TNF antagonists may offer therapeutic potential in solid tumors, but patients who have high serum levels of TNF-α fail to respond to infliximab, suggesting consumption of the circulating antibody and loss of transmembrane TNF-α (tmTNF-α) on tumors by ectodomain shedding. Addressing this possibility, we developed a monoclonal antibody (mAb) that binds both full-length tmTNF-α and its N-terminal truncated fragment on the membrane after tmTNF-α processing but does not cross-react with soluble TNF-α. We documented high levels of tmTNF-α expression in primary breast cancers, lower levels in atypical hyperplasia or hyperplasia, but undetectable levels in normal breast tissue, consistent with the notion that tmTNF-α is a potential therapeutic target. Evaluations in vitro and in vivo further supported this assertion. tmTNF-α mAb triggered antibody-dependent cell-mediated cytotoxicity against tmTNF-α–expressing cells but not to tmTNF-α–negative cells. In tumor-bearing mice, tmTNF-α mAb delayed tumor growth, eliciting complete tumor regressions in some mice. Moreover, tmTNF-α mAb inhibited metastasis and expression of CD44v6, a prometastatic molecule. However, the antibody did not activate tmTNF-α–mediated reverse signaling, which facilitates tumor survival and resistance to apoptosis, but instead inhibited NF-κB activation and Bcl-2 expression by decreasing tmTNF-α–positive cells. Overall, our results established that tmTNF-α mAb exerts effective antitumor activities and offers a promising candidate to treat tmTNF-α–positive tumors, particularly in patients that are nonresponders to TNF antagonists. Cancer Res; 73(13); 4061–74. ©2013 AACR.
TNF-α exists in 2 bioactive forms: a 26-kD transmembrane form and a 17-kD soluble form that is released after proteolytic cleavage. Both forms of TNF-α display bioactivities via TNF receptors, TNFR I and TNFRII. TNF-α is a 2-edged sword for tumors. On the one hand, there is evidence that it induces tumor necrosis and apoptosis as its name implies. On the other hand, TNF-α is also reported to promote tumor development. Although our and others' previous studies reveal that the ectopic expression of TNF-α at the site of malignancy induces strong and long-term tumor regression (1, 2), increasing evidence indicates that TNF-α functions in promotion and progression of tumors, rather than in protection against tumors, including proliferation, transformation, angiogenesis, invasion, and metastasis in many cancers (3). Knockout of TNF-α or TNFR1 creates mice resistant to chemical carcinogenesis of the skin or liver, respectively (4, 5). Conversely, pretreatment with TNF-α enhances lung metastases (6), and overexpression of TNF-α confers invasive properties on xenograft tumors (7). In the clinic, TNF-α production by tumors is related with a poor prognosis, loss of hormone responsiveness, and cachexia (8, 9). Most studies on TNF-α and tumors refer to the soluble TNF-α (sTNF-α), a typical proinflammatory cytokine; thus, sTNF-α is considered to bridge inflammation and cancer.
Although sTNF-α originates from the extracellular sequence of transmembrane TNF-α (tmTNF-α), the bioactivities of both forms of TNF-α are not quite the same. Our previous study shows that tmTNF-α is able to kill some tumor cell lines that are resistant to sTNF-mediated cytotoxicity (10). In contrast, tmTNF-α expressed by tumor cells protects them from apoptosis by inducing constitutive activation of NF-κB via its reverse signaling, as the ectopic expression of the leader sequence of tmTNF-α [namely deletion of the sTNF-α sequence, retaining N-terminal truncated fragment (NTF) including the intracellular part of tmTNF-α] is sufficient to activate NF-κB and render tumor cells resistant to apoptosis induced by sTNF-α (11, 12). Conversely, interruption of the reverse signaling mediated by tmTNF-α increases the sensitivity of tumor cells to sTNF-α–induced cytotoxicity (12). These data suggest that tmTNF-α may influence the opposite actions of sTNF-α, anti- or protumor. Furthermore, it may be a good tumor surface target for immunotherapy.
TNF-α antagonists such as etanercept, infliximab, and adalimumab are used to treat a variety of autoimmune diseases, such as rheumatoid arthritis (13) and Crohn's disease (14). Because TNF-α is believed to be a target for solid tumor therapy, the efficiency of TNF antagonists has been tested in phase I and II clinical cancer trials as single agents. Although some patients with advanced cancer indeed benefit from TNF-α antagonist therapy (15–17), the patients who have high serum levels of TNF-α failed to respond to the antibody infliximab (16). One of the reasons is the consumption of antibody in the circulation by soluble serum TNF-α so that the antibody reaching the tumor site decreases. Even though the antibody can reach the tumor, its tumoricidal effect must be affected, owing to its ability to bind local sTNF-α (18) that leads to the competitive inhibition of the antibody targeting the tmTNF-α on the surface of tumor cells. Another reason is that tumor cells lose tmTNF-α because of increased ectodomain shedding by overexpressed ADAM 17 in cancers (19). In the present study, we made a monoclonal antibody (mAb) that not only binds to full-length tmTNF-α, but also to its still-anchored membrane NTF after tmTNF-α processing without cross-reaction to sTNF-α. This antibody was shown to have effective antitumor activities, in vitro and in vivo.
