• TNFα;
  • certolizumab pegol;
  • infliximab;
  • adalimumab;
  • etanercept


  1. Top of page
  2. Abstract

Background: Inhibitors of tumor necrosis factor α (TNFα) have demonstrated significant efficacy in chronic inflammatory diseases, including Crohn's disease (CD). To further elucidate the mechanisms of action of these agents, we compared the anti-TNFα agents certolizumab pegol, infliximab, adalimumab, and etanercept in several in vitro systems.

Methods: The ability of each anti-TNFα agent to neutralize soluble and membrane-bound TNFα; mediate cytotoxicity, affect apoptosis of activated human peripheral blood lymphocytes and monocytes; induce degranulation of human peripheral blood granulocytes, and modulate lipopolysaccharide (LPS)-induced interleukin (IL)-1β production by human monocytes was measured in vitro.

Results: All 4 agents neutralized soluble TNFα and bound to and neutralized membrane TNFα. Infliximab and adalimumab were comparable in their ability to mediate complement-dependent cytotoxicity and antibody-dependent cell-mediated cytotoxicity, and to increase the proportion of cells undergoing apoptosis and the level of granulocyte degranulation. Etanercept generally mediated these effects to a lesser degree, while certolizumab pegol gave similar results to the control reagents. LPS-induced IL-1β production was inhibited by certolizumab pegol, infliximab, and adalimumab, but only partially inhibited by etanercept.

Conclusions: In contrast to the other anti-TNFα agents tested, certolizumab pegol did not mediate increased levels of apoptosis in any of the in vitro assays used, suggesting that these mechanisms are not essential for the efficacy of anti-TNFα agents in CD. As certolizumab pegol, infliximab, and adalimumab, but not etanercept, almost completely inhibited LPS-induced IL-1β release from monocytes, inhibition of cytokine production may be important for efficacy of anti-TNFα agents in CD.

(Inflamm Bowel Dis 2007)

Tumor necrosis factor α (TNFα) is a pluripotent cytokine that is present as both soluble (17 kDa) and membrane-bound (26 kDa) forms and mediates a diverse range of biologic effects. TNFα monomers associate to form a trimeric complex that is capable of binding to both TNFα receptor (TNFR)-1 and TNFR-2. Although the precise function of TNFα in health and disease is unclear, elevated levels of TNFα have been found in patients experiencing inflammatory diseases, and clinical data on the efficacy of anti-TNFα agents suggest that it plays a key role in the pathogenesis of a number of inflammatory diseases such as Crohn's disease (CD) and rheumatoid arthritis (RA). However, differences between the clinical profiles of the currently marketed anti-TNFα agents suggest that neutralization of soluble TNFα may not be the only mechanism of action of these agents. The availability of a number of anti-TNFα therapies with different structures and in vitro properties offers the opportunity to investigate the potential role of these properties in clinical efficacy.

Certolizumab pegol (Cimzia™; UCB, Belgium) is a PEGylated Fab′ fragment of a humanized monoclonal antibody that binds and neutralizes human TNFα. The pharmaco-kinetic properties of Fab′ in vivo are usually poor. However, attachment of a 40 kDa polyethylene glycol (PEG) moiety to the Fab′ fragment markedly increases the half-life of certolizumab to a value comparable with that of a whole antibody product. The Fab′ fragment was engineered with a single free-cysteine residue in the hinge region, which enables site-specific attachment of PEG without affecting the ability of the Fab′ fragment to bind and neutralize TNFα.

The unique structure of certolizumab pegol differs from that of other anti-TNFα agents that have been tested for the treatment of CD, i.e., infliximab (a mouse–human chimeric whole antibody) (Remicade™; Centocor, Horsham, PA, USA), adalimumab (a recombinant human whole antibody) (Humira; Abbott Laboratories, Abbott Park, IL, USA), and etanercept (a TNFR2–human IgFc fusion protein) (Enbrel™; Amgen, Thousand Oaks, CA, USA). The latter 3 agents are all based on the human IgG1 Fc (which has the capability of fixing complement and binding to Fc receptors) and therefore have the potential for Fc-mediated effects. Infliximab has demonstrated efficacy in the management of symptoms in patients with active CD.1 In contrast, etanercept failed to show efficacy in a double-blind, placebo-controlled Phase II study.2 A number of reasons have been proposed to explain these observations including variations in the penetration of the diseased tissue among the anti-TNF agents or differences in stability of the immune complexes with TNFα but, importantly, these data suggest that neutralization of soluble TNFα (sTNFα) is not sufficient to induce and maintain remission in active CD. Other mechanisms of action proposed to be important for the efficacy of anti-TNFα agents in CD are: induction of apoptosis of both monocytes3, 4 and T cells5, 6; neutralization of membrane TNFα (mTNFα)5; antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC)7; and reverse signaling via mTNFα.8

To investigate whether the structure of certolizumab pegol could result in a different mechanism of action from that of infliximab, adalimumab, and etanercept, we assessed the 4 agents in a series of in vitro studies: neutralization of sTNFα and mTNFα; induction of ADCC, CDC, and apoptosis; degranulation of neutrophils; and inhibition of lipopolysaccharide (LPS)-induced cytokine production.


  1. Top of page
  2. Abstract

Anti-TNFα Reagents and Control Antibodies

Infliximab, adalimumab, etanercept, and certolizumab pegol were prepared according to the manufacturers' recommended methods. Control reagents were a human IgG1κ (Sigma, Poole, UK) and a control Fab′ PEG reagent (UCB Celltech, Belgium).

