Escherichia coli induced matrix metalloproteinase-9 activity and type IV collagen degradation is regulated by progesterone in human maternal decidual

Abstract

Background

Escherichia coli (E. coli) is one of the main bacteria associated with preterm premature rupture of membranes by increasing pro-matrix metalloproteinase 9 (proMMP-9) and degradation of type IV collagen in human feto-maternal interface (HFMi). proMMP-9 is regulated by progesterone (P4) but it is unclear whether P4 inhibits proMMP in human maternal decidual (MDec). This study aimed to determine a role of P4 on proMMP-2 and − 9 and type IV collagen induced by E. coli infection in MDec.

Methods

Nine HFMi were mounted in a Transwell system. MDec was stimulated with P4 or E. coli for 3-, 6-, or 24-hours. proMMP-2, -9 and type IV collagen were assessed.

Results

Gelatin zymography revealed an increase in proMMP-9 after 3, 6, and 24 h of stimulating MDec with E. coli. Using immunofluorescence, it was confirmed the increase in the HFMi tissue and a reduction on the amount of type IV collagen leading to the separation of fetal amniochorion and MDEc. The degradative activity of proMMP-9 was reduced by 20% by coincubation with P4.

Conclusions

P4 modulates the activity of proMMP-9 induced by E. coli stimulation but it was unable to completely reverse the degradation of type IV collagen in human MDec tissue.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-024-06847-8.

Background

Matrix metalloproteinases (MMPs) are a family of zinc and calcium-dependent neutral endopeptidases produced by a wide variety of cell types, including polymorphonuclear innate immune cells [1–3], human placental cells [4–6] and the feto-maternal interface [7–10]. These cell types release MMPs in the form of inactive zymogens or proMMPs and are activated by small extracellular proteins known as tissue inhibitor of metalloproteinases [11, 12].

MMPs are triggered by cellular oxidative stress processes [13–15] and inflammatory responses [16, 17]. These proteins induce degradation of various structural components of the extracellular matrix including collagen type I, IV, V, VII, X, fibronectin, elastin, proteoglycan [18, 19], basement membrane [20] and cell-binding adhesion proteins [21, 22]. MMPs are classified in 6 groups according to the substrate degraded: stromelysins (MMP-3, MMP-10 and MMP-11); collagenases (MMP-1, MMP-8 and MMP-13); gelatinases (MMP-2 and MMP-9); matrilysins (MMP-7 and MMP-26); membranal type 1 (MMP-14, -15, -16, and − 24) and type II (MMP-23) [23–25]. The chromosomal location of all MMPs has been determined [26]. These MMPs have been detected in human feto-materna interface (HFMi) and amniotic fluid [27, 28]. Differential expression between MMP-2 and MMP-9 indicate that MMP-2 is constitutively expressed while MMP-9 is considered an inducible metalloprotease.

HFMi is composed of two compartments: the fetal amniochorion (FAC) layer formed by a single sheet of cuboidal amnion epithelial cells, chorionic trophoblasts, and fibroblasts, embedded in the extracellular matrix and containing integrins, laminins, microfibrils, elastin [29, 30] and collagen type I, II, III, IV, V and VI [27, 31]. This connects with the second compartment, the maternal decidua (MDec) layer that is a mucus membrane of the pregnant uterus containing antigen presenting cells (macrophages, leukocytes, NK) [32–36] and collagen type I, III, IV, V, and VI [37].

In full-term pregnancy, the degradation of the extracellular matrix is induced by a synchronized activity of MMPs [31, 38–40]. Asynchronous changes in MMPs expression, activity and degradation could lead to different obstetric pathologies such as spontaneous abortions [41, 42], preeclampsia [43, 44] chorioamnionitis [45–48] and preterm premature rupture of the membranes (PPROM).