Materials and Methods
Generation and purification of mAbs against tmTNF-α
We obtained the hybridoma-secreting tmTNF-α mAb using conventional hybridoma technology. Briefly, a fragment from the signal peptide of tmTNF-α containing an epitope served as an immunogen. The peptide is conjugated to keyhole limpet hemocyanin (KLH) or to bovine serum albumin (BSA). BALB/c mice were immunized 4 times with the peptide–KLH complex. The hybridoma cells were obtained by fusion of the resulting splenocytes and SP2/0-Ag14 murine myeloma, in the presence of polyethylene glycol (PEG) 4000 (Sigma), selected in HAT medium (hypoxanthine-aminopterin-thymidine medium; Sigma) and then screened by specific binding ELISA. The tmTNF-α mAb-secreting hydridoma cells were injected intraperitoneally into BALB/c mice for large-scale production of mAb. The mAb was purified by HiTrap protein G HP (Amersham Biosciences).
ELISA for screening the hydridoma plus analysis of isotype and specificity of tmTNF-α mAb produced
In the indirect ELISA, the peptide–BSA complex was used to coat ELISA plates in a concentration of 0.5 μg/mL. After overnight coating, these plates were blocked with PBS solution containing 1% BSA, followed by addition of hybridoma culture supernatants for a 1-hour incubation. After washing, antimurine immunoglobulin conjugated with horseradish peroxidase (HRP) was added into each well. After another 1-hour incubation, the plates were washed and we added the substrate O-phenylenediamine dihydrochloride (Sigma). Absorbance at a wavelength of 450 nm was read in a microplate reader (Bio-Rad).
We detected the isotypes of tmTNF-α mAbs by a commercial Mouse Immunoglobulin Isotype Kit (Invitrogen), according to the manufacturer's instructions.
For the specificity analysis of tmTNF-α mAb, sandwich-ELISA was adopted using tmTNF-α mAb as a capture antibody and an HRP-labeled mAb to TNF-α that binds to another epitope (Shanghai Yaji Biotechnology Co. Ltd.) as a detecting antibody. The peptide–BSA complex, interleukin (IL)-2, BSA, and IFN-γ (Sigma) were used as antigens.
For the inhibitory competitive assay, 200 μg of either a control peptide (acetylcholine receptor 97–116) or a tmTNF peptide was used to coat the plates. Purified tmTNF-α mAb was incubated for 30 minutes at 4°C with the tmTNF peptide–BSA complex or sTNF-α in the indicated concentrations, before addition to the plates. The secondary antibody was HRP-conjugated antimurine immunoglobulin.
Noncompetitive enzyme immunoassay for determination of mAb affinity
The affinity constant (Kaff) of the antibody was measured by ELISA as described by Beatty and colleagues (20). A microtiter plate was first coated with peptide–BSA in different concentrations (5, 2.5, 1.25, 0.65 mg/L) and then incubated with the indicated concentrations of tmTNF-α mAb. The plate was sequentially incubated with HRP-conjugated antimurine immunoglobulin, followed by the substrate as mentioned earlier. Kaff was calculated by the following formula, where n = Ag/Ag′: Kaff = (n − 1)/2(nAb′ − Ab).
We purchased the human breast carcinoma cell lines MCF-7 and MDA-MB-231 from American Type Culture Collection, and a human malignant B cell line derived from Burkitt lymphoma, Raji cell, from China Center for Type Culture Collection, Wuhan University (Wuhan, Hubei, China). The breast cancer cell lines T47D, MDA-MB-435, and MDA-MB-453 were a gift from Prof. Jian Lu (Department of Pathophysiology, Second Military Medical University, Shanghai, China). Cells were cultured at 37°C in 5% CO2 in RPMI-1640 medium (GIBCO) supplemented with 10% heat-inactivated (56°C, 30 minutes), pyrogen-free fetal calf serum (FCS; Sijiqing), 1.0 mmol/L sodium pyruvate, 2.0 mmol/L l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, and 5 × 10−5 mol/L 2-mercaptoethanol (2-ME). Fresh vials of cells were periodically thawed and used in experiments to ensure that cell lines were unchanged in culture during the study.
Peripheral blood mononuclear cells (PBMC) from healthy volunteers were prepared by Ficoll-Hypaque density gradient centrifugation. Human monocytes and peritoneal macrophages from mice were purified by their adherence of cells onto 24- or 96-well plates (Costar; Thomas Scientific) in 10% FCS RPMI-1640 medium for 1 hour, and then washed to remove the nonadherent cells. tmTNF-α expression on the surface of the monocytes and macrophages was induced by incubation with 100 ng/mL lipopolysaccharide (LPS) for 16 hours.