Isolation and Activation of Peripheral Blood Cell Populations

Venous blood from healthy donors was collected by venipuncture into heparinized tubes (BD Biosciences, San Jose, CA, USA) and the peripheral blood mononuclear cells (PBMC) were isolated by density-gradient centrifugation with Ficoll Paque (Amersham Biosciences, UK). After washing, the PBMC were then depleted of monocytes by magnetic separation using CD14 MicroBeads (Miltenyi Biotec, Germany) and an LS column (Miltenyi Biotec): CD14-negative cells (peripheral blood lymphocytes [PBL]) passed through the column and were collected, washed, and resuspended in Roswell Park Memorial Institute (RPMI) 1640 culture medium. Retained monocytes (CD14-positive cells) were recovered from the column, washed, and resuspended in culture medium. Polymorphonucleocytes (PMN; predominantly neutrophils) were also separated from venous blood by density-gradient centrifugation with Ficoll Paque and the contaminating erythrocytes were lyzed with erythrocyte lysis buffer.

To activate PBL, 106 cells/mL were cultured at 37°C for 48 hours in RPMI 1640 culture medium containing 10% (v/v) fetal calf serum (FCS; PAA Labs, Austria), 2 mM L-glutamine, 50 U/mL penicillin, and 50 μg/mL streptomycin (complete medium) in the presence of 2 μg/mL each of anti-CD3 monoclonal antibody (OKT3, UCB Celltech, Belgium) and anti-CD28 monoclonal antibody (BD PharMingen, San Diego, CA, USA). To activate monocytes, the cells were cultured at 106 cells/mL for 72 hours at 37°C in complete medium in the presence of 300 U/mL granulocyte-macrophage colony-stimulating factor (R&D Systems, Minneapolis, MN, USA) and 300 U/mL interleukin-4 (R&D Systems).

Neutralization of sTNFα

The relative abilities of the anti-TNFα agents to neutralize sTNFα were assessed using 2 cell-based in vitro assay systems: measurement of sTNFα-induced death of murine L929 cells and sTNFα-induced luciferase activity in a human reporter cell line.

The TNF6.5 cell line used for these studies is a stably transfected NS0 cell line expressing human TNFα both on the cell surface and in secreted form. An expression plasmid was transfected into NS0 cells. The plasmid contained the glutamine synthetase selection marker,9 in which the expression of the full-length human TNFα gene is under the control of the human cytomegalovirus major immediate early promoter. Stable cell lines growing in glutamine-free media were selected for surface TNF expression by flow cytometry. Cells were maintained in Iscove's Modified Dulbecco's medium (Sigma) containing the following additions per 500 mL bottle: 10% dialyzed FCS (Sigma), 3.5 mg each of adenosine, guanosine, cytidine, and uridine (Sigma), 30 mg each of glutamate and asparagine (Sigma), 50 U/mL of penicillin, and 50 mM of streptomycin (Invitrogen, UK). A single batch of TNF6.5 supernatant containing sTNFα was used for all the experiments. L929 cells were prepared at 2 × 105 cells/mL in complete medium, plated out at 100 μL/well in 96-well flat-bottomed plates (Becton Dickinson, UK), and cultured at 37°C for 24 hours before use. The medium was removed and TNF6.5 supernatant (final sTNFα concentration ≈100 pg/mL), with or without varying concentrations of the anti-TNF agents, was added in a total volume of 100 μL per well in duplicate. The assay medium comprised complete medium with the addition of 1 μg/mL actinomycin D (Sigma) to make the cells more sensitive to TNFα-mediated cell death. Plates were incubated for a further 16 hours at 37°C and then 10 μL/well of methylthiazoletetrazolium (Sigma) at 5 mg/mL in culture medium was added for a further 4 hours. The reaction was stopped by the addition of 100 μL of solubilization buffer containing 20% sodium dodecyl sulfate dissolved in 50% N,N-dimethylformamide and 50% deionized water. After overnight incubation at 37°C to allow the dye to dissolve, the plates were read using a Multiskan EX plate reader (Labsystem, Finland) at 570 nm with subtraction at 630 nm. The amount of sTNFα not neutralized in the samples (residual sTNFα) was calculated from a recombinant human TNFα (Strathmann Biotec, Germany) standard curve using Genesis software. The relative activity of the anti-TNFα agents was determined by calculating the concentration of each reagent required to neutralize 90% of the sTNFα in the assay (IC90).

To assess the ability of the test agents to neutralize recombinant sTNFα in a bioassay with human cells, A549-Luc cells were used. The A549-Luc cells are a human lung carcinoma cell line that expresses luciferase under control of the E-selectin promoter, which is activated by TNFα binding to cell surface TNFα receptors. A549-Luc cells were maintained in complete medium containing 1 mg/mL geneticin. For assays, A549-Luc cells were plated out at 1.5 × 104 cells/well in 96-well flat-bottomed plates and incubated overnight at 37°C to form a confluent monolayer. The plates were aspirated and sTNFα at 3 ng/mL, with or without various concentrations of the anti-TNF agents, was added in a final volume of 100 μL per well. All dilutions were made with RPMI 1640 medium without phenol red. The plates were incubated for 4 hours at 37°C before being processed for assessment of luciferase activity.

Luciferase Measurement

A vial of luciferase substrate (Luclite; Perkin Elmer, Boston, MA, USA) was reconstituted with 10 mL of Luclite buffer solution, mixed gently, and allowed to equilibrate at room temperature. Reconstituted substrate (100 μL/well) was added to the plate containing 100 μL/well of cells in RPMI 1640 medium without phenol red. The plates were incubated in the dark for at least 10 minutes before transferring 180 μL of culture medium to a white opaque plate for measurement of luminescence using an LJL Analyst plate reader.

