process is impaired, pregnancy failure results. While restricted invasion is associated with pre-eclampsia or fetal growth retardation, extreme invasiveness could result in placenta percreta (Lyall, 2006; Papadakis and Christodoulou, 2008; Pijnenborg et al., 2008).

The maternal microenvironment within the placental bed seems to play an important role in the regulation of CTB's migration. At the implantation site, matrix-producing decidual stromal cells (DCS) and maternal leucocytes, such as uterine natural killer (uNK) cells, elicit some degree of control over the fetal trophoblast (Burrows et al., 1996; Lash et al., 2010). Both DCS and maternal leucocytes produce a broad variety of cytokines and chemokines that might support CTB invasion (Saito, 2001; Lash et al., 2010). CTB even exert some paracrine effects by secretion of cytokines/chemokines and expression of the corresponding receptors (Bischof et al., 2000; Sato et al., 2003).

Invasive processes in general are facilitated by the expression and activity of matrix metalloproteinases (MMPs). MMPs belong to a family of structurally related zinc-dependent endopeptidases capable of degrading and remodelling specific components of the ECM. In total, 23 members of the human MMP gene family have been identified so far (Murphy and Nagase, 2008). MMPs are grouped into different classes, according to substrate specificity and primary structure (Murphy and Nagase, 2008). Most MMPs are produced as proenzymes and secreted into the ECM as inactive zymogens, where they are subsequently processed and activated by proteolytic cleavage (Kessenbrock et al., 2010). Considering their destructive capabilities, the transcription level of MMPs is tightly regulated by growth factors, hormones, cell–cell and cell–matrix interaction or cytokines (reviewed by Yan and Boyd, 2007). It is known that several proinflammatory cytokines such as interleukins IL-6, IL-1β and IL-8 are able to stimulate MMPs (Bischof et al., 2000; Dubinsky et al., 2010; Jovanović et al., 2010). Specific endogenous inhibitors such as α2-macroglobulin, or tissue inhibitors of MMPs (TIMPS) also regulate MMP activity by processing the pro-zymogens and inhibiting active enzymes (Kessenbrock et al., 2010).

Although the volume of literature on MMPs in pregnancy has increased in recent years, the few systematic studies involve only the most representative members of the MMP family. Expression of MMP-1, -2, -9 and -14 was found in decidua during pregnancy (Qin et al., 1997; Huppertz et al., 1998; Hurskainen et al., 1998; Goldman et al., 2003). Expression of TIMPs was also studied in decidual samples and TIMP-1, -2 and -3 were found in trophoblasts and DSC of first trimester placental beds (Hurskainen et al., 1996).

In order to lay the foundation for further functional studies, we screened a panel of healthy human deciduas during different weeks of gestation (first, second and third trimester) for the expression of all known human MMPs and TIMPs on mRNA and protein level. Expression and activity of MMPs and TIMPs were also analysed in the most prominent decidual cell populations, stromal cells and uNK cells, as well as in invasive CTB. This background knowledge of MMP expression during human placentation will be a basis to determine whether their expression correlates to the invasive potential of CTBs.

Materials and Methods

Tissue samples

Samples of human decidual tissues (n = 31) were obtained at different weeks of pregnancies (Table I) with the approval of the local Ethics Committee of the University of Würzburg Hospital. All patients signed a consent form, agreeing to the use of their decidua for this investigation. Decidual sections originated from therapeutic abortions of early gestations, from Caesarean sections for medical reasons at very early stages (which were unlikely to affect placental structure and function) and from delivery or Caesarean sections of healthy pregnancies at term. Decidual fragments were dissected free of conception products and blood clots. One part was fixed in 4% buffered formalin for subsequent paraffin embedding and the major part was either snap-frozen in liquid nitrogen and stored at −80°C for RNA and protein extraction or used immediately for isolation of specific cell populations. Prior to RNA/protein isolation, tissue samples were cut into sections on a cryostat. One representative section of each sample was stained with haematoxylin to define representative tissues without signs of necrosis. Tissue was selected to obtain approximately equal amounts of stromal tissue, glands and other structures. Another section was stained immunohistochemically with cytokeratin to identify CTB. Only those decidual samples with invasive trophoblasts and no signs of necrosis were chosen for RNA and protein extraction.

Table I

Information about decidua samples.

Tissue number	cDNA dilution	Week of pregnancy	Trimester of pregnancy
1	1:4	6	First
2	1:2	6
3	1:1	6
4	1:1	8
5	1:3	8
6	1:3	8
7	1:2	10
8	1:1	10
9	1:4	10
10	1:4	12
11	1:1	12
12	1:4	12
13	1:4	13	Second
14	1:2	14
15	1:3	15
16	1:2	16
17	1:1	16
18	1:2	18
19	1:3	18
20	1:2	20
21	1:1	20
22	1:1	24
23	1:2	27	Third
24	1:3	27
25	1:2	27
26	1:2	37
27	1:4	37
28	1:2	37
29	1:2	40
30	1:1	40
31	1:2	40