The plasmids pcDNA3.0, containing wild-type (wt) TNF-α or TNF-NTF(LS) were constructed in our laboratory as described previously (21). For stable transfection, we generated pIRES2-EGFP/wtTNF-α or pIRES2-EGFP/TNF-NTF plasmids containing human cDNA that codes for wtTNF-α or TNF-NTF, respectively, by inserting the corresponding PCR product into the vector pIRES2-EGFP. This PCR product was amplified from TNF-NTF/pcDNA3.0 or wtTNF-α/pcDNA 3.0, using 2 pairs of primers (TNF-NTF: forward TAAGGAATTCATGAGCACTGAAAGCATGATCCGGGA and reverse ATATGGATCCTATGCCTGGGCCAGAGGGCTGATTAGA; wtTNF: forward TAAGGAATTCATGAGCACTGAAAGCATGATCC and reverse GTAGGATCCTCACAGG GCAATGATCCCAAAG) for 30 cycles (94°C for 20 seconds, 58°C for 20 seconds, 72°C for 30 seconds). The gene-specific short hairpin RNA (shRNA) and control shRNA (“HuSH 29mer shRNA constructs against TNF in pGFP-V-RS vector”) were obtained from OriGene.
A total of 3 × 105 parental MCF-7 or MDA-MB-231 cells were cultured in 6-well culture plates until 70% confluent. Two microgram of recombinant plasmid was transfected into cells using the Lipofectamine 2000 transfection reagent (Invitrogen), according to the manufacturer's instructions. For stable transfection, selection was conducted in 800 μg/mL G418 for 2 weeks and the bulk-transduced cells were subcloned by limiting dilution at 0.3 cells/100 μL/well in 96-well microtiter plates and maintained thereafter in 400 μg/mL G418.
Binding analysis of tmTNF-α mAb for the expression of tmTNF-α on the surface of human breast cancer cell lines, human monocytes or murine peritoneal macrophages was conducted by flow cytometry. Cells were detached with trypsin–EDTA (Invitrogen) and resuspended in PBS containing 1% BSA and 0.1% sodium azide. The samples were incubated for 1 hour on ice with tmTNF-α mAb, anti-sTNF-α mAb (Becton Dickinson), antimurine TNF-α antibody (R&D Systems), or their corresponding isotype antibodies as negative controls. After washing with PBS, cell staining was carried out by another 1-hour incubation with fluorescein isothiocyanate (FITC)-labeled antimurine or anti-rabbit immunoglobulin G (IgG; Jackson ImmunoResearch Laboratories). Stained cells were analyzed on a FACSCalibur 440E (Becton Dickinson) using Cell Quest software (BD Biosciences Immunocytometry Systems).
Immunofluorescence and confocal microscopy
A total of 1 × 105 transfected cells were seeded onto coverslips placed in a 12-well culture plate overnight. The cells were washed twice with cold PBS and blocked with 5% nonfat dry milk in PBS for 1 hour at room temperature. Next, these cells were incubated overnight at 4°C with tmTNF-α mAb. After further washes with PBS, the secondary antibody phycoerythrin (PE)-conjugated anti-mouse IgG was applied. The cells were then observed under a confocal microscope FU5000 (Olympus).
ADCC and CDC assays
For antibody-dependent cell-mediated cytotoxicity (ADCC), peritoneal macrophages from mice served as effector cells. The target cells were MDA-MB-231 and MCF-7 breast carcinoma cells including wtTNF-α- or NTF-transfected MCF-7 cells. Both effector and target cells were incubated together for 24 hours at an effector/target ratio of 20:1. The target cells were incubated with either tmTNF-α mAb or a mAb to the surface antigen of hepatitis B virus (HBVs mAb, as an isotype control, kindly gifted by Prof. Dongliang Yang, Department of Infection, Union Hospital, Huazhong University of Science and Technology, Wuhan, Hubei, China) for 30 minutes before the addition of effector cells.
For complement-dependent cytotoxicity (CDC), target cells were plated at a density of approximately 5,000 cells per well in the presence of the above 2 antibodies and 5% fresh guinea pig serum with active complement, and then incubated at 37°C for 4 hours. Cell viability was detected by incubation with CCK-8 (Dojindo) for 4 hours and then the absorbance (A) was measured at 450 nm, with a reference wavelength at 630 nm. Assays were conducted in triplicate. The percentage of cytotoxicity for each test type was calculated as follows:
Xenotransplantation of MDA-MB-231 cells into nude mice
Our animal studies were approved by the Animal Care and Use Committee of Huazhong University of Science and Technology. Six-week-old female BALB/c nude mice (weight 16–20 g; Shanghai SLAC Laboratory Animal Co., Ltd.) were housed in a specific pathogen-free environment with a 12-hour light–dark cycle and allowed ad libitum access to food and water. A total of 2 × 106 human MDA-MB-231 cells were inoculated into the right mammary fat pads of these mice (7 per group). Tumor size was measured every 5 days with microcalipers and it was calculated by assuming the following equation (22): length × width2 × π/6.
tmTNF-α mAb (120 or 240 μg per mouse) was injected intraperitoneally every 3 days. The treatment began at day 14 after transplantation and lasted 6 weeks. As a control, 120 μg of HBVs mAb or the same volume of normal saline was inoculated intraperitoneally.
