Identification of novel trophoblast invasion-related genes: heme oxygenase-1 controls motility via peroxisome proliferator-activated receptor gamma

Abstract

Invasion of cytotrophoblasts (CTBs) into uterine tissues is essential for placental development. To identify molecules regulating trophoblast invasion, mRNA signatures of purified villous (CTB, poor invasiveness) and extravillous trophoblasts (EVTs) (high invasiveness) isolated from first trimester human placentae and villous explant cultures, respectively, were compared using GeneChip analyses yielding 991 invasion/migration-related transcripts. Several genes involved in physiological and pathological cell invasion, including A disintegrin and metalloprotease-12, -19, -28, as well as Spondin-2, were up-regulated in EVTs. Pathway prediction analyses identified several functional modules associated with either the invasive or noninvasive trophoblast phenotype. One of the genes that was down-regulated in the invasive mRNA pool, heme oxygenase-1 (HO-1), was selected for functional analyses. Real-time PCR analyses, Western blotting, and immunofluorescence of first trimester placentae and differentiating villous explant cultures demonstrated down-regulation of HO-1 in invasive EVTs as compared with CTBs. Modulation of HO-1 expression in loss-of as well as gain-of function cell models (BeWo and HTR8/SVneo, respectively) demonstrated an inverse relationship of HO-1 expression with trophoblast migration in transwell and wound healing assays. Importantly, HO-1 expression led to an increase in protein levels and activity of the nuclear hormone receptor peroxisome proliferator activated receptor (PPAR) gamma. Pharmacological inhibition of PPARgamma abrogated the inhibitory effects of HO-1 on trophoblast migration. Collectively, our results demonstrate that gene expression profiling of EVTs and CTBs can be used to unravel novel regulators of cell invasion. Accordingly, we identify HO-1 as a negative regulator of trophoblast motility acting via up-regulation of PPARgamma.

FIG. 1

Trophoblast invasion as a model to identify invasion-regulating genes. A, Diagram of the histological organization of the human maternal-fetal interface at early gestation. CTBs, which are specialized (fetal) epithelial cells of the placenta, differentiate into EVTs and invade the uterine wall. EVTs were collected from most distally outgrowing cells of first trimester villous explants growth factor-reduced Matrigel after 72 h culturing, whereas villous CTBs were purified from first trimester placentae as described in Materials and Methods. AV, Anchoring villus. B, Localization of integrins α1 and α6, HLA-G1, Kip2/p57, cytokeratin 7 (Cyto) (depicted in green), vimentin (Vim) (depicted in red) in first trimester villous explant cultures seeded on Matrigel. To visualize nuclei, sections were counterstained with DAPI. The staining pattern is indicative for highly invasive EVTs (marked by arrows) or poorly invasive CTBs, respectively. Representative examples are shown. CC, Cell column; VC, villous core. C, GeneChip-derived expression patterns of well-known markers for CTBs and EVTs.

FIG. 2

Microarray analysis of EVT vs. CTB identifies multiple differentially expressed genes. A, SAM analysis between one group of EVT (n = 6) and one group of CTB (n = 5) samples. The scatter plot of the observed difference d(I) vs. the expected relative difference d_E(I) is shown. The broken lines are drawn at a distance of Δ= 0.30 from the solid line that indicates d(I) = d_E(I). Genes outside the broken lines are regarded as genes with significant changes in expression, yielding 991 genes. The median estimated FDR is 5%. Dots above the broken line indicate genes induced in EVTs (370), and dots below the broken line indicate genes that are suppressed in EVTs (621). B, A heat map of top 50 genes that differentiate EVT and CTB groups as judged by SAM analysis. Down-regulated genes are shown in green and up-regulated genes in red.

FIG. 3

Multiple functional gene sets are altered between EVTs and CTBs. Top-scoring pathways displayed as heat maps showing relative gene expression. One exemplary pathway is shown for each category according to Table 3. Down-regulated genes are shown in blue and up-regulated genes in red. Genes with statistically significant core enrichment are indicated by the black bar on the left of the heat maps. A, Example of a top-scoring pathways enriched in EVTs. B, Example of a top-scoring pathways enriched in CTBs.

FIG. 4

HO-1 expression is reduced upon differentiation of CTBs into EVTs. A, Real-time PCR of EVT and CTB mRNAs (each n = 3) was performed as described previously. B, Immunofluorescence of first trimester placental tissues. Two representative examples (left panel: 40-fold magnification; right panel: 20-fold magnification) of five different placentae analyzed are shown. Serial sections of first trimester placental tissues were incubated with specific antibodies against HO-1 (upper panels) and Ki67/Kip2-p57 (red/green; middle panels) and counterstained with DAPI (lower panels). Note that most differentiated, p57-positive EVTs are largely devoid of HO-1. Dotted line, Direction of invasion. VC, Villous core. C, Immunofluorescence of tissue sections of first trimester explant cultures seeded on Matrigel (48 h). A representative example of 15 different explants of three different placentae is depicted. Dotted arrow denotes direction of trophoblast invasion. D, HO-1 and PPARγ protein expression in primary trophoblast cells and HTR-8/SVneo and BeWo cells. Western blot analysis demonstrated high HO-1 expression in CTB and BeWo and no/low HO-1 protein levels in EVT and HTR-8/SVneo cells, respectively. Note higher PPARγ protein levels in CTBs compared with EVTs. β-Actin was used as a loading ctrl. The blot is representative for three independent experiments.