Single-cell isolation

For isolation of specific decidual cell populations, total decidua samples (2–6 g) were washed twice in phosphate-buffered saline (PBS) and minced into fragments of ∼1 mm³. Tissue fragments were then subjected to digestion in PBS containing 200 U/ml of hyaluronidase (Sigma, Taufkirchen, Germany), 1 mg/ml collagenase type IV (Seromed, Berlin, Germany), 0.2 mg/ml DNase I (2500 Kunitz U/mg, Sigma) and 1 mg/ml bovine serum albumin/fraction V (Sigma) under slight agitation at 37°C for 20 min. The obtained cell suspension was then filtered through sterile stainless steel wire mesh (50 μm), washed once in PBS and separated by gradient centrifugation (lymphocyte separation medium, PAA Laboratories, Pasching, Austria) following the manufacturer's protocol. Afterwards, the cells were washed twice in PBS and used immediately for subsequent cell isolation. Different cell populations were selected using magnetic cell sorting (MACS) technique (Table II). uNK cells, DSC and trophoblasts (CTB) were purified by positive selection using anti-CD56 microbeads (uNK), anti-fibroblast microbeads (DCS) and anti-ErbB2 microbeads (CTB), respectively (all Miltenyi, Bergisch Gladbach, Germany), in the magnetic cell separator Vario-MACS (MACS, Miltenyi) following the manufacturer's instructions. The positive uNK-, fibroblast-or CTB-enriched fractions were stored at −20°C and used for subsequent expression analysis. Purity of isolated cells was verified by fluorescence-activated cell sorting-staining and was higher than 90% of the liberated cell type (not shown).

Table II

Information about decidua cells.

Sample number	Week of pregnancy	Cell type	Microbeads
1	8	uNK	Anti-CD56
2	6
3	6
4	8
5	8	Fibroblasts	Anti-fibroblast
6	8
7	8
8	8
9	7
10	7	CTB	Anti-ErbB2
11	7
12	10
13	8
14	6
15	7

Total RNA purification

Total RNA was isolated from cryostat cuts of frozen decidual samples using the Nucleo-Spin RNA/Protein Kit (Macherey-Nagel, Düren, Germany). In total, 30 mg of tissue slices were homogenized using 350 μl RP1 lysis buffer following the manufacturer's protocol. Genomic DNA contaminations were removed by DNase I during RNA extraction. Total RNA was eluted in 60 μl of RNase-free water and purified samples stored at −20°C.

Semiquantitative RT–PCR analysis

First strand cDNA was generated from 1 to 5 μg of total RNA using the RevertAid H minus first strand cDNA synthesis kit (Fermentas, St. Leon Rot, Germany) and the provided oligo(dT)₁₈ primer. Reverse transcription was carried out at 42°C for 60 min in a final reaction volume of 20 μl and terminated by heating the samples at 70°C for 10 min. Resulting cDNAs were stored at −20°C. PCR was performed in 25 μl volume containing template cDNA, 2.5 U of Taq-polymerase (Eppendorf, Hamburg, Germany), 10X buffer with 1.5 mM MgCl₂ (Eppendorf), 200 μM dNTPs (Fermentas), 0.4 μM of both forward and reverse primers and formamide at a final concentration of 4%. Primer sequences were designed in flanking exons using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi), according to common nucleotide sequences (see Table III for primer sequences and PCR conditions for TIMP primers; for MMPs, see Köhrmann et al., 2009). PCR reactions were performed using a Px2 thermal cycler (Techne, Staffordshire, UK). PCR conditions consisted of an initial incubation at 94°C for 5 min, followed by 28–32 cycles at 94°C for 30 s and at optimized annealing temperatures for 30 s. The amount of cDNA was normalized to the strength of the PCR products of the ubiquitously expressed gene hydroxymethylbilane synthase (HMBS; Table III). Presence of CTB in decidua samples was confirmed by amplification of cErbB2 (Table III). PCR products were electrophoresed on 1% agarose gels and visualized using ethidium bromide (Sigma). The intensity of ethidium bromide luminescence was measured using the ImageJ software (NIH, Bethesda, USA; http://rsb.info.nih.gov/ij/).

Table III

Primers and conditions used for semiquantitative RT–PCR screening.

Gene	Forward primer	Reverse primer	Tm (°C)	Cycles	cDNA size (bp)	Genomic DNA (bp)
TIMP-1	5′-CTGCGGATACTTCCACAGGT-3′	5′-GCAGGATTCAGGCTATCTGG-3′	59	32	356	1.420
TIMP-2	5′-GATGCACATCACCCTCTGTG-3′	5′-TCCAGGAAGGGATGTCAGAG-3′	59	32	385	2.042
TIMP-3	5′-GTGCAACTTCGTGGAGAGGT-3′	5′-AGTGTCCAAGGGAAGCTCAG-3′	59	32	345	1.386
TIMP-4	5′-CGAGCCCTGGGAGGACCTGT-3′	5′-GGCAACAGGCTGAGGGCAGG-3′	65	32	465	1.079
cErbB2	5′-ACCTTCTACCGCTCACTGCT-3′	5′-CAAGTACTCGGGGTTCTCCA-3′	57	32	585	1.017
HMBS	5′-GGAAGTGCGAGCCAAGGACCAGGA-3′	5′-AGGGACATGGATGGTAGCCTGCAT-3′	60	28	241	598

Tm, annealing temperature; bp, base pair.

