Fresh decidual samples were analyzed by immunohistochemistry. As expected, tissue contained cytokeratin 7⁺ placental cells and CD34⁺ endothelial cells. A high number of immune decidual cells were also visualized in isolated tissue (Figure 1); CD56⁺ NK cells, CD14⁺ antigen presenting cells and CD3⁺ T cells. After the digestion of the tissue, mononuclear cells were analyzed by flow cytometry. Immune cell populations present within the decidua are shown on Figure 2 from one representative experiment. As previously described 20 25 26 , decidual CD3^-/CD56⁺ NK cells represent the main leukocyte population in the decidual tissue (mean 58% ± 7.8). Decidual leukocytes are also composed of CD14⁺ antigen presenting cells (mean 19% ± 4.7) and CD3⁺ T lymphocytes (mean 8% ± 5), including CD4⁺ and CD8⁺ T lymphocytes (n = 21). Altogether, these results confirmed that the studied tissue was the decidua basalis, a maternal tissue in direct contact with the placenta. Flow cytometry analyzes show that both leukocytes (CD45⁺) and non-leukocytes (CD45^-) cells were present in decidual mononuclear cells and that dNK cells are the main leukocyte population.

To identify soluble factors secreted by decidual cells, we applied Luminex technology and ELISA methods to culture supernatants of decidua basalis mononuclear cells. Cytokines were quantified after 24 hours of culture without stimulation. Figure 3 shows the cytokines detected according to their abundance: low (Figure 3A), medium (Figure 3B), and high (Figure 3C). Cytokines detected at low levels included IL-10 (mean 277 pg/ml ± 72), IL-12 (456 pg/ml ± 46), IL-15 (118 pg/ml ± 14), TNF-α (372 pg/ml ± 107), IFN-α (71 pg/ml ± 6), IFN-γ (86 pg/ml ± 8) and CXCL-12 (148 pg/ml ± 36). Chemokines detected at moderate or high levels included CCL-3 (11460 pg/ml ± 2367), CCL-4 (7272 pg/ml ± 1760), CCL-5 (1492 pg/ml ± 300), CXCL-10 (11300 pg/ml ± 2260) and CCL-2 (106 ng/ml ± 1.7). The pro-inflammatory cytokines IL-6 (33 ng/ml ± 7.6) and IL-8 (3.10³ ng/ml ± 953) were also abundant. The proinflammatory cytokine IL-2 was undetectable (data not show), in keeping with its known absence from the healthy maternofetal interface 27 .

The detected soluble factors were also present in similar proportions, but at lower levels in decidua basalis and decidua parietalis histoculture supernatants (data not shown). IL-6 and IL-8 were significantly more abundant in decidua basalis supernantants than in decidua parietalis supernatants (p = 0.0012 and p = 0.01 respectively).

These results showed that decidual cells secreted soluble factors known to regulate HIV-1 infection, including β-chemokines known to inhibit R5 viral entry.

Decidua basalis culture supernatants were analyzed with Luminex technology 14 days after infection, at a time of sustained HIV-1 replication. As the culture medium was renewed every 3 days, reported cytokine levels are those having accumulated between day 11 and day 14. CCL-3 and CCL-4 were significantly more abundant (p = 0.026 and p = 0.027) in the supernatants of R5 HIV-1-infected cells than of non-infected cells. CCL-5 was not significantly more abundant in R5 HIV-1-infected cell supernatants 14 days after infection (p = 0.06)(Figure 4A).

In contrast to R5 HIV-1-infected cells, cytokine secretion was not significantly modulated by X4 HIV-1 infection (Figure 4A and 4B). Production of IL-12, IL-6, CCL-2, CXCL-10 and CXCL-12, that were also detected at day 14, was not significantly affected by either R5 or X4 HIV-1 infection (data not show).

Thus, CCL-3 and CCL-4 release by cultured decidua basalis mononuclear cells was enhanced by R5 HIV-1 infection.

Luminex analysis showed that CCL-3 and CCL-4 release was increased by R5 HIV-1 infection. To identify the source of these chemokines, freshly isolated, HIV-1-uninfected decidua basalis mononuclear cells were analyzed by flow cytometry after intracellular staining. CCL-3 and CCL-4 staining was observed in non leukocytic cells (CD45^-), natural killer cells (CD56⁺ dNK), and antigen-presenting cells (CD14⁺ dAPC) (Figure 5A). The mean fluorescence indexes (MFI) of CCL-3 and CCL-4 in cells from decidua from 4 different women were higher in dAPC than in non leukocytic and dNK cells (Figure 5B). CD4⁺ T cells and CD8⁺ T cells both had very low MFIs for CCL-3 and CCL-4.

To confirm the results of flow cytometry, CCL-3 and CCL-4 were quantified by Luminex analysis in 3-day culture supernatants of purified dAPC and dNK cells. Large amounts of both CCL-3 and CCL-4 were detected in dAPC supernatants (23 454 pg/ml ± 12 214 and 10 496 pg/ml ± 4 898, respectively) (Figure 5C). CCL-3 and CCL-4 were also found in dNK cell supernatants, but at lower levels (717 pg/ml ± 314; and 1 445 pg/ml ± 285, respectively). Small amounts of CCL-5 were found in both dAPC and dNK supernatants (452 pg/ml ± 325 and 291 pg/ml ± 50. respectively).