Binding of Anti-TNFα Agents to Membrane TNFα

The ability of the anti-TNFα agents to bind mTNFα was assessed by fluorescence-activated cell sorter (FACS) analysis. TNF6.5 cells, which express mTNFα, were washed to remove any sTNFα from the culture supernatant and resuspended at 1.25 × 105 cells/mL of FACS buffer (Dulbecco's phosphate-buffered saline without Ca2+ and Mg2+ [Invitrogen Ltd., Gibco, UK], 5% [v/v] FCS, sodium azide 0.5% [w/v] [BDH, UK]) in the presence of the anti-TNFα agents. The cells were left at 4°C for 30 minutes, then washed twice in FACS buffer, incubated with phycoerythrin-labeled secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) (100 μL of 1/250 dilution) at 4°C for 30 minutes, then washed again before final resuspension in FACS buffer and analysis by flow cytometry (FACSCalibur system; Becton Dickinson). An anti-Fab2 second-layer polyclonal antiserum was used to reveal specific binding of certolizumab pegol, while an anti-H+L polyclonal antiserum was used to label the other 3 anti-TNFα agents (which are all based on human IgG1).

Neutralization of Membrane TNFα

To assess the ability of the test agents to neutralize mTNFα, we conducted co-culture experiments using the NS0 clone 23 cell line (engineered to express a noncleavable form of TNFα) and A549-Luc cells, as described above. For assays, A549-Luc cells were plated out at 104 cells/well in 96-well flat-bottomed plates and incubated overnight at 37°C to form a confluent monolayer. The plates were washed once with RPMI 1640 medium without phenol red, then 105 NS0 clone 23 cells in 80 μL were added per well, together with 20 μL of anti-TNFα agent at varying concentrations. All dilutions were made with RPMI 1640 medium without phenol red. The plates were incubated for 18 hours at 37°C before being processed for assessment of luciferase activity (described above). Control wells were also included containing A549-Luc cells incubated with recombinant sTNFα or clone 23 cells alone.

Cell-dependent Cytotoxicity and Antibody-dependent Cell-mediated Cytotoxicity

To compare the ability of the 4 anti-TNFα agents to mediate CDC and ADCC, TNF6.5 cells were used as target cells, because of their high expression of mTNFα, and cell damage (breakdown in membrane integrity) was determined by the cellular uptake of the vital stain propidium iodide (PI).

To assess CDC, 10 mL of TNF6.5 cells at 2 × 106 cells/mL in RPMI 1640 medium without phenol red (Gibco) were mixed with 500 μL of baby rabbit complement (Serotec, UK). The anti-TNFα agents were diluted in Dulbecco's phosphate-buffered saline (Gibco) to the required concentrations and plated onto a 96-well round-bottomed plate in 50 μL volumes in duplicate, then 100 μL of the cell–complement mix was added per well, and the plate incubated at 37°C for 4 hours. Cells were stained with PI at 3 μg/mL for 10 minutes at room temperature, and then uptake of PI was determined by flow cytometry.

To assess ADCC, 6 × 104 TNF6.5 target cells and 1.4 × 105 effector cells (CD14-positive cell-depleted PBMC) were incubated per well in 200 μL RPMI 1640 medium in 96-well round-bottomed culture plates, with or without anti-TNFα agents (25 μL/well), for 4 hours at 37°C. Cells were stained by adding 25 μL/well of a solution containing 100 μg/mL antimouse CD45 fluorescein isothiocyanate (FITC) and 24 μg/mL PI. Plates were incubated on ice for 15 minutes and then the target cells were analyzed by flow cytometry (gated based on positive staining with an antimouse CD45 FITC).

Induction of Apoptosis by Anti-TNFα Agents

PBL and monocyte populations were isolated and activated as described above. The respective cell populations were then cultured at 106 cells/mL for 24 hours at 37°C in the presence of anti-TNFα agent (1–100 μg/mL) or controls. To assess the proportion of apoptotic cells, 50 μL of a solution containing 4 mL Hank's Buffered Salt Solution, 1 mL Annexin V Buffer, 150 μL Annexin V-FITC, and 60 μg PI was added to 200 μL of cells and incubated for 10 minutes at room temperature. The proportion of apoptotic cells per well was determined by the binding of Annexin V-FITC (BD PharMingen) to the cell surface, as determined by flow cytometry and by cellular uptake of PI (data not shown). For the purposes of these studies, all cells staining positive for Annexin V were considered apoptotic.

Incubation of Granulocytes with Anti-TNFα Agents

PMN were washed once in RPMI 1640 medium, then 200 μL of PMN were plated out at 106 cells/mL in 96-well round-bottomed culture plates in RPMI 1640 medium without phenol red with 50 μL of anti-TNFα agent or isotype controls at final concentrations of 0.5–100 μg/mL. Cells were incubated for 12 hours at 37°C before addition of PI at 3 μg/mL for the final 10 minutes. Uptake of PI by the target cells was assessed by flow cytometry and degranulation by colorimetric measurement of the release of myeloperoxidase (MPO) into the supernatant. Maximum MPO release was determined by detergent-mediated lysis of cells in appropriate control wells. Briefly, 10 μL of a 10× lysis solution (Promega, UK) per 250 μL of culture medium (giving a concentration of ≈0.3% Triton X-100) was added to the wells for 45 minutes prior to harvesting the supernatants. This would be expected to result in maximal release of MPO from the cells. Following incubation, a colorimetric assay was used to determine neutrophil primary granule release by measurement of MPO using 3,3′,5,5′-tetramethylbenzidine (Sigma) as a substrate. Corrected values obtained were used in the following formula to calculate percent degranulation:

  • equation image

Effect of Preincubation of Human Peripheral Blood-derived Monocytes with Anti-TNFα Agents on Lipopolysaccharide-induced Cytokine Release

The anti-TNFα agents were tested for their ability to inhibit subsequent LPS-induced production of interleukin (IL)-1β by human PBMC. Purified monocytes in 200 μL volumes of RPMI 1640 medium at 5 × 106 cells/mL were pre-incubated for 1 hour at 37°C in 96-well round-bottomed plates with concentrations of the anti-TNF agents from 0 μg/mL to 100 μg/mL or appropriate controls. After washing the monocytes were resuspended in complete medium. The cells were then incubated with LPS (Sigma; 100 ng/mL) for 4 hours at 37°C, prior to the collection of supernatants for cytokine measurement. Human IL-1β was measured in the supernatants using an enzyme-linked immunosorbent assay kit (R&D Systems) according to the manufacturer's instructions.


  1. Top of page
  2. Abstract

Neutralization of TNFα

All 4 anti-TNFα agents tested effectively neutralized sTNFα in both assay systems (Figs. 1A,B). In the L929 assay system, etanercept was the most potent, with an IC90 of 0.7 ng/mL, followed by certolizumab pegol, with an IC90 of 3 ng/mL, then infliximab and adalimumab, each with an IC90 of 9 ng/mL (Fig. 1A). Similar results were obtained using the A549-Luc TNFα-responsive reporter cell assay and recombinant TNFα (Fig. 1B), although certolizumab pegol showed similar potency to etanercept in this assay. In this experimental system, infliximab did not completely neutralize recombinant TNFα, suggesting that this agent may not bind to recombinant TNFα as effectively as the other agents.

thumbnail image

Figure 1. In vitro neutralization of soluble TNFα. (A) Measurement of TNFα cytotoxicity to L929 cells. (B) Induction of luciferase activity in A549-Luc cells using (●) certolizumab pegol, (▴) infliximab, (▪) adalimumab, (⧫) etanercept, (○) control Fab′ PEG (B only), and (□) control IgG (B only). Results are the mean of 3 experiments ± standard error of the mean.

Download figure to PowerPoint

All 4 test agents appeared to bind mTNFα (Fig. 2). In contrast to published data involving different assay systems,5 etanercept bound to mTNFα to a similar extent as infliximab and adalimumab.

thumbnail image

Figure 2. Binding of anti-TNFα agents to membrane TNFα on TNF6.5 cells in vitro. Levels of bound (▴) infliximab, (▪) adalimumab, (⧫) etanercept, and (□) control IgG were revealed using a phycoerythrin-labeled anti-heavy and light chain polyclonal antiserum (left-hand Y axis) and levels of bound (●) certolizumab pegol and (○) control Fab′ PEG were revealed using an anti-Fab2 polyclonal antisera (right-hand Y axis). Both datasets are expressed as fluorescence intensity. Results are representative of 3 experiments.

Download figure to PowerPoint

All 4 anti-TNFα agents inhibited mTNFα-mediated effects in a concentration-dependent manner (Fig. 3). The abilities of certolizumab pegol, adalimumab, and infliximab to neutralize mTNFα-mediated signaling were comparable, while etanercept appeared to be about 2-fold less potent. To verify the assay system used to assess neutralization of mTNFα, supernatants collected from NS0 clone 23 cells cultured alone at 105 cells/well were tested and it was shown that soluble TNF did not contribute significantly to the induction of luciferase activity in A549-Luc cells (data not shown).

thumbnail image

Figure 3. In vitro neutralization of membrane TNFα-expressing NS0 clone 23 cell stimulation of A549-Luc cells by anti-TNFα agents: (●) certolizumab pegol, (▴) infliximab, (▪) adalimumab, (⧫) etanercept, (□) control IgG, or (○) control Fab′ PEG. Results are representative of 3 experiments. Optical density measurements were ≈2500 for supernatants from equivalent numbers of NS0 clone 23 cells incubated with the A549-Luc cells, which was the same level as seen with cells alone.

Download figure to PowerPoint

Cell-dependent Cytotoxicity and Antibody-dependent Cell-mediated Cytotoxicity Activity

Infliximab, adalimumab, and etanercept all mediated CDC (Fig. 4A) and ADCC (Fig. 4B) of TNF6.5 cells compared to the control human IgG1, although the level of ADCC mediated by etanercept was apparently reduced compared to infliximab and adalimumab (≈30% versus 45% specific killing; Fig. 4B). Certolizumab pegol does not have an Fc region and, therefore, as expected did not mediate cell killing either directly through fixation of complement or via recruitment of effector cells (Figs. 4A,B).

thumbnail image

Figure 4. (A) Mediation of complement-dependent cytotoxicity of TNFα-expressing TNF6.5 cells by infliximab (▴), adalimumab (▪), etanercept (⧫), and certolizumab pegol (●) in vitro in the presence of complement. Cytotoxicity was assessed by uptake of PI; values shown are the percentages of PI-positive cells with the test agent minus the percentage of PI-positive cells achieved with the control. Results are the mean of 3 experiments ± SEM. (B) Mediation of antibody-dependent cell-mediated cytotoxicity of TNFα-expressing TNF6.5 cells (targets) by infliximab (▴), adalimumab (▪), etanercept (⧫), and certolizumab pegol (●) in vitro in the presence of CD14-positive cell-depleted peripheral blood mononuclear cells (effectors), as assessed by uptake of PI by the target cells. Values shown are the percentages of PI-positive cells with the test agent minus the percentage of PI-positive cells achieved with the control. Results are the mean of 3 experiments ± SEM.

