Research ArticleReproductive Biology

Reversible EMT and MET mediate amnion remodeling during pregnancy and labor

See allHide authors and affiliations

Science Signaling  11 Feb 2020:
Vol. 13, Issue 618, eaay1486
DOI: 10.1126/scisignal.aay1486
  • Fig. 1 Human amnion membranes from normal term labor show evidence of EMT.

    (A) Transmission electron microscopy (TEM) of amnion membranes from TL (term labor) compared to TNIL (term not in labor) fetal membranes showing tight junctions (arrows). Insets show electron-dense desmosomes. Images are representative of three independent biological replicates. Scale bar, 0.05 nm. (B) Histological analyses of TL and TNIL membranes: Masson’s trichrome stain to show collagen, dual immunohistochemical (IHC) staining for CK-18 (pink) and vimentin (brown), and single IHC for N-cadherin, E-cadherin, and MMP9. Images are representative of three biological replicates. Scale bar, 50 μm. (C and D) Western blot analysis (C) and quantification (D) of TWIST, vimentin, N-cadherin, and E-cadherin in TL and TNIL membranes. Actin is a loading control. Cropped Western blots are shown; full blots are provided in the Supplementary Materials (fig. S5A). n = 5 biological replicates. Error bars represent means ± SEM. P = 0.004 (TWIST), P = 0.009 (vimentin), and P = 0.031 (N-cadherin/E-cadherin). Linear adjustment of contrast and brightness has been applied to the bright-field images throughout the figure. a.u., arbitrary units.

  • Fig. 2 TNIL human amnion membranes exposed to oxidative stress show evidence of EMT.

    (A) TEM of cultured TNIL amnion membranes either untreated or exposed to cigarette smoke extract (CSE). Tight junctions are noted with arrows. Insets show a close-up of one tight junction. Images are representative of three biological replicates. Scale bar, 0.05 nm. (B) Histological analyses of untreated TNIL and CSE-treated TNIL membranes: Masson’s trichrome stain to show collagen, dual IHC for CK-18 (pink) and vimentin (brown), and single IHC for N-cadherin, E-cadherin expression, and MMP9. Images are representative of three biological replicates. Scale bar, 50 μm. (C and D) Western blot analysis (C) and quantification (D) of ZEB1, vimentin, N-cadherin, and E-cadherin in untreated and CSE-treated TNIL membranes. Actin is a loading control. Cropped Western blots are shown; full blots are provided in the Supplementary Materials (fig. S5B). n = 3 biological replicates. Error bars represent means ± SEM. P = 0.036 (ZEB1), P = 0.664 (vimentin), and P = 0.0005 (N-cadherin/E-cadherin). Linear adjustment of contrast and brightness has been applied to all bright-field images throughout the figure.

  • Fig. 3 Term labor and oxidative stress induce p38 MAPK–dependent EMT in mouse amnion membranes.

    (A) Masson’s trichrome staining, dual IHC staining for CK-18 (pink) and vimentin (brown), and single IHC staining for N-cadherin, E-cadherin, and MMP9 in D18 (pregnancy) and D19 (parturition) mouse amniotic sacs. Images are representative of three biological replicates. Scale bar, 50 μm. (B) Western blot analysis and quantification of SNAIL, vimentin, N-cadherin, and E-cadherin in D18 and D19 amniotic sacs. Actin is a loading control. Cropped Western blots are shown; full blots are provided in the Supplementary Materials (fig. S6, A to C). n = 5 biological replicates. Error bars represent means ± SEM. P = 0.001 (SNAIL), P = 0.005 (vimentin), and P = 0.009 (N-cadherin/E-cadherin). (C) Dual IHC staining for CK-18 (pink) and vimentin (brown) and single IHC staining for N-cadherin, E-cadherin, and MMP9 in amniotic sacs from pregnant mice injected with PBS (control), CSE, CSE and the p38 MAPK inhibitor SB203580 (SB), or SB alone. Images are representative of three biological replicates. Scale bar, 50 μm. Linear adjustment of contrast and brightness has been applied to all bright-field images throughout the figure.

  • Fig. 4 TGF-β induces EMT in a TAB1- and p38 MAPK–dependent manner in human AECs.

