The mechanisms underlying the chronic intestinal inflammation that is a hallmark of inflammatory bowel diseases (IBD) are complex. Components of the pathological response include the adaptive and innate immune systems, as well as the intestinal epithelium and endothelium. Advances in the understanding of the roles of each of these components have resulted in the development of multiple biological agents that all represent an alternative to the use of current therapies in patients with refractory Crohn's disease or ulcerative colitis. This study systematically reviews the mechanisms of action, efficacy and safety of new and emerging therapies that are currently in clinical trials and discusses future directions in the treatment of IBD.
- Experimental colitis
- IBD basic research
- IBD clinical
- inflammatory bowel disease
- mucosal immunity
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The inflammatory bowel diseases (IBD) are characterised by chronic intestinal inflammation involving a pathological response in both the innate and adaptive immune systems. The pathogenesis of both types of IBD is caused by the interplay of many factors including genetic susceptibility, the external environment, infectious agents, the commensal enteric flora and immune system dysfunction.1 2 The diverse array of possible triggers and the complex immune response that results offers many therapeutic targets, which is reflected in the wide array of drugs for IBD that have been developed, are in development, or have failed to demonstrate efficacy.
The most popular approach to IBD has been to target the surplus or excessive activity of the adaptive immune system using biological agents such as monoclonal antibodies against tumour necrosis factor alpha (TNFα). However, although treatment with these anti-TNFα agents is successful in many patients, only a third or less will achieve remission and many of those who do will eventually lose their response.3–5 Because of this fact and the life-long nature of the disease, new therapies need to be developed. In addition to the inhibition of elements of the adaptive immune system, other approaches to the treatment of IBD address the apparent impairment of the innate immune system observed in patients with Crohn's disease (CD),6 7 boosting the innate immune system and blocking the infiltration of leucocytes into the intestinal mucosa.8 New molecules include either biological agents or small molecules, which both have advantages or not. Compared with small molecules, biological agents may seem more attractive because of the potency of action and their selectivity, which might include a limited or almost absence of off-target toxicity. However, their limitations are immunogenicity, extremely high costs and the long half life, which has to be taken into account in particular when they carry immunosuppressive activity.
This review comprehensively describes the mechanism of action, efficacy and safety of new biomolecules under development for CD and ulcerative colitis (UC). Tables 1 and 2 and figures 1–3 summarise all the new treatments in development for IBD.
An electronic search of English publications in the databases of PubMed up to August 2011 and Embase up to August 2011 was performed for the keywords: ‘Crohn's disease’, ‘ulcerative colitis’, ‘inflammatory bowel disease’, ‘treatment’, ‘biological therapy’, ‘cytokine’, ‘T cell’, ‘adhesion’, ‘growth factors’, ‘biomolecules’. Also a hand-search of abstracts from the yearly meetings of Digestive Disease Week and United European Gastroenterology Week between 2003 and 2011 was performed. In addition, clinical trials status was checked on http://www.clinicatrials.gov and new drug names were also searched and matched on google and on the website of the pharmaceutical companies developing new drugs.
Targeting the adaptive immune system
Extensive studies into T-cell polarisation in IBD have attempted to ascribe an easily addressed pathological mechanism to both CD and UC. Research shows that the patchy transmural inflammation of the gastrointestinal tract characteristic of CD is associated with the activation of types 1 and 17 T-helper (Th) cells in response to the production of interleukin (IL)-12, IL-18, IL-23 and transforming growth factor beta by antigen-presenting cells (APC) and macrophages. In turn Th1 and Th17 cells increase secretion of the pro-inflammatory cytokines IL-2, IL-17, interferon (IFN)-γ and TNFα. These cytokines feed into a self-sustaining cycle whereby they stimulate APC, macrophages, fibroblasts and endothelial cells to produce TNFα, IL-1, IL-6, IL-8, IL-12 and IL-18.2 9–12
In contrast, the uniform mucosal inflammation of the colon found in patients with UC is suggested to be associated with a Th2 immune response mediated by IL-4, IL-5 and IL-13 that results in an increase in levels of IL-13.13 14 In both scenarios T cells are also activated by direct contact with antigens presented on the surface of APC and can be regulated by anti-inflammatory cytokines such as IL-10. Therefore, the majority of new biomolecules for IBD aim to reduce pathogenic T-cell activation and its effects by inhibiting the actions of pro-inflammatory cytokines, increasing anti-inflammatory cytokines, blocking T-cell co-stimulation, or inducing T-cell apoptosis.