PPROM is produced by several pathogenic microorganisms located in the vaginal and cervical tract that upon entering the uterine cavity via an ascending route, colonize and infect the fetal membranes inducing their rupture [49, 50]. Bacteria or their structural components are recognized by Toll-like receptors (TLRs) located in the FAC and MDec layers [51, 52] activating both an inflammatory response and the expression of MMP [48, 53]. Recently, Zuo et al. [10], found a positive correlation between an increase in MMP-9 and a decrease in type IV collagen in serum of patients with PPROM, suggesting an important role for MMPs during pregnancy.

PPROM start prior to 37 weeks of completed gestation, affecting between 8 and 10% of all preterm birth pregnancies [54–57]. Despite remarkable improvements in prenatal care over the past decades, rates of PPROM and subsequent preterm delivery have been steady or worsened [58]. The etiology of PPROM is unclear, some factors that increase the likelihood has been identified, for example cervical shortening [59–61] and intraamniotic infection [8, 48, 62]. Infection caused by Escherichia coli (E. coli) has been previously associated with PPROM and the bacteria found in amniotic fluid [63, 64], neonates blood [65, 66] and maternal blood in up to 10% of cases [67, 68]. Furthermore, it has been reported that E. coli infection activates the secretion of MMP-9 in fetal membranes [8, 69–72].

Despite continued research efforts to develop new preventive strategies, there have been not major advances, and progesterone (P4) continues as the most researched prophylactic agent for spontaneous preterm birth. P4 is a natural steroid hormone that is critical for pregnancy maintenance due to its multiple actions including, myometrial quiescence, uterine contractility inhibition, and maternal immunomodulatory functions. This hormone drives the polarization of immune cells towards an anti-inflammatory phenotype to promote a tolerogenic microenvironment, a process that involves downregulation of pro-inflammatory mediators [73]. It has been found that LPS-induced pro-inflammatory cytokines production in human amniotic epithelium and in placental explants and this cytokines production is reduced by incubation with P4 [74, 75]. Interestingly, P4 decreased E. coli and LPS-induced IL-1β, IL-6, IL-8, and TNF-α expression through inhibiting NF-κB and MAPK pathways in bovine endometrial epithelial cells [76].

Recently, Palacios-Luna et al. [77] from our laboratory, reported that P4 reduced the secretion of extracellular heat shock proteins (eHsp)-60, -70 and IL-1β induced by E. coli infection in human MDec tissue. Intriguingly, the reduction of IL-1β was partial, suggesting that the remaining secretion could be associated with adverse effects during pregnancy [77]. In parallel, Allen et al. [78] demonstrated that medroxyprogesterone acetate reduced in 41% the activity of MMP-9 induced by TNFα, in HTR-8/SVneo cells, a cytotrophoblast cell line while Cortes-Reynosa et al. [79] demonstrated an association between MMP-9 activity and the degradation of type IV collagen in HMF-7 cells, a human breast cancer cell line. Despite recent advances, it is not yet known whether P4 regulates proMMP-9 and contribute to the degradation of collagen in human MDec tissue during infection. Thus, this study aimed to determine whether P4 has a role on the secretion of proMMP-9 activity and type IV collagen degradation in human MDec tissue during E. coli infection. The Transwell system was chosen because it mimics in vivo conditions and provide both quantitative and qualitative insights. This system allows direct stimulation of the MDec tissue and evaluating the response produced without the participation of the FAC. The model has been extensively validated by our research group [8, 77, 80] and it is considered one of the most prominent and time-tested approach in placental transport [81, 82].

Methods

Biological sample collection and human feto-maternal interface culture

The Institutional Review Board and the Research Ethics Committee approved this study (protocol number 2017-2-82, Instituto Nacional de Perinatología). Figure 1 shows the schematic study design and MDec tissue stimulation procedure.

Fig. 1.

