Abstract
Purpose: Natural killer (NK) cells express an activating Fc receptor (FcγRIIIa) that mediates antibody-dependent cellular cytotoxicity (ADCC) and production of immune modulatory cytokines in response to antibody-coated targets. Cetuximab is a therapeutic monoclonal antibody directed against the HER1 antigen. We hypothesized that the NK cell response to cetuximab-coated tumor cells could be enhanced by the administration of NK cell–stimulatory cytokines.
Experimental Design: Human NK cells stimulated with cetuximab-coated tumor cells and interleukin-2 (IL-2), IL-12, or IL-21 were assessed for ADCC and secretion of IFN-γ and T cell–recruiting chemokines. IL-21 and cetuximab were given to nude mice bearing HER1-positive xenografts.
Results: Stimulation of human NK cells with cetuximab-coated tumor cells and IL-2, IL-12, or IL-21 resulted in 3-fold to 10-fold higher IFN-γ production than was observed with either agent alone. NK cell–derived IFN-γ significantly enhanced monocyte ADCC against cetuximab-coated tumor cells. Costimulated NK cells also secreted elevated levels of chemokines (IL-8, macrophage inflammatory protein-1α, and RANTES) that could direct the migration of naive and activated T cells. IL-2, IL-12, and IL-21 enhanced NK cell ADCC against tumor cells treated with cetuximab. The combination of cetuximab, trastuzumab (an anti-HER2 monoclonal antibody), and IL-21 mediated greater NK cell cytokine secretion and ADCC than any agent alone. Furthermore, administration of IL-21 enhanced the effects of cetuximab in a murine tumor model.
Conclusions: These results show that cetuximab-mediated NK cell activity can be significantly enhanced in the presence of NK cell–stimulatory cytokines. These factors, therefore, may be effective adjuvants to administer, in combination with cetuximab, to patients with HER1-positive malignancies.
- cetuximab
- HER1
- NK cell
- Fc receptor
- interleukin-2
- interleukin-12
- interleukin-21
The epidermal growth factor receptor belongs to the erbB receptor tyrosine kinase family, which consists of four related transmembrane receptors: erbB1 (epidermal growth factor receptor or HER1), erbB2 (HER2/neu), erbB3 (HER3), and erbB4 (HER4). Upon ligand binding, HER1 homodimerizes or heterodimerizes with other erbB family members and initiates signaling through the phosphatidylinositol 3-kinase/AKT and Ras/mitogen-activated protein kinase pathways, which promote survival and proliferation (1). HER1 overexpression is observed in multiple human malignancies, including colorectal, lung, breast, gastric, and pancreatic carcinomas. In tumor cell lines, HER1 overexpression is associated with increased proliferation and decreased apoptosis, and in human cancer patients, tumor expression of HER1 has been correlated with poor clinical outcome (2–5). These observations suggested that therapeutic measures designed to disrupt HER1 function might lead to reduced tumor growth in patients with HER1-overexpressing malignancies. Cetuximab (Erbitux) is a monoclonal antibody (mAb) that binds to the extracellular domain of the HER1 molecule and inhibits ligand binding. Cetuximab has been approved by the Food and Drug Administration for the treatment of patients with metastatic colorectal cancer when given in combination with the topoisomerase inhibitor irinotecan. Promising results have also been achieved in phases I and II clinical trials for non–small cell lung cancer, head and neck cancer, and other carcinomas in which cetuximab was combined with chemotherapy or kinase inhibitors, such as erlotinib. However, response rates to cetuximab remain low overall (6, 7).
Natural killer (NK) cells are large, granular lymphocytes that participate in innate immune responses to viruses, bacteria, and neoplastic cells (8). Although most innate immune cells express both inhibitory and activating FcR, NK cells are unique in that they constitutively express only a low-affinity, activating FcR (FcRIIIa or CD16), which enables them to interact with antibody-coated targets. In addition to their ability to mediate antibody-dependent cellular cytotoxicity (ADCC), FcR-activated NK cells have also been shown to secrete factors, such as IFN-γ, tumor necrosis factor-α, and chemokines, such as macrophage inflammatory protein-1α/β, IL-8, and RANTES, that inhibit tumor cell proliferation, enhance antigen presentation, and aid in the chemotaxis of T cells (9). The activity of cetuximab and other antibodies directed against tumor antigens has largely been attributed to the direct, antiproliferative, or proapoptotic effects of the antibodies on the tumor cells. However, in a murine xenograft model, the antitumor effects of trastuzumab (an anti-HER2/neu mAb) were at least partially dependent upon the presence of FcR-bearing immune cells, including NK cells (10). The observation that FcR-dependent mechanisms contributed to the effects of antitumor mAbs suggested that their efficacy could be enhanced via the administration of immune modulatory cytokines with the capacity to activate NK cells.