Download figure to PowerPoint

Apoptosis of Activated Human Peripheral Blood Lymphocytes and Monocytes In Vitro

Under the conditions used, even at concentrations of up to 100 μg/mL, certolizumab pegol did not increase the proportion of apoptotic cells above control levels in either PBL (Fig. 5A) or monocyte cultures (Fig. 5B). However, both infliximab and adalimumab did increase the proportion of apoptotic cells above control levels at concentrations higher than 5 μg/mL with PBL and higher than 1 μg/mL with monocytes. At the maximum tested concentrations of infliximab and adalimumab (100 μg/mL), ≈20% of PBL (Fig. 5A) and 35% of monocytes (Fig. 5B) appeared to be apoptotic with either agent, as determined by Annexin V binding. Etanercept also increased the proportion of apoptotic cells compared with controls in both PBL and monocyte systems, although the percentages of apoptotic cells were lower than those observed following incubation with infliximab or adalimumab: ≈15% of PBL (Fig. 5A) and 20% of monocytes (Fig. 5B) bound Annexin V.

thumbnail image

Figure 5. Effect of (●) certolizumab pegol, (▴) infliximab, (▪) adalimumab, or (⧫) etanercept on (A) apoptosis of activated peripheral blood lymphocytes and (B) monocytes in vitro. Results are expressed as percentages of cells binding Annexin V-FITC after subtraction of the levels of binding observed with the appropriate control reagent (around 5%). Results are the mean of 3 experiments ± SEM.

Download figure to PowerPoint

In Vitro Degranulation of Polymorphonuclear Cells

Prior to assessment of the effects of the anti-TNFα agents on polymorphonuclear granulates (PMN), binding of the 4 anti-TNFα agents to peripheral blood-derived PMN was confirmed by FACS analysis (data not shown). Incubation with certolizumab pegol did not appear to result in changes to PMN integrity, while degranulation and loss of cell membrane integrity (as determined by PI uptake) was dramatically increased after incubation with infliximab, adalimumab, or etanercept (Fig. 6A). To determine whether the loss of cell viability was associated with release of material from intracellular granules, levels of MPO were determined in the culture supernatants. MPO levels were elevated in a concentration-dependent manner in cultures from PMN incubated with infliximab, adalimumab, or etanercept (Fig. 6B); concentrations of these agents of 1 μg/mL and above were sufficient to induce PI uptake (Fig. 6A) and MPO release above control levels (Fig. 6B). Infliximab and adalimumab induced very similar levels of MPO release at all drug concentrations tested, with ≈75% of the maximal achievable release induced at the maximum concentration tested (100 μg/mL) (Fig. 6B). Levels of both PI uptake and MPO release were slightly lower after incubation with etanercept versus infliximab or adalimumab, with ≈60% of maximal release (Figs. 6A,B). Consistent with the level of PI uptake, certolizumab pegol did not induce any elevation in MPO release above control levels even at the highest tested concentration of 100 μg/mL (Fig. 6B). This effect was not mediated by binding to Fc receptors, as the Fab2 fragment of infliximab had the same effect on PMNs as the IgG (data not shown).

thumbnail image

Figure 6. Effect of anti-TNFα agents on granulocyte membrane integrity and myeloperoxidase release in vitro. (A) Uptake of PI and (B) myeloperoxidase release after culture with (●) certolizumab pegol, (▴) infliximab, (▪) adalimumab, or (⧫) etanercept. Results are expressed relative to maximal loss of membrane integrity and myeloperoidase release induced by incubation of the cells with lysis buffer, and are the mean of 3 experiments ± SEM.

Download figure to PowerPoint

These data suggest that incubation of PMN with infliximab, adalimumab, or etanercept induced a loss of cell viability and release of intracellular granular material into the tissue culture medium. Due to the limitations of the culture systems used, it was not possible to determine whether loss of cell viability preceded release of material from the granules, or whether the loss of viability resulted from increased levels of granular material in the medium.

In Vitro Inhibition of Cytokine Production by Human Monocytes

Pre-incubation of human monocytes with certolizumab pegol at concentrations of 1 μg/mL and above completely inhibited subsequent production of IL-1β in response to LPS (Fig. 7). Certolizumab pegol appeared to inhibit IL-1β release more potently than the other 3 agents, although concentrations of 1 μg/mL and above of infliximab or adalimumab also completely inhibited LPS-induced cytokine production. In contrast, etanercept showed only partial inhibition even at concentrations up to 100 μg/mL (Fig. 7).

thumbnail image

Figure 7. Effect of anti-TNFα agents (●) certolizumab pegol, (▴) infliximab, (▪) adalimumab, or (⧫) etanercept, (○) control Fab′ PEG, and (□) the control IgG1 on LPS-induced production of IL-1β by human monocytes in vitro. Results are representative of 3 experiments.

Download figure to PowerPoint

Despite extensive washing, the carry-over of anti-TNF agents during the LPS stimulation is a potential issue in this assay. However, the addition of mouse TNF (which is not neutralized by any of the antibody reagents) to this assay did not mediate any IL-1β production over 4 hours, whereas IL-1β production was mediated over 24 hours by mouse TNF, showing that it is active in this system (data not shown). Therefore, the observed inhibition of IL-1β appears not to be due to neutralization of sTNF in this assay.