Blockade of pro-inflammatory cytokines
The development of anti-TNFα therapies such as infliximab, adalimumab and certolizumab pegol have made a significant difference to the health-related quality of life of many patients with IBD, but among primary responders only a third will maintain remission after 1 year, and the risk of later surgery is still high in patients receiving anti-TNF therapy.5 15 The development of anti-TNFα therapies continues with novel TNF inhibitors such as golimumab, dersalazine, HMPL-004 and ozoralizumab (ATN-103) currently in various phases of the clinical trial process, but researchers are also exploring other avenues. One such closely related avenue is the development of a vaccine against TNFα consisting of the TNFα derivative TNFα kinoid (Debio-01512; Novacs, Paris, France), a complex of inactivated human TNFα and the carrier protein keyhole limpet haemocyanin.16 Having shown promise in animal models of chronic inflammation and rheumatoid arthritis (RA),17 18 TNFα kinoid was moved into phase I/II clinical trials in patients with moderate to severe CD. A recent report of the results of the trial indicates that TNFα kinoid was well tolerated with no serious adverse effects and with 76% of patients showing a clinical response and 43% achieving remission.19 20
In addition to developing alternative ways to target TNFα, researchers have also turned their attention towards targeting other pro-inflammatory cytokines such as IL-12/IL-23, the α-chain (CD25) of the IL-2 receptor (IL-2R), and IFN-γ. Apilimod mesylate, ustekinumab, briakinumab (ABT-874) and SCH-900222 all act to inhibit the effects of IL-12/IL-23 and have shown limited results in early phase clinical trials. Apilimod is a small molecule inhibitor of the transcription of IL-12 and IL-23, ustekinumab and briakinumab both target the p40 subunit common to both IL-12 and IL-23, whereas SCH-900222 targets the p19 subunit that is specific to IL-23. An initial open-label dose-escalating trial with apilimod mesylate in patients with active CD indicated that the drug demonstrated clinical activity and was well tolerated, with the most common adverse events being dizziness, nausea, headache and fatigue.21 However, a subsequent randomised controlled trial showed that apilimod did not have a significantly greater efficacy than placebo treatment.22
Ustekinumab was also subjected to phase II trials in patients with moderate to severe CD with slightly more promising results. Following 6 weeks of treatment, clinical response rates were significantly higher for the patients given ustekinumab at 53% compared with 30% for placebo. However, at week 8 the response rate was not significantly different, at 49% for ustekinumab and 40% for placebo. The long-term response to the drug is unknown but it is interesting that a subgroup of patients who had previously responded to infliximab had a greater response to ustekinumab.23 This effect was examined in a further trial in which 59% of patients who had previously been treated with infliximab responded to ustekinumab in comparison with 26% of patients who received placebo. The same trial also indicated that patients with higher baseline C-reactive protein (CRP) levels tend to have a greater response to ustekinumab.24 A more recent report of a multicentre, larger phase IIB trial in patients with moderate to severe CD who were unresponsive to infliximab indicated that at week 6 39.7% of patients treated with ustekinumab showed a clinical response compared with 23.5% of patients administered placebo (p<0.05), but the rates of clinical remission were similar. Those patients who responded to ustekinumab by week 6 were given maintenance therapy and at week 22 41.7% of patients achieved remission compared with 27.4% of patients given placebo. In addition, 69.4% of patients treated with ustekinumab had a clinical response at week 22 compared with 42.5% administered placebo, indicating that a significant number of patients who respond to ustekinumab remain in remission.25 In all three trials no serious adverse events were reported.23–25 Overall, given the results of those trials, it seems that ustekinumab may be particularly useful in patients who have previously failed to respond to anti-TNF therapy.
