Effect of nuclear factor κB inhibition on tumor cell sensitivity to natural killer-mediated cytolytic function

2.1 Induction of NF-κB activation in target cells upon interaction with NK cells

It has been suggested that NF-κB may mediate aspects of programmed cell death. In these studies, we first demonstrated that co-incubation of MCF7 target cells with freshly isolated NK cells (> 90 % CD3^–, see Sect. 4.2) for 15 or 30 min resulted in the translocation of NF-κB (Fig. 1). Similar results were obtained with a TCRγ δ T cell clone, L6-1, displaying an MHC-unrestricted, "NK-like", lysis (Fig. 1 A 11. Integrity of the nuclear extracts was controlled by electrophoretic mobility shift assays (EMSA) using a probe that interacts with Oct 1 which is not inducible (Fig. 1 B). These data suggest that target cell interaction with NK cell effectors may utilize surface structures in initiating transmembrane signaling. In this regard, it should be noted that NK cells interact with tumor cells via some target cell membrane proteins, including MHC molecules. Whether there is a subset of receptors (i. e. TNFR, CD48, TRAIL / R) initiating second messengers in target cells that may regulate NK-induced cell killing has to be determined. Immunofluorescence analysis indicated Fas and TNFR expression on all tumor cel lines (data not shown). Recently, it has been reported that fresh NK cells express the membrane form of TNF 12, which may represent a potential structure involved in the induction of NF-κB. Future studies are, however, needed to delineate the exact structure involved and to characterize the putative target surface receptors that initiate NF-κB activation. Experiments are in progress to further delineate the signaling pathways activated in target cells and to examine their possible interference with the control of susceptibility to NK-induced cell death.

2.2 Inhibition of NF-κB activation by cell transfection with mutated IκB-α cDNA

Mutations of the serine 32 and 36 of IκB-α protein have been described 13. These mutations, by disrupting the two main inducible IκB-α phosphorylation sites, abolish IκB-α degradation and thus prevent NF-κB activation. To inhibit NF-κB translocation, we therefore introduced the dominant negative human IκB-α mutant construct coding for a mutated IκB-α that is resistant to both phosphorylation and proteolytic degradation into MCF7, HCT116 and OVACAR-3 cells. As demonstrated in gel retardation assays performed with nuclear extracts from representative stable clones, NF-κB / DNA binding activity was potently induced in all cells tested after stimulation with TNF for 30 min (Fig. 2). In sharp contrast, TNF failed to induce the nuclear translocation of NF-κB in cells expressing mutated IκB (MCF7 / MT, HCT116-MT and slightly OVCAR-MT). No further activation of NF-κB could be detected after 24 h treatment with TNF, suggesting that NF-κB was inhibited by a stabilized association with the mutated IκB-α. Similarly, no NF-κB translocation was observed in Iκ-B-α-transfected cells when incubated with NK cells for 30 min (data not shown).

2.3 Functional inhibition of NF-κB in mutated IκB-α-expressing cells

We then tested the functional effect of the inhibition of NF-κB translocation in IκB-α-transfected cells by studying the expression of one of the TNF-inducible genes, mitochondrial manganous superoxide dismutase (MnSOD). This gene presents potential κB sites in its promoter region and the induction of its expression is closely associated with NF-κB activation by TNF. Northern blot analysis indicated that, as opposed to MCF7, OVCAR and HCT116 parental cells, TNF failed to induce MnSOD transcription in mutated IκB-α-transfected equivalents demonstrating a functional inhibition of NF-κB in these cells (Fig. 3). Similar results were obtained with another NF-κB-regulated gene, ICAM-1 (data not shown).

2.4 NF-κB inhibition in target cells does not increase their sensitivity to NK-mediated killing

It is now well established that NK cell-mediated cytotoxicity is inhibited upon specific recognition of MHC class I molecules on target cells 14, 15. Given the fact that NF-κB has been shown to participate in the control of the MHC class I gene basal expression 16, initial experiments were performed to examine the effect of NF-κB inhibition on constitutive MHC class I expression on the target cells. Basal MHC class I expression was thus compared in parental cells and in their derivatives stably transfected with the mutated IκB-α gene. FACS analysis indicated comparable fluorescence intensity in HCT, MCF7 and OVCAR parental cells and in their equivalent transfected with mutated IκB-α (data not shown). These results indicate that stable transfection of the mutated IκB-α does not alter the constitutive expression of MHC class I in these cells.

Cytotoxicity assays were then performed to determine the sensitivity of the mutated IκB-α -transfected cells to NK-induced cytotoxicity. For this purpose, fresh NK cells (> 90 % CD3^–) were isolated from healthy donor PBMC after depletion of T cells, B cells and macrophages. These non-stimulated NK cells were able to efficiently lyse all tumor cell lines in a non-MHC class I-restricted manner. Data shown in Fig. 4 A indicate the absence of significant difference in the cytotoxic response between the control tumor cells and the mutated IκB-α-transfected cells. This suggests that the activation of NF-κB by NK in target cells is not protective and fits with our previous work indicating the failure of NF-κB inhibition to increase sensitivity to cytotoxic drugs in these cells 17. To further confirm that inhibition of NF-κB does not influence the non-MHC-restricted cytotoxicity, we used the TCRγ δ T cell clone, L6-1. This clone displays an HLA-A1-restricted TCR-mediated cytotoxic activity toward its specific target cells as well as a MHC-unrestricted lysis against K562 and Daudi target cells 11. As shown in Fig. 4B, L6-1 T cell clone killed parental target cells and their equivalents, in which NF-κB activation was inhibited, in a similar manner.