  1. Top of page
  2. Abstract

Therapeutic agents targeting TNFα were evaluated in clinical trials in a number of inflammatory disease settings including CD, RA, ankylosing spondylitis, and psoriasis.10 However, the diverse range of biologic effects mediated by TNFα has hampered efforts to define the mechanisms of action of these agents.

Recently released positive clinical data from the Phase III PRECiSE studies of certolizumab pegol in CD11, 12 have provided an opportunity to reassess available data from studies of infliximab, adalimumab, and etanercept. Certolizumab pegol, infliximab, and adalimumab appear to be similar in terms of induction and maintenance of response and remission, as reported in the similarly designed PRECiSE 2,11 ACCENT I,13 and CHARM14 trials. Infliximab is also effective in ulcerative colitis and for fistulizing disease and mucosal healing in CD; studies have not yet been published for certolizumab pegol in these indications.

The structure of certolizumab pegol differs significantly from the other agents that have undergone clinical evaluation in CD. Therefore, it is becoming possible for the first time to begin to determine which facets of an anti-TNFα agent may be required for clinical efficacy in CD and which may be unnecessary and potentially undesirable.

The studies described here were undertaken to compare the 4 agents in a range of in vitro experimental systems. Most of these studies used the anti-TNFα agents at a concentration of up to 100 μg/mL to reflect the concentration of these drugs found in the blood of patients, which are all in the μg/mL range. As expected, all the agents potently neutralized the bioactivity of sTNFα. These data are consistent with published data with infliximab and etanercept.5 Although neutralization of sTNFα is likely to play an important role in the efficacy profile of anti-TNFα agents across a range of inflammatory diseases, these data suggest that this mechanism alone does not predict clinical efficacy in CD, because of the apparent failure of etanercept in Phase II clinical trials for CD.2

TNFα expressed on the membrane of cells appears to be capable of mediating a range of effects, and mTNFα can bind and stimulate TNFα receptors on other cells.15 In the studies described here, certolizumab pegol bound to and neutralized effects mediated by mTNFα. It has been suggested previously that etanercept does not bind to mTNFα to the same extent as infliximab5, 16, 17 and the experiments described in this report generally support these observations. However, etanercept was clearly capable of some degree of binding to mTNFα, as suggested by its ability to neutralize mTNFα-mediated stimulation of a reporter cell line and to mediate CDC and ADCC (Figs. 4A,B). With the exception of CDC, etanercept generally appeared less potent in these assays than infliximab and adalimumab.

Certolizumab pegol did not mediate either CDC or ADCC (Figs. 4A,B), as would be predicted from the absence of an Fc region in the molecule. Both infliximab and adalimumab effectively bound and neutralized mTNFα (Figs. 2, 3) and mediated CDC and ADCC (Figs. 4A,B), consistent with data from other experimental systems.7, 16 It is possible that the different results obtained with etanercept in the experimental systems used in these studies and previously published data may relate to differences in the proportions of monomeric and trimeric TNFα expressed by the respective cell lines used. Based on these data, binding to monomeric TNFα cannot therefore be excluded as a potential requirement for efficacy in CD. Etanercept is also different from the other reagents in that it neutralizes lymphotoxin in addition to TNFα,16 although it is unclear if this fact is significant with regard to the lack of efficacy in CD. It would, however, appear that cell depletion via CDC and/or ADCC is not a requisite for efficacy in this disease, as demonstrated by the positive clinical data from the PRECiSE studies.11, 12

Resistance of cells, such as T cells, to apoptosis has been proposed to play an important role in inflammatory bowel disease,18 and TNFα is a known survival factor for certain cell types. A number of published studies have examined the ability of the anti-TNFα agents to mediate apoptosis, in particular infliximab and etanercept. The data suggest that infliximab is capable of inducing apoptosis in human T cell and monocytic cell lines,4, 6 activated peripheral blood T cells or lymphocytes from normal donors,5, 19 lamina propria T cells,5, 6 and monocytes from patients with CD.3 Adalimumab was also observed to induce apoptosis of PBMC.4 In contrast, etanercept was found to be less effective at inducing apoptosis in PBMC, activated peripheral blood lymphocytes from normal donors, and lamina propria T cells from patients with CD,4, 5 suggesting that induction of apoptosis could be a key mechanism of action for anti-TNFα agents in CD. The data generated in our experimental systems, however, suggest that induction of apoptosis via binding to mTNFα may not be required for clinical efficacy in CD. Consistent with published reports, the proportion of activated T cells and monocytes from normal donors showing binding of Annexin V increased following incubation with either infliximab or adalimumab. Exposure to etanercept did result in an increased proportion of both cell types binding Annexin V, but at a reduced level compared with infliximab and adalimumab. In contrast to the other agents, certolizumab pegol did not induce any detectable increase in the proportion of either cell type binding Annexin V. It is important to recognize that the ability of certolizumab pegol to induce apoptosis of lamina propria cells derived from patients with CD has not been addressed in these studies. Hence, the potential for a different result under these conditions cannot be excluded because there is some evidence that lamina propria cells are more susceptible to infliximab-induced apoptosis.20 Whether any of these cell-killing mechanisms occur in vivo is also unclear, although it has been shown21–23 that there is a prolonged reduction in the numbers of PBMC after infliximab treatment.

It is not clear why certolizumab pegol does not induce apoptosis of cells bearing mTNF, as do the other anti-TNF agents. It may be because certolizumab pegol binds to a different epitope from the other agents, which leads to a different signaling pattern inside the cell. An increase in apoptosis has been reported in tissue sections from diseased bowel 24 hours6 and 28 days20 after infliximab treatment, although the cause of the increase in apoptosis is difficult to determine. It is possible that in this situation the anti-inflammatory effect of infliximab leads indirectly to apoptosis of activated cells.