A phase I study of the response to briakinumab in patients with active CD showed that by week 8 the drug produced a response in 75% of patients compared with 25% given placebo. By week 18, however, the difference in response rates was no longer significant.26 A more recent study in patients with moderate to severe CD also concluded that briakinumab was not effective for the induction or maintenance of remission.27 In both trials the rates of serious adverse events were comparable between briakinumab and placebo.26 27 In the more recent study one patient treated with briakinumab died of respiratory failure due to pancreatitis, although this was thought to be unrelated to treatment.27 The final compound targeting the IL-12/IL-23 pathway, SCH-900222, is very early in development with the only trial in progress being a phase I/II dose-finding study in psoriasis.28
IL-2 is a crucial cytokine for T-cell activation and proliferation, and has been investigated as a very attractive target for therapeutic intervention. Inhibitors of IL-2R include basiliximab and daclizumab, which are both monoclonal antibodies against CD25. An initial study with basiliximab in 10 patients with steroid-resistant UC indicated that nine out of 10 patients achieved clinical remission in 8 weeks. Eight of the nine responders then relapsed but remission was re-achieved with corticosteroids and azathioprine. At 24 weeks, seven patients were in full remission and tests indicated that patients were rendered steroid sensitive by the presence of basiliximab.29 A further open-label trial, in which patients with steroid-resistant UC were given a single dose of basiliximab in combination with steroids, resulted in 65% of the patients achieving clinical remission at week 24.30 Adverse events noted in the two trials included two cases of herpes zoster and two cases of fever that were resolved with treatment, along with other minor adverse events that did not require treatment. As there was no placebo control group it was not possible to determine if the events were attributable to the administration of basiliximab.29 30 In contrast to the results with basiliximab, although an initial open-label study of 10 patients with refractory UC indicated that daclizumab produced some clinical benefit,31 a larger randomised controlled trial in patients with moderate UC showed that those treated with daclizumab were not more likely to be in remission after 8 weeks of treatment compared with patients given placebo.32 Adverse events were few with daclizumab, with the most common being nausea in patients simultaneously treated with azathioprine31 as well as nasopharyngitis and fever.32
Fontolizumab is the only biological therapy targeted against IFN-γ currently being tested in patients with IBD. An initial study in 45 patients with moderate to severe CD indicated that fontolizumab appeared to have a biological effect and was reasonably well tolerated, although there were slightly more reports of chills, flu-like symptoms and asthenia compared with placebo as well as two cases of worsening of CD symptoms.33 In a further study in 133 patients with CD, fontolizumab produced a clinical response in patients, but only after two doses given a month apart, with an increased response rate (69%) seen at day 56.34 Analysis of baseline CRP levels also indicated that elevated CRP was associated with pronounced clinical benefits. Fontolizumab was also well tolerated in this study and there were no evident patterns of adverse events attributable to fontolizaumab.34 A more recent study, in which the primary endpoint was clinical response on day 29, failed to show any significant clinical benefits in patients with CD treated with fontolizumab compared with placebo. However, at time points beyond 29 days patients treated with fontolizumab showed a significantly greater improvement in the CD activity index score and CRP levels,35 suggesting that the actions of fontalizumab are more gradual.