In a phase I clinical trial of trastuzumab and IL-12 conducted by our group, as well as in a follow-up phase I trial of trastuzumab, IL-12, and paclitaxel, clinical outcome correlated with NK cell production of IFN-γ (11, 12). Elevated levels of the NK cell–derived chemokines IL-8, macrophage inflammatory protein-1α, and RANTES were also detected within the sera of responding patients. These NK factors could induce the chemotaxis of naive and activated T cells, and their presence correlated with the infiltration of tumor tissue by CD8+ cytotoxic T cells. These data suggested that the combination of trastuzumab and IL-12 could lead to NK cell cytokine and chemokine production in a subset of patients and that this was associated with a favorable clinical outcome.
In the current report, we show that cetuximab acts as a potent NK cell FcR stimulus and that NK cell activation in response to cetuximab is enhanced in the presence of immune modulatory cytokines. NK cells costimulated with cetuximab-coated tumor cells, and IL-2, IL-12, or IL-21 secreted elevated levels of IFN-γ and several T cell–recruiting chemokines compared with NK cells stimulated with either agent alone. NK cell ADCC against cetuximab-coated tumor cells was also enhanced in the presence of IL-2, IL-12, or IL-21. Coadministration of IL-21 and cetuximab to mice bearing HER1-positive xenografts resulted in decreased tumor growth compared with mice receiving cetuximab or IL-21 alone. These results suggest that that immune modulatory cytokines would be effective adjuvants to administer in combination with cetuximab.
Materials and Methods
Cytokines and antibodies. Recombinant human IL-12 was provided by Genetics Institute, Inc. Recombinant human IL-21 was provided by ZymoGenetics, Inc. Recombinant human IL-2 was provided by Roche Pharmaceuticals. Recombinant human IFN-γ was purchased from Biosource/Invitrogen. Cetuximab (Erbitux) was purchased from ImClone Systems, Inc.. Trastuzumab (Herceptin) and rituximab (Rituxan) were provided by Genentech Corporation. Polyclonal human IgG1 was purchased from Sigma-Aldrich.
Cell lines. The SKBR3 and MDA-MB-468 human breast adenocarcinoma lines and the A549 and H460 human non–small cell lung carcinoma lines were obtained from the American Type Culture Collection. The human pancreatic cell lines BxPC3, HPAC, and PANC-1 were a gift from Dr. Mark Bloomston (Ohio State University). All cell lines were propagated in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 300 μg/mL of l-glutamine, penicillin (100 units/mL), streptomyocin (100 μg/mL), amphotericin B (0.25 μg/mL), and anti-PPLO agent (0.06 mg/mL; all from Life Technologies-Bethesda Research Laboratories), except for the PANC-1 and HPAC cell lines, which were propagated in DMEM medium supplemented as described above.
Isolation of human cells. Human NK cells were isolated directly from fresh peripheral blood leukopacks (American Red Cross) by 30-min incubation with RossetteSep cocktail (Stem Cell Technologies) before Ficoll Hypaque (Sigma) density gradient centrifugation. NK cells (>96% CD56+) were cultured in RPMI 1640 supplemented with 10% heat-inactivated pooled human antibody serum (C-Six Diagnostics), 100 units/mL penicillin, 100 μg/mL streptomyocin, and 0.25 μg/mL amphotericin B. To isolate monocytes, human peripheral blood mononuclear cells were first isolated by density gradient centrifugation over Histopaque (Sigma). Monocytes were purified from the peripheral blood mononuclear cells by negative selection using the MACS monocyte isolation kit (Miltenyi Biotec), as previously described (13). Monocytes were cultured in RPMI 1640 supplemented with 10% fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomyocin, and 0.25 μg/mL amphotericin B.
Flow cytometry of tumor cell lines. To evaluate tumor cell line expression of HER1 and HER2, tumor cells were harvested by trypsanization, washed once in PBS, and incubated for 30 min on ice in PBS containing 10 μg/mL cetuximab, trastuzumab, or an isotype-matched irrelevant control antibody (rituximab, anti-CD20). Cells were washed once, resuspended in PBS, and labeled for 30 min on ice with a PE-conjugated antihuman IgG secondary antibody (BD Biosciences). Cells were then fixed in 1% formalin and stored at 4°C until analysis. Nonspecific staining in rituximab-treated cells was subtracted from the staining in the presence of cetuximab or trastuzumab to determine the mean fluorescence intensity (MFI) of the positively staining population.