    (A) Bright-field microscopy and confocal fluorescence imaging showing cell morphology, CK-18, and vimentin in untreated (Ctrl) and TGF-β–treated human AECs. Nuclei are labeled with DAPI (4′,6-diamidino-2-phenylindole). Image, n = 3. Scale bar, 50 μm. (B) Quantification of CK-18 and vimentin in (A). For CK-18, 23.81 ± 7.124 (control) and 17.73 ± 1.416 (TGF-β treated); P = 0.0002. For vimentin, 12.66 ± 10.63 (control) and 23.33 ± 4.751 (TGF-β treated); P < 0.0001. Values are expressed as mean intensity ± SD. Error bars represent means ± SEM. n = 3 biological replicates. (C) Bright-field microscopy showing morphology of human AECs that were untreated (Ctrl) or treated with the indicated combinations of nontargeting (NT) siRNA, TAB1 siRNA, TGF-β, and the p38 MAPK inhibitor SB. Images are representative of five biological replicates. Scale bar, 50 μm. (D and E) Western blots (D) and quantification (E) of SLUG, SNAIL, N-cadherin, and E-cadherin in AECs treated as indicated. Actin is a loading control. n = 3 biological replicates. Error bars represent means ± SEM. Top graphs, TGF-β only compared to control: P = 0.027 (SNAIL), P = 0.013 (SLUG), and P = 0.008 (N-cadherin/E-cadherin); TGF-β only compared to TGF-β + SB: P = 0.022 (SNAIL), P = 0.025 (SLUG), and P = 0.046 (N-cadherin/E-cadherin). Bottom graphs, TGF-β only compared to control: P = 0.023 (SNAIL), P = 0.015 (SLUG), and P = 0.015 (N-cadherin/E-cadherin); TGF-β only compared to TGF-β + TAB1 siRNA: P = 0.184 (SNAIL), P = 0.080 (SLUG), and P = 0.159 (N-cadherin/E-cadherin). Linear adjustment of contrast and brightness has been applied to all bright-field or fluorescent images throughout the figure.

  • Fig. 5 P4 induces MET in human AMCs in a PGRMC2-dependent manner.

    (A) Bright-field microscopy and confocal imaging showing cell morphology, CK-18, and vimentin in untreated (Ctrl) and progesterone (P4)–treated primary human AMCs. Nuclei are labeled with DAPI. Scale bar, 50 μm. (B) Quantification of CK-18 and vimentin in (A). Error bars represent means ± SEM. n = 3 biological replicates. For CK-18, 226.7 ± 29.01 (control) and 348.4 ± 111 (P4); P = 0.002. For vimentin, 467.6 ± 106.6 (control) and 448.5 ± 82.1 (P4); P = 0.995. (C) Bright-field microscopy showing morphology of untreated AMCs (Ctrl) and AMCs treated with the indicated combinations of P4, NT siRNA, and PGRMC2 siRNA. Images are representative of three independent experiments. Scale bar, 50 μm. (D and E) Western blot analysis (D) and quantification (E) of c-MYC, SNAIL, SLUG, N-cadherin, and E-cadherin in AMCs treated as indicated. Error bars represent means ± SEM. P4 only compared to control: P = 0.010 (c-MYC), P = 0.490 (SNAIL), P = 0.130 (SLUG), and P = 0.030 (N-cadherin/E-cadherin); P4 only compared to P4 + siRNA PGRMC2: P = 0.398 (c-MYC), P = 0.084 (SNAIL), P = 0.381 (SLUG), and P = 0.490 (N-cadherin/E-cadherin). n = 3 biological replicates. Linear adjustment of contrast and brightness has been applied to all bright-field images throughout the figure. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

  • Fig. 6 P4 induces MET in human AMCs in a c-MYC–dependent manner.

    (A) Bright-field microscopy showing morphology of primary human AMCs that were untreated or treated with P4, the c-MYC inhibitor 10058, or both P4 and 10058. Images are representative image of three independent experiments. Scale bar, 25 μm. (B and C) Western blot analysis (B) and quantification (C) of c-MYC, SNAIL, SLUG, N-cadherin, and E-cadherin in AMCs treated with P4, 10058, or both. Actin is a loading control. n = 3 biological replicates. Error bars represent means ± SEM. P4 only compared to control: P = 0.049 (c-MYC), P = 0.0.326 (SNAIL), P = 0.176 (SLUG), and P = 0.047 (N-cadherin/E-cadherin); P4 only compared to P4 + 10058: P = 0.050 (c-MYC), P = 0.609 (SNAIL), P = 0.201 (SLUG), and P = 0.512 (N-cadherin/E-cadherin). Linear adjustment of contrast and brightness has been applied to all bright-field images throughout the figure.

  • Fig. 7 P4 maintains the epithelial state of human AECs.