Flow diagram of the study design. Nine healthy pregnant patients with a singleton pregnancy greater than 37 weeks of gestation were selected and underwent selective cesarean section. In each case, the human feto-maternal interface (HFMi) was recovered, and a portion was analyzed for the search of pathogenic microorganisms and only negative samples were included in the study. HFMi were mounted in the Transwell system (n = 9), and the maternal decidua tissue (MDec) stimulated for 3, 6, and 24 h. After each stimulation period, tissue was recovered, and immunofluorescence staining performed. Culture medium was recovered after each stimulation period for analysis of proMMP-2 and proMMP-9 activity

HFMi tissues were collected from nine healthy non-laboring women with normal singleton pregnancy and with indication of elective cesarean section at term prior to the onset of labor (age: 31.2 ± 8.4, body weight at delivery 76.5 ± 10.8 kg, parity 4.5 ± 1.0). The number of patients enrolled in the study was based on previous publications by our research group [8, 77, 80] and by other groups using the same Transwell model [83, 84]. Written voluntary informed consent was obtained from all patients after the study protocol was explained. Patients with evidence of infection, those who received antimicrobial treatment at least 20 days before cesarean section, or other obstetric pathology associated with an inflammatory response were not included in the study. General microbiologic analyses were conducted on all samples, and those with Gram-negative, Gram-positive or atypical bacteria were excluded.

HFMi tissues were dissected from the placenta immediately after delivery and transported to the laboratory in sterile saline solution. HFMi tissues were washed with phosphate-buffer saline (PBS) to eliminate blood clots, cut into 18-mm diameter pieces and mounted in a Transwell culture system (Costar, New York, NY, USA) as previously described [8, 80, 83]. The amniotic tissue faced the upper chamber of the Transwell system while the maternal decidua tissue faced to the lower chamber. Both compartments were incubated in Dulbecco’s modified Eagle medium (DMEM; Gibco, Bethesda, MD, USA) supplemented with 10% fetal calf serum (Gibco) and antibiotic-antimycotic solution (penicillin 100 U/mL, streptomycin 100 µg/mL; Gibco) at 37 °C in 5% CO2 for 18 h. After the incubation period, both compartments were washed with PBS to remove DMEM-fetal calf serum and antibiotic-antimycotic solution. Next, both compartments were incubated with 1 mL of DMEM with 0.2% lactalbumin hydrolysate (Gibco) at 37 °C in 5% CO₂.

Human maternal decidua (MDec) stimulation

To determine the effect of P4 on proMMP-2 and proMMP-9 activity and type IV collagen degradation after E. coli infection, MDec tissues were incubated for different periods (3, 6, and 24 h) with 4.8 × 10^− 7 M of P4 (progesterone water soluble P-7556; Sigma-Aldrich, St Louis, MO, USA) in the presence or absence of 10² and 10⁶ CFU/mL of E. coli. At the end of each incubation period, the medium of the lower chamber was collected using a sterile transfer pipette (Falcon, Becton-Dickinson, Franklin Lakes, NJ, USA) and centrifuged at 1400 rpm at 4 °C for 5 min, the sediment discarded, and the supernatant stored at -70 °C until further analysis. At the end of each incubation period, tissues were washed twice with PBS and fixed in 10% methanol. proMMP-9 and type IV collagen (n = 3) were assessed by immunodetection while proMMP-2 and proMMP-9 activity were determined by zymography (n = 9). P4 concentration and incubation periods were selected based on a previous report [85]. The strain of E. coli used herein was previously isolated from a patient diagnosed with premature rupture of fetal membranes and chorioamnionitis [80]. Phylogenetic group B2 was determined by PCR using gene-specific primers [86]. To determine the effectiveness of P4 in regulating the activity of E. coli-induced proMMP-2 and proMMP-9, we used a minimum of 10² CFU/mL and maximum of 10⁶ CFU/mL, both concentrations have been previously validated by our research group [8, 80]. CFU were determined based on turbidity equivalent to a 0.5 McFarland standard.

