Background: Thalidomide, one of whose activities is to inhibit Tumour Necrosis Factor (TNF)-α production, has been reported to be an effective treatment for refractory inflammatory bowel disease (IBD). TNF-α driven production of matrix metalloproteinase (MMP)-3 by gut lamina propria mononuclear cells (LPMCs) is a major pathway of tissue injury in IBD; however the effect of thalidomide and newer more potent immunomodulatory derivatives on this pathway has not been studied.
Aim: To investigate the effect of thalidomide, CC-4047 (pomalidomide), CC-5013 (lenalidomide), and CC-10004 (apremilast) on gut LPMC TNFα and MMP-3 production in patients with IBD.
Methods: Gut LPMCs and myofibroblasts were isolated from patients with IBD, and cultured with thalidomide, CC-4047, CC-5013, and CC-10004. MMP-3 and TIMP-1 levels were determined by western blotting and real-time PCR, and TNF-α levels by ELISA.
Results: CC-10004 significantly reduced both TNF-α production and MMP-3 production by cultured LPMCs. Thalidomide and CC-4047 and CC-5013 had no significant effect on the production of TNF-α or MMP-3 by LPMCs.
Conclusion: These results provides a mechanistic rationale for both the failure of lenalidomide (CC-5013) in a recent randomised controlled trial in Crohn's disease, and for the evaluation of CC-10004 as a novel oral therapy in the treatment of CD and UC.
Tumour Necrosis Factor (TNF)-α is a pleiotropic cytokine which plays a central role in the pathogenesis of Crohn's disease (CD) and ulcerative colitis (UC). High levels of TNF-α are found in the stools, lamina propria, and blood of patients with active disease, 1 – 3 and inhibition of TNF-α with neutralising monoclonal antibodies reduces disease activity in both CD and UC. 4 , 5 There is now strong evidence that one of the major pathways through which TNF-α damages the gut is via upregulation of matrix metalloproteinase (MMP) production by gut myofibroblasts, leading to extracellular-matrix (ECM) breakdown, tissue destruction, and ulcer formation. 6 MMPs are key biological mediators of tissue degradation and restitution, and are involved in many disease processes including rheumatoid arthritis, liver fibrosis, cardiovascular disease, periodontal disease, cancer, and IBD. 7 In IBD, high levels of MMP-1, MMP-2, MMP-3, and MMP-9 are found in areas of inflammation and ulceration in the gut in both CD and UC. 8 – 13 MMP-3 and MMP-9 are also abundant in fistulous tracts in CD. 14 The importance of MMP-3 has been demonstrated in functional ex-vivo studies using a human fetal gut model of IBD, where addition of recombinant MMP-3 results in severe tissue damage and T cell-driven mucosal degradation is inhibited by a p55 TNF receptor-human IgG fusion protein. 15 , 16 Inhibition of MMP-3 also reduced the severity of inflammation in a mouse model of colitis induced by dextran sulphate sodium (DSS). 17
Thalidomide is an immunomodulatory drug that, among other properties, is a powerful inhibitor of TNF-α production by blood monocytes, 18 which has led to its recent use in a wide variety of inflammatory conditions. In uncontrolled trials it has proved effective in the treatment of refractory CD and UC. 19 – 25 Its effectiveness is presumed to relate to inhibition of gut TNF-α production. However, its usefulness is limited by both its teratogenicity, and by a high rate of adverse reactions in clinical practice.
Several other immunomodulatory compounds have recently been developed that are considerably more potent inhibitors of TNF-α than thalidomide, and have an improved safety profile, making them attractive potential oral agents for the treatment of IBD. These can be split into two groups, the IMiDs®, immunomodulatory imide drugs, that have a wide range of immunomodulatory actions, including inhibition of TNF-α, IL-1β and IL-12 production, and the highly selective PDE4 inhibitors, which are potent inhibitors of PDE4 (phosphodiesterase 4) activity, and TNF-α production. 26 Phosphodiesterases are enzymes that regulate intracellular levels of cyclic nucleotides by catalising their breakdown to inactive metabolites. PDE4 cAMP phosphodiesterases exclusively hydrolyse cAMP. PDE4 inhibition results in the intracellular accumulation of cAMP, downstream activation of protein kinase A (PKA), and subsequent phosphorylation of the transcription factor cAMP-response element binding protein (CREB). 27 This modulates gene transcription of many cytokines including suppressing TNF-α production by mononuclear cells, thereby inhibiting their tissue destructive and pro-inflammatory properties. Lenalidomide (CC-5013), the lead investigational drug in the immunomodulatory imide class, has recently been reported to be ineffective in the treatment of CD in a randomized placebo controlled trial. 28 In contrast, though the group of PDE4 inhibitors have been reported to be effective inhibitors of TNF-α production by normal gut LPMCs, 29 their use has not yet been reported in IBD. Accordingly, in this study we investigated the effect of thalidomide and two immunomodulatory drugs, CC-5013 and CC-4047, and the PDE4 inhibitor CC-10004, on in-vitro production of TNF-α and MMP-3 by cultured LPMCs.