Western blot

Proteins were extracted from 30 mg of cryo-cut tissue samples, which were lysed in precooled Ripa-buffer (Pierce, Rockford, Ilinois) containing phosphatase inhibitors (Phosphatase Inhibitor Cocktails Set II, Calbiochem, Germany), protease inhibitors (complete, Roche, Germany) and reducing agent 2.5 mM DTT (dithiothreitol, Sigma). The mixture was incubated on ice for 30 min, combined with vortexing every 10 min. The cell lysates were clarified of cell debris by centrifugation at 14 000 g for 5 min through a QIAshredder spin column assembly (Qiagen, Hilden, Germany). Afterwards, the samples were mixed in 5X loading buffer (Fermentas), denatured at 95°C for 5 min, chilled quickly on ice and stored at −20°C for further analysis.

Protein concentration was determined by the Bradford-method (Bradford, 1976) using Coomassie Brilliant Blue (Roti-Quant; Roth, Karlsruhe, Germany). Samples were subjected to electrophoresis on 10% polyacrylamide gel (SDS–PAGE), blotted onto nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany) for 45 min at 10 V using a semi-dry-transfer unit (PeqLab, Erlangen, Germany). The membrane was stained with Ponceau-red to verify that the proteins were blotted. To avoid unspecific binding the membrane was then blocked with 5% non-fat milk protein in PBS/Tween at RT for 1 h. Subsequently, the membrane was incubated with the appropriate primary antibody diluted in 2% non-fat milk and PBS/Tween at 4°C for 18 h. Clones, sources and dilutions of the primary antibodies used in this study are summarized in Table IV. After washing with PBS, the membrane was incubated with species-specific horseradish peroxidase (HRP)-conjugated secondary antibodies (Table IV) for 60 min at RT. A monoclonal mouse anti-β-actin primary antibody, diluted 1:10 000 (Abcam, Cambridge, USA) was used as internal control. Immunoblots were visualized by home made ‘enhanced chemiluminescence’ (Haas, 2005) with subsequent exposure on an X-ray film (Fuji Super RX medical X-ray films; Fuji Photo Film, Düsseldorf, Germany).

Table IV

List of primary antibodies for western blot.

Gene	Application	Protein forms detected by WB	Species	Type/clone	Dilution in WB	Dilution in IHC-P	Company
MMP-1	WB, IHC-P	Latent and active	Rabbit	Polyclonal	1:750	1:100	Biozol
MMP-2	WB	Latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-3	WB	Latent and active	Mouse	SPM 293	1:500		Abcam
MMP-7	WB	Latent and active	Mouse	111433	1:500		Abcam
MMP-8	WB	Latent and active	Mouse	115-13D2	1:1000		Chemicon
MMP-8	IHC-P	Latent and active	Mouse	115-13D2		1:100	Chemicon
MMP-9	WB	Latent and active	Mouse	9D4.2	1:500		Chemicon
MMP-10	WB	Latent and active	Mouse	1VC5	1:500		Chemicon
MMP-11	WB	Latent and active	Mouse	SL 3.01	1:500		Abcam
MMP-11	IHC-P	Latent and active	Mouse	SPM 199		Prediluted	Biozol
MMP-12	WB	Latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-13	WB	Latent and active	mouse	87512	1:500		R&D
MMP-14	WB	Latent and active	Rabbit	Polyclonal	1:500		Abcam
MMP-15	WB	Latent and active	Rabbit	Polyclonal	1:500		Abcam
MMP-16	WB	Latent and active	Rabbit	Polyclonal	1:2500		Abcam
MMP-19	WB	Latent and active	Rabbit	Polyclonal	1:3000		Biozol
MMP-19	IHC-P	Latent and active	Rabbit	Polyclonal		Prediluted	Biozol
MMP-23	WB	Latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-23	IHC-P	Latent and active	Rabbit	Polyclonal		Prediluted	Abcam
MMP-24	WB	Latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-26	WB	latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-27	WB	Not specified	Rabbit	Polyclonal	1:1000		Abcam
MMP-28	WB	Not specified	Rabbit	Polyclonal	1:1000		Abcam
ß-actin	WB	β-actin	Mouse	M/Abcam 8226	1:10.000		Abcam
Cytokeratine	IHC-P	Cytokeratin	Mouse	AE1/AE3	1:100		DAKO
CD34	IHC-P	CD34	Mouse	QBEnd10	1:100		DAKO
CD45	IHC-P	CD45	Mouse	2B11 PD7/26	1:50		DAKO
CD56	IHC-P	CD56	Mouse	1B6	1:100		Novocastra
MMP-2	WB	Latent and active	Rabbit	Polyclonal	1:1000		Abcam
MMP-3	WB	Latent and active	Mouse	SPM 293	1:500		Abcam

WB: Western blot; IHC-P: Immunohistochemistry (formalin-fixed paraffin-embedded section)