In the following experiments, we examined the mechanisms used by both NK and TCRγ / δ cells to mediate their cytotoxic activity toward all tumor cell lines, transfected or not with IκB-α. For this purpose, blocking experiments were performed in the presence of ZB4 (anti-Fas neutralizing mAb), EGTA (used to inhibit the Ca²⁺-dependent perforin / granzyme-mediated lysis) or NKTa control mAb. Results presented in Fig. 5 clearly showed that both effector cells mainly use a granule exocytosis pathway to kill tumor targets expressing or not NF-κB. Furthermore, experiments performed either with agonistic anti-Fas mAb (CH11) or soluble FasL indicated that all IκB-α-transfected cells were as efficiently lysed as parental tumor cells (Fig. 6 A). This cytotoxicity was inhibited in the presence of ZB4 anti-Fas blocking mAb (Fig. 6 B). It should be noted that previous immunofluorescence experiments indicated that both parental and transfected tumor cells express similar intensity of CD95 on their surface (data not shown).

4.1 Target cell lines and reagents

HCT116 human colon carcinoma cells (ATCC CCL 247), MCF7 breast cancer cells, and OVCAR-3 ovarian carcinoma cells were described previously 19. MCF7 / MT cells were transfected with IκB-α cDNA as described 20. OVCAR-MT and HCT116-MT were obtained by stable transfection of parental cells with a linearized pcRV-CMV plasmid encoding the IκB-α protein mutated at Ser-32 and Ser-36 (a gift from Dr. A. Israël, Institut Pasteur, Paris, France) as described previously 17. L6-1 TCRγ / δ T cell clone was obtained from a healthy individual (A1 / A24, BW57 / B27, CW6) as previously reported 11. The human recombinant TNF (specific activity 6.63 × 10⁶ U / mg protein) was kindly provided by Dr. I. Apfler (Bender Wien, Austria).

4.2 NK cell preparation and FACS analysis

PBL were isolated from healthy donors (Banque du sang, Hôpital Saint-Louis, Paris, France) by Ficoll-Hypaque gradient centrifugation followed by plastic adherence. NK cells were enriched by Percoll density gradient and incubated for 30 min at 4 °C with a cocktail of OKT3, OKT4 and OKT8 antibodies (Ortho Diagnostics Systems, Inc, Westwood, MA), which recognized CD3, CD4 and CD8 molecules, respectively, and MY4 (anti-CD14) and B4 (anti-CD20) mAb, (Coulter / Immunotech, Marseille, France). The cells were resuspended in RPMI complete medium supplemented with 10 % human serum, incubated with magnetic beads (Dynal, Oslo, Norway) for 2 min, and passed over a magnet. This separation was performed twice and the resulting cell preparation contained more than 90 % CD3^– cells. For flow cytometry analysis of MHC class I expression, tumor cells were trypsinized, washed and resuspended in cold buffer consisting of PBS with 2 % FCS. Cells (2 × 10⁵) were resuspended in 100 μl W6 / 32 mAb at a concentration of 1 μg / ml and then were incubated on ice for 25 min, washed twice in the above buffer and resuspended in 100 μl of a 1 / 200 dilution of FITC-labeled goat anti-mouse IgG.

4.3 Nuclear extracts and EMSA

Tumor cells (3 × 10⁶) were incubated in medium in the presence of 50 ng / ml TNF or in the presence NK cells or TCRγ / δ T cells (9 × 10⁶) for 15 or 30 min. When tumor cells were cocultured with the NK cells or L6-1 CTL clone, only adherent cells were used for nuclear extract preparation. The cells were then trypsinized and washed with PBS. Nuclear extracts were prepared according to the procedure of Dignam et al. 21.

4.4 Cytotoxicity assay

Serial dilutions of effector cells were distributed in triplicates (0.1 ml per well) into round-bottom microwell plates. Target cells were labeled with 200 mCi Na₂⁵¹CrO₄ (5 μCi / ml; Amersham) for 1 h at 37 °C and washed three times. Cells (5.000 per well in 0.1 ml) were distributed in round-bottom microwells with 0.1 ml of effector cells at various E / T ratios. After 4 h of incubation at 37 °C, the plates were centrifuged at 2,000 g for 2 min, and cell-free supernatants were collected using a cell harvester (Skatron Inc, Sterling, VA). Supernatant radioactivity was assayed using an automated gamma counter (Packard Instrument Co, Meriden, CT). Spontaneous release was determined by incubating target cells in medium alone. Maximum release was determined by adding 0.1 ml of 1 M HCl to the target cell suspension. The percentage of specific lysis was calculated as follows: 100 × (experimental ⁵¹Cr release – spontaneous ⁵¹Cr release) / (maximum ⁵¹Cr release – spontaneous ⁵¹Cr release). SD were < 10 %. Inhibitory effect of the antibodies was tested by preincubating target cells for 1 h in the presence of ZB4 (anti-Fas mAb), isotype control mAb (NKTa) or EGTA + MgCl₂, the effector cells were then added.