Total protein preparation and Western blot analysis
Total protein was extracted from cultured cells or frozen tumor tissue in ice-cold lysis buffer (50 mmol/L Tris–HCl pH 7.5, 150 mmol/L NaCl, 5 mmol/L EDTA, 1% Nonidet P-40,) containing a protease inhibitor cocktail (Calbiochem) on ice for 15 minutes. Twenty microgram protein was fractionated by 12% SDS-PAGE and then transferred from the gel onto a polyvinylidene difluoride (PVDF) membrane. After blocking with 5% nonfat milk in PBS–Tween-20 (0.05%) overnight at 4°C, the membrane was then probed with antibodies specific to p-p65, Bcl-2, caspase-8, or β-actin (Santa Cruz Biotechnology), followed by a HRP-conjugated secondary antibody against rabbit/mouse IgG. The enhanced chemiluminescence (ECL) from NEN LIFE Science was used to visualize the antibody reaction. Bands were quantified by a calibrated imaging densitometer (GS-710; Bio-Rad) and analyzed by “Quantity One” software (Bio-Rad).
For identification of the degree of purification of tmTNF-α mAb, 20 μg of purified tmTNF-α mAb was subjected to electrophoresis on 12.5% SDS-PAGE and then stained with 0.05% Coomassie brilliant blue.
Immunohistochemistry and TUNEL staining
Diseased breast tissues were collected from patients in Tongji Hospital, Huazhong University of Science and Technology, including 39 patients with breast cancer, 15 patients with atypical dysplasia, 10 patients with hyperplasia, and 5 normal breast tissues that were adjacent to breast tumors. Archival hematoxylin and eosin (H&E)–stained slides for each case were reviewed by a pathologist. This study was approved by the Clinical Research Committee of Tongji Medical College (Wuhan, Hubei, China).
Sections of 4-μm thick, formalin-fixed and paraffin-embedded tissues from the patients or animal experiments were deparaffinized in xylene and rehydrated in graded ethanol. Following heat-mediated antigen retrieval, tmTNF-α, phosphorylated NF-κB p65, Bcl-2, or CD44v6 were detected in the diseased tissue by the avidin–biotin complex method. The sections were incubated for 1 hour with the corresponding specific primary antibodies (Santa Cruz Biotechnology), followed by washing, another 1-hour incubation with biotin-labeled antimurine IgG antibody (Boyao Biotechnology), washing again, and then incubation with peroxidase-labeled streptavidin for 20 minutes. Immunostaining was visualized by color reaction to diaminobenzidine for 5 minutes. As a negative control, isotype antibody was substituted for the primary antibody. Then, the percentages of brown-stained cells indicating the presence of tmTNF-α, phosphorylated NF-κB p65, Bcl-2, or CD44v6 were determined. The positive cell ratio (integral optical density value/integral area) was calculated by Tongji Qianping Image Analysis Software.
Tumor invasion area in the lymph node was measured at ×200 magnification in 10 fields for each sample, namely the area occupied by tmTNF-α–positive tumor cells was analyzed using the Image-Pro Plus software version 6.1 (Media Cybernetics) and expressed as a percentage of total lymph node area.
To evaluate apoptosis in tumors that was induced by treatment with tmTNF-α mAb, the terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay was carried out with a commercial kit (Beyotime Biotechnology) according to the manufacturer's protocol. Nuclei were counterstained with hematoxylin. TUNEL-stained cells were counted at ×400 magnification, in 5 high-power fields, by the Image-Pro Plus software version 6.1. The apoptotic index (AI) was calculated as follows: AI = (number of positive cells/total number of cells) × 100%.
Paraffin-embedded sections (4 μm) of tumors were stained with H&E and evaluated under a microscope in a blinded manner. Tumor emboli in the blood vessels or lymphatic vessels were counted at ×200 magnification in 10 fields for each sample and represented as the average number of tumor emboli in 5 fields.
A total of 5 × 103 cells were seeded in a 96-well plate and incubated overnight at 37°C, in 5% CO2. The following day, 200 U/mL sTNF-α (PeproTech) was added and incubated for 24 hours. Cell viability was then detected by staining for 4 hours with CCK-8. The photometric measurement was conducted at 450 nm, with a reference wavelength set at 630 nm. TNF-α–induced cytotoxicity was calculated by using the following formula: cell death rate (%) = (1 − ODsample/ODcontrol) × 100%.
Data are represented as the mean ± SEM. The differences were analyzed using one-way ANOVA test. A P value less than 0.05 was considered statistically significant.