Immunohistochemistry

Immunohistochemical analyses of tissue samples were performed to localize the different MMPs in decidua. Tissues were sectioned (2 µm) from formalin-fixed, paraffin-embedded tissue blocks, mounted on glass-slides (Superfrost, Langenbrink, Emmendingen, Germany) and dried overnight at room temperature. Paraffin sections were dewaxed twice with xylene and rehydrated in a graded series of ethanol and in distilled water. The sections were predominantly stained without pretreatment for antigen demasking. Only in the case of MMP-19 were slides pretreated in the microwave oven in a 10 mM sodium citrate buffer solution (pH 6.0) for 10 min (750 W/s). Endogenous peroxidase activity was then blocked with 0.3% hydrogen peroxide in methanol for 10 min. To reduce possible unspecific binding capacity of the tissue, slides were treated with a solution of 1% human immunoglobulin (Beriglobin; Aventis Behring, Marburg, Germany) in phosphate-buffered saline for 15 min at room temperature. The sections were then incubated overnight at 4°C with one of the respective primary polyclonal antibodies directed against MMP-1, -8, -11, -19, -23, cytokeratin, CD34, CD45 and CD56 (Table IV) diluted in antibody diluent (DAKO, Hamburg, Germany). After washing with PBS, the sections were incubated with the HRP-labelled LSAB2 kit (streptavidin–biotin system; DAKO). Peroxidase activity was then developed with diaminobenzidine (DAB; DCS, Hamburg, Germany) as a substrate for 5 min, which resulted in brown staining. The slides were counterstained with haematoxylin and analysed using a light microscope Othoplan (Leica, Germany). Negative control experiments were carried out by staining with a mixture of the appropriate isotype-control antibodies.

For double immunohistochemical staining, the sections were pretreated in the microwave with 10 mM sodium citrate buffer solution (pH 6.0) for 15 min (750 W/s). Endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 10 min, as mentioned above. The sections were then incubated with the first monoclonal antibody against MMP-19 or MMP-23 at appropriate dilutions followed by incubation with HRP-labelled LSAB2 kit. Peroxidase activity was developed with DAB. After that the sections were incubated with the second antigen-specific antibody CD56 (clone 123C3, Monosan, the Netherlands; dilution 1:100) or cytokeratin, respectively, followed by a second incubation with the HRP-labelled LSAB2 kit. The detection reaction for the second colour (green) was developed using the Histogreen peroxidase substrate kit (Linaris, Wertheim, Germany). Negative control experiments were carried out by staining with a mixture of the appropriate isotype-control antibodies.

Gelatin zymography

MMP-2 and MMP-9 enzymatic activity of isolated decidual cells (Table II) was measured by gelatin zymography (as described in detail in Schröpfer et al., 2010). After a 24 h incubation of the cells in serum-free medium, the cell supernatants were collected and cells were briefly lysed in precooled Ripa-buffer. Conditioned media (20 μl) or cell lysates (10 μg protein) were incubated with SDS gel sample buffer (Invitrogen, Heidelberg, Germany) for 10 min at room temperature and then electrophoresed on 10% ‘Novex precast zymogram gelatine’ gels (Invitrogen). The gels were run, renatured and developed according to the manufacturer's instructions. Following electrophoresis, gels were rinsed twice with ‘Novex Zymogram Renaturing Buffer’ (30 min per wash at room temperature) and then once with fresh ‘Novex Zymogram Developing Buffer’ (Invitrogen). Afterwards, gels were incubated in the same buffer for 18 h at 37°C and briefly rinsed in distilled water. Gelatine activity was visualized by staining the gels with Coomassie Brilliant Blue G250 for 2 h. Molecular weight of both gelatinases was determined using size marker SeeBlue plus2 (Invitrogen) and recombinant MMP-2 and MMP-9 proteins (R&D Systems, Wiesbaden, Germany).

Data analysis and statistics

The intensity of Ethidium bromide luminescence and protein expression in western blot images was quantified densitometrically using the ImageJ software (NIH, Bethesda, USA, http://rsb.info.nih.gov/ij/) and normalized in respect to the corresponding fragment concentration of the housekeeping genes HMBS or β-actin, respectively. Boxplots were generated using GraphPad Prism 4.0 Software (La Jolla, USA). Comparison of expression values between the groups: first (1°), second (2°) and third (3°) trimester were compared using the non-parametric two-tailed Mann–Whitney U-test and P-values of <0.05 were considered statistically significant.

Results

Expression of MMPs in decidua on RNA level

The results of the semiquantitative RT–PCR of MMPs in decidual samples are summarized in Fig. 1; boxplot analyses of the densitometrically quantified PCR products are shown in Fig. 2; and P-values are listed in Table V.

Table V

P-values obtained by comparison of the MMP-mRNA expression between decidua samples of the first (1°), second (2°) and third (3°) trimester, respectively (Man–Whitney U-test).

	1°–2°	2°–3°	1°–3°
MMP-1	P < 0.05	P < 0.05
MMP-2			P < 0.05
MMP-3	P < 0.05	P < 0.001
MMP-7	P < 0.05
MMP-10	P < 0.05		P < 0.01
MMP-16		P < 0.05	P < 0.01
MMP-24		P < 0.05	P < 0.05
MMP-26			P < 0.01
MMP-28			P < 0.05

MMPs with no significant values obtained are not listed.

Empty fields: not significant results.