In addition to targeting IL-12/23, IL-2R and IFN-γ, more recent approaches have expanded to address almost all the pro-inflammatory cytokines shown to be elevated in IBD. Biological agents now entering phase I trials include monoclonal antibodies against IL-6, IL-6R, IL-13, IL-17, IL-18 and IL-21. As a pleiotropic cytokine IL-6 contributes to Th17 differentiation, and increased levels of IL-6 and soluble IL-6R are associated with increased disease severity in IBD.36 A polymorphism within the IL-6 gene has also been linked with early-onset CD,37 and persistent activation of the IL-6 signalling pathway plays a role in the development of colon cancer.38–40 Biological therapies targeting IL-6 include C326, currently undergoing a phase I trial in patients with CD;28 sirukumab (CNT0136), which has just been shown to be well tolerated in healthy subjects;41 CDP6038, currently being tested in patients with RA;28 and PF-04236921, which is going to be used in a phase I/II trial that is currently recruiting patients with CD who are unresponsive to anti-TNFα.28 In contrast, tocilizumab is a monoclonal antibody against IL-6R that has been through extensive trials in patients with RA,42 43 and has been approved as a second line monotherapy for patients with RA who have failed other approved therapies.44 Trials of tocilizumab in patients with IBD have been limited. The only trial completed to date was a placebo-controlled phase I study in 36 patients with active CD. The results of the study indicated that 80% of the patients given biweekly intravenous infusions of tocilizumab for 12 weeks had a clinical response compared with 31% of the placebo-treated patients, but only 20% of the patients given tocilizumab went into remission.45 Although the rates of adverse events were similar between tocilizumab and placebo in the study in patients with CD,45 the larger studies in patients with RA indicate that tocilizumab is associated with an increase in the incidence of reversible grade 3 neutropenia, elevated lipids, abnormal liver function and an increased risk of infection.42 43 These adverse events are generally mild and can be resolved with treatment but the longer term effects of increased lipids and abnormal liver function in response to tocilizumab need to be investigated.42 43
The cytokine IL-13 is produced by naive T cells and activates natural killer T (NK-T) cells, which then also produce IL-13. The IL-13 produced by NK-T cells has been shown to be pathogenic in UC as it impairs the function of the epithelial barrier and also causes apoptosis of epithelial cells.13 46 47 Inhibition of IL-13 production with the administration of IFN-β1a also results in the suppression of inflammation in UC,48 which makes IL-13 an appropriate target for biological therapy in IBD. Some researchers have argued that enhancing IL-13 activity would also result in a therapeutic response in IBD as IL-13 suppresses the activity of the pro-inflammatory cytokines IFN-γ and IL-17 in murine models.49 Given the weight of the evidence that IL-13 plays a significant role in the pathogenesis of UC,13 46 48 it is unlikely that enhancing the levels of IL-13 will have a therapeutic effect in UC.47 Monoclonal antibodies against IL-13 currently in clinical trials for IBD include anrukinzumab, currently in a phase II trial in patients with UC,28 and QAX567, currently being used in phase I/II trials in patients with CD.28
Similar to IL-6 and IL-13, monoclonal antibodies have also been developed to target IL-17, IL-18 and IL-21, which have all been shown to be present at elevated levels in the inflamed intestinal mucosa of patients with IBD.50–55 AMG827 and secukinumab (AIN457) both target IL-17, GSK1070806 targets IL-18, and ATR-107 (PF-05230900) targets IL-21, and all are currently in phase I/II clinical trials either in patients with CD or healthy volunteers.28 Of these therapies initial data are only available for secukinumab. In a double-blind, placebo controlled, proof-of-concept study in 59 patients with CD patients given secukinumab had disease worsening compared with placebo. In addition, there were some data to suggest that the inhibition of IL-17A exacerbated CD in a subset of patients and increased infection risk.56 The reason for such unexpected and opposite clinical effects are obscure because in psoriasis and RA blockade of IL-17A has been beneficial in proof-of-concept studies, but probably an effect of IL-17A in protecting barrier integrity and regulatory T-cell function is crucial in IBD compared with other inflammatory diseases.57 On the contrary, vidofludimus (4SC-101/SC12267), a small molecule inhibitor of the release of IL-17, has shown promise. Results from the single-arm, open-label ENTRANCE study indicate that vidofludimus is safe, well tolerated, with only mild to moderate drug-related adverse events reported, and may be useful in maintaining clinical remission. Following the administration of vidofludimus for 12 weeks in 26 steroid-dependent patients in remission with CD or UC 53.9% of patients remained in steroid-free remission, 34.6% were in remission at a lower corticosteroid dose than their baseline dose, and 11.5% showed no response.58 This suggests that attempts to modulate the function of IL-17 in IBD deserve more investigation.