NK cell FcR stimulation assays. For immobilized antibody experiments, wells of a 96-well flat-bottomed plate were coated with 100 μg/mL of cetuximab in cold PBS overnight at 4°C and then plated with human NK cells (2 × 105 per well) and 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21. For the in vitro tumor coculture assays, tumor cells were grown to confluence in a 96-well flat-bottomed culture plate and then treated with 100 μg/mL cetuximab or 100 μg/mL polyclonal human IgG for 1 h at 37°C. Purified NK cells were then added at 2 × 105 per well in unsupplemented 10% human antibody serum medium or in medium containing 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21. At the indicated time points, cell-free culture supernatants were harvested and analyzed for levels of various cytokines and chemokines using commercially available ELISA kits (R&D Systems). NK cells were checked for viability at the conclusion of each experiment. Cell viability in all conditions was routinely >95%. In FcγRIIIa blocking experiments, NK cells were incubated for 30 min at 4°C in PBS containing 100 μg/mL anti-CD16 blocking antibody (clone 3g8, Beckman Coulter) or an isotype-matched irrelevant control IgG before use in tumor coculture assays. In IFN-γ neutralization experiments, culture supernatants derived from NK cells cultured with control IgG-treated tumor cells or cetuximab-coated tumor cells and IL-21 were incubated for 30 min at 4°C with 10 μg/mL anti–IFN-γ neutralizing antibody (R&D Systems) or an isotype-matched irrelevant control IgG before use.
ADCC assays. For NK cell ADCC, NK cells from normal human donors were plated in 96-well V-bottomed plates in 10% human antibody serum media supplemented with 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21, and incubated overnight at 37°C. Cetuximab-treated or control IgG-treated 51Cr-labeled tumor cells were added to the NK cells at various effector/target (E:T) ratios. The percentage of lysis was calculated as previously described (14). The percentage of ADCC was calculated for each condition as the percentage of lysis in the presence of cetuximab minus the percentage of lysis in the presence of control IgG. After a 4-h incubation at 37°C, supernatants were harvested for quantification of chromium release. For monocyte ADCC, the incubation period was extended to 18 h.
Cytokine antibody arrays. Human NK cells were cocultured with cetuximab-coated MDA-MB-468 cells in the presence of IL-2, IL-12, or IL-21, as described above. Control conditions included NK cells cultured with control IgG-treated tumor cells, NK cells cultured with cetuximab-coated tumor cells, and NK cells cultured with control IgG-treated tumor cells in media containing IL-2, IL-12, or IL-21. After 24 h, culture supernatants were harvested and analyzed using the human cytokine antibody array following the manufacturer's instructions (Panomics, Inc.).
Measurement of T-cell chemotaxis. Human T cells were isolated directly from fresh peripheral blood leukopacks by 30-min incubation with RossetteSep cocktail before Ficoll Hypaque density gradient centrifugation. A portion of the T cells were left untreated and used immediately (naive T cells), whereas the remainder were activated for 5 days by sequential culture with phytohemagglutinin (Sigma-Aldrich) and IL-2, as described previously (15). Transwell inserts for the migration experiments (pore size, 5 μm; Corning) were prepared as described (15). Migration experiments were done by placing 2 × 105 purified naive or activated T cells in 100 μL of 10% human antibody serum medium in the upper chambers of fibronectin-coated Transwell inserts. Inserts were then placed into the wells of a 24-well plate. NK cell culture supernatants (400 μL) were placed in the lower chambers of the wells. After a 3-h incubation at 37°C, the fluid in the lower chambers was collected, and migrated cells were quantitated by trypan blue exclusion.
Murine tumor model. Age-matched, female, BALB/c nude mice (The Jackson Laboratory) were inoculated s.c. with 2 × 106 A549 tumor cells in a 200-μL volume. When the tumors had reached a volume of ∼200 mm3 (5-7 days), mice were randomly allocated to treatment with PBS, 10 μg IL-21 per mouse, 1 mg/kg cetuximab or cetuximab and IL-21 (n = 5 mice per group). Doses were chosen on the basis of titration experiments examining the effects of cetuximab or IL-21 alone against A549 xenografts (data not shown). All treatments were given i.p. thrice weekly. Tumor dimensions were measured twice weekly with calipers, and tumor volume was calculated as follows: tumor volume = 0.5 × [(large diameter) × (small diameter)2]. All protocols were approved by the Ohio State University Animal Care and Use Committee, and mice were treated in accordance with institutional guidelines for animal care.