Figure 1

Expression analysis of MMPs in human decidua samples by semiquantitative RT–PCR. Gel analysis of the PCR products as obtained by semiquantitative RT–PCR, visualized on 1% agarose gels and stained by Ethidium bromide. Decidua samples were obtained from different weeks of pregnancy (first, second and third trimester). Total RNA from decidua specimens was used as template for RT–PCR analysis. PCR amplification, performed with primers specific for each transcript, resulted in longer amplicons if human genomic DNA was amplified [positive control (+)] and in shorter amplicons representing cDNAs. The various cDNA concentrations were normalized to that of the HMBS housekeeping gene (last line), which was used as internal loading control. According to these initial results, the PCR reaction was performed in a total volume of 25 μl containing 1 μl of aqueous dilution of cDNA. Presence of CTB in decidua samples was confirmed by amplification of cErbB2 mRNA (second last line).

Figure 2

Boxplot analysis of densitometrically quantified expression of MMP mRNAs. The expression level of each gene was normalized to the corresponding expression of HMBS. For each MMP three boxes were mapped: expression in first (1°) trimester decidua samples (white box); expression in second (2°) trimester decidua samples (light grey box); expression in third (3°) trimester decidua samples (dark grey box). The line within the boxplot corresponds to the median value, the box length to the interquartile range and bars indicate the smallest and largest observations. Significances are not shown herein but summarized in Table V.

There were no differences in the median expression levels of MMP-8, -15, -17, -21, -23 and -27 in the three gestation stages. There was virtually no detectable expression of MMP-20 in any of the samples investigated. MMP-25 was not expressed in the first two trimester samples, but a very weak expression in some of the third trimester decidual samples was observed.

MMP-1, -3, -7 and -10 showed a significantly increased expression from the first to the second trimester of pregnancy. By the third trimester MMP-1, -3 and -7 expression levels dropped back to the base levels of the first trimester, whereas a further increase was observed for MMP-10. MMP-2 showed the same expression pattern as MMP-1 and -3, but only the differences in expression levels between first and third trimester reached statistical significance. The differences in the expression levels of MMP-9, -11, -12, -14 and -19, although clear in tendency and comparable to the previous MMPs, were statistically not significant. A continuously increasing expression of MMP-28 mRNA was observed, with a significantly higher value in the third compared with the first trimester.

In contrast, a significant decrease in the expression of MMP's mRNA in decidual samples from the first to the third trimester was observed for MMP-16, -24 and -26. While the mRNA levels for MMP-16 and -24 were relatively stable throughout the first two trimesters, a continuous drop in the mRNA content was detected for MMP-26. A comparable tendency, albeit not significant, was seen for MMP-13.

Expression of MMP proteins in decidua

Results obtained in western blots are summarized in Fig. 3 and boxplot analyses of the densitometrically quantified expression of MMP proteins are shown in Fig. 4. Corresponding P-values are listed in Table VI.

Table VI

P-values obtained by comparison of the MMP protein expression between decidua samples of the first (1°), second (2°) and third (3°) trimester, respectively (Man–Whitney U-test).

	1°–2°	2°–3°	1°–3°
proMMP-3	P < 0.001		P < 0.01
proMMP-8	P < 0.01
MMP-9	P < 0.05		P < 0.05
MMP-10			P < 0.01
MMP-12	P < 0.05	P < 0.05
MMP-13	P < 0.05		P < 0.01
proMMP-14; MMP-14		P < 0.01	P < 0.01; P < 0.01
proMMP-16; MMP-16		P < 0.05; P < 0.001	P < 0.01; P < 0.01
proMMP-19		P < 0.05	P < 0.05
proMMP-23; MMP-23	P < 0.01; P < 0.01		P < 0.01
MMP-26	P < 0.05		P < 0.01
MMP-28 (62 KDa); MMP-28 (46 KDa)	P < 0.05		P < 0.01; P < 0.01

MMPs with no significant values obtained are not listed.

Empty fields: not significant results.

Figure 3

Expression analysis of MMP proteins in human decidua samples by western blot. Autoradiography films representing the western blot results are presented. Protein lysates were isolated from decidua samples derived from different weeks of pregnancy. The specific antibodies detected inactive (longer) and active (shorter) forms of MMP proteins. β-actin was used as the internal loading control.

Figure 4

Boxplot analysis of densitometrically quantified expression of MMP proteins. Protein levels were normalized to the corresponding expression of β-actin. For each MMP form detectable, three boxes were mapped: expression in first (1°) trimester decidua samples (white box); expression in second (2°) trimester decidua samples (light grey box); expression in third (3°) trimester decidua samples (dark grey box). The black line within the box represents the median, boxes show the quartiles and bars indicate the minimum and maximum value.

A relatively stable protein level from the first to the third trimester of gestation was observed for MMP-1, -2 (latent and active form), -11 and the 74 kDa form of MMP-27. However, expression of the 50 kDa form of MMP-27 tends to increase from the first to the third trimester. The expression level of MMP-13 was significantly lower in both second and third trimester, compared with the first trimester, and the weak expression of pro-MMP-3 decreased significantly from the first to the second trimester of pregnancy and then remained stable.

