Selection of Epstein-Barr virus specific cytotoxic T lymphocytes can be performed with B lymphoblastoid cell lines created in serum-free media

Donors and B lymphoblastoid cell lines (BLCL)

Fifteen ml of heparinized blood were collected from three seropositive donors: donor no. 8 (Do8), donor no. 29 (Do29) and donor no. 24 (Do24). Peripheral blood mononuclear cells (PBMC) were separated by Ficoll Density centrifugation (lymphocyte separation medium; Eurobio, Paris, France). For BLCL establishment, two supernatants were prepared from the EBV-producing B95·8 cell line; one using the serum-free culture medium X-VIVO15 (Cambrex Biosciences, Paris, France), and the other using RPMI 1640 (Sigma-Aldrich, Les Ulis, France) supplemented with 10% FCS, 2 mM glutamine and 50 µg/ml gentamicin. For each infection, 2 × 10⁶ PBMC in 100 µl culture medium were incubated overnight in 24 well culture plates in the presence of and 500 µl/well B95·8 supernatant and 1 µg/ml cyclosporin A. The next day, 1·5 ml of culture medium was added supplemented with 1 µg/ml cyclosporin A.

Generation and expansion of EBV-specific cytotoxic cell lines

Donor PBMC were plated in 24-well culture plates in RPMI 1640 (Sigma-Aldrich) culture medium (with or without 10% fetal calf serum (FCS), or 8% pooled human serum (HS) or 8% autologous serum (AS) and further supplemented with 1%l-glutamine, 100 U/ml penicillin and 0·1 µg/ml streptomycin) at 2 × 10⁶ cells/well and stimulated with 5 × 10⁴ 35 Gray-irradiated autologous BLCL (PBMC : BLCL ratio of 40 : 1). The culture conditions are summarized in Table 1. After 10 days, T cells were collected and restimulated at a T : B ratio of 4 : 1 (5 × 10⁵ T cells and 1·25 × 10⁵ BLCL/well). IL-2 was added 4 days after the second stimulation (40 IU/ml) and a third stimulation was performed 7 days after the second in the presence of IL-2 and with the same T : B ratio (4 : 1).

Magnetic CD4+ and CD8+ cell separation

Twenty-five × 10⁶ CTL were magnetically labelled with CD4 MicroBeads (Miltenyi Biotec, Paris, France) and separated using an MS column (Miltenyi Biotec) according to the manufacturer's instructions.

Cytotoxicity assay

Cytotoxic activity was tested using a standard ⁵¹Cr release assay. Briefly, target cells were labelled with 100 µCi of Na₂⁵¹CRO₄ for 1 h at 37 °C, washed four times, and then plated at the indicated E : T ratio, in 96-well U-bottom plates. After 4 h of incubation at 37°C, 25 µl of supernatant from each well was removed and counted in a liquid β scintillation counter. Each test was performed in triplicate. Results were expressed as a percentage of lysis according to the following formula:

where experimental release represents mean count per minute (c.p.m.) releases from target cells in the presence of effector cells, spontaneous release represent the mean c.p.m. for target cells incubated without effectors, and maximum release represents the mean c.p.m. for target cells incubated with 0·5% triton.

Proliferation assays

Responder T lymphocytes were cocultured with 35 Gray-irradiated autologous or allogenic target BLCL for 72 h in 96-microwell flat-bottom culture plates at a 1 : 1 responder : stimulator ratio (10 000–25 000 of each). One µCi of ³H-thymidine was added to each well for the final 18 h of culture and ³H-thymidine uptake was then measured in a liquid β scintillation counter. Results were expressed as the mean of triplicate cultures.

Antibodies

The following TCRBV region-specific mAbs (from Immunotech Products – Beckman Coulter, Roissy, France) were used for flow cytometry: UN7 (anti-BV1S1), E2·2E7·2 (anti-BV2S1), LE89 (anti-BV3S1), IMMU157 (anti-BV5S1), 36213 (anti-BV5S2), 3D11 (anti-BV5S3), CRI304·3 (anti-BV6S1), 3G5D15 (anti-BV7S1), 56C5·2 (anti-BV8S1/S2), FIN9 (anti-BV9S1), C21 (anti-BV11S1), VER2·32·3 (anti-BV12S2), IMMU222 (anti-BV13S1), JU74·3 (anti-BV13S6), CVAS1·1.3 (anti-BV14S1), TAMAYA1·2 (anti-BV16S1), E17·5F3 (anti-BV17S1), BA62·6 (anti-BV18S1), ELL1·4 (anti-BV20S1), IG125 (anti-BV21S3) and IMMU546 (anti-BV22S1). Their binding was revealed by FITC-conjugated anti-mouse IgG antiserum (goat anti-mouse (GAM)-FITC (CALTAG Laboratories, Burlingame, CA, USA). The following mAbs were used for BLCL phenotyping: J4·119 (anti-CD19-APC), BL1a (anti-CD5-FITC), anti-CD79b-PE, 25·3.1 (anti-CD11a-PE), anti-CD40-PE, B6 (anti-CD23-FITC), anti-CD80-PE, anti-CD86-PE and j33 (anti-CD45-FITC) from Immunotech Products, HB7 (anti-CD38-PE), Leu-16 (anti-CD20-FITC), S-HCL-1 (anti-CD22-PE), 2A3 (anti-CD25-PE), W8E7 (anti-CD10-FITC), and anti-CD2-PE from BD Pharmingen, Le Pont de Claix, France, and FMC7-FITC from Dako, Trappes, France.