Identification of mAb to tmTNF-α
To make a mAb to tmTNF-α, the extracellular part of the signal peptide of the molecule was synthesized and conjugated to the carrier KLH, and then used to immunize BALB/c mice. The hybridoma cells were made by fusion of the resulting murine splenocytes and myeloma cells. After screening, we obtained several mAbs to tmTNF-α, which were identified as IgG1 and κ light chain. We purified one of them by HiTrap protein G HP and proved purification by SDS-PAGE (Fig. 1A). The affinity constant of the mAb was 1.00 × 109, as determined by ELISA noncompetitive assay (Fig. 1B). The specificity of the mAb was detected by binding ELISA and inhibitory competitive ELISA. Results showed that the antibody only bound to tmTNF-α peptide conjugated with BSA, but did not cross-react to BSA, IL-2, or IFN-γ (Fig. 1C). We showed, by preincubation of the antibody with tmTNF-α peptide conjugated with BSA or sTNF-α in different concentrations, that there was specific binding of tmTNF-α mAb with the tmTNF-α peptide, but not with control peptide, as when coated on the plate it was significantly competitively inhibited by the peptide–BSA complex (>8 ng; Fig. 1D), and yet, it was unaffected by sTNF-α (Fig. 1E). This indicated that the specificity of the antibody is directed to tmTNF-α, instead of sTNF-α. To detect whether the antibody could bind to native tmTNF-α on the surface of cells, we used LPS to activate human monocytes to express tmTNF-α and a human B lymphoma cell line, Raji, which was previously confirmed as expressing tmTNF-α at a high level (12). Indeed, flow cytometry showed that tmTNF-α mAb, like the antibody to sTNF-α, could bind specifically to the human tmTNF-α on the surface of Raji cells (Fig. 1F) and LPS-activated monocytes (Fig. 1G), but it could not bind to murine tmTNF-α on the surface of LPS-activated peritoneal macrophages (Fig. 1G).
Expression of tmTNF-α on clinical tissue samples of breast cancer and hyperplasia of mammary gland, and on breast cancer cell lines
We used this mAb to detect tmTNF-α expression on tissue sections from patients with hyperplasia, atypical hyperplasia, or primary breast cancer. We found that tmTNF-α staining was strong positive on breast cancer, and weaker on atypical hyperplasia and hyperplasia, but it was negative in normal breast tissue (Fig. 2A and B). The positive rate was 92.3% for breast cancer, 80% for atypical hyperplasia, and 33.3% for hyperplasia, indicating that the expression of tmTNF-α becomes increased from benign breast mass to malignant breast neoplasm. Furthermore, we also detected tmTNF-α expression on 5 breast cancer cell lines. As is shown in Fig. 2C, tmTNF-α expression was at higher levels in MDA-MD-435 and MDA-MD-231 cells (52.94% and 89.14%, respectively), at lower levels in T47D and MDA-MD-453 cells, but undetectable in MCF-7 cells. Therefore, we used the MDA-MD-231 cell line as the target cells for tmTNF-α mAb in in vitro and in vivo experiments. To further confirm the specific binding of the mAb to the full length of tmTNF-α or to its NTF that is still bound to the surface of cells after ectodomain shedding of tmTNF-α, we stably transfected tmTNF-positive MDA-MD-231 or tmTNF-negative MCF-7 cells with shRNA or wtTNF-α and its NTF, respectively. We confirmed by fluorescence-activated cell sorting (FACS) that the mAb could detect the downregulation of tmTNF-α expression on MDA-MD-231 cells transfected by shRNA (Fig. 2D) and the upregulation of tmTNF-α expression on MCF-7 cells that were transfected with wtTNF-α (Fig. 2E). Similarly, the results of confocal microscopy using this mAb showed that tmTNF-α staining on the MDA-MD-231 cell line became weakened because of shRNA (Fig. 2F), whereas tmTNF-α staining on the MCF-7 cells became positive, owing to transfection of wtTNF-α (Fig. 2G). In addition to detection of the full-length tmTNF-α, this mAb could also recognize NTF at the surface of MCF-7 transfectants, as shown by FACS (Fig. 2E) and confocal microscopy (Fig. 2G).
tmTNF-α mAb mediated ADCC and CDC toward tmTNF-α–expressing breast cancer cell line
Next, we explored the ADCC and CDC effects of our mAb to tmTNF-α on MDA-MD-231 target cells. As shown in Fig. 3A, the antibody effectively mediated the cytotoxicity of macrophages toward tmTNF-α–expressing breast cancer cells in a dose-dependent manner, with its effect reaching a peak at a concentration of 2 μg/mL (P < 0.001). In contrast, this antibody failed to mediate ADCC to a tmTNF-α–negative cell line, MCF-7 (Fig. 3B). In addition, a mAb to HBV surface antigen (HBVs) had no effect on the tmTNF-α–positive and -negative breast cancer cell lines (Fig. 3B).