Compared with the first trimester, we observed a significant increase in MMP protein in the third trimester for MMP-10, pro-MMP-19 (while for MMP-19 the increase did not reach significance), MMP-26, as well as for both the pro- and active forms of MMP-14 and -16 and two protein forms of MMP-28 (62 and 46 kDa). The continuous increase of MMP-24 and -7 throughout the course of pregnancy did not reach significance. After an initial increase in protein from the first to the second trimester, the expression levels of pro-MMP-8, MMP-9 and -15 remained stable to the end of pregnancy. A significant initial increase was also observed for MMP-12, which then dropped to the first trimester's base level in the third trimester. Its expression showed, however, considerable sample-to-sample variation of expression in all three gestation stages.

The expression of MMP-23 decreased significantly from the first to the second trimester and then increased in the third trimester of pregnancy to a median expression level lower than in the first trimester. However, a converse expression pattern was detected for the latent form of MMP-23.

Immunostaining of tissue samples

In order to identify the cell types producing MMPs within the decidua in situ, we performed immunostaining on tissues derived from the 7th or 8th week of pregnancy. Specific antibodies resulting in reliable and reproducible staining were available for MMP-1, -8, -11, -19 and -23 only (Fig. 5). Tissue incubated with a mixture of the appropriate isotype-control antibodies served as the negative control.

Figure 5

Immunostaining of MMP expression on paraffin-embedded decidua tissue sections. (A) MMP-1, (B) MMP-8, (C) MMP-11, (D) MMP-19 and (E) MMP-23 expression (brown colour) in decidua tissue samples, deriving from the first trimester of pregnancy (7th or 8th week of pregnancy), was visualized using specific antibodies (Table IV). (F) Presence of invasive trophoblasts (CTB) was detected by staining of decidua samples with antibody directed against cytokeratin; (G) CD34 + endothelial cells; (H) CD45 + lymphocytes; (I) CD56 + uNKs. (J) Tissue incubated with a mixture of the appropriate isotype-control antibodies served as the negative control. Magnification: ×400 for all pictures.

MMP-1, -8, -11, -19 and -23 were clearly positive in the cytoplasm of DSC and CTB (Fig. 5A–E). MMP-1 was also found in endothelial cells of blood vessels (Figure 5A, star). In addition, MMP-8, -11 and -23 were found in decidual glands (Fig. 5B, C and E, open arrowhead). Strong expression of MMP-19 was also identified in uNK cells (Fig. 5D, arrows) and MMP-23 macroscopically allocated to the cytoplasm of uNK cells (Fig. 5E, arrows).

Double staining using anti-MMP-19 or -23 and anti-CD56 or anti-cytokeratin, respectively, was performed to confirm the expression of MMP-19 by uNK cells and MMP-23 by uNK and CTB. This resulted in clear MMP-19 staining of single uNK cells in decidual tissue (Fig. 6A, arrows). Furthermore, the double-staining confirmed the previous finding of MMP-23 expression by uNK cells (Fig. 6B, arrows) and CTBs (Fig. 6C, arrows).

Figure 6

Double staining of the uNK cells and cytotrophoblasts. (A) Double staining of uNK cells with antibodies directed against MMP-19 and CD56 (a marker for uNKs); (B) double staining of uNK cells with antibodies against MMP-23 and CD56; (C) double staining of cytotrophoblasts with antibodies against MMP-23 and cytokeratin (positive marker for cytotrophoblasts); green: CD56 or cytokeratin. (D) Presence of invasive trophoblasts (CTB) was detected by staining of decidua samples with antibody directed against cytokeratin; (E) CD56 + uNKs. (F) Tissue incubated with a mixture of the appropriate isotype-control antibodies served as the negative control. Magnification: ×400 for all pictures.

Expression of MMPs and TIMPs in isolated decidua cells

We found mRNA expression of all MMPs investigated, at different expression levels, in uNK cells, DSC and CTB, with the exception of MMP-20 and -25 (Fig. 7A). While this finding corresponds to the lack of MMP-20 expression in whole decidual tissue, MMP-25 was detected—albeit at very low levels—in whole decidual tissue. Therefore, it seems to be expressed by other cell types not investigated herein. MMP-2, -11 and -19 were expressed at high levels in all three cell types investigated. However, in tendency, the uNK cells seem to have slightly weaker expression levels of MMP-11 and -19 compared with DSC and CTB. MMP-7, -9, -12, -14, -17 and -23 were also detectable in all cell types investigated, however, at a median level. MMP-7 and -12 showed the strongest expression in CTB. MMP-9 was detected in CTB, too, but at a lower level. In the case of MMP-14 and -23, the RT–PCR results for one cell type were quite different and did not show a clear tendency. MMP-1 and -10 expression was high in DSC and CTBs. Furthermore, for MMP-13, -16 and -28, the expression of mRNA was detected only in one sample of uNK cells, the same sample being highly positive for MMP-1 compared with the other isolates. In general, uNKs seem to be negative or only slightly positive for MMP-1, -3, -8, -13, -16, -20, -21, -24, -25, -26, -27 and -28. In contrast, CTB seem to be highly and/or specifically positive for MMP-3, -15, -27 and -28, while MMP-13, -16 and -24 predominate in DSC.