These results showed that dAPC were the main sources of CCL-3 and CCL-4 in the decidua.

To examine the role of the cytokine environment in the inhibition of viral replication, decidua basalis mononuclear cells were cultured for 24 hours before infection with R5 HIV-1 or X4 HIV-1. The cells were then infected, following a washing step or without a washing step (i.e. in the presence or absence of their respective 24-hour supernatants). R5 HIV-1 infection of decidual mononuclear cells was inhibited, as shown by the p24 antigen assay 7 days post-infection, in experiments with 6 out of 9 donors (range 0 to 80%, mean 28.64% ± 11.6; p = 0.039) (Figure 6). Inhibitory activity was lower 10 days post-infection (mean 18.69% ± 9.6; p = 0.088). Inhibition of X4 HIV-1 infection was observed in experiments with 6 out of 9 donors, 10 days post-infection (mean 11.25% ± 11.9; p = 0.375)(Figure 6); however, in contrast to the effect on R5 HIV-1 infection, the mean percentage of inhibition was not statistically significant.

To determine whether decidual soluble factors inhibited HIV-1 entry, the HeLa P4P cell line, which expresses the CD4 HIV receptor and also the co-receptors CCR5 and CXCR4, were infected by HIV-1 pseudotypes in the presence or absence of 24 h decidual conditioned medium (dCM). Efficiency of infection was measured in terms of luciferase activity (representative experiment in Figure 7A). dCM from 7 out of 8 donors significantly reduced R5 HIV-1_BaL pseudotype infection (mean 32.9% ± 7, range 0-58.4%; p = 0.002), while HIV-1_VSV-G pseudotype infection was unaffected (mean 5.25% ± 10.7, p = 0.640) (Figure 7B). dCM from 5 out of 7 donors reduced X4 HIV-1_HxB2 pseudotype infection, but not significantly (mean 15.2% ± 13.1, range 0-68%; p = 0.289).

Altogether, these results indicated that decidual culture supernatants participated in the inhibition of HIV-1 entry.

Decidual tissue samples were obtained from healthy women undergoing voluntary termination of pregnancy during the first trimester (6-10 weeks) at Antoine Béclère Hospital, Clamart, France and Tenon Hospital, Paris, France. All the women gave their written informed consent to the use of their tissues. The study was approved by the ethics committee of Hôtel Dieu, Paris, France, the Assistance Publique des Hôpitaux de Paris (n° VAL/2006/06-41/01) and the Biomedical Research Committee of the Institut Pasteur, Paris, France (n° RBM/2005.024).

Histocultures of decidua parietalis and decidua basalis were performed as previously described by Marlin et al 24 .

For cell isolation, freshly isolated decidual tissue was minced into small fragments and digested for 1 hour at 37°C under agitation in RPMI 1640 culture medium (Gibco) with 1 mg/ml collagenase IV (Sigma, St Quentin Fallavier, France) and 50 U/ml recombinant DNase I (Roche, Meylan, France). The cell suspension was successively filtered through 100 μm, 70 μm and 40 μm pore-size sterile nylon cell strainers (BD Biosciences, Le pont de Claix, France). The mononuclear cell population was isolated with Lymphocyte Separation Medium (PAA) and cultured in the rich culture media Ham F12: DMEM Glutamax (v:v) (Gibco) supplemented with 15% foetal calf serum (PAA, Les Mureaux, France), penicillin (0.1 U/l) and streptomycin (1 × 10^-8 g/l) (Gibco). CD14⁺ and NK cells were purified with anti-CD14 and anti-CD56 magnetic beads, respectively, as recommended by the manufacturer (Miltenyi, Paris, France).

Decidual histocultures and decidual mononuclear cells were infected with two HIV-1 primary isolates, HIV-1_BaL and HIV-1_LAI (with R5 or X4 tropism respectively). The isolates were amplified on PHA-stimulated PBMC from two blood donors for 10 days. PBMC cultures were maintained in RPMI 1640 Glutamax (Gibco) supplemented with 10% fetal calf serum, penicillin, streptomycin and 100 U/ml recombinant interleukin-2 (Chiron-Nederlands). Virions were concentrated by centrifugation on Vivaspin 100 000 Kda (Sartorius, Palaiseau, France) at 1400 g for 40 minutes.

Luciferase-expressing viral pseudotypes were based on the NL4-3 HIV-1 and VSV-G (amphotropic), BaL (R5) or HxB2 (X4) env plasmids 53 54 55 . The viral pseudotypes were generated by transfecting HEK-293T cells with the corresponding cDNA plasmid (pNL4-3ΔEnvLuc+ plus the appropriate env cDNA) using the transfection reagent SuperFect transfection reagent (Qiagen) as directed by the manufacturer. The pNL4-3ΔEnvLuc+ lacks the nef gene and the resulting pseudotype is non replicative. Supernatants were collected 72 hours after transfection, assayed for the viral pseudotypes by means of p24 antigen ELISA (Zeptometrix) and titrated on HeLa P4P cells. The efficiency of infection by the viral pseudotypes was determined by measuring luciferase expression with Luciferase Reagent (Promega) and a Glomax luminometer.

