Activation of peroxisome proliferator-activated receptor gamma by human cytomegalovirus for de novo replication impairs migration and invasiveness of cytotrophoblasts from early placentas

Abstract

Human cytomegalovirus (HCMV) contributes to pathogenic processes in immunosuppressed individuals, in fetuses, and in neonates. In the present report, by using reporter gene activation assays and confocal microscopy in the presence of a specific antagonist, we show for the first time that HCMV infection induces peroxisome proliferator-activated receptor gamma (PPARgamma) transcriptional activity in infected cells. We demonstrate that the PPARgamma antagonist dramatically impairs virus production and that the major immediate-early promoter contains PPAR response elements able to bind PPARgamma, as assessed by electrophoretic mobility shift and chromatin immunoprecipitation assays. Due to the key role of PPARgamma in placentation and its specific trophoblast expression within the human placenta, we then provided evidence that by activating PPARgamma human cytomegalovirus dramatically impaired early human trophoblast migration and invasiveness, as assessed by using well-established in vitro models of invasive trophoblast, i.e., primary cultures of extravillous cytotrophoblasts (EVCT) isolated from first-trimester placentas and the EVCT-derived cell line HIPEC. Our data provide new clues to explain how early infection during pregnancy could impair implantation and placentation and therefore embryonic development.

FIG. 1.

HCMV infection induces PPARγ transcriptional activity. (A to D) U373MG cells were transfected with a PPRE-luciferase plasmid, left for 24 h in culture, then mock infected for 24 h (n.i) or infected with HCMV VHLE strain (MOI, 3) for the times indicated (h p.i) (A) or for 24 h before treatment or not (NT) with PPARγ agonist 15d-PGJ₂ (15d) in DMSO or rosiglitazone (Ro) in DMSO, with an inhibitor of Cox-2 (NS398) in DMSO, antagonists of PPARγ G3335 (G33) in distilled water or Gw9662 (Gw) in DMSO, at the indicated final concentrations (B to D). Alternatively, HCMV was inactivated by UV irradiation (CMV-UV). Luciferase activity (Luc) was quantified by using a Mithras luminometer. Histograms representing means of duplicates are representative of five (A and B) or three (C and D) independent experiments with standard deviations. The right histogram in panel D corresponds to transfection of cells with control plasmid deprived of PPRE sequences (p0RE-Luc). MRC5 and U373MG cells were cultured on coverglasses and either left uninfected (n.i) or infected (CMV) with HCMV VHLE strain for 24 h at an MOI of 3 and alternatively treated with G3335. (E) Cells were labeled with anti-IE or anti-PPARγ antibodies and submitted to confocal microscopy. (F) The function of PPARγ was monitored in U373MG cells through accumulation of lipid droplets, revealed by Oil Red O staining, as exemplified (arrowheads). Identical pictures were obtained in three independent experiments.

FIG. 2.

PPARγ takes part in regulation of IE2 expression and HCMV replication. U373MG cells were infected (VHLE, MOI of 3) for the times indicated in the presence (+) or absence (−, NT) of PPARγ antagonist G3335 (G33). (A and B) At the times as indicated (h p.i) RT-PCR (A) and Western blotting (B) for IE1, IE2, and β-actin were performed on cell extracts. Experiments shown are representative of three independent ones. (C) Virus titers were determined every day (d p.i) in supernatants of U373MG cells, as PFU/ml/24 h on MRC5 cells. Histograms are presented as means with standard deviations of three independent experiments and virus titers are means of duplicates. (D) Hek-293 cells were either uninfected (lane a, mock) or infected with control adenovirus (lane b, Ad-Vector) or adenovirus encoding PPARγ (lane c, Ad-PPARγ) at an MOI of 0.1, and expression of PPARγ (◃) was revealed by Western blotting. At 24 h after infection with adenovirus, RT-PCR was performed on uninfected (n.i) or HCMV-infected cells (MOI, 3) for the times as indicated (h p.i). (E) Supernatants from Hek-293 cell cultures were harvested after 48 h of infection with HCMV, and viral titers (in PFU/ml) were determined using standard methods. Identical results were obtained in three independent experiments.

FIG. 3.