Statistics. Statistical significance in all experiments, except the murine tumor model, was determined using Student's two-tailed t test with data from at least three independent experiments (the exact number of experiments done is indicated in each figure legend). A P value of <0.05 was considered significant. Changes in tumor volume over time were assessed via a longitudinal model. Tumor values were first log-transformed, and then a mixed effects model was applied to the data. Estimated slopes (changes in tumor volume over time) were calculated with 95% confidence intervals, and estimated differences in tumor volume were calculated at baseline and at the end of each study.
Results
Immune modulatory cytokines enhance IFN-γ secretion from NK cells cocultured with cetuximab-coated tumor cells. We were interested in determining the ability of immune stimulatory cytokines to enhance NK cell IFN-γ production in response to immobilized cetuximab. NK cells from normal human donors secreted modest amounts of IFN-γ after culture with cetuximab alone. Costimulation of NK cells with cetuximab plus IL-2, IL-12, or IL-21 synergistically enhanced IFN-γ production (P = 0.004; Fig. 1A ). NK cell IFN-γ production was detectable within 6 h of exposure to cetuximab and cytokines and peaked near 24 h (data not shown). In dose-response experiments, elevated IFN-γ production was observed even in response to the lowest concentration of cetuximab tested (12.5 μg/mL), whereas the antibody concentration of 100 μg/mL seemed to elicit maximal stimulation (data not shown). Of note, the cytokine that served as the most potent costimulus for NK cell IFN-γ production varied between donors, as did the overall levels of IFN-γ production that were observed. However, for all donors tested, costimulation of NK cells with cetuximab and either IL-2, IL-12, or IL-21 resulted in 3-fold to 10-fold higher levels of IFN-γ production compared with stimulation with either agent alone.
Human NK cells secrete high levels of IFN-γ in response to cetuximab-coated tumor cells. A, human NK cells were cultured on plates precoated with cetuximab (100 μg/mL) in the presence of IL-2 (0.1 nmol/L), IL-12 (10 ng/mL), or IL-21 (10 ng/mL). Control conditions consisted of NK cells cultured with medium alone, cetuximab alone, or cytokine alone. Cell-free culture supernatants were harvested at 24 h and analyzed for IFN-γ content by ELISA. B, HER1-positive breast cancer cell lines were treated with cetuximab and cultured with human NK cells in medium supplemented with IL-2, IL-12, or IL-21 (Cetux + Cytokine; ref. 14). Additional conditions included NK cells cultured with control IgG-treated tumor cells in unsupplemented medium (Medium), NK cells cultured with control IgG-treated tumor cells in media containing IL-2, IL-12, or IL-21 (Cytokine), and NK cells coated with cetuximab-coated tumor cells in media lacking cytokine (Cetux). Supernatants were harvested for quantification of IFN-γ release at 24 h. Similar experiments were conducted using HER1-positive non–small cell lung carcinoma cell lines (C) and human pancreatic cancer cell lines (D). E, NK cells were pretreated with a CD16 blocking antibody or an isotype-matched control IgG and then plated with cetuximab-coated MDA-468 cells in the presence of IL-2, IL-12, or IL-21. Cell-free culture supernatants were harvested at 24 h and analyzed for IFN-γ content by ELISA. Each graph depicts results from one of five representative donors. Columns, mean of triplicate ELISA wells; bars, SD.
We next examined NK cell IFN-γ production in response to cetuximab-coated tumor cells. An in vitro assay was used, in which purified human NK cells were cocultured with cetuximab-coated human tumor cells in the presence of IL-2, IL-12, or IL-21. Control conditions consisted of NK cells cultured with control IgG-treated tumor cells, NK cells cultured with IgG-treated tumor cells in medium containing IL-2/IL-12/IL-21, and NK cells cultured with cetuximab-coated tumor cells in unsupplemented medium. After 24 h of coculture, supernatants were harvested and analyzed for IFN-γ content. The HER1 expression of human tumor cell lines was determined by flow cytometry. The cell lines used included the breast cancer cell lines SKBR3 (30.8% HER1+, MFI 6.8; HER1low) and MDA-MB-468 (81.5% HER1+, MFI 20.8; HER1high), the non–small cell lung carcinoma lines H460 (49.3% HER1+, MFI 7.6; HER1low) and A549 (95.7% HER1+, MFI 21.9; HER1high), and the pancreatic cancer cell lines BxPC3 (44.5% HER1+, MFI 7.7; HER1low), PANC-1 (38.4% HER1+, MFI 7.0; HER1low), and HPAC (92.4% HER1+, MFI 31.5; HER1high). Culture of NK cells with cetuximab-coated tumor cells alone or with IL-2, IL-12, or IL-21 alone resulted in relatively low levels of IFN-γ production. IFN-γ production was synergistically enhanced when NK cells were cocultured with cetuximab-coated tumor cells and IL-2, IL-12, or IL-21 (P < 0.001; Fig. 1B-D). This result was not observed when NK cells were treated with an anti-CD16 blocking mAb before costimulation, confirming that the ability of cetuximab to stimulate NK cell IFN-γ production is dependent on activation of FcγRIIIa (Fig. 1E). Of note, synergistic IFN-γ production was observed even when HER1low cell lines were used (e.g., H460).