Decidual sections were obtained by embedding freshly isolated decidual tissue in Tissue-Tek (Sakura, Gentaur, Paris, France) and snap-frozen in an isopentane/liquid nitrogen bath. Frozen Tissue-Tek blocks were cut with a cryostat and frozen sections (5 μm thick) were fixed in acetone and rehydrated in TBS (Dako, Trappes, France). Tissue sections were stained for CD14 (RMO52, Beckman Coulter), CD3 (F7.2.38, Dako), CD34 (QbEnd10, Beckman Coulter), CD56 (N901, Beckman Coulter) or Cytokeratin 7 (OV-TL 12/30, Dako). Endogenous peroxidase and alkaline phophatase (AP) were blocked for 10 minutes by addition of hydrogen peroxide and levamisol (Dako). Surface markers were visualized using the Envision+ dual link system (Dako), or using biotin-streptavidin-alkaline phosphatase complex and Vector Red (Abcys, Paris, France), as an AP substrate. Tissue sections were counterstained with haematoxylin (Labonord, templars, France), mounted in permanent medium and analysed with a Nikon Eclipse 80i microscope.

The main cytokines involved in the regulation of HIV-1 infection, namely IL-2, IL-10, IL-12, IL-15, TNF-α, IFN-α, IFN-γ, CCL-3 (MIP1-α), CCL-4 (MIP1-β), CCL-5 (RANTES), IL-6, IL-8, CXCL-10 (IP10) and CCL-2 (MCP1) were measured by Luminex assay (Human Cytokine 25-plex antibody bead kit, Invitrogen Corporation, Carlsbad, California). CXCL-12 (SDF1) was measured with the Human SDF1 Quantikine Immunoassay (R & D System, Minneapolis) in the manufacturer's recommended conditions. The cytokines were measured in supernatants of unstimulated decidual mononuclear cells collected after 24 hours of culture, or after 72 hours in the case of purified dAPC CD14⁺ and dNK cells, when cytokine concentrations were maximal in the culture supernatants. To compare cytokine production by infected and non-infected decidual mononuclear cells, 3-days of cultured supernatants (day 11 to day 14) were assayed 14 days after infection with or without HIV-1_BaL or HIV-1_LAI.

Decidual mononuclear cells were tested for chemokine production by flow cytometry. After isolation, cells were cultured for 16 hours with Brefeldin A (Sigma-Aldrich) at 5 μg/ml, then labelled with anti-CD45-Amcyan, anti-CD14-Pacific Blue, anti-CD4-PE-Cy7, anti-CD56-Alexa 700 (Becton Dickinson), anti-CD3-PE-TexasRed and anti-CD8-APC (Beckman Coulter) before fixation and permeabilisation with the Intraprep reagent (Beckman Coulter). The cells were then labelled with anti-CCL-3-FITC, anti-CCL-4-FITC, anti-CCL-5-FITC (R & D System, Minneapolis) or IgG-FITC (control). Flow cytometry was performed with an LSRII device (Becton Dickinson) and FlowJo 9.0.1 software.

Decidual mononuclear cells were incubated for 1 hour with HIV-1 isolates or uninfected PBMC supernatant, at 10^-3 MOI, then washed 3 times and cultured at 10⁶ cells/ml. Viral production was measured every 3 or 4 days by detection of p24 antigen in cell culture supernatants by ELISA (Zeptometrix).

To examine the influence of the cytokine environment on the inhibition of viral replication, decidual mononuclear cells were cultured for 24 hours before infection with HIV-1_BaL or HIV-1_LAI. After 24 hours, the cells were infected with 10^-4 MOI for 1 hour at 37°C following or not a washing step. After 3 washes post-infection, the cultured cells were resuspended in culture medium and HIV-1 p24 antigen was titrated by ELISA in the culture supernatants at day 7 and day 10.

HeLa P4P cells (CD4⁺ CCR5⁺ CXCR4⁺) 56 were seeded at 10⁴ cells/well and cultured for 24 hours in 96-well plates. Entry inhibition assays were performed by incubating the cells for 1 hour with 100 μl of decidual conditioned medium collected after 24 h of decidual mononuclear cell culture, or with fresh culture medium (Mock medium), followed by treatment with 6 ng of p24 equivalent of each HIV pseudotype for 1 hour. The cells were then washed in 1X phosphate buffered saline (PBS) and cultured for 72 hours at 37°C. HeLa P4P cells were washed in 1X PBS and lysed with 100 μl of Cell Lysis Buffer (Promega). The cell lysates (10 μl) were used to determine luciferase activity as described above. Results are expressed as Relative Light Units (RLU)/100 μl of lysate. Each experimental condition was performed in triplicate.

Each condition was tested in at least four independent experiments. The one-sample t test was used. The hypothetical value was 1 for the fold-change graph and 0 for the percentage inhibition graph. Significance was assumed at p < 0.05. Prism 4 software (Graph Pad) was used for all analyses.