Materials and methods
This work was approved by the Southampton and South West Hampshire local ethics committee. Macroscopically inflamed colonic mucosal tissue was obtained from patients with CD ( n = 8) and UC ( n = 12) undergoing either surgical resection, or an endoscopic procedure on clinical grounds. In all cases the diagnosis was confirmed using established clinical, radiological, and endoscopic criteria with confirmatory histology. Two of the patients with CD patients were receiving corticosteroids alone, two were receiving corticosteroids and azathioprine, one was also taking a 5-ASA and four were on no treatment. One of the patients with UC were receiving corticosteroids and a 5-ASA, two were on corticosteroids and azathioprine, two were on azathioprine and a 5-ASA, two were taking a 5-ASA alone, and five were on no treatment. Control samples ( n = 10) were obtained from subjects undergoing endoscopic evaluation for IBS, polyp surveillance, or rectal bleeding or from uninvolved areas of normal mucosa at least 5 cm away from any visible lesion in patients undergoing surgery for colon cancer. Whole blood was collected from six healthy donors into green top, sodium heparin containing Vacutainer® glass tubes (BD, Cowley, Oxford). Following removal, all samples were placed in ice-cold medium and used within 1 h.
Purification and culture of peripheral blood mononuclear cells
Peripheral blood mononuclear cells were isolated from whole blood using a Ficoll gradient according to the manufacturer's instructions. Briefly, whole blood was diluted to 30 ml with RPMI 1640 (Sigma Chemical, Dorset, UK) containing 100 μg/ml streptomycin, 100 U/ml penicillin, 50 μg/ml gentamicin and layered over 15 ml Ficoll Paque. Samples were then centrifuged for 25 min, after which PBMCs were harvested from the interface, washed, and resuspended. Cell count and viability were determined using Trypan blue exclusion staining. Cell isolations with less than 90% viability were discarded.
Purification and culture of lamina propria mononuclear cells
Lamina propria mononuclear cells (LPMCs) were isolated using a modified dithiothreitol/EDTA and collagenase method as previously described. 30 In brief the epithelial layer was removed by agitating the biopsy specimens in 1 mM EDTA. The denuded mucosal sample was then mechanically disrupted and incubated at 37 °C in RPMI 1640 containing 10% fetal calf serum (FCS), 100 μg/ml streptomycin, 100 U/ml penicillin, 50 μg/ml gentamicin, and 1 mg/ml collagenase V for 90 min. The resulting cell fraction was washed three times in RPMI and resuspended. In all cases sample viability was determined as > 90% by Trypan blue staining. For culture experiments the resulting cells were resuspended in RPMI 1640 containing 100 μg/ml streptomycin, 100 U/ml penicillin, 50 μg/ml gentamicin, 1 mg/ml amphoterecin, and HL-1 (Biowhittaker, Wokingham, UK)., a serum replacement agent.
Western blotting was performed according to a modified method previously described. 15 In brief, 20 μl of cell culture supernatant was loaded into each lane and run on 10% SDS-PAGE gels (Novex Bis-Tris; Invitrogen, Paisley) under reducing conditions. Following electrophoresis, protein was transferred onto nitrocellulose (BioRad) and then blocked with 5% non-fat dry milk in 0.1% Tween TBS. Ponceau S solution was used to reveal uniformity of protein loading and transfer. A sheep antihuman stromelysin-1 polyclonal antibody (1:500 dilution, The Binding Site Ltd, Birmingham, UK) and a monoclonal mouse antihuman TIMP-1 antibody (1 μg/ml Oncogene Research, Nottingham, UK) were used primary antibodies. Rabbit anti-sheep or rabbit anti-mouse HRP conjugated antibodies (DAKO) were used as secondaries. Reactive bands were identified using the ECL plus kit (Amersham Biosciences, Little Chalfont, UK).