Tissue sectioning for immunohistochemistry

HFMi tissues were fixed in 10% methanol (Sigma-Aldrich) at -20 °C for 15 min, next washed twice with PBS, embedded in Tissue-TeK (Sakura, Finetek, CA, USA) and stored at -70 °C. A Leica CM1850 microtome (UV. Wetzlar, Germany) at -20 °C was used for sectioning with a thickness of 5 μm and sections were adhered to a salinized glass slide (Fisher brand, Thermo Fisher Scientific Inc., Waltham, MA, USA). To assess proMMP-9 and type IV collagen staining, tissue sections were permeabilized with 0.3% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) for 45 min and blocked with 5.0% BSA and 0.05% Tween-20 for 45 min. Next, washed and incubated with anti-MMP-9 (AF911, R&D System) and human collagen-IV (MAB6308, R&D System) primary antibodies at 1:1000 in 1% bovine serum albumin (BSA, AMRESCO, Solon, OH, USA) at 4 °C overnight. At the end of the incubation period, sections were washed twice with PBS and incubated with anti-goat (1:1000; Alexa-Flour, 568) and antirabbit (1:1000; Alexa-Flour, 588) secondary antibodies, diluted in 1% BSA (AMRESCO), at room temperature for 95 min. Next, sections were washed twice with PBS and the cell nuclei stained using 4´,6-diamino-2-phenylindole (DAPI; 1:100) at room temperature for 7 min. Negative controls consisted in sections without primary antibody (data not showed). The immunostained sections were analyzed using an epi-fluorescence microscope (Olympus, IX-81, Tokyo, Japan) and photographed with a CCD camera (Hamamatsu, ORCA-Flash 2.8, Tokyo, Japan).

Zymography gel activity

A gelatin-specific zymography was used for determination of proMMP-2 and proMMP-9 activity in culture media of maternal decidua tissue. SDS-polyacrylamide gels, containing porcine gelatin (1 mg/mL) as substrate, were used as described previously [8]. Each well was loaded with 0.75 µg of protein; previously measured by the Bradford colorimetric assay [87]. As standard for proMMP-2 and proMMP-9 activity, we used medium from U937 promyelocyte cells (ATCC Manassas, VA, USA). Experimental samples were diluted with sample buffer (0.5 M Tris (pH 6.8), 10% glycerol,1% bromophenol blue). After electrophoresis was ran, at a constant electric current of 10 mA at 4 °C, SDS was removed by incubation with 1% Triton X-100 (twice) for 15 min at room temperature under constant agitation. Next, gels were incubated overnight at 37 °C in activation buffer: 50 mMTris (pH 7.4), 0.1 M CaCl₂, 0.15 M NaCl, and 0.2 mg/mL NaN₃. At the end, gels were stained with 0.2% Coomassie brilliant blue R-250 (Sigma-Aldrich, St Louis, MO, USA) in methanol: acetic acid: glycerol: water (10:10:10:70) for 1 h at room temperature. For destaining we used a solution containing methanol: acetic acid: water (10:10:80) until intense bands were seen. Bands for proMMP-2 and proMMP-9 were quantify by densitometry using the EpiChemi Darkroom gel documentation system (UVP; CA, USA) and results expressed in relative densitometric units. Gel activity assays were performed in nine independent experiments.

Statistical analysis

All statistical analyzes were performed using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA, USA). Results are presented as mean ± SD of proMMP-9 relative activity. Data were analyzed using Shapiro-Wilk test to confirm normal distribution and Tukey’s test for means comparison between all treatment groups. A p-value of < 0.05 was considered statistically significant. The proMMP activity gel assay was performed in 9 independent experiments with duplicates.

Results

E. coli-induced secretion of proMMP-2 activity in human maternal decidua

MDec tissue incubated under control conditions and during different incubation periods showed non-change in proMMP-2 activity (p = 0.9816; Fig. 2B, D, and F).