Similarly, tmTNF-α mAb also exerted CDC toward the tmTNF-α–expressing MDA-MD-231 cell line, but not to the tmTNF-α–negative MCF-7, in a dose-dependent manner, exhibiting maximal effect at 2 μg/mL in the presence of complement (Fig. 3C). In comparison with the effective CDC of tmTNF-α antibody, the antibody to HBVs had no CDC observed toward either the tmTNF-α–positive or -negative breast cancer cell lines, pointing out the specificity of action by that tmTNF-α antibody (Fig. 3D).
Furthermore, we examined whether the mAb mediates ADCC and CDC toward NTF-carrying tumor cells. As is shown in Fig. 3E and F, the ectopic expression of either wtTNF-α or NTF rendered the MCF-7 cells sensitive to tmTNF-α antibody-mediated ADCC and CDC, indicating that this mAb also has functionality after tmTNF-α processing.
In vivo antitumor activity of tmTNF-α mAb on tmTNF-α–expressing human breast cancer in a mouse model
To observe the antitumor effect of the mAb in vivo on tmTNF-α–expressing breast cancer, we first inoculated MDA-MD-231 cancer cells into nude mice. The tumor formation rate was about 95%. After 2 weeks, the tumors formed grew to a visible size (∼0.05 cm3) and we then treated the animals with the mAb every 3 days for 6 weeks. We showed that tmTNF-α mAb significantly inhibited tumor growth (Fig. 4A; P < 0.001) and that tumor size became markedly smaller (Fig. 4B; P < 0.01), as compared with the controls. Furthermore, tmTNF-α mAb induced complete tumor regression in 3 mice. Although the inhibitory rate of tumor growth was shown to be 70% or more by treatment with tmTNF-α mAb in either of 2 doses (Fig. 4C), the effect of the antibody in the lower dose seemed to be better. In contrast, treatment with HBVs mAb had no influence on tumors. Moreover, we found that tmTNF-α mAb treatment mainly induced tumor apoptosis, as there were many brown-stained apoptotic cells by TUNEL (Fig. 4D) and obvious activation of caspase-8 (Fig. 4E) in tumors treated with tmTNF-α mAb, as compared with tumors treated with saline or HBVs mAb controls.
Histologic examination revealed that in addition to the pathologic active karyokinesis aspect, tumor emboli in blood or lymphatic vessels were obviously observed in control tumors, whereas fibrosis and inflammatory cell infiltration were evidently seen in the tumors treated with tmTNF-α mAb, accompanied with a significantly decreased number of tumor emboli in blood and lymphatic vessels (Fig. 5, top). Furthermore, treatment with tmTNF-α mAb suppressed tumor lymph node metastasis as manifested by a clearly decreased invasion area in lymph nodes and a declined rate of lymph node metastasis, as compared with those in control tumors (Fig. 5, middle). The expression of CD44v6, a molecule associated with tumor metastasis, also declined following treatment with tmTNF-α mAb, but not with HBVs mAb or saline (Fig. 5, bottom). The data indicated that tmTNF-α mAb had apparent antitumor effects in vivo.
tmTNF-α mAb markedly suppressed tmTNF-α–mediated reverse signaling
Previously, we proved that tmTNF-α–expressing tumor cells are resistant to sTNF-α–mediated apoptosis, via activation of NF-κB through its reverse signaling (11, 12). Therefore, we checked whether tmTNF-α mAb affects tmTNF-α–mediated reverse signaling, although the epitope recognized by this antibody is not located at the receptor-binding site of tmTNF-α. As expected, our results showed that treatment with tmTNF-α mAb did not activate tmTNF-α reverse signaling, instead, it inhibited this pathway in vivo. As shown in Fig. 6, tmTNF-α mAb treatment resulted in a significant reduction of tmTNF-α staining (top) that was accompanied with a marked suppression of phosphorylation of NF-κB p65 (middle) and its target gene Bcl-2 expression (bottom) in the tumor tissue, as compared with those in the tumors treated with HBVs mAb or saline.
Deletion of tmTNF-α strong positive tumor cells by tmTNF-α mAb contributes to suppression of tmTNF-α reverse signaling
To determine whether tmTNF-α mAb directly influences tmTNF-α reverse signaling, we used different concentrations of the antibody to treat MDA-MD-231 cells. In contrast to the suppression of tmTNF-α reverse signaling observed in vivo, tmTNFα mAb failed to affect the levels of p-p65 and Bcl-2 in vitro, in the cell line (Fig. 7A), indicating that there was no direct effect of the antibody on tmTNF-α reverse signaling. We speculated that the deletion of tmTNF-α strong positive cancer cells in vivo might contribute to the inhibition of tmTNF-α reverse signaling. To test this hypothesis, we checked for tmTNF-α expression on MDA-MD-231 cells after the mAb-mediated ADCC. Indeed, the tmTNF-α expression was significantly reduced (Fig. 7B), similar to that on shRNA-transfected MDA-MD-231 cells (Fig. 2D). Therefore, we used stably shRNA-transfected MDA-MD-231 cells to examine changes in tmTNF-α reverse signaling. As expected, downregulation of tmTNF-α expression decreased levels of p-p65 and Bcl-2 in these cells, as compared with control (Fig. 7C), and rendered originally sTNF-resistant MDA-MD-231 cells sensitive to the cytotoxicity of sTNF-α (Fig. 7D).