As neutrophils commonly occur at sites of inflammation and are able to produce TNFα upon stimulation,24 we also examined the ability of the anti-TNFα agents to bind to human PMNs. After demonstrating that all 4 agents bound to the surface of these cells, functional effects resulting from binding were examined. Although certolizumab pegol bound to human PMNs, no increase in cell death or release of MPO was observed (Fig. 6), possibly because of a difference in the way in which it signals through membrane TNF. In contrast, the other 3 anti-TNFα agents all induced both cell death and release of MPO.

The implications of these findings are currently unclear. Previous studies investigating the effect of infliximab,25, 26 adalimumab,27 or etanercept26, 28 on neutrophil functionality did not report any significant functional alterations. However, these studies were conducted on the circulating neutrophil pool, and therefore did not assess potential decreases in the neutrophil population after dosing. Some evidence for decreases in circulating neutrophil numbers has been reported after treatment with infliximab.20, 29

There is evidence that mTNFα can mediate reverse signaling back into the cell30, 31 and this is a potential mechanism by which infliximab, adalimumab, and etanercept increase the proportion of apoptotic cells. A potential role for reverse signaling has been proposed in the subsequent downregulation of cytokine production in response to LPS stimulation8, 19, 32 by monocytes. Intriguingly, in these studies incubation with infliximab resulted in subsequent inhibition of LPS-induced cytokine production, whereas incubation with etanercept did not.8 Comparable data have been generated in the present studies, in which certolizumab pegol also effectively inhibited subsequent LPS-induced cytokine production. Although production of TNFα is inhibited, any potential effects of carry-over of the antibodies were addressed by measuring IL-1β. The structural differences between certolizumab pegol and the other anti-TNFα agents may provide important insights into the mechanisms underlying these observations. Bacteria and bacterial products are increasingly thought to play a role in inappropriate activation of the immune system in CD.33 This provides a potential link between these in vitro findings and mechanisms that may underlie the disease process, particularly as etanercept neither effectively inhibits this process, nor has been shown to be efficacious in CD. LPS stimulation of cells of the innate immune system is a very powerful signal suggesting that the inhibition mediated by some of the anti-TNFα agents may be a potent anti-inflammatory mechanism. Further ex vivo studies are ongoing using cells from the inflamed bowel of patients with CD in order to determine the relevance of this effect in clinical disease.

In summary, a series of comparative in vitro studies was conducted to explore the mechanism of action of certolizumab pegol in CD. The unique structure of this molecule, coupled with positive clinical data in this disease indication, have provided important new information to increase our understanding of the key requirements for an anti-TNFα agent in CD. While these studies are all based on in vitro systems, they nevertheless suggest that mechanisms such as induction of apoptosis, CDC, and ADCC may not be required for clinical efficacy of an anti-TNFα agent in CD. If these mechanisms are not required for efficacy, further studies are required to determine whether they may, in fact, be associated with undesirable effects. In contrast, inhibition of bacterially stimulated cytokine production from macrophages may be a necessary function for an anti-TNFα agent to produce efficacy in CD. A better understanding of the differences in mechanisms of action between the available anti-TNFα agents could enable the selection of only those functions required for clinical efficacy in the development of the next generation of anti-TNFα agents.