Cetuximab-specific ADCC is enhanced in the presence of immune modulatory cytokines. The effect of overnight activation of human NK cells with IL-2, IL-12, or IL-21 was examined in an ADCC assay using cetuximab-coated human tumor cells as targets. Human NK cells were cultured overnight in unsupplemented medium or medium containing 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21. Target cells were labeled with 51Cr and treated with either polyclonal human IgG or cetuximab before use in a standard 4-h ADCC assay (14). For all cell lines tested, cytokine-activated NK cells exhibited a low level of cytolytic activity against cells treated with a control antibody (<5% lysis; data not shown). NK cells incubated overnight in unsupplemented medium were able to mediate significant levels of ADCC against cetuximab-coated targets (P = 0.026 at an E:T ratio of 50:1). In addition, cytokine activation markedly enhanced cetuximab-specific ADCC activity (P = 0.001 at an E:T ratio of 50:1; P < 0.05 at all E:T ratios; Fig. 2A-C ). Notably, significant cetuximab-specific ADCC was detected even against cell lines with low HER1 expression (e.g., H460).
IL-2, IL-12, and IL-21 enhance NK cell ADCC against cetuximab-treated human cancer cells. Purified human NK cells were incubated overnight in medium alone or in medium supplemented with 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21. Lytic activity was then assessed in a standard 4-h chromium release assay using cetuximab-coated breast cancer cell lines (A), non–small cell lung cancer cell lines (B), or pancreatic cancer cell lines (C) as targets. The percentage of ADCC was calculated for each condition as the percentage of lysis in the presence of cetuximab minus the percentage of lysis in the presence of control IgG (14). Each graph depicts results from one representative donor. A minimum of five normal donors was tested per tumor cell line. Points, mean of three replicates; bars, SD.
Cetuximab exhibits a similar capacity to induce NK cell FcR activation as trastuzumab. We were next interested in comparing the relative abilities of cetuximab and trastuzumab to stimulate NK cell FcR effector functions. To address this question, we used the H460 cell line, which was found to express approximately equivalent (low) levels of HER1 and HER2 protein on the cell surface (Fig. 3A ). To compare the effects of cetuximab and trastuzumab on NK cell IFN-γ production, NK cells from a single donor were cocultured with H460 tumor cells that had been pretreated with 100 μg/mL of cetuximab or trastuzumab. NK cells secreted similar amounts of IFN-γ when costimulated with cetuximab-coated H460 tumor cells as with trastuzumab-coated H460 cells (P = 0.43; Fig. 3B). However, NK cell ADCC against cetuximab-coated H460 cells was significantly higher than NK cell ADCC against trastuzumab-coated H460 cells (P = 0.038 at an E:T ratio of 50:1; Fig. 3C). This effect was even more pronounced with NK cells that had been activated with IL-12 or IL-21 (P < 0.001 at the 50:1 E:T ratio). Furthermore, enhanced NK cell ADCC and IFN-γ production were observed when H460 tumor cells were treated with both cetuximab and trastuzumab (Fig. 3D). These results show that cetuximab is at least as potent an NK cell FcR stimulus as trastuzumab.
NK cell IFN-γ production and ADCC in response to trastuzumab-coated tumor cells and cetuximab-coated tumor cells are equivalent. A, HER1 and HER2 expression of H460 human non–small cell lung cancer cell lines. B, purified human NK cells from a single donor were cocultured with 10 ng/mL IL-12 (top) or IL-21 (bottom) and trastuzumab-coated H460 tumor cells or cetuximab-coated H460 tumor cells. IFN-γ production was measured after 24 h. Similar results were obtained when IL-2 was used as the costimulus. C, NK cells from a single donor were incubated overnight in unsupplemented medium or in medium containing 10 ng/mL IL-12 (top) or IL-21 (bottom). NK cells were then used in a standard 4-h chromium release assay using cetuximab-coated H460 cells or trastuzumab-coated H460 cells as targets. Similar results were obtained using IL-2. D, NK cell IFN-γ production (top) and ADCC (bottom) were assessed using H460 tumor cells treated with trastuzumab, cetuximab, or the combination. Results from one of three independent experiments. Points and columns, mean from three replicates; bars, SD.