In addition to inhibiting either cytokines or cytokine receptors an alternative means of reducing the inflammatory response in IBD is to block the downstream signalling pathways mediated by cytokines. As signalling molecules that interact with cytokine receptors the Janus kinases (JAK), JAK1, JAK2 and JAK3, play a crucial role in cell growth, survival, development and differentiation of immune cells. JAK1 and JAK2 are ubiquitously expressed but JAK3 is found only in haematopoietic cells and is part of the signalling pathways activated by IL-2, IL-4, IL-7, IL-9, IL-15 and IL-21.59 Tofacitinib (CP-690,550) is a first-in-class small molecule inhibitor of JAK360 currently in phase II clinical trials for various immune-related disorders including CD and UC. In-vitro studies of the activity of the drug indicate that it interferes with Th2 and Th17 cell differentiation, blocks the production of IL-17 and IL-22 and offers the possibility of effective suppression of the pathological immune response present in IBD.61 Recent results from a multicentre, double-blind, placebo controlled study in patients with moderate to severe active CD show that although tofacitinib is well tolerated it produced no clinically significant response following 4 weeks of treatment compared with placebo.62 Tofacitinib was, however, associated with a significant decrease in CRP and fecal calprotectin levels.62 In contrast, a study in patients with moderate to severe UC showed that 8 weeks of treatment with tofacitinib was associated with a dose-dependent improvement in both clinical response and remission rates.63 In trials both in CD and UC the overall incidences of adverse events were similar between patients administered tofacitinib and those given placebo. However, as has been seen to a lesser degree with tocilizumab, there was a dose-dependent increase in low-density lipoprotein cholesterol observed in patients treated with tofacitinib,62 63 which will require further investigation to determine the long-term effects.
Administration of anti-inflammatory cytokines
An alternative and logical approach to reducing the inflammation present in IBD would be to restore the balance between pro-inflammatory and anti-inflammatory cytokines by increasing the amount of anti-inflammatory cytokines present in the body. Anti-inflammatory cytokines that have been developed as biological therapies for IBD include IL-10, IL-11 and IFN-β. Unfortunately the results from clinical trials administering these compounds have not been very useful.
IL-10 first came to light as an anti-inflammatory cytokine when it was shown that IL-10-deficient mice spontaneously develop colitis.64 Since then it has been shown that IL-10 decreases intestinal inflammation65 and that variants of the IL-10 gene are linked to susceptibility to developing UC.66 This led to the development of recombinant human IL-10 that could be administered subcutaneously. An initial trial in 46 patients with steroid-resistant CD showed that 50% of patients administered IL-10 achieved remission compared with 23% in the placebo group.67 Unfortunately, further larger clinical trials showed that although the therapy was well tolerated and there was a tendency towards clinical improvement, remission was not achieved in any patients included in the studies. The only notable adverse effects observed in the two trials were dose-dependent, asymptomatic and reversible anaemia and thrombocytopaenia.68 69 IL-10 therapy was also tested in patients with CD who had undergone ileal or ileocolonic resection, but treatment did not prevent the recurrence of inflammation when compared with placebo, even when IL-10 was given daily.70 One possibility for the failure of the treatment may be due to the method of drug delivery, with systemic administration of IL-10 resulting in very little of it reaching the site of inflammation. A novel way of bypassing the systemic circulation has been developed using oral administration of genetically modified IL-10 producing Lactococcus lactis to target the cytokine directly to the gut mucosa. This approach has been shown to be safe,71 and has gone into phase II clinical trials, which showed lack of clinical efficacy. Similar to IL-10, recombinant human IL-11 has also been developed for subcutaneous administration. Although initial trials showed evidence that the approach was well tolerated and produced a response in more than 33% of patients with CD,72 73 when it was compared with prednisolone, IL-11 produced a significantly inferior response rate and rate of short-term remission.74 The most significant adverse effect noted in the three trials of IL-11 was a dose-dependent increase in platelet count,72 73 along with mild adverse events such as fever, rash, nausea/vomiting and headache.74