Fig. 2.

Activity of proMMP-2 in human maternal decidua after stimulation with E. coli. Representative gelatin-gel zymography (A, C, E) showing enzymatic activity of proMMP-2 secreted into the culture medium. U937 (72 KDa) standard activity of medium from promyelocytes cell cultures. Each proMMP-2 lysis band was quantified by densitometric analysis after 3 (B), 6 (D), and 24 h (F). Data represent the mean ± standard deviation of seven experiments, each in duplicate

Incubation of MDec with E. coli (10² or 10⁶ UFC/mL) for 3 (p = 0.6570), 6 (p = 0.9568), and 24 h (p = 0.9309) induced a non-significant increase in proMMP-2 activity with respect to the control group (Fig. 2B, D, and F). Finally, co-incubation of MDec with P4 did not modify the activity profile of proMMP-2 (Fig. 2B, D, and F).

E. coli-induced secretion of proMMP-9 activity in human maternal decidua

MDec tissue incubated under control conditions and during different incubation periods showed no change in proMMP-9 activity (p = 0.3152; Fig. 3B, D, and F).

Fig. 3.

Activity of proMMP-9 in human maternal decidua after stimulation with E. coli. Representative gelatin-gel zymography (A, C, E) showing enzymatic activities of proMMP-9 secreted into the culture medium. U937 (92 KDa) standard activity of medium from promyelocytes cell cultures. Each proMMP-9 lysis band was quantified by densitometric analysis after 3 (B), 6 (D), and 24 h (F). The data represent the mean ± standard deviation of seven experiments, each in duplicate

Incubation of MDec with 10² UFC/mL of E. coli for 3, 6, and 24 h induced a significant increase of 2.2- (p = 0.0002), 1.8- (p = 0.0203), and 1.8-fold (p = 0.0005) in proMMP-9 activity compared with their respective control group (Fig. 3B, D, and F). Incubation of MDec with 10⁶ UFC/mL of E. coli for 3, and 6 h also induced a significant increase of 2.0- (p = 0.0021), and 2.0- (p = 0.0030) in proMMP-9 activity respectively with respect to the control group (Fig. 3B, and D). In contrast, after 24 h of incubation, the activity of proMMP-9 increased 1.3-fold; that it is not statistically significant with respect to the control group (p = 0.0893; Fig. 3F).

Progesterone modulates proMMP-9 activity induced by E. coli infection in human choriodecidua

Co-incubation of MDec with P4 and E. coli (10² CFU/mL) for 3, 6 and 24 h reduced 1.2- (p = 0.3572), 1.1- (p = 0.8245) and 1.3-fold (p = 0.0893) the proMMP-9 activity. These reductions were not statistically significant with respect to the group incubated with E. coli alone (Fig. 3B, D and F). In addition, after 24 h of stimulating the MDec tissue with E. coli (10⁶ CFU/mL) the activity of proMMP-9 decreased 1.2-fold (Fig. 4B) compared to 3 and 6 hours (Fig. 3D and F). In the next phase of this study, we set out to determine the immunolocalization of proMMP-9 and its participation in the degradation of type IV collagen.

Fig. 4.

Double immunofluorescence staining for proMMP-9 and type IV collagen in human maternal decidua. A bright signal from actMMP-9 and type collagen was detected after 3, 6, and 24 h. The nucleus was stained using 4´,6-diamino-2-phenylindole (DAPI, blue) actMMP-9 (red) and type IV collagen (green). The assay was performed in five independent experiments. The magnification is 100x, scale bars 100 μm

Immunofluorescence of proMMP-9 and type IV collagen detection in human feto-maternal interface tissue

Under control conditions, we observed a positive immunodetection of proMMP-9 (red staining islow) and type IV collagen (green) in the extracellular matrix of MDec tissue (Fig. 4).