NK cell–derived IFN-γ enhances monocyte ADCC against cetuximab-coated tumor cells. We next wanted to determine whether IFN-γ could enhance monocyte ADCC against cetuximab-coated targets. Peripheral blood monocytes (PBM) from normal donors were incubated in unsupplemented medium or in medium containing increasing concentrations of IFN-γ. After an overnight incubation period, PBMs were assessed for ADCC activity in response to cetuximab-coated MDA-MB-468 cells. As shown in Fig. 4A , IFN-γ-activated PBMs mediated significant ADCC against cetuximab-coated tumor cells (P = 0.022). In contrast, incubation of PBMs with IL-12, IL-2, or IL-21 had no effect on ADCC. PBM secretion of IL-12 after stimulation with cetuximab-coated tumor cells was also enhanced in the presence of IFN-γ, but not IL-2, IL-12, or IL-21 (P = 0.008; Fig. 4B). To determine whether the ADCC activity of PBMs could be enhanced in response to IFN-γ derived from cetuximab-stimulated NK cells, PBMs were incubated overnight in cell-free culture supernatants derived from NK cells that had been stimulated with cetuximab and IL-21. As observed previously, PBMs incubated with culture supernatants from costimulated NK cells exhibited considerable ADCC against cetuximab-coated tumor cells. However, this effect was significantly decreased in the presence of an IFN-γ neutralizing antibody (P < 0.001; Fig. 4C). PBMs incubated with supernatants from costimulated NK cells also secreted significant quantities of IL-12 after stimulation with cetuximab-coated tumor cells, an effect that was also dependent on IFN-γ (P < 0.001; Fig. 4D). These findings suggest that IFN-γ secreted by NK cells could enhance monocyte activation in response to cetuximab-coated targets.
NK cell–derived IFN-γ enhances monocyte ADCC and cytokine release in response to cetuximab-coated tumor cells. A, human PBMs were incubated overnight in medium alone or medium supplemented with IFN-γ (5 or 25 ng/mL). Lytic activity was then assessed in an 18-h chromium release assay against cetuximab-coated MDA-MB-468 cells. PBMs activated with 0.1 nmol/L IL-2, 10 ng/mL IL-12, or 10 ng/mL IL-21 served control conditions. B, PBMs were pretreated as described above and then cocultured with cetuximab-coated MDA-MB-468 cells for an additional 18 h. Culture supernatants were harvested and analyzed for IL-12 p40 content by ELISA. C, culture supernatants harvested from NK cells that had been stimulated with medium alone or with cetuximab and IL-21 were treated with an IFN-γ neutralizing antibody (10 μg/mL) or with an isotype-matched control IgG. PBMs were cultured overnight in the NK cell supernatants and then used in an ADCC assay against cetuximab-coated MDA-MB-468 cells. D, PBMs cultured in NK cell supernatants were assessed for IL-12 secretion after the addition of cetuximab-coated MDA-MB-468 cells. Columns, mean of one of three representative experiments; bars, SD.
Supernatants from NK cells costimulated with cetuximab-coated tumor and IL-2, IL-12, or IL-21 induce chemotaxis of naive and activated T cells. We next wanted to determine whether NK cells stimulated with cetuximab-coated tumor cells could secrete factors that could induce T-cell migration. NK cells were cocultured with cetuximab-coated MDA-MB-468 tumor cells in the presence of IL-21. After 24 h, cell-free culture supernatants were harvested and used in a human cytokine antibody array. Costimulated NK cells were found to secrete elevated levels of IL-8, RANTES, macrophage inflammatory protein-α, and macrophage inflammatory protein-1β, compared with NK cells from control conditions (P < 0.05; Fig. 5A ). Elevated secretion of these factors was confirmed by ELISA (data not shown). Enhanced secretion of the same chemokines was observed when NK cells were cocultured with cetuximab-coated MDA-MB-468 cells and IL-2 or IL-12 (data not shown). We next wanted to determine whether the chemokines that NK cells secreted in response to cetuximab-coated tumor cells could induce T-cell chemotaxis. Naive or activated T cells were placed in the upper chambers of Transwell inserts. The lower chambers contained NK cell culture supernatants from a tumor coculture experiment that used the MDA-MB-468 cell line. Culture supernatants derived from NK cells stimulated with cetuximab-coated tumor cells or with IL-2, IL-12, or IL-21 induced the chemotaxis of both naive and activated T cells (Fig. 5B). Enhanced T-cell chemotaxis was observed in response to supernatants derived from NK cells that had been costimulated with cetuximab-coated tumor cells and IL-2, IL-12, or IL-21 (P = 0.003).