The third approach using anti-inflammatory cytokines is that of the administration of IFN-β, which has been tested both in patients with CD and UC. IFN-β was first administered subcutaneously to patients with steroid-refractory UC, in which it produced an 88% remission rate with a mean length of remission of 13 months.75 Subsequent open-label and double-blind, placebo controlled studies in patients with UC produced a variety of results, with some studies indicating that IFN-β did have a therapeutic effect48 76 77 while others reported that IFN-β did not produce a significant therapeutic effect compared with placebo.78 79 It is interesting to note that the positive reports of an effect of IFN-β were from open-label trials, whereas the negative results were from randomised controlled trials. In randomised controlled trials higher rates of remission were always observed with IFN-β compared with placebo, but rates of remission in patients given placebo were also high enough to confound the results. A more recent trial with IFN-β in patients with CD also came to the conclusion that there was no difference between the effects of the administration of IFN-β and placebo.80 There is some suggestion that response to IFN-β is dependent on the levels of circulating IL-17 and IL-6, with increased levels of the cytokines being associated with no response to IFN-β.48 An additional complication of IFN-β therapy is that there have been reports that the administration of the cytokine in patients with multiple sclerosis was associated with the development of IBD.81 Furthermore, recent investigations in a mouse model of colitis indicated that the administration of IFN-β exacerbated colitis.82 Data from all the clinical trials conducted so far in patients with IBD do not confirm this suggestion, with the most common adverse effect being flu-like symptoms and no exacerbation of IBD symptoms reported.48 75–80
Blockade of T-cell stimulation and induction of T-cell apoptosis
During the process of physiological inflammation the population of activated T cells that responds to pathogens is constantly balanced by the occurrence of an equal amount of proliferation and apoptosis. During chronic pathological inflammation proliferation of T cells far outweighs apoptosis of T cells.1 2 One approach to controlling pathological inflammation would be to increase the T-cell apoptosis that occurs in response to a lack of antigen-specific and costimulatory signals from APC.83 Visilizumab is a monoclonal antibody against the CD3 chain of the T-cell antigen receptor (TCR), blockade of which leads to T-cell apoptosis. An initial phase I study in patients with severe steroid-resistant UC reported that all the patients achieved remission with the administration of visilizumab.84 This was followed by a further study in which 84% of patients with steroid-refractory UC achieved a clinical response.85 However, although a phase I/II dose escalation trial of visilizumab in severe UC showed evidence of clinical response, it also resulted in 100% of patients reporting adverse events including abdominal abscess, atrial fibrillation, cytomegalovirus infection and herpes zoster.86 Following on from that study a placebo controlled trial indicated that visilizumab was not effective in patients with thromboembolic adverse events.87 A trial in patients with CD also indicated that visilizumab is associated with cytokine release syndrome that results in transient liver injury.88 Therefore, although biologically attractive, excessive T-cell targeting often leads to serious adverse events.
As an alternative to targeting the proliferation of T cells, rituximab targets the CD20 component of the TCR on B cells with the aim of inducing apoptosis in circulating B cells. A phase II randomised controlled trial of rituximab in 24 patients with steroid-refractory moderate UC showed that although there was some clinical response with rituximab the rate of induction of remission was no different from placebo. In addition, any observed response was short term and was only maintained to week 12 following the administration of rituximab.89 Similar to the use of visilizumab, some severe adverse reactions have been reported with rituximab although they have generally been seen in other disease states. In isolated cases, including a patient with follicular lymphoma,90 a patient with Grave's disease91 and a child with nephrotic syndrome,92 the administration of rituximab was associated with the development of UC. In one patient with refractory UC, rituximab caused an exacerbation of symptoms that was associated with the depletion of IL-10-producing B cells.93 Although few adverse reactions were reported in the trial of Leiper et al,89 with the most notable being one chest infection and three mild infusion reactions, it remains to be seen if, during more prolonged larger studies, targeting TCR on B cells leads to the same severe reactions as it does with T cells.
In addition to TCR–antigen interaction T cells require separate costimulatory signals for full activation. CD28, which is expressed on T cells, interacts with its ligands CD80 and CD86 on APC, but the cytotoxic T-lymphocyte-associated antigen 4 can interrupt this interaction to induce T-cell anergy and apoptosis. Abatacept is a soluble recombinant fusion protein containing cytotoxic T-lymphocyte-associated antigen 4 and IgG1 that is approved for use in RA and is currently being trialled for IBD. However, trials both in patients with CD and UC at doses superior to those approved for RA indicated that although abatacept was well tolerated with no significant difference in adverse effects compared with placebo, the rate of remission with abatacept was not significantly different either.94 95
These disappointing results including the many failures in the attempt at blockade of T-cell function in IBD are difficult to interpret. However, one could speculate that general blockade of T cells is not key or could even be detrimental if it includes T-regulatory cell blockade, and perhaps specific blockade of the pathogenic T cells secreting inflammatory cytokines should be aimed at.