FIG. 5

HO-1 inhibits trophoblast migration in loss-of as well as in gain-of function cell models. A, A miRNA targeting human HO-1 was shuttled from pSM2 into the retroviral vector LMP used to construct HO-1 knockdown cells as described in Materials and Methods. B, Western blot analysis of ctrl (LMP) or HO-1 miRNA-adapted short hairpin RNA transduced BeWo cells (miHO-1) demonstrates efficient HO-1 knockdown. β-Actin indicates equal loading. A representative example is shown. C, Transwell migration assays of LMP and miHO-1 BeWo cells were performed as described in Materials and Methods. Cell migration of ctrl-infected cells (LMP) through FN-coated filters was set to 100%. Note that miHO1 cell migration is increased in cells lacking HO-1. D, HTR-8/SVneo cells were retrovirally transduced to express the human HO-1 cDNA under the control of a DOX-sensitive promoter (Tet-On) as described in Materials and Methods. HO-1 protein expression in HTR8-HO-1 cells after 24 h DOX addition (0, 0.1, 1.0, and 10.0 μg/ml) was analyzed by Western blotting. β-Actin was used as a loading ctrl. E, Immunofluorescence of mock- or HO-1 transduced cells after 24 h DOX (1 μg/ml) induction using specific HO-1 antibodies. F, Transwell migration assays of DOX-induced HTR8 and HTR8/HO-1 cells were performed as described in Materials and Methods. Cell migration of ctrl-infected cells (HTR8) was set to 100%. Note that HTR8/HO1 cell migration through FN-coated filters is reduced in cells expressing HO-1. Bars represent mean values ± sem of three different experiments. *, P < 0.05 vs. HTR8. G, Effects of HO-1 expression on cell migration in wound healing assays. Wound healing assays of DOX-treated, mock- and HO-1-transduced HTR-8/SVneo cells were performed as described in Materials and Methods. Micrographs of wounded, nonfixed cell layers at 0 and 8 h after addition of culture medium are depicted. Initial sizes of wounds are marked by arrowheads. Representative examples of three different experiments are shown.

FIG. 6

HO-1 inhibits trophoblast migration via PPARγ. A, Protein levels of PPARγ were investigated in LMP or miHO1 BeWo cells by Western blotting. HO-1 and β-actin demonstrate HO-1 knockdown with equal protein loading, respectively. Representative examples are shown. B, Luciferase assays of BeWo LMP and miHO-1 cells after cotransfection with PPARγ-responsive and phRL ctrl plasmids. Firefly luciferase activity was determined in protein extracts and normalized to Renilla luciferase activity as described previously. For comparison, values of ctrl (LMP) infected cells were arbitrarily set to 100% in each experiment. Bars represent mean values of three independent transfections of three different cultures. Error bars indicate SD.*, P < 0.05. C, DOX was added to HTR8 and HTR8/HO-1 cells, and PPARγ protein expression was investigated 24–36 h later by Western blotting. HO-1 and β-actin demonstrate increased HO-1 expression with equal protein loading, respectively. D, Luciferase activity of the PPARγ-responsive reporter after 36 h DOX treatment of HTR8 and HTR8/HO-1 cells. Luciferase activity was determined as described in Materials and Methods. For comparison, values of ctrl (HTR8) infected cells were arbitrarily set to 100% in each experiment. Bars represent mean values of three independent transfections of three different cultures; error bars indicate SD.*, P < 0.05. E, Transwell migration assays toward FN using LMP or miHO1 cells in the absence (ctrl) or presence of GW9662. Migration of LMP cells was set to 100%. Bars represent mean values ± sem of three different experiments. Note that the inhibitory effect of endogenous HO-1 expression in LMP cells was abrogated when GW9662 was present. *, P < 0.01 vs. miHO1;^#, P < 0.05 vs. LMP plus GW9662. F, Transwell migration assays of DOX-induced HTR8 and HTR8/HO-1 cells toward FN in the absence (ctrl) or presence of GW9662. Note that the inhibitory effect of HO-1 expression in DOX-treated HTR8/HO-1 cells was abrogated when GW9662 was present. *, P < 0.05 vs. HTR8;^#, P < 0.05 vs. HTR8/HO-1 plus GW9662.