This profile changed when MDec tissue was stimulated with E. coli; we observed a significant increase in proMMP-9 and a decrease in type IV collagen immunoreactivity in the compacta layer after 3, 6, and 24 h compared to the control group (Fig. 4). In the same assay, we observed that co-incubation of MDec with P4 and 10² or 10⁶ CFU/mL of E. coli for 3, 6 and 24 h did not reduce the relationship of proMMP9/type IV collagen immunodetection with respect to the group stimulated with E. coli alone (Fig. 4).

Discussion

PPROM is a pregnancy complication, where several epidemiological and clinical factors are considered as precursors including obstetric problems, behavioral habits, and maternal reproductive tract infections. Although the cause of membrane rupture in most cases is unknown, asymptomatic intrauterine infection is a frequent precursor leading to an inflammatory response of the HMFi and making the pregnancy difficult to maintain. Patients with PPROM reportedly have a higher rate of microbial infections, especially E. coli. The endotoxins enriched in the cell wall of this gram-negative bacteria are recognized by Toll-like receptors [88] and this interaction activates the MyDD88 signal pathway and IRAK1-2 proteins that in turn are recruited by TRAF6, activating various proinflammatory cytokines [89, 90] and the activation of proMMP-9 (Fig. 5).

Fig. 5.

Progesterone and the degradative activity of proMMP-9 induced by E. coli in the feto-maternal inter-face. (A) Histological section of human feto-maternal interface consisting of fetal amniochorion (FAC) formed by amniotic epithelium (ae), basal layer (bm), and compact layer (cl), connects with maternal decidua (MDec, top panel). When the MDec tissue is stimulated with E. coli, we observe the degradation of the extracellular matrix in the region of the FAC and MDec and a separation between the FAC/MDec (bottom panel). (B) Immunohistochemistry staining of the human feto-maternal interface, actMMP-9 (red), type IV collagen (green), and nucleus (blue). MDec stimulated with E. coli induces proMMP-9 activity, reducing the detection of type IV collagen (bottom panel). Images were captured at 100x. (C) Proposed of model of proMMP-9 activation. During normal conditions, the proMMP-2 and − 9 are in a zymogen state. A minimal production is transformed to its active forms (Fig. 4) that is regulated by the production of specific tissue inhibitors (TIMPs) (12, 130) and MT1-MMP (128). In infection, simulated with E. coli, the regulatory factors are modified so that the MT1-MMP removes the propeptide associated with the active for of proMMP, inducing a conformational change towards its actMMP form [131, 132]. Some key references are given (⊣: negative regulation; →: positive regulation)

MMPs have been detected in the human feto-maternal interface [91–93] and play an important role in the remodeling of the extracellular matrix [94–96], embryo implantation [97], trophoblast invasiveness [98, 99], morphogenesis and rupture of fetal membranes at term [100–102]. Deficiencies in the regulation of MMPs have been associated with spontaneous abortions [41, 42], preeclampsia [44, 64], and PPROM [48, 102, 103].

Flores-Herrera et al. [8], demonstrated, in an experimental model of infection, that stimulation of MDec tissue with E. coli increased the secretion of the proinflammatory cytokines IL-1β, TNFα and induced the secretion of proMMP-9 in a dose- and time-dependent manner. In a recent report, Fuentes-Zacarias et al. [21], reported that proMMP-9 activity degraded epithelial cadherin binding proteins; critical molecules for maintaining the epithelial structure and tissue homeostasis.

Progesterone is essential for the establishment and maintenance of pregnancy, is produced by the human placenta [104], and its receptors have been found in the placenta and the feto-maternal interface [105–107]. During human pregnancy, this hormone has been shown to modulate, the secretion of different molecules associated with inflammatory cytokines [108, 109], cell stress [110], extracellular heat shock-proteins [111], and the degradative response of extracellular matrix metalloproteases [112]. During pregnancy, normal physiologic levels of P4 lower the risk of miscarriage [113] and prevents preterm birth [114, 115]. To our knowledge, there are no reports on progesterone regulation and the degradative activity of proMMP-9 and type IV collagen during E. coli infection in MDec.