Supernatants from NK cells costimulated with cetuximab-coated tumor cells plus IL-2, IL-12, or IL-21 induce chemotaxis of naive and activated T cells. A, NK cells were cocultured with cetuximab-treated MDA-MB-468 cells in medium containing 10 ng/mL IL-21 (Cetux + IL-21). Control conditions included NK cells cultured with control IgG-treated tumor cells in unsupplemented media, NK cells cultured with IgG-treated tumor cells in media containing IL-21, and NK cells cultured with cetuximab-coated tumor cells in unsupplemented medium. Culture supernatants were harvested at 24 h and used in a human cytokine antibody array (see Materials and Methods). The fold induction of the various chemokines was determined by densitometry based on internal positive controls, with chemokine levels in unstimulated NK cells set to 1.0. Up-regulation of the same chemokines was observed when IL-2 or IL-12 was used as the costimulus. B, naive or activated T cells from a single donor were assayed for chemotaxis in response to culture supernatants derived from NK cells costimulated with cetuximab-coated tumor plus IL-2, IL-12, or IL-21. Supernatants from control-stimulated NK cells were also used. Results are representative of three independent experiments. Columns, mean of three replicates; bars, SD.
IL-21 enhances the effect of cetuximab in a murine tumor model. To determine whether IL-21 could enhance the effect of a cetuximab in vivo, nude mice bearing A549 xenografts were treated thrice a week with PBS, murine IL-21 (10 μg per mouse), cetuximab (1 mg/kg), or the combination. Based on a longitudinal model using log-transformed values, no significant differences in tumor volume were found between the four groups at baseline. Furthermore, at day 30 of treatment, the average tumor volumes of mice receiving either IL-21 or cetuximab alone were not significantly smaller than those of the PBS-treated mice (P = 0.361 for cetuximab versus PBS treatment and P = 0.097 for IL-21 versus PBS treatment; Fig. 6 ). However, the average tumor volumes for mice receiving the combination of cetuximab and IL-21 were significantly less than mice receiving PBS (P ≤ 0.001), IL-21 alone (P = 0.011), or cetuximab alone (P = 0.002). Because the only cell populations to express the IL-21R are T cells, B cells, and NK cells, and nude mice lack the B-cell and T-cell populations, NK cells can be presumed to be the effectors in this model (16). These data show that IL-21 can enhance the effects of cetuximab in a solid tumor model.
IL-21 enhances tumor regression in mice treated with cetuximab. BALB/c nude mice bearing s.c. A549 xenografts were treated thrice per week i.p. with PBS, cetuximab (1 mg/kg), IL-21 (10 μg per mouse), or the combination (same doses). Tumor dimensions were measured, and tumor volumes were calculated as described in Materials and Methods. Points, mean for n = 5 mice per group; bars, SE.
Discussion
In the current report, we show that cetuximab, a therapeutic antibody specific for HER1, serves as a potent stimulus for NK cell FcR-mediated effector functions, including IFN-γ production and ADCC. Furthermore, NK cell IFN-γ production and ADCC in response to cetuximab was markedly enhanced in the presence of immune modulatory cytokines, such as IL-2, IL-12, or IL-21. IFN-γ secreted by costimulated NK cells enhanced monocyte ADCC and cytokine production in response to cetuximab-coated tumor cells, and costimulated NK cells also secreted a number of factors that could induce the chemotaxis of naive and activated T cells. Administration of IL-21 enhanced the effects of cetuximab in a murine xenograft model. These findings are the first to examine NK cell IFN-γ and chemokine production in response to cetuximab and show that these activities can be enhanced in the presence of a costimulatory signal.