    (A) Bright-field microscopy showing morphology of primary human AECs treated with P4 for 6 days and confocal immunofluorescence showing vimentin and CK-18. Nuclei are labeled with DAPI. Images are representative of three biological replicates. Scale bars, 50 μm. (B) Quantification of CK-18 and vimentin in (A). Error bars represent means ± SEM. For vimentin, 223 ± 15.47 (control) and 130.1 ± 15.68 (P4); P < 0.0001. For CK-18, 523 ± 57.5 (control) and 480.3 ± 52.01 (P4); P = 0.166. n = 3 independent experiments. (C and D) Western blot analysis (C) and quantification (D) of the indicated proteins in untreated and P4-treated AECs. Cropped Western blots are shown; full blots can be found in the Supplementary Materials (fig. S7A). n = 5 biological replicates. Error bars represent means ± SEM. c-MYC (P = 0.0211), SNAIL (P = 0.023), SLUG (P = 0.007), and N-cadherin/E-cadherin ratio (P = 0.005). Linear adjustment of contrast and brightness has been applied to all bright-field images throughout the figure.

  • Fig. 8 P4 expedites wound healing of human AECs in vitro.

    (A) Healing of scratch wounds in primary human AECs in normal cell culture conditions (control) or treated with P4 or TGF-β. A red mask was applied to aid visualization of the wound field. Images are representative of three biological replicates. Scale bar, 100 μm. (B) Immunofluorescence showing vimentin (green) and CK-18 (red) in scratch-wounded primary AECs treated as indicated. A black mask was applied to aid visualization of the wound field. Images are representative of three biological replicates. Scale bar, 100 μm. Linear adjustment of contrast and brightness has been applied to all bright-field and fluorescent images throughout the figure.

  • Fig. 9 Schematic of amnion membrane maintenance and disruption due to cellular transitions.

    The amnion membrane is composed of amnion epithelial cells (AECs) and amnion mesenchymal cells (AMCs) that are embedded in ECM and separated from the AECs by a basement membrane (BM) that is rich in type IV collagen. During gestation, progesterone (P4) helps maintain the epithelial state of AECs, and AMCs have a fibroblastoid morphology. We hypothesize that the plasticity of amnion membrane cells allows remodeling of the membrane through cycles of MET and EMT. This maintains inflammatory homeostasis during gestation through a balance of P4-mediated MET and TGF-β–mediated EMT pathways to limit localized inflammation and repair microfractures. At term and preterm, increased oxidative stress and senescence cause an increase in TGF-β both in AMCs and in amniotic fluid and a functional withdrawal of P4 signaling in AMCs due to a decrease in the P4 receptor PGRMC2. This induces a static state of EMT, increased numbers of AMCs, and inflammation, all of which contribute to labor-associated outcomes such as membrane rupture. TGFBR1, TGFB receptor 1.

Supplementary Materials

  • stke.sciencemag.org/cgi/content/full/13/618/eaay1486/DC1

    Fig. S1. EMT-associated transcription factors and p38 MAPK at term labor in human and mouse amnions.

    Fig. S2. TGF-β–associated changes in AECs.

    Fig. S3. P4- and P4 receptor–associated changes in AMCs and AECs.

    Fig. S4. AMC qualities at term.

    Fig. S5. Complete Western blots for human tissues.

    Fig. S6. Complete Western blots for mouse tissue.

    Fig. S7. Complete Western blots for cultured primary cells.

    Table S1. Oxidative stress–induced changes in pregnant CD-1 mice.

    Table S2. siRNA sequences.

    Table S3. Plasmid sequence for GFP-PGRMC2.

    Table S4. qRT-PCR primer sequences.

  • This PDF file includes:

    • Fig. S1. EMT-associated transcription factors and p38 MAPK at term labor in human and mouse amnions.
    • Fig. S2. TGF-β–associated changes in AECs.
    • Fig. S3. P4- and P4 receptor–associated changes in AMCs and AECs.
    • Fig. S4. AMC qualities at term.
    • Fig. S5. Complete Western blots for human tissues.
    • Fig. S6. Complete Western blots for mouse tissue.
    • Fig. S7. Complete Western blots for cultured primary cells.
    • Table S1. Oxidative stress–induced changes in pregnant CD-1 mice.
    • Table S2. siRNA sequences.
    • Table S3. Plasmid sequence for GFP-PGRMC2.
    • Table S4. qRT-PCR primer sequences.

    [Download PDF]

Stay Connected to Science Signaling

Navigate This Article