The following mAbs and their isotype controls were also used: anti-CD4-PE-Cy5, anti-CD8-PE-Cy5, anti-CD16-PE anti-CD8-APC, anti-CD4-PE and anti-IFNγ-FITC (Immunotech Products).

Cytokine measurements

The Cytokine Bead Array kit (BD Biosciences, San Diego, CA, USA) was used for simultaneous measurement of IL-2, IL-4, IL-6, IL-10, TNF-α, and IFN-γ, in BLCL supernatants [13]. Fifty microliters of each sample (24 h culture supernatants from 10⁶ non irradiated or 35 Gray-irradiated BLCL) were added to an equal volume of the cytokine bead mixture and detection reagent followed by a 3 h incubation at room temperature in the dark. Ten additional tubes, each containing equal volumes of beads, detection reagent, and graded amounts of the six cytokines, were prepared in parallel to generate a standard curve for each cytokine. At the end of the incubation period, the beads were washed with the buffer provided in the kit, centrifuged at 200 × g for 5 min, and the supernatants were carefully aspirated. The pellets were resuspended in 150 µl of the wash buffer provided in the kit and analysed immediately using a FACSArray; cytokine concentrations were determined using the FCAPArray software provided.

T-cell line phenotype analysis

A two-step procedure was used for TCRBV region-staining: CTL were incubated for 30 min at 4°C in V-bottom microtiter plates in the presence of 25 µl of the first TCRBV region-specific mAb at an optimal concentration. Cells were further incubated in the presence of GAM-FITC for 30 min. Between each step, plates were centrifuged, the supernatant was discarded by flicking and wells were washed twice with 200 µl ice-cold PBS/5% HS (PBS-HS). Labelled cells were analysed using a FACSCalibur flow cytometer (Beckton Dickinson, Mountain View, CA, USA).

Expression vectors

The expression vectors encoding 10 lytic proteins (BZLF1, BMLF1, BRLF1, BCRF1, BMRF1, BHLF1, BALF2, BLLF1, BALF4 and BHRF1), the eight latent EBV proteins (EBNA-1, −2, −3a, −3b, −3c, and –LP, LMP1, and LMP2), and the HLA-class I alleles (HLA-A*0101, HLA-A*0201, HLA-A*0301, HLA-B*0801, HLA-B*3501, HLA-B*4402, HLA-Cw0401, HLA-Cw0304 and HLA-Cw0701) have been described previously [14,15].

COS transfection and T-cell stimulation assay

COS cell transfection was performed using the DEAE–dextran chloroquine method, as already described [14–16]. Briefly, 15 000 COS cells were cotransfected with 100 ng of an expression vector coding for an EBV protein and 100 ng of an expression vector coding for one of the HLA class-I alleles. Transfected COS cells were used 48 h after transfection for CTL stimulation assays. T-lymphocytes to be tested (5 × 10⁴) were added to transfected COS cells. Culture supernatants were harvested 6 h later and tested for TNF-α content by measuring culture supernatant cytotoxicity for WEHI 164 clone 13 in a colourimetric assay [17]. OD was calculated from duplicate samples.

Peptides

The following 59 peptides (23 mers) from eight EBV proteins were tested: BZLF1 [18] (positions: 1–23, 12–34, 23–45, 34–56, 67–89, 78–100, 111–133, 122–144, 133–155, 144–166, 166–188, 177–199, 188–210, 199–221, 210–232 and 221–243), BRLF1 (position: 67–89), BHRF1 [19–22] (positions: 45–67, 56–78, 122–144 and 169–191), EBNA1 [23–27] (positions 67–89, 397–419, 419–441, 430–452, 452–474, 474–496, 485–507, 507–529, 518–540, 529–551, 540–462, 551–573, 573–595, 584–606 and 595–617), EBNA2 [3,11,28] (position: 243–265), EBNA3C [23,29] (positions: 56–78, 67–89, 100–122, 133–155, 364–386, 386–408, 540–562, 584–606, 617–639, 628–650, 639–661, 479–771 and 958–980), LMP1 [23] (positions: 122–155, 210–232, 331–353 and 342–364) and LMP2 [23] (positions: 144–166, 166–188, 221–243, 375–397 and 386–408).