Figure 7

Analysis of MMPs in isolated decidua cells by semiquantitative RT–PCR, western blot and gelatin zymography. Decidua cells including uNK cells, fibroblasts and invasive cytotrophoblasts (CTB), were isolated from first trimester tissue samples (Table II). (A) Extracted total RNA from the cell was used for RT–PCR analysis of 23 MMPs, as described in Fig. 1. PCR amplification was performed with primers specific for each of the MMP transcripts (Table III). (B) Protein lysates were isolated from decidua cells and separated by polyacrylamide gel electrophoresis. Expressed gelatinases (inactive and active MMP-2 and MMP-9) were visualized using specific antibodies (Table IV). β-actin was used as internal loading control. (C) Decidua cells were incubated in serum-free RPMI medium for 24 h. Cell lysates as well as samples of conditioned medium were assayed for MMP-2 and MMP-9 activity by gelatin zymography. Gelatinolytic activity of pro- and active MMP-2 and active MMP-9 are visible as a clear area on the gel, indicating where the gelatine has been digested. In addition, two MMP complexes of ∼120 and 130 kDa were detected. Mw, Molecular weight marker SeeBlue plus2 (Invitrogen); Positive control (+), recombinant MMP-2 and MMP-9 proteins (R&D Systems, Wiesbaden, Germany).

Expression of both the pro- and active form of MMP-2 was detected in all three isolated decidual cell types, with the highest level in DSC (Fig. 7B). These data are consistent with the RT–PCR analysis. Using gelatin zymography we showed that all three of these cell types were secreting corresponding amounts of both forms of MMP-2 protein into serum-free medium, while in cell lysates gelatinolytic activity of MMP-2 was detected in DSC only (Fig. 7C). The highest expression level of active MMP-9 was detected in DSC by western blot analysis (Fig. 7B). Gelatinolytic activity of its active protein form was identified in all three analysed cell types (Fig. 7C). In addition, gelatinolytic activity of two MMP complexes of ∼120 and 130 kDa could also be identified in uNKs and trophoblasts in both cell lysates and cell supernatants (Fig. 7C). No active MMP-9 was detected in any of the cell lysates.

We also investigated expression of four TIMPs in isolated decidua cells by semiquantitative RT–PCR. Their transcripts displayed equal levels of expression in the examined cells (Supplementary data S1).

Discussion

To the best of our knowledge, the present study is the first complete compilation of the expression pattern of all human MMPs on mRNA and protein levels in human decidua samples obtained from different weeks of pregnancy (first, second and third trimester). MMPs, along with other proteases, are crucial for the invasion processes of many different types of cells, amongst them cytotrophoblasts invading the uterus in human placentation (Huppertz et al., 1998). MMPs also play a crucial role in the activation of labour at term (Weiss et al., 2007). Our analyses showed that all known human MMPs, with the exception of MMP-20, were expressed at mRNA level in human decidua in the course of pregnancy. Furthermore, the synthesis of the proteins for most of the MMPs was elevated at the start of labour, although expression of their transcripts was lower towards the end of gestation. Levels of MMP mRNA do not necessarily have to correlate with their protein level since the protein-per-mRNA ratio is influenced by different factors like, e.g. mRNA stability, transcription or translation (reviewed by de Sousa Abreu et al., 2009). Thus, the cells that possess MMP transcripts have the capacity for their translation to produce proteins, which they require in the local decidual matrix at certain gestation stage.

The three collagenases MMP-1, -8 and -13 are responsible for the degradation of the collagens I and III, which predominate in fetal membranes (Amenta et al., 1986). Our data show that MMP-1 is expressed at varying levels at all gestational ages. It has been shown that chorioamnionitis, which could cause preterm delivery, is accompanied by elevated levels of MMP-1 (Oner et al., 2008), which could fit with its higher expression at the beginning of second trimester. Furthermore, expression of MMP-1 was localized to the cytoplasm of trophoblasts and their surrounding matrix, decidual stroma cells, as well as endothelial cells of blood vessels in early and term decidua specimens; as suggested by our study and others (Vettraino et al., 1996; Huppertz et al., 1998; Oner et al., 2008). In contrast to our findings, Huisman et al. (2004) could not detect gelatinolytic activity of MMP-8 and -13 in decidua, which might be due to different sensibility of the applied methods. For MMP-8 we also found staining in the cytoplasm of early trophoblasts and decidual stroma cells, as well as in decidual glands, whereas expression of the MMP-13 transcript was specific for maternal first trimester fibroblasts.