TNF was measured using the Quantikine® HS Human TNF ELISA immunoassay (R&D Systems, Abingdon, Oxford, UK) according to the manufacturer's instructions, and read using a spectrophotometric microplate reader capable of measuring absorbance at 490 nm. Minimum assay sensitivity was 0.5 pg/ml.
Cells were cultured in RPMI 1640 containing 100 μg/ml streptomycin, 100 U/ml penicillin, 50 μg/ml gentamicin, 1 mg/ml amphoterecin, and HL-1. Cells were resuspended at a concentration not exceeding 2 × 10 6 cells/ml in 12 well or 96 well culture plates according to experimental design. PWM and LPS (Sigma, Poole, UK) were used in some experiments at a concentration of 1 ug/ml. Cells were incubated for 24–72 h according to experimental protocol at 37 °C in a constant 5% CO 2 .
Fibroblast cell culture
Aliquots of isolated LPMCs were cultured in six-well plates in 1 ml Dulbecco's Modified Eagle Medium (DMEM)/Hams Nutrient F12 Mix, supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mg/ml amphoterecin and 2 mM glutamine (all from Invitrogen). Cells were grown with weekly media changes for 3–4 weeks after which cells were removed with trypsin-EDTA and placed in 25 cm 2 flasks (BD Biosciences, UK). Thereafter, cells were trypsinised when confluent and flasks reseeded at 100 cells/mm 2 . Cells were grown up to at least passage 8 before there were enough to use in the experiments and produced type I and IV collagens, laminin, and expressed smooth muscle cell actin, vimentin, and TGF-ß receptors I and II (not shown).
Thalidomide (Sigma), the immunomodulatory drugs CC-5013 and CC-4047, and the PDE4 inhibitor CC-10004 (kind gift of Celgene Corporation, New Jersey, USA) were all dissolved in dimethyl-sulphoxide (DMSO) to make stock solutions of 10 mg/ml. Immediately prior to use these were further diluted in 100% DMSO to the appropriate concentrations to be added to culture experiments. This resulted in a final concentration of 0.1% DMSO in all culture experiments unless otherwise stated.
Data was analysed using SPSS v12.0 (SPSS Inc, Chicago, Illinois) and SigmaPlot v8.0 (Systat Software UK ltd, London, UK). Differences between groups were compared using the Student's t test or the Mann–Whitney U test as appropriate. A level of p < 0.05 was considered statistically significant.
Thalidomide, CC-5013, CC-4047, and CC-10004 inhibit TNFα production by LPS stimulated PBMCs
Assessment of the effect of thalidomide, CC-5013, CC-4047, and CC-10004 on TNFα production by peripheral blood mononuclear cells (PBMCs) from normal volunteers was first undertaken. All four of the compounds significantly reduced TNFα production; with a mean reduction of 17% (CI − 26.6 to − 7.4; p = 0.02) over vehicle control (DMSO 0.1%). CC-4047 caused a mean reduction of − 71.7% (CI − 93.6 to − 49.8; p = 0.04), CC-10004 a reduction of − 81.5% (CI − 87.4 to − 75.6; p < 0.001), compound C − 64.5 (− 83.7 to − 45.3; p = 0.004) ( Fig. 1 A).
Only CC-10004 significantly reduced TNFα production by gut LPMCs
The effect of thalidomide, CC-5013, CC-4047, and CC-10004 on TNFα production by LPMCs was next investigated. LPMCs from normal patients were stimulated with PWM to induce TNFα production. PWM 1 μg/ml caused an approximately 10-fold rise in TNFα production ( p < 0.01). CC-10004 caused a mean reduction of 63.2% (CI − 51.1 to − 75.2; p < 0.001) in TNFα production. There was no significant difference in TNFα levels between LPMCs treated with thalidomide, CC-5013, and CC-4047 (compounds A and C) and the vehicle control ( Fig. 1 B).
CC-10004 inhibits production of MMP-3 by LPMCs from patients with CD and UC
The effect of thalidomide, CC-5013, CC-4047 and CC-10004 on the spontaneous production of MMP-3 from LPMCs of IBD patients was investigated next. CC-10004 significantly inhibited MMP-3 production by LPMCs from patients with CD or UC. In contrast, thalidomide, CC-5013, and CC-4047 did not inhibit MMP-3 production ( Fig. 2 ). Further studies revealed that CC-10004 significantly inhibited both LPMC MMP-3 RNA and protein production in a dose dependent manner at a concentration range of 0.1–50 μg/ml ( Fig. 3 A–C). No effect on LPMC TIMP-1 RNA was seen ( Fig. 3 D).