The best molecular target for antibody treatment is a membrane molecule on the surface of a tumor, but not in the normal tissue, because of antibody accessibility to antigenic determinants for exerting ADCC and CDC toward the tumor. Although it is reported that sTNF-α can be released by solid tumors, including breast cancer (3, 23), there is no clinical data about tmTNF-α expression on the surface of breast cancer. Here, we show that tmTNF-α is a surface marker for a part of breast cancers, but is absent in normal mammary tissue. Interestingly, the results from our limited survey of patients showed that the degree of tmTNF-α expression was different in benign breast mass and malignant breast neoplasms, namely, the order of its expression rate and expression intensity was: breast cancer > atypical hyperplasia III (precancerous lesion) > hyperplasia (benign breast mass). It is likely that tmTNF-α expression becomes upregulated during the development from breast mass to breast cancer. Whether tmTNF-α–positive (++∼+++) atypical hyperplasia has a tendency to develop carcinoma was unclear, so it should be further investigated. Furthermore, our results indicated that the level of tmTNF-α expression on cancer cells, rather than serum TNF-α, is a good biomarker for determining the antibody-targeting therapy. In this regard, patients with tmTNF-α–positive tumors may benefit more from TNF antagonists than those with tmTNF-α–negative tumors.
As tmTNF-α can be cleft by ADAM17, releasing sTNF-α and leaving NTF (namely, the leader sequence of tmTNF-α) membrane-anchored until intramembrane cleavage occurs (21, 24), we developed a mAb that not only binds to full-length tmTNF-α, but also to NTF (Fig. 2), increasing its opportunities to attack tumor cells by targeting more molecules on the surface of breast cancer cells. Especially when tmTNF-α processing is increased because of ADAM17 overexpression, NTF becomes important for antibody targeting and efficacy. Indeed, our results showed that tmTNF-α antibody mediated ADCC and CDC not only toward tmTNF-α–positive MDA-MD-231 and wtTNF-transfected MCF-7 cells, but also to NTF-carrying transfectants of MCF-7. However, our mAb had no ADCC and CDC on tmTNF-α–negative parental MCF-7 cells. The treatment of tmTNF-α–positive tumor-bearing mice with this antibody significantly suppressed tumor growth and even led to total tumor regression in 3 mice, displaying an obvious antitumor effect. In addition, tmTNF-α mAb had no cross-reaction to sTNF-α, thus it avoids wasting of the antibody in the circulation and may diminish the treatment dosage and possible side effects.
Another possible advantage of this antibody is that its affinity constant was about 109. The avidity of the antibody is not so high that it may only attack tumor cells with a high density of tmTNF-α, but neither neutralizes sTNF-α nor influences normal cells (including immune cells) with a low density of tmTNF-α. This may leave a physiologic concentration of TNF-α non-neutralized and benefit the resistance of patients to infection and tumorigenicity.
Another feature of the best molecular targets on tumors for antibody therapy is the requirement of this target molecule for tumor growth, resistance to apoptosis, and promotion of metastasis. If the tumor loses this molecule because of cancer immunoediting, it would be unable to grow well. Our previous results show that once a tumor expresses tmTNF-α, it becomes resistant to sTNF-α–induced apoptosis. This is mediated by constitutive activation of NF-κB, through tmTNF-α reverse signaling (11, 12). Constitutive activation of NF-κB is a main cause responsible for the resistance of tumor cells to apoptotic stimuli, including chemotherapy and radiotherapy (25). In line with our previous study, these study results showed that tmTNF-α–positive breast cancer was accompanied with NF-κB activation (phosphorylation of NF-κB p65) and the expression of Bcl-2, one of its target genes encoding antiapoptotic molecules. The treatment with tmTNF-α mAb did not activate, but instead impaired reverse signaling of tmTNF-α, as NF-κB activation decreased and its targeted gene Bcl-2 was consequently downregulated. It is unlikely that tmTNF-α mAb directly affected reverse signaling by interference with the interaction between tmTNF-α and TNF receptors, because tmTNF-α exists as a compact conical-shaped trimer and the epitope that tmTNF-α mAb recognizes is near the cell membrane, under the base of this bell-shaped trimer (26) and is thereby not included in the binding site for TNFR. As expected, our in vitro results showed that tmTNF-α mAb had no effect on the levels of p-p65 and Bcl-2 in tmTNF-α–positive MDA-MD-231 cells. As decreased tmTNF-α expression in the tumor by tmTNF-α mAb treatment was accompanied with the suppression of tmTNF-α reverse signaling in vivo, we presumed that tmTNF-α mAb treatment decreased at least the number of tumor cells with high-density tmTNF-α by ADCC and CDC, so that the reverse signaling of tmTNF-α in the rest of the tumor cells with lower density of this membrane molecule was consequently crippled. This hypothesis was confirmed by evidence that tmTNF-α expression on MDA-MD-231 cells was significantly decreased after ADCC and that downregulation of tmTNF-α decreased the levels of p-p65 and Bcl-2 in the cells. Furthermore, we proved that tmTNF-α antibody induced the apoptosis of tumor cells, showing increased TUNEL-stained cells and the activation of caspase-8 in the tumors. In addition to ADCC, another possible reason for the observed increased apoptosis was inhibition of NF-κB activation through the antibody-induced downregulation of tmTNF-α, which may in turn sensitize tumor cells to apoptosis induced by sTNF-α (12) or by other proapoptotic molecules, such as FASL (27) and TRAIL (28, 29). Our evidence that downregulation of tmTNF-α by shRNA rendered originally sTNF-resistant MDA-MD-231 cells sensitive to sTNF-α–mediated cytotoxicity supports this hypothesis. It is consistently reported that the deficiency of TIMP3, a physiologic inhibitor of the TNF sheddase ADAM 17, leads to an increase in sTNF-α release (thus less tmTNF-α) and activation of caspase-8 in regressed mammary tissue, which is reversed by reconstitution with recombinant TIMP3, inducing more tmTNF-α expression as a result of the inhibition of its processing (30).