The expression of the two gelatinases MMP-2 and MMP-9 in decidua has been well investigated in many studies using different methods (Huppertz et al., 1998; Xu et al., 2000; Isaka et al., 2003; Huisman et al., 2004; Jones et al., 2006). Our western blot analysis showed that the intensity of active MMP-2 remained stable during pregnancy, whereas its inactive protein form displayed the highest expression in first and third trimester deciduas (preterm and term labour). Furthermore, expression of pro-MMP-2 was up-regulated after termination of pregnancy in the 13th week, due to infections. It has been shown that in first trimester deciduas invasive trophoblasts and their surrounding extracellular matrixes, as well as decidual cells, were the main sites of MMP-2 immunoreactivity (Huppertz et al., 1998; Huisman et al., 2004; Jones et al., 2006). We were able to detect high levels of gelatinolytic activity for pro-MMP-2 secreted by cultured early trophoblasts, which is in accordance to previous in vitro studies (Xu et al., 2000; Isaka et al., 2003). In addition, we found activity for both pro- and active MMP-2 in medium conditioned with first trimester uNK cells and DSC, as well as in DSC lysates. Expression of active MMP-9 significantly increased from first to second trimester and then remained stable to the end of pregnancy. Its expression was previously found in first and third trimester trophoblasts and their ECM (Polette et al., 1994; Hurskainen et al., 1996; Huppertz et al., 1998; Huisman et al., 2004). Our zymography analysis showed that cultured first trimester CTB secrete high amounts of active MMP-9, which is in accordance to previous in vitro analysis of those cells (Xu et al., 2000). In addition, we showed that first trimester uNKs and fibroblasts also secrete active MMP-9, which had not been previously investigated. Furthermore, zymography analysis detected two additional gelatinase bands of ∼120 and 130 kDa in CTB and uNK cell lysates, as well as in medium conditioned by those two cell types. A MMP complex of 130 kDa has already been described in periodontal ligament and in term deciduas, and its activity was elevated at the start of labour (Goldman et al., 2003; Snoek-Van Beurden and Von den Hoff, 2005). Goldman et al. (2003) suggest that this complex could be a hetero-dimer of MMP-9 and lipocalin.

The three stromelysins MMP-3, -10 and -11 and the two matrilysins MMP-7 and MMP-26 were found in almost all analysed decidual samples on mRNA level, although Pilka et al. (2003) could not detect MMP-26 mRNA in early decidual tissue, probably due to fewer analysed tissue samples or different primers used. Pro-MMP-7, MMP-10 and -11 were all expressed by early CTBs and DSCs. Literature data, available for MMP-3 and -7 only, are in line with our findings (Vettraino et al., 1996; Huppertz et al., 1998; Jones et al., 2006; Nishihara et al., 2008; Oner et al., 2008).

All membrane-type MMP mRNAs were expressed in decidua in the course of pregnancy, with the exception of MMP-25 that was identified at a very low level in some term deciduas only. MMP-14 mRNA was detected in all three cell types isolated from the first trimester decidua. Our data are in agreement with earlier findings that MMP-14 is produced by pure trophoblasts propagated in culture as well as in first trimester decidual tissue (Hurskainen et al., 1998; Xu et al., 2000; Plaisier et al., 2008). We found expression of MMP-15 mRNA to be specific for early CTB, which is in line with findings obtained by in situ hybridization of decidualized endometrium (Bjørn et al., 2000). Two protein forms of MMP-27 (50 and 74 kDa) were expressed at an equally high level in all analysed decidua samples, whereas the level of expression of MMP-28 (46 and 62 kDa protein forms) increased continuously towards the end of pregnancy. To date, the precise sizes of breakdown products and putative active and inactive forms of MMP-27 and -28 are still not known. Furthermore, mRNAs of both MMPs were expressed predominantly in CTB.

Production of specific MMPs by CTB is important for the selective degradation of matrix proteins in the decidua, which then enables their migration into the decidual stroma and vasculature. Our immunohistochemical analysis of first trimester tissue samples showed clear expression of MMP-1, -8, -11, -19 and -23 by CTBs. The maternal microenvironment is also involved in the production of MMPs and thus, plays an important role in the processes of implantation. MMP-1, -8, -11, -19 and -23 are also examples for the possible cross-talk between mother and embryo during early pregnancy, since their staining was also identified in decidual stroma cells, decidual glands, uNK cells or blood vessels. Our RT–PCR analyses also suggest that additional MMPs are expressed by either trophoblasts and/or maternal decidual cells in first trimester tissue samples, as discussed above.

During normal placentation natural inhibitors of MMPs (TIMPs) are responsible for the precise regulation of CTB's invasive potential. It has been shown that TIMPs are produced by CTBs as well as by the cells of fetal membranes (Hurskainen et al. 1996; Marzusch et al., 1996; Ruck et al. 1996), which are of importance for control of CTB invasion. Our RT–PCR analysis detected mRNA expression of all four TIMPs in first trimester CTBs, as well as in uNKs and fibroblasts, and thus, confirmed previous findings.

The present study suggests that the diversity and the amount of produced and secreted MMPs are of great importance for the development of pregnancy. MMPs participate in tissue turnover at the beginning of the pregnancy, important for the invasion of CTBs, and play also a crucial role in the activation of labour at term. They are expressed mostly on demand across gestation which explains detected changes in their expression. Our findings could be thus the basis for further functional studies to clarify the special functional importance of MMPs for physiological placentation.

Authors’ roles

J.A. set up the experiments, collected the data, analysed and interpreted the results and drafted the manuscript. S.E.S. helped in collecting the samples and preparation of the manuscript. C.H. participated in the study design and finalization of the manuscript. S.F., M.K. and R.B. performed the laboratory experiments. U.K. participated in the study design, experimental concept, interpretation of the results and revision of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by a grant of the ‘Deutsche Forschungsgemeinschaft’ (KA-1253/2-3) to U.K. The content of this publication is solely in the responsibility of the authors.

Acknowledgements

We thank L. Sevenson-Knebel for language correction of the manuscript.