The principal findings of this study were: (1) proMMP-9 was immunodetected at the feto-maternal interface after 24 h of stimulation with E. coli; (2) progesterone reduced in 20% the activity of proMMP-9 induced by E. coli stimulation and (3) proMMP-9 was associated with the degradation of type IV collagen in MDEc tissue.

Progesterone and its analogues have been used in in vitro studies as well as in clinical trials with variable results. In in vitro studies Oner et al. [116] and Cakmak et al. [117] demonstrated that medroxyprogesterone acetate (10^− 7 M) reduced the expression of MMP-1, MMP-3, and IL-11 induced by IL-1β and TNFα in decidual cells. Also, Allen et al. [78] observed that the stimulation of amnion or chorion cells with TNFα (10 ng/mL) increased MMP-9 activity in both cell types, but co-incubation with medroxyprogesterone acetate (10^− 6 M) reduced MMP-9 activity in amnios cells only. In the same study, the authors found that progesterone or 17α hydroxyprogesterone acetate did not reduce MMP-9 activity. On the other hand, Hung et al. [118], demonstrated that homogenates explants of human fetal membranes incubated with progesterone reduced the activity of MMP-2 and MMP-9 induced by LPS and E. coli; but they also pointed out this is a tissue dependent effect. In contrast, in our infection model we did not find changes in the activity of proMMP-2 after stimulation with E. coli alone or during coincubation with P4; this is in line with a previous report [8], where and we did not find that progesterone did not decrease proMMP-2 activity (Fig. 3). Flores-Espinosa et al. (2021) showed that LPS-E. coli (500ng/mL) did not increase MMP-2 activity in choriodecidual tissue and amniotic epithelium [119] while Maymon et al. (2001) observed no change in MMP-2 expression in amniotic fluid in patients with infection and premature rupture of fetal membranes [120]. In addition, we found that E. coli induced proMMP-9 activity in the MDec tissue and coincubation with P4 reduced proMMP-9 activity by 20% with respect to group stimulated with E. coli alone (Fig. 3).

Interestingly, Zuo et al. [10] found an increase in MMP-9 and a decrease in type IV collagen in serum of patients with premature rupture of fetal membranes. Our results extended that investigation and show that proMMP-9 is immunolocalized at the feto-maternal interface explaining the degradation of type IV collagen is not inhibited by progesterone (Fig. 4) it posits the low efficiency of the progesterone treatment during premature rupture of fetal membranes in clinical trials.

Fibulins are critical components to maintain a healthy HMFi and are a large family of extracellular matrix glycoproteins that provides elastic and fibrillar properties. They are also responsible for the integrity of the basement membrane [121] and human fetal membranes [122]. Recently Orvik et al. [123], demonstrated in 9 patients, who developed PPROM, a decrease in the concentration of fibulin-1 which is degraded by MMP-9 [124, 125]. Nakamoto et al. [126], and Okada et al. [127], reported that progesterone regulates the expression of fibulin-1 in the human endometrium; however, this regulation has not been evaluated in human fetal membranes during an infection process; a step that will be determined in the next phase of our study.

Regulation of the MMPs is controlled mainly by the tissue inhibitor of extracellular matrix metalloproteinases (TIMP 1–4) and the MMP/TIMP ratio [25, 128, 129]. A decrease in TIMP-2 and an increase in the activity of MMP-2 and MMP-9 have been demonstrated in patients with PPROM [7, 120, 130]. However, Krzysztof Szymanowski et al. [131], found no association between progesterone and the expression of MMP-2, MMP-9 and TIMP-1.

Progesterone in clinical trials.