The HCMV MIEP contains functional PPRE sequences in the distal enhancer. (A) Organization of the MIEP enhancer, with locations of recognition sequences for NF-κB, AP-1, and RARE sequences (15). The numbers refer to the nucleotide position relative to the initiation site of transcription. (B) PPRE sequences PP1, PP2, and PP3, identified in the distal part of the enhancer by using the PPREfinder program. (C) U373MG cells were transfected with control plasmid (mock) or plasmids containing PP1, PP2, or PP3 cloned upstream of a luciferase gene. Luciferase activity (Luc) was quantified from cells either uninfected (n.i) or infected with HCMV (CMV) for 24 h at an MOI of 3 and alternatively treated with PPARγ antagonist G3335 (G33). Histograms are representative of five independent experiments with standard deviations of duplicates.

FIG. 4.

PPARγ binds to functional PPRE sequences identified in the MIEP and is associated with the MIEP in infected cells. (A) A gel mobility shift assay was performed with ³²P-end-labeled probes from PP1, PP2, and PP3 sequences incubated (+) or not (-) with nuclear extracts (NE) recovered from HCMV-infected U373MG cells. Alternatively, an excess of unlabeled PP1 (UP) or anti-PPARγ antibody (PP-Ab) or control antibody (Cl-Ab) was added to the mixture. Positions of PPRE-(PPARγ) (◃) and anti-PPARγ-conjugated PPRE-PPARγ (▸) complexes are indicated. Experiments shown are representative of three independent ones. (B) Lysates from HCMV-infected (24 h; MOI, 3) U373MG cells either untreated (VC; left panel) or treated with 35 μM G3335 (right panel) were divided into four equal parts. Three parts were immunoprecipitated with anti-H3 (H3), control IgG (IgG), or anti-PPARγ (PPARγ) antibody and submitted to PCR. One part (input) was submitted to PCR before chromatin immunoprecipitation. Results are expressed as the percentage of recovery from the input and are representative of three independent experiments.

FIG. 5.

PPARγ is expressed in trophoblasts in situ, in primary (EVCT) and in transformed (HIPEC) cytotrophoblasts. (A) Schematic representation of a chorionic villous at the implantation site. EVCT (in red) invade the decidua up to the first third of the myometrium and the uterine arteries (vct, villous cytotrophoblast; st, syncytiotrophoblast; evct, extravillous cytotrophoblast; sc, stromal core; is, intervillous space; ua, uterine artery; d, decidua). (B and C) Immunohistochemical staining for CK07, a specific trophoblast marker of invasive EVCT (B) and for PPARγ (C) in placental sections at the implantation site (week 16 of pregnancy). (D and E) Immunostaining for CK07, RXRα, PPARγ, and nucleus with hematoxylin and DAPI in EVCT (D) and HIPEC (E). These experiments were extensively reproduced.

FIG. 6.

Activation of PPARγ by HCMV impairs migration of HIPEC. (A) Monolayers of HIPEC either uninfected (n.i) or infected with HCMV (VHLE; MOI, 3) and treated or not with G3335 (G33) were maintained in culture for 24 h, and accumulation of lipid droplets was revealed by Oil Red O staining (arrowheads). (B) At the times as indicated (h p.i) Western blotting was performed for IE1, IE2, and β-actin on protein extracts from either uninfected (n.i) or infected (CMV) and treated cells (G33). Pictures and Western blot assay results were reproduced in three independent experiments. (C) HIPEC monolayers either uninfected (n.i) or infected with HCMV (VHLE; MOI, 3) and alternatively treated with G3335 (G33) were scraped and migration was quantified as indicated in Materials and Methods. Results for migration indices from five independent experiments are expressed as means with standard deviations.

FIG. 7.

HCMV-induced activation of PPARγ impairs the invasiveness of primary EVCT. (A) Schematic diagram of the experimental design for quantification of pseudopodia crossing a matrigel-coated transwell and sites of the inferior side following CK07 immunostaining of the porous membrane, as recovered from uninfected (n.i) or infected (CMV) cells. (B and C) Invasion indices (variations relative to controls; mean percent ± the standard deviation) are shown for primary EVCT treated for 48 h with increasing concentrations of rosiglitazone (Ro) or 15d-PGJ₂ (15d) (B) or left untreated (Col) or treated with Ro (1 μM), G3335 (G33 at 35 μM), and infected with HCMV for 48 h (VHLE; MOI, 3) and treated or not, as indicated (C). Results are expressed as means with standard deviations from five (B) or three (C) independent experiments, respectively.