MMP-3 production by cultured myofibroblasts is not inhibited by thalidomide, CC-5013, CC-4047, or CC-10004
Myofibroblasts are the major source of MMP-3 in gut inflammation and are abundant in LPMC suspensions. Inhibition of MMP-3 in LPMCs may therefore result from a direct effect of the drugs on myofibroblast MMP-3 production, or indirectly through TNF-α inhibition. To distinguish between these two possibilities, the effect of thalidomide and analogues on myofibroblast MMP-3 production was investigated. Myofibroblasts were cultured for 48 h in serum free media in the presence or absence of thalidomide or analogues at a concentration of 10 μg/ml. All cultures including controls contained a final concentration of 0.1% DMSO. Myofibroblast production of MMP-3 was not inhibited by thalidomide or CC-10004, CC-5013, and CC-4047 ( Fig. 4 ).
In this study we have demonstrated that CC-10004 is an effective inhibitor of both PBMC and gut LPMC TNF-α production, and of gut LPMC MMP-3 production in an ex-vivo cell culture system. In contrast, though thalidomide, CC-4047 and CC-5013 were capable of inhibiting PBMC TNF-α production, they had no effect on gut LPMC TNF-α or MMP-3 production. This illustrates the fundamental difference in pharmacological activity between CC-10004, thalidomide, and the immunomodulatory imide compounds CC-4047, and CC-5013.
TNF-α is a key mediator in both UC and CD, and it is now well established that anti-TNF-α therapy is effective in the treatment of both diseases. 4 , 5 There is strong circumstantial evidence that in IBD TNF-α and other pro-inflammatory cytokines mediate tissue destruction through upregulation of matrix degrading MMPs such as MMP-3. In inflamed bowel, levels of TNF-α and IL-1β correlate strongly with levels of MMP-3, 11 while in cultured mucosal explants from patients with IBD, infliximab downregulates production of MMP-3 along with MMP-1 and MMP-9. 31 These studies suggest that MMP-3 inhibition is an important biological target in IBD, and may act as a surrogate marker for the promotion of mucosal healing. The clinical efficacy of thalidomide has been attributed to its ability to inhibit production of TNF-α production by enhancing the degradation of TNF-α mRNA, 32 though this remains unproven. Bauditz et al reported that in short term culture experiments, thalidomide inhibited production of both TNF-α and IL-12 from pokeweed mitogen (PWM) stimulated gut LPMCs. 23 In contrast, we found that though thalidomide inhibited production of TNF-α by LPS stimulated PBMCs, it had no effect on TNF-α or MMP-3 production by PWM stimulated gut LPMC in short-term culture. This is in keeping with the study by Prehn et al, who also found that thalidomide had no inhibitory effect on TNF-α production by anti-CD2 stimulated gut LPMC. 29 Taken in conjunction with other in-vitro studies, these findings indicate that cell type, activation status, and stimulus all influence the ability of thalidomide to inhibit cellular TNF-α production. Overall, the interpretation of these results is that if thalidomide is active in IBD, the mechanism of action is not through inhibition of mucosal LPMC TNF-α production. Thalidomide does have multiple other immunological properties including the ability to inhibit NFκB, 33 enhance production of IL-4 and IL-5, 34 induce apoptosis in monocytes, 35 and inhibit angiogenesis. 36 , 37 It is possible that these play a role in its mechanism of action though this requires further evaluation. However, it is also possible that the clinical effectiveness of thalidomide in IBD has been overstated. Reported response rates to treatment with thalidomide range from 64–84%, which though superficially impressive, have all been from small, uncontrolled case-series. 19 – 21 , 23 – 25 As a note of caution there are many examples in the literature where novel therapies for IBD (such as IL-10 and etanercept) have initially yielded encouraging results in uncontrolled pilot studies that have not been borne out in subsequent large randomised controlled trials.
In contrast, the discovery that the selective PDE4 inhibitor CC-10004 (apremilast), is a powerful in-vitro inhibitor of gut LPMC TNF-α and MMP-3 production is an exciting finding that provides a strong rationale for investigating its activity in-vivo in the treatment of IBD. Our findings are supported by those of a previous study by Prehn et al, who found that three different related PDE4 inhibitors all inhibited production of TNF-α by normal gut LPMCs, though to different degrees. 29 It is of particular interest that in our study, the IMiDs, CC-5013 (lenalidomide) and CC-4047 (pomalidomide) had no effect on gut LPMC TNF-α or MMP-3 production, despite all compounds inhibiting blood PBMC TNF-α production. This disparity suggests different mechanistic pathways in operation in gut LPMCs which are in a highly activated state, compared with resting PBMCs. CC-10004 inhibits TNF-α production through PDE4, while the inhibitory activity of the IMiDs is independent of this pathway. Thus, the pathway through which IMiDs inhibit PBMC TNF-α production, does not appear to play an important role in LPMC TNF-α production.