Previous studies have suggested that FcR-mediated effector functions might contribute to the antitumor actions of cetuximab. Masui et al. compared the abilities of cetuximab and the F(ab)′2 fragment of cetuximab to inhibit the growth of HER1-positive xenografts in vivo (17). Although both antibodies were able to inhibit tumor growth, the antitumor activity of the unmodified cetuximab antibody was stronger than that of the F(ab)′2 fragment. Because both molecules exerted similar effects on HER1-mediated signaling within the tumor cells, the authors concluded that ADCC might contribute to the mechanism of action of cetuximab. Ciardiello et al. have generated a cetuximab-resistant colon carcinoma cell line. Although this cell line was insensitive to the antiproliferative effects of cetuximab in vitro, a modest inhibition of tumor growth was observed when cetuximab was given mice to bearing xenografts of the cetuximab-resistant cell line (18). These results suggest that both FcR-dependent and FcR-independent mechanisms contribute to the antitumor effects of cetuximab. The ability of IL-21 to augment the actions of cetuximab in a murine xenograft model supports this contention. Furthermore, in a recent report, IL-2–activated NK cells, but not monocytes, could mediate ADCC against cetuximab-coated lung cancer cells (19). Our work shows that monocytes can be induced to mediate high levels of ADCC against cetuximab-coated targets when primed with NK cell–derived IFN-γ. IFN-γ produced by NK cells could also exert direct antiproliferative effects against tumor cells via a STAT1-dependent pathway (20), and NK cell–derived chemokines could promote T-cell infiltration of tumor deposits.
In a recent report, cetuximab was unable to inhibit proliferation or induce apoptosis of non–small cell lung cancer lines expressing common HER1 kinase domain mutations, whereas non–small cell lung cancer lines expressing wild-type HER1 were sensitive to the direct effects of cetuximab (21). Although data is limited, it is possible that HER1 kinase domain mutations might also confer resistance to cetuximab in human cancer patients. Also, Lièvre et al. have reported that failure to respond to cetuximab was associated with mutations in K-ras that lead to constitutive mitogen-activated protein kinase signaling (22). Cetuximab should still be able to bind to cell lines with mutations of the HER1 kinase domain and to tumor cells with K-ras mutations, because the extracellular portion of HER1 is unaffected by these alterations (21). Therefore, expression of these mutations should not affect the FcR-mediated elimination of tumor cells after administration of cetuximab plus cytokine.
Recently, it has been shown that high HER1 expression levels correlated with cetuximab resistance (4, 23). This result is controversial, however, as other groups have found no association between HER1 expression level and clinical outcome (24, 25). In the present study, treatment of cancer cells with increasing amounts of immobilized cetuximab led to increased NK cell production of IFN-γ. Similarly, cotreatment of HER1/HER2-positive cancer cells with both cetuximab and trastuzumab led to superior activation of NK cells, presumably due to the increased availability of FcR moieties. However, low amounts of cetuximab remained relatively effective in this model: simultaneous stimulation of NK cells with cetuximab-coated tumor cells and cytokine led to significant NK cell ADCC and IFN-γ production even when cell lines with low HER1 levels were used. These results suggest that coadministration of NK cell–activating cytokines with cetuximab would be effective regardless of the expression levels of the HER1 antigen.
It is evident from the experimental data presented in Fig. 1 (C-E) that individuals vary significantly in their NK cell response to costimulation with mAbs and cytokines. In our previous studies, NK cell activation correlated with a favorable clinical outcome in a phase I clinical trial of trastuzumab and IL-12 with paclitaxel (12). Elevated NK cell production of IFN-γ was detected in each of the 11 (of 21) patients that exhibited stable disease or a clinical response while on this regimen. These observations suggested that the addition of IL-12 to the trastuzumab/paclitaxel regimen could induce an antitumor immune response in a subset of patients that correlated with reduced tumor growth. The variable response to this treatment regimen could be related to genetic factors that govern IFN-γ production. The IFN-γ gene contains an A→T polymorphism at position 874 of intron 1 that is associated with increased IFN-γ production (26). IFN-γ production in response to IL-21 is also affected by a polymorphism in the IL-21R gene (27). In addition, investigators have identified a dimorphism in the gene encoding FcγRIIIa (Phe→Val at position 158) that influences the binding affinity of the receptor (28). Higher levels of NK cell ADCC and IFN-γ production are observed in donors expressing the high-affinity FcγRIIIa (29), and clinical outcome in patients receiving rituximab has been shown to correlate with expression of the high-affinity FcγRIIIa variant (30). It is therefore possible that genetic factors that influence IFN-γ production and FcR affinity for IgG might play a role in determining the clinical outcome of mAb therapy when cytokine adjuvants are used. Further research is needed to ascertain the predictive power of these polymorphisms.
In summary, stimulation of NK cells with IL-2, IL-12, or IL-21 enhanced the NK cell response to cetuximab-coated tumor cells. These results suggest that the administration of exogenous cytokines could be an effective means of enhancing the immune response in patients receiving cetuximab.
Footnotes
-
Grant support: NIH grants P01 CA95426 and T32 CA090223 (J. Butchar).
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Accepted July 25, 2007.
- Received April 12, 2007.
- Revision received June 27, 2007.