Different multicenter, randomized, double-blind clinical trials in which patients with singleton pregnancies were stratified for different gestational ages with a risk factor for the development of PPROM and treated with 17-hydroxyprogesterone weekly did not show a decrease in PPROM compared to patients who received placebo. Additionally, in the group that received the hormonal treatment, the neonates presented low birthweight, high incidence of morbidity, respiratory failure and sepsis [132, 133]. Likewise, Lindsay Wood et al. [134], demonstrated that the intramuscular administration of 17-hydroxyprogesterone in 409 patients with a short cervix did not prevent spontaneous labor compared to the control group. In contrast, Facchinetti et al. [135], demonstrated a beneficial effect of the administration of 17-hydroxyprogesterone to reduce preterm delivery compared to patients who did not receive the hormonal treatment. Although progesterone is generally accepted as essential for the establishment and maintenance of pregnancy, its role during PPROM is unclear. Further advances on how P4 work during PPROM are clearly needed. Figure 5 shows the main points of regulation of progesterone in the degradative activity of proMMP-9 induced by E. coli at the feto-maternal interface.

Progesterone receptors

It has been shown a high expression of the progesterone receptor type B (PR-B) during pregnancy. However, towards the activation of labor, the expression of the PR-A isoform increases, which has been associated with an inflammatory response [105, 106]. Allen et al. (2014) demonstrated that a HTR8/SVneo cell line previously co-stimulated with TNFα and medroxyprogesterone acetate showed an increased expression of the progesterone receptor membrane component 1 (PGRMC1) which has been associated with the reduction in relative activity of MMP-9 [78]. In that same study, they showed that progesterone and 17-hydroxyprogesterone have low affinity to interact with PGRMC1. Feng et al. [108], demonstrated that infecting amnion cells, chorion and decidua with Ureaplasma parvum increases the expression of COX-2 and MMP-9, which were associated with a decrease of the PGMC1 receptor, contributing to the PPROM. Despite the use of progesterone in medical practice to reduce the incidence of PPROM, the PRA/PRB ratios have not been evaluated during this condition, reducing our understanding of how this therapy may work. Finally, our results suggest the importance of exploring variations in the expression of the PR in patients with PPROM before and after progesterone treatment.

Conclusions

In conclusion, our study demonstrates that progesterone reduced proMMP-9 activity induced by E. coli infection in human maternal decidual tissue. Moreover, the activity of proMMP9 was not completely reversed by P4 treatment, maintaining a degradative type IV collagen content at the human feto-maternal interface, these results posit the low efficiency of P4 in multicenter trials to reduce obstetric complications such as premature rupture of fetal membranes.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CFU

Colony-forming unit

DMEM

Dulbecco’s modified Eagle medium

proMMP-9

Pro Matrix metalloproteinase-9

MDec

Human Maternal decidual

HFMi

Human feto-maternal interface

P4

Progesterone

Author contributions

Obtained the voluntary consent and reviewed clinical records: SS-Z and MO-C. Fetal membranes and culture system stimulation, and zymography gel activity: GB-B. immunohistochemistry: GB-B and GG-L. Histological interpretation: MLP-V. Managed and maintained the E. coli strain. Performed microbiological analysis: GV-G. Formal data analysis and discussion of the results: GG-L, MO-C, , RJA-G, IL-P, OD-R, and HF-H. Designed the study, wrote the manuscript, and obtained financial support: HF-H.

Funding

This study was supported by a grant (2017-2-82 to HF-H) from the Instituto Nacional de Perinatología Isidro Espinosa de los Reyes, Ciudad de México, México.

Data availability

All the relevant information from the study is described in the manuscript.

Declarations

Ethics approval and consent to participate

Each patient was informed of the intention of the present study, that the fetal membranes would be taken after the birth of their baby and the procedure would not affect the health of their baby. In all cases the signature of the consent was obtained. The protocol was reviewed and approved by the Ethics and Research Committees of the Instituto Nacional de Perinatología Isidro Espinosa de los Reyes (2017-2-82).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.