In recent years there has been increasing interest in the use of PDE4 inhibitors in a wide variety of inflammatory conditions including asthma, psoriasis, COPD and more recently rheumatoid arthritis and IBD. 38 Rolipram, a selective PDE4 inhibitor, have been shown to inhibit TNF-α and abrogate inflammation in a variety of animal models of colitis though its use has not been reported in human studies. 39 , 40 More recently, tetomilast, a novel thiazole compound that inhibits PDE4, has shown potential efficacy in a phase II study in the treatment of UC, even though the trial's primary endpoint was not met, and a phase III study is now underway. 41 However, in an earlier open label study, pentoxifylline, a weak PDE4 inhibitor was not found to be effective in the treatment of CD, though this study was uncontrolled and only contained 16 subjects. 42 One possible explanation for these seemingly contradictory results is that individual PDE4 inhibitors may vary in their relative ability to inhibit mucosal cytokine production. It is clear from the Prehn study that this can vary widely even between closely related compounds with CC-10004 being considerably more potent than the other two PDE4 inhibitors which were tested in the study. 29 CC-10004 has also been shown to inhibit vascular endothelial growth factor (VEGF) and IL-6 production by co-cultures of myeloma and endothelial cells, and to be capable of inducing myeloma cell apoptosis. 43 It is possible these properties may also play a role in IBD. Finally, the finding that CC-10004 appears to suppress MMP-3 production indirectly through inhibition of leucocyte TNF-α production, rather than through a direct effect on gut myofibroblasts, also provides a mechanistic basis to explain how TNF-α inhibition in IBD promotes mucosal healing.
The IMiDs®, immunomodulatory compounds, of which CC-5013 (lenalidomide) is the lead investigational compound, have diverse mechansims of action including TNF-α and other cytokine inhibition, inhibition of angiogenesis, inhibition of cell adhesion, induction of apoptosis, and stimulation of T-cells and NK cells. 44 Lenalidomide has already been proved effective clinically in the treatment of relapsed multiple myeloma, where it is thought to exert its effect through modulating cytokine production by bone-marrow stromal cells, reducing adhesion of malignant cells to the stroma, and inducing apoptosis or cell growth arrest. 45 Our findings that CC-5013 (lenalidomide) and CC-4047 (pomalidomide) do not inhibit gut mucosal TNF-α and MMP-3 production provides clear evidence that it is crucially important to identify the effect of investigational drugs on gut mucosa cytokine production rather that simply blood monocytes. Though lenalidomide is a powerful inhibitor of TNF-α production by PBMCs, our study findings that it does not suppress mucosal TNF-α production would predict that it would not be effective in the treatment of IBD. This helps explain the results of the recent randomised controlled trial of lenalidomide in the treatment of moderately severe active CD. 28 In this study 89 patients were randomised to receive lenalidomide 25 mg daily, lenalidomide 5 mg daily, or placebo. Response rates in both lenalidomide treatment groups were not significantly different from placebo at any point during the 12 week trial.
In summary, this is the first study to demonstrate that the PDE4 inhibitor CC-10004 (apremilast) effectively inhibits gut LPMC TNF and MMP-3 production in a dose dependant manner providing a clear rationale to investigate its clinical effectiveness as an oral agent in patients with IBD. This study illustrates a fundamental difference in the pharmacological activity between apremilast and the immunomodulatory imide drugs, lenalidomide and pomalidomide, and also provides a mechanistic reason why lenalidomide was shown to be ineffective in the treatment of CD.
JG carried out the studies and data analyses and drafted the manuscript. JG, JP, CT, KP, and AD carried out the sample analyses. JG, SP, and TT participated in the design of the study and performed the statistical analysis. JG, PM, and TT conceived the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
JG was in receipt of an NHS R&D training fellowship during the period of this study. The immunomodulatory drugs CC-5013 and CC-4047, and the PDE4 inhibitor CC-10004 were kind gifts of the Celgene Corporation, New Jersey, USA.
fetal calf serum
inflammatory bowel disease
lamina propria mononuclear cell
nuclear factor kappaB
peripheral blood mononuclear cell
tissue inhibitor of metalloproteinase