Autopresentation of class II restricted peptides and detection of intracellular IFN-γ accumulation

Immuno-selected CD4+ T lymphocytes were seeded in 96-well V-bottom plates (200 000 per well) and incubated for 2 h (37°C, 5% CO₂) in 50 µl X-Vivo15 medium alone or supplemented with 40 µg/ml of a peptide, or 40 ng/ml of PMA (Sigma, St Quentin Fallavier, France) + 4 µg/ml of Ionomycin (Sigma) for positive controls. Brefeldin A (10 µg/ml) (Sigma) was then added for 4 additional hours. After centrifugation, cells were incubated for 15 min at RT with 2 mM EDTA, centrifuged once, incubated for 10 min at RT with 4% paraformaldhehyde, washed once with PBS/5% HS and then stored overnight at 4 °C in 1% paraformaldhehyde. The next day, after washing in PBS/5% HS, cells were resuspended in PBS/5% HS + Saponin (0·1%) (Sigma). After 10 min at RT, Saponin-treated cells were washed twice with PBS/5% HS. For IFN-γ detection, lymphocytes were incubated in the dark for 30 min at RT with a mixture of IFNγ-FITC and CD4-PE-Cy5. For the negative control, cells incubated without peptide were incubated with a mixture of each isotype control (IgG1-FITC and IgG1-PE-Cy5). After staining, cells were washed once with PBS/5% HS, once with PBS alone and finally stored in PBS/1% paraformaldhehyde until acquisition using a FACSCalibur flow cytometer (Beckton Dickinson).

Production of BLCL⁰ and BLCL^FCS

To eliminate biological risks arising from the presence of FCS, we first adapted the B95·8 cell line to a serum-free culture medium and used its supernatant as a serum-free source of EBV. We next attempted to transform PBMC without any addition of serum: after the 24 h incubation period in serum-free EBV-containing supernatant, X-VIVO15 was used again instead of RPMI + FCS for BLCL dilution. In general, although variable from one EBV supernatant to another, and between donors, BLCL cultures in serum-free medium were successful but had a doubling time longer than that observed in the presence of FCS. Under these conditions, it usually took 2 months to establish a serum-free BLCL in culture (BLCL⁰), compared to the 2–3 weeks it usually takes to obtain a ‘standard’ BLCL^FCS using FCS-supplemented medium. These results showed that serum was dispensable for B95·8 growth and EBV production as well as for BLCL infection and culture.

BLCL immunophenotyping and cytokine production

Previous studies have illustrated the potential variable expression of B cell surface markers as well as the variable level of cytokine production by BLCL ([8]). To test whether in vitro production of BLCL in the absence of serum could have biased their phenotype, the BLCL^FCS and BLCL⁰ were first characterized for 20 haematopoietic cell surface markers by direct immunostaining and FACS analysis. These markers included those that are routinely used to characterize leukaemia and lymphoma cells (CD2, CD5, C18, CD19, CD20, FMC7 (a CD20 epitope), CD22, CD23, CD25, CD38, CD45, CD79b, sIgK and sIgL. We also analysed the expression of the following CD that may be important during stimulation: CD11a, CD40, CD80 and CD86. As one can see in Table 2, with the exception of sIgK and sIgL expression, all BLCL appeared remarkably homogeneous in their expression of the markers tested. In fact, differences between BLCL^FCS and BLCL⁰ could be detected only when analysing the low or bright expression of CD23 (the low-affinity immunoglobulin E receptor, FcepsilonRII) and CD38 (an ectoenzyme that catalyses the synthesis of cyclic ADP ribose). The unique surface expression of IgK (for Do8 BLCL^FCS and BLCL⁰ and Do24 BLCL⁰) or IgL (for Do24 BLCL^FCS) strongly suggests that a particular clone had emerged within these cultures. We also examined cytokine production by non irradiated or irradiated (35 Gray) BLCL in supernatants after a 24 h culture period. The results are also shown in Table 2. Consistently low but nevertheless detectable amounts of IL-4, IL-6, IL-10, TNF-α and IFN-γ were produced by several cell lines. In accordance with the comparison of BLCL^FCS and BLCL⁰, one can note that pair BLCLs appeared almost identical in terms of cytokine production. Taken together, the above set of results demonstrate that omission of serum during BLCL production does not significantly bias their phenotype or cytokine profile.