Our results showed that tmTNF-α mAb significantly suppressed blood or lymphatic metastasis of breast cancer, one of antitumor effects of this antibody. It is well known that sTNF-α is associated with tumor metastasis (3), as sTNF-α induces angiogenesis by promoting VEGF expression (31, 32) and stimulates production of chemokines and adhesion molecules (33, 34) that are involved in tumor metastasis. High serum concentration of TNF-α in patients with breast cancer correlates with invasive tumors (35). Because the tmTNF-α mAb did not neutralize sTNF-α, its inhibition of tumor metastasis may be attributed to (i) interference with constitutive NF-κB activation, reducing activation of its targeted prometastatic genes such as MMP9 and VEGF (33); and (ii) removal of tumor cells with a high density of tmTNF-α by ADCC and CDC, as those tumor cells may have high metastatic and invasive potential (further validation is needed). In addition, the CD44 molecule plays an important role in distant malignant metastasis. It exists in a standard form (CD44s) or a variant form (CD44v). CD44v6 contributes to metastasis and its expression reaches 100% in metastatic carcinoma present in the lymph nodes of patients with breast cancer (36). It has been reported that sTNF-α upregulates CD44s, CD44v3, and CD44v6 expression via the p38 mitogen-activated protein kinase (MAPK) pathway in MDA-MB-231 cells (37). Our results showed that tmTNF-α mAb decreased CD44v6 expression, suggesting that tmTNF-α may be responsible for the induction of this molecule, which is involved in the metastasis of breast cancer.
Although anti-TNF-α agents are used to treat solid tumors including breast cancer (15–17), higher circulating concentrations of TNF-α are observed in the nonresponding patients (16), suggesting that the infused antibody must be quickly bound by sTNF-α and cleared from the blood, thereby markedly reducing arrival of efficient amounts of the antibody to the tumor. Even though the antibody enters into the tumor site, a portion of these can be also consumed by sTNF-α in that microenvironment and the tumor may lose the membrane molecule because of ectodomain shedding, thus weakening the biologic efficacy of the antibody. Besides, if antibody reaching the tumor cannot effectively kill tumor cells, it may enhance reverse signaling by binding to tmTNF-α (18), which may favor tumor progression through activation of NF-κB. In contrast, tmTNF-α mAb neither binds to sTNF-α nor activates reverse signaling of tmTNF-α. In addition, our results showed that the antibody induced a decrease of tmTNF-α expression and its reverse signaling, which sensitized tumor cells to apoptosis (11). Our data indicated that the tmTNF-α mAb is a promising therapeutic antibody and may be suitable for targeted cancer therapy, although we have a long way to go before it is humanized and further evaluated for clinical purposes.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: Z. Li
Development of methodology: M. Yu, X. Zhou, L. Niu, G. Lin, H. Gan
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Yu, X. Zhou, L. Niu, G. Lin, J. Huang, H. Gan
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Yu, G. Lin, J. Huang, W. Zhou
Writing, review, and/or revision of the manuscript: M. Yu, J. Wang, Z. Li
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Yu, G. Lin, W. Zhou, J. Wang, X. Jiang, B. Yin
Study supervision: M. Yu, Z. Li
This work was supported by the Major Research Plan of the National Natural Science Foundation of China (Grant no. 91029709) and by the National Natural Science Foundation of China (Grant no. 30872376 and No. 30901308).
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.
The authors thank Kathleen Kite-Powell for editing the article.
- Received October 24, 2012.
- Revision received March 13, 2013.
- Accepted April 4, 2013.
- ©2013 American Association for Cancer Research.