Anti-CD3 antibody (F101.01) recognizes a conformational determinant on the TcR/CD3 complex [15]. The antibody was a kind gift from Professor C. Geisler, University of Copenhagen and was used for T cell activation. The following mAbs were purchased from Becton-Dicki nson,New Jersey, U.S.A. as mouse anti-human Abs: FITC-conjugated anti-CD3; APC-conjugated anti -CD8, anti-CD14, and anti-CD56. Mouse anti-human FITCconjugated anti-CD4 and APC-conjugated anti C-D19 an tibodies were purchased from Elektra and PE-con jugated mouse anti-human B7-H4 (clone H74) was purchased from eBioscience (California, U.S.A.). B7 -H4-Ig (human B7-H4 extracellular domain fused with human IgG1) fusion proteins were produced in suspension culture in an animal protein-free-adapted CHOK- 1SV (Lon za Biologics, Allendale, NJ) cell line utilizing the glutamine synthetase gene expression system, and purified using Amplimmune’s protein purification process. KS is an inactive mutant of B7-DC-Ig (provided by Amplimmune Inc.), a fusion protein consisting of the Fc portion of human IgG1 protein and extracellular domain of B7-DC, in which a missense mutation in the active site changed lysine residue at position 113 to serine. The KS mutant was used as a negative, isotype control for B7- H4-Ig. Synagis (also known as palivizumab) was used as another negative isotype control for B7-H4-Ig, as well as a Fc-blocker for flow cytometry experiments. Synagis is a humanized (IgG1 Fc domain) monoclonal antibody against respiratory syncytial virus.

Human PBMCs were obtained from buffy coats of healthy donors in Danish Blood Donor Corps by Ficoll gradient centrifugation and stored at −140˚C. Thawing of the cells followed a standard protocol. Cells were cultured on 96-well round (MLCs) or flat-bottomed plates (anti-CD3 stimulated T cells) in 220 μl of RPMI-1640 medium supplied with Glutamax, 10% human serum (Valley Biomedical), 0.5% 2-mercaptoethanol and 0.5% penicillin/streptomycin. All cultures were conducted with a total of 0.2 million cells per well in an incubator at 37˚C, 5% CO₂. One-way MLCswere prepared by irradiating one set of cells (“stimulator”) with 2000 rad from Cs-137 for 10 minutes. CD8 T cells were depleted using Dynabeads-CD8 (Invitrogen) following manufacturer’s manual. T cells were enriched 2-3 fold by two-step nonadherence to plastic. All cells were cultured for 72 hours and their proliferation assessed by tritium thymidine incorporation assay.

Immobilization of B7-H4-Ig, KS, anti-CD3 antibody and Synagis or a mixture of them was carried out with 40 μl of the agent incubated in flat-bottomed plates overnight at 4˚C and washed twice. In case of experiments shown in Figures 5-8 the indicated agent (B7-H4-Ig, KS or Synagis) was either added as “immobilized” (incubated overnight together with anti-CD3) or “soluble” (added together with the cells after the plate had been washed). In experiments shown in Figures 9 and 10 the plate was washed and B7-H4-Ig was added at 40 μl for two hours at 37˚C followed by washing and plating of cells essentially as described by Sica et al. [2].

Cells for FACS analysis were incubated with 300 μg/ml Synagis for 15 min on ice (to block Fc receptors) followed by incubation for 1h on ice with either 100 μg/ml of B7-H4-Ig or PBS. After wash 1 x 106 Cells were stained with fluorochrome labeled antibodies containing 1% bovine serum albumin (BSA) in PBS for 20 minutes on ice. After washing twice, cells were fixed with 1% paraformaldehyde. Isotype control antibodies were included to define appropriate cut-off levels. For labelling of non-adherent, anti-CD3 activated T cells, these were incubated for 15 minutes with 300 μg/ml Synagis, washed, and incubated for 1 hour with 200 μg/ml B7-H4-Ig or PBS, then washed and incubated for 30 minutes in FITC-conjugated mouse anti-CD3 (BD Pharmingen) and PE-conjugated mouse anti-B7-H4 antibodies (eBioscience,clone H74), in PBS supplied with 0,5% BSA. Cells were then washed, resuspended in either FACS buffer or 1% paraformaldehyde and analysed using a FACScalibur flow cytometer (Becton-Dickinson). Data were analysed using cellquest software (BectonDickinson).

Two plates were incubated with a range of anti-CD3 concentrations overnight at 4˚C. They were washed twice with PBS, incubated with 40 μl of 50 μg/ml of B7-H4-Ig or PBS for 2 hours at 37˚C and washed twice with PBS. ELISA assay for anti-CD3 binding was performed on the first plate. To the second plate enriched human T cells were added and incubated at 37˚C/5% CO₂ for 3 days. Tritium thymidine incorporation was performed to assess the proliferation.

Freshly harvested PBMCs from healthy donors were washed twice and resuspended in medium containing 10% FCS, at 5 × 10⁶ cells/ml and directly tested for cytotoxicity or incubated for 24 h min at 37˚C in the atmosphere of 5% CO₂ in air in the presence of a stimulatingcytokine cocktailof (IL-12 and IL-15: 50 ng/ml, IL-2: 2000 U/ml and IFN-α: 10.000 U/ml) (ImmunoTools GmbH; Germany) to generate LAK cells. B7-H4 -Ig was added to the effector cells at 50µg/ml and the formulation buffer and PBS was added as controls.K562 and Daudi cells were used as targets for NK and LAK respectively. Target cells were labeled with 30μCi/mil of 51Cr (Perkin Elmer, Billerica, MA) for 1 h at 37˚C. The labeled cells were washed and resuspended in medium. Cells were co-incubated at effector to target (E:T) ratios at 100:1 - 12.5:1. Co-cultures were set up in 96 well V-bottom plates (Nunc; Denmark) in four replicates and incubated for 4h at 37˚C in the atmosphere of 5% CO₂ in air. Controls included target cells incubated in medium alone for spontaneous release or lysed with 10% triton X-100 for maximal release. Radioactivity was measured by Wallac Wizard 1470 Automatic Gamma Counter. Specific release was calculated using the following formula:

% specific lysis = (sample cpm − spontaneous cpm)/ (maximal cpm − spontaneous cpm) × 100%.

Data were analysed in R, using ggplot2 package for visualization. Statistical significance was assessed by onesample t-test on medium control normalized data. A posthoc Benjamini-Hochberg correction was employed to control for multiple testing. The results of simultaneous ELISA and proliferation assays were analysed by fitting a four-point logistic model over a range of anti-CD3 concentrations and response values (c.p.m. and OD values in proliferation and ELISA assays, respectively) in control samples. Using this model anti-CD3 concentrations corresponding to response values in B7-H4-Ig treated samples were reversely predicted. Ratio of predicted and originally applied anti-CD3 antibody concentrations was calculated (equalling 1 for control samples) and plotted.

Figure 1 shows the FACS gating strategy for PBMC. The frequencies of T cells (CD3+), B cells (CD19+), NK cells (CD56+) and monocytes (CD14+) are shown in the FACS plot of normal PBMC and day 3 responder cells in an allogeneic two-way MLC stained with cell type specific antibodies. As shown, the frequency of the four major cell types changes only slightly during the three days of MLC. Numbers are the percent of labeled cells in the respective gates.