  1. Top of page
  2. Abstract
  • 1
    Rutgeerts P, D'Haens G, Targan S, et al. Efficacy and safety of retreatment with anti-tumor necrosis factor antibody (infliximab) to maintain remission in Crohn's disease. Gastroenterology. 1999; 117: 761769.
  • 2
    Sandborn WJ, Hanauer SB, Katz S, et al. Etanercept for active Crohn's disease: a randomized, double-blind, placebo-controlled trial. Gastroenterology. 2001; 121: 10881094.
  • 3
    Lugering A, Schmidt M, Lugering N, et al. Infliximab induces apoptosis in monocytes from patients with chronic active Crohn's disease by using a caspase-dependent pathway. Gastroenterology. 2001; 121: 11451157.
  • 4
    Shen C, Assche GV, Colpaert S, et al. Adalimumab induces apoptosis of human monocytes: a comparative study with infliximab and etanercept. Aliment Pharmacol Ther. 2005; 21: 251258.
  • 5
    Van den Brande JM, Braat H, van den Brink GR. Infliximab but not etanercept induced apoptosis in lamina propria T-lymphocytes from patients with Crohn's disease. Gastroenterology. 2003; 124: 17741785.
  • 6
    ten Hove T, van Montfrans C, Peppelenbosch MP, et al. Infliximab treatment induces apoptosis of lamina propria T lymphocytes in Crohn's disease. Gut. 2002; 50: 206211.
  • 7
    Scallon BJ, Moore MA, Trinh H, et al. Chimeric anti-TNF-α monoclonal antibody cA2 binds recombinant transmembrane TNF-α and activates immune effector functions. Cytokine. 1995; 7: 251259.
  • 8
    Kirchner S, Holler E, Haffner S, et al. Effect of different tumor necrosis factor (TNF) reactive agents on reverse signaling of membrane integrated TNF in monocytes. Cytokine. 2004; 28: 6774.
  • 9
    Bebbington CR, Renner G, Thomson S, et al. High-level expression of a recombinant antibody from myeloma cells using a glutamine synthetase gene as an amplifiable selectable marker. Biotechnology. 1992; 10: 169175.
  • 10
    Feldmann M, Brennan FM, Foxwell BM, et al. Anti-TNF therapy: Where have we got to in 2005? J Autoimmun. 2005; 25: 2628.
  • 11
    Schreiber S, Khaliq-Kareemi M, Lawrence I, et al. Certolizumab pegol, a humanised anti-TNF PEGylated Fab′ fragment, is safe and effective in the maintenance of response and remission following induction in active Crohn's disease: a Phase III study (PRECiSE). Gut. 2005; 54( Suppl VII): A82 (Abstract).
  • 12
    Sandborn WJ, Feagan BG, Stoinov S, et al. Certolizumab pegol administered subcutaneously is effective and well tolerated in patients with active Crohn's disease: results from a 26-week, placebo-controlled Phase III study (PRECiSE 1). Gastroenterology. 2006; 130( Suppl 2): A1A912 Abstract 745.
  • 13
    Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance infliximab for Crohn's disease: the ACCENT I randomised trial. Lancet. 2002; 359: 15411549.
  • 14
    Colombel J, Sandborn WJ, Rutgeerts P, et al. Adalimumab induces and maintains clinical response and remission in patients with active Crohn's disease: results of the CHARM trial. Gastroenterology. 2006; 130( Suppl 2): A1A912 Abstract 686.
  • 15
    Grell M, Douni E, Wajant H, et al. The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell. 1995; 83: 793802.
  • 16
    Scallon B, Cai A, Solowski N, et al. Binding and functional comparisons of two types of tumor necrosis factor antagonists. J Pharmacol Exp Ther. 2002; 301: 418426.
  • 17
    Mitoma H, Horiuchi T, Tsukamoto H. Binding activities of infliximab and etanercept to transmembrane tumor necrosis factor-α. Gastroenterology. 2004; 126: 934935.
  • 18
    Peppelenbosch MP, van Deventer SJ. T cell apoptosis and inflammatory bowel disease. Gut. 2004; 53: 15561558.
  • 19
    Ringheanu M, Daum F, Markowitz J, et al. Effects of infliximab on apoptosis and reverse signaling of monocytes from healthy individuals and patients with Crohn's disease. Inflamm Bowel Dis. 2004; 10: 801810.
  • 20
    Di Sabatino A, Ciccocioppo R, Cinque B, et al. Defective mucosal T cell death is sustainably reverted by infliximab in a caspase dependent pathway in Crohn's disease. Inflamm Bowel Dis. 2003; 53: 7077.
  • 21
    Ohshima S, Saeki Y, Mima T, et al. Long-term follow-up of the changes in circulating cytokines, soluble cytokine receptors, and white blood cell subset counts in patients with rheumatoid arthritis (RA) after monoclonal anti-TNF alpha antibody therapy. J Clin Immunol. 1999; 19: 305313.
  • 22
    Lorenz HM, Antoni C, Valerius T, et al. In vivo blockade of TNF-α by intravenous infusion of a chimeric monoclonal TNF-α antibody in patients with rheumatoid arthritis. J Immunol. 1996; 156: 16461653.
  • 23
    Lorenz HM, Grunke M, Hieronymus T, et al. In vivo blockade of tumour necrosis factor-α in patients with rheumatoid arthritis: longterm effects after repeated infusion of chimeric monoclonal antibody cA2. J Rheumatol. 2000; 27: 304309.
  • 24
    Balazovich KJ, Suchard SJ, Remick DG, et al. Tumor necrosis factor-α and FMLP receptors are functionally linked during FMLP-stimulated activation of adherent human neutrophils. Blood. 1996; 88: 690696.
  • 25
    den Broeder AA, Wanten GJ, Oyen WJ, et al. Neutrophil migration and production of reactive oxygen species during treatment with a fully human anti-tumour necrosis-α monoclonal antibody in patients with rheumatoid arthritis. J Rheumatol. 2003; 30: 232237.
  • 26
    Hartmann P, Franzen C, Rubbert A, et al. Blockade of TNF does not alter oxygen burst and phagocytosis of human neutrophils in patients with rheumatoid arthritis. Immunobiology. 2005; 209: 669679.
  • 27
    Capsoni F, Sarzi-Puttini P, Atzeni F, et al. Effect of adalimumab on neutrophil function in patients with rheumatoid arthritis. Arthritis Res Ther. 2005; 7: R250R255.
  • 28
    Moreland LW, Bucy RP, Weinblatt ME, et al. Immune function in patients with rheumatoid arthritis treated with etanercept. Clin Immunol. 2002; 103: 1321.
  • 29
    Agnholt J, Kelsen J, Brandsborg B, et al. Increased production of granulocyte-macrophage colony-stimulating factor in Crohn's disease — a possible target for infliximab treatment. Eur J Gastroenterol Hepatol. 2004; 16: 649655.
  • 30
    Smith CA, Farrah T, Goodwin RG. The TNF receptor superfamily of cellular and viral proteins: activation, costimulation, and death. Cell. 1994; 76: 959965.
  • 31
    Watts AD, Hunt NH, Wanigasekara Y, et al. A casein kinase I motif present in the cytoplasmic domain of members of tumour necrosis factor ligand family is implicated in “reverse signalling”. EMBO J. 1999; 18: 21192126.
  • 32
    Eissner G, Kirchner S, Lindner H, et al. Reverse signaling through transmembrane TNF confers resistance to lipopolysaccharide in human monocytes and macrophages. J Immunol. 2000; 164: 61936198.
  • 33
    Cominelli F. Cytokine-based therapies for Crohn's disease — new paradigms. N Engl J Med. 2004; 351: 20452048.