The effect of serum on CTL-line growth in the presence of BLCL⁰ or BLCL^FCS

While serum addition was found to be dispensable for BLCL infection and culture, it was absolutely essential for efficient T-cell expansion (no significant amplification was observed when BLCL and PBMC were cocultured in X-VIVO15 only). Accordingly, for comparison, CTL-lines (1–4) were produced by coculture with BLCL⁰ or BLCL^FCS in the presence of FCS or pooled human serum (HS), or autologous serum (AS) in the case of CTL3 and CTL4 from Do24 as indicated in Table 1. In the case of donor Do24, all CTL lines were expanded in the presence of human serum and no differences were observed wether we used a pool (HS) or autologous (AS) serum (not shown). In contrast, for Do8 and Do29, considerable differences were observed: the four cultures performed in the presence of HS grew faster than those performed in the presence of FCS (Fig. 1). Concerning the effect of serum deprivation on BLCL, no difference in CTL-line growth was noted between selection performed with BLCL⁰ or BLCL^FCS (Fig. 1).

CTL-line diversity

A first ‘crude’ analysis of diversity was performed on CTL-lines 1–4 from Do8, Do24 and Do29 through determination of CD4+, CD8+ and CD16+ proportions. For Do29 and Do24 these proportions remained very similar from one culture condition to another. In contrast, for Do8 the proportion of the CD4+ subset was increased when FCS was present in the culture; either during BLCL production (CTL2), during CTL selection (CTL3), or both (CTL4) (data not shown). This initial analysis of CTL-line composition indicated that for Do8, FCS may have favoured the proliferation of a CD4+ subset, which was indeed the case, as demonstrated below. The T-cell repertoire diversity of the Do29 and Do24 CTL-lines (that appeared similar after CD4, CD8, CD16 determination) was further probed using a panel of 22 mAbs recognizing different TCRBV T-cell subsets. As can be seen in Fig. 2, for each donor, the proportion of the TCRBV populations selected during the culture seemed remarkably constant from one cell line to another. This was the case for the percentages of TCRBV1, TCRBV12, TCRBV14, TCRBV17 and TCRBV21 for the Do29 cell lines and of TCRBV2, TCRBV17 and TCRBV21 for the Do24 cell lines. Because all lines were found to be specific for the autologous BLCL (see below), these results suggested not only that serum had no drastic effect on CTL line composition (even though HS clearly had a positive effect on CTL-line expansion) but also that the selection procedure appeared to be fairly reproducible.

Specificity testing and recognition of BLCL⁰ and BLCL^FCS

All cell lines were tested against the autologous BLCL⁰, BLCL^FCS and against unrelated BLCL^FCS. Although some variability between donors and between experiments was observed in the cytotoxicity scores as well as in the levels of proliferation, all lines were found to be specific. Moreover, in line with the presence or absence of serum during culture of the target BLCL, no significant difference was observed between the CTL lines in their ability to recognize BLCL⁰ or BLCL^FCS (see example of two independent determinations for Do29 in Fig. 3). Because the presence of xenogenic antigens derived from the serum during immunization is expected to have a preferential effect on the CD4 compartment [30], we further tested the effect of FCS on the reactivity of CD4 purified CTL. Poor levels of reactivity were observed for most lines, suggesting that the small CD4 fraction within each line was not entirely composed of EBV specific T-cells (not shown). In contrast, for Do8, a clear-cut difference was observed between the level of proliferation against the BLCL⁰ and the BLCL^FCS. As would have been expected if an immunization had occurred against the FCS component, the line selected against BLCL^FCS showed a high level of proliferation against BLCL^FCS but not against BLCL⁰. This recognition was most likely restricted by autologous HLA-antigens, since it was not observed against an unrelated BLCL^FCS (Fig. 4a). Further evidence of the presence of FCS-specific CD4 T-lymphocytes within this cell line was provided by the demonstration that incubation of BLCL⁰ with FCS, but not HS, restored its recognition by the CD4 purified CTL (Fig. 4b).

EBV epitope mapping

Finally, we used the COS transfection assay (see Methods section) to test the precise EBV specificity of four cell lines from Do29 and Do24 that had been selected against either BLCL⁰ or BLCL^FCS. The potential recognition of 18 EBV-proteins was tested in the context of 5 HLA class-I molecules: three specificities were detected for Do29 CTL1: BZLF1, BMRF1 and EBNA3A, all of which were presented within HLA-B*3501. The BZLF1 and EBNA3A specificities were also detected from Do29 CTL2. Two specificities were detected for Do24 CTL1 (BZLF1/HLA-B*0801 and BML1/HLA-A*0201), that were also found for Do24 CTL2. Determination of class II specificities after autopresentation of HLA class-II restricted peptides (see Materials and methods section) allowed identification of a response against the EBNA-1-(394–419) peptide for the two CTL-lines CTL1 and CTL2 from Do29 (data not shown). Taken together, these data suggest that dominant specificities, which were either class I or class II restricted, were preserved in the different culture conditions tested. Thus, at least for these few dominant specificities, we did not find example suggesting a competition between FCS and EBV epitopes.