Enhancing the innate immune system
The innate immune system mounts a non-specific defence against pathogens that relies on monocytes, dendritic cells and granulocytes. One hypothesis for the development of CD is that patients have an impaired acute inflammatory response that results in delayed clearance of bacteria that penetrate the gut wall.6 7 Evidence for this comes from the identification of mutations in the gene for the bacterial sensing protein nucleotide-binding oligomerisation domain containing 2 (NOD2) as being associated with CD.96 97 Three particular variations of the NOD2 gene are associated with up to a 40-fold increase in the likelihood of developing ileal CD. Expression of the NOD2 protein controls the expression of antimicrobial peptides including the defensins, another key immune molecule that has been identified as being reduced in CD.98 The therapeutic implication of these defects in innate immunity is that an entirely new therapeutic approach can be envisioned whereby the acute immune response can be boosted by the stimulation of innate immunity, rather than immune suppression for a selected subgroup of patients. The ideal drug would stimulate the innate immune system to produce defensins against antimicrobial peptides and also induce tolerance.
Stimulators of the innate immune system that have been tested so far include granulocyte colony-stimulating factor (G–CSF) and granulocyte monocyte colony-stimulating factor (GM–CSF). Filgrastim and lenograstim are recombinant human forms of G–CSF and sargramostin is a recombinant human form of GM–CSF, all are currently or have been subject to clinical trials. In a pilot study of G–CSF in five patients with clinically inactive CD, but with severe endoscopic ileitis following intestinal resection, it was associated with mucosal healing in two of the five patients. The time course of healing covered 3–9 months but there were no adverse effects and neutrophil counts and IL-1RA levels were observed to increase.99 A slightly larger open-label trial of filgrastim in patients with active CD produced a clinical response in 55% of patients and remission in 25%, with the only notable adverse effect being mild bone pain.100 Currently no further results of the effects of G–CSF have been published.
Sargramostin (GM–CSF) was initially tested in an open-label study with 15 patients with moderate to severe CD, which indicated that GM–CSF produced a clinical response and was well tolerated.101 A subsequent placebo controlled study indicated that sargramostin therapy decreased disease severity and improved quality of life in patients with active CD.102 More recently, a phase II trial in patients with steroid-resistant CD indicated that sargramostim resulted in a significantly higher rate of remission (18.6%) compared with placebo (4.9%).103 An open-label study in paediatric patients with moderate to severely active CD also produced encouraging results, with 88% of patients achieving remission or response with few side effects beyond mild injection site reactions.104 In contrast to these trials a phase III multicentre double-blind, placebo controlled study in 286 patients with active CD showed no significant difference in clinical efficacy between sargramostim and placebo.105 In all five trials the adverse effects were generally mild to moderate, with bone pain and injection site reactions the most commonly reported events along with elevated neutrophil counts necessitating dose reduction.101–105 However, in the trial by Korzenik et al102 serious adverse events related to sargramostim treatment included a case of migraine, a man with ischaemic heart disease who developed anorexia, weakness and lethargy, and a woman who developed transient right-sided weakness consistent with a possible demyelinating event.
Preventing leucocyte infiltration of endothelium
Once T cells and neutrophils have been activated they migrate from the systemic circulation into the intestinal mucosa. The process of leucocyte infiltration is governed by the expression of integrins and chemokine receptors (CCR) on leucocytes and adhesion molecules such as intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), and mucosal addressin cell adhesion molecule 1 (MAdCAM-1) on endothelial cells,106 107 molecules that are upregulated in IBD.108 A major new approach to biological therapies in IBD has been the development of inhibitors of the various elements in the leucocyte infiltration process. The predominant targets of the novel biological agents in this group are the integrins α4β1, α4β7 and α2β2, which interact with VCAM-1, MAdCAM-1 and ICAM-1, respectively, to mediate the interactions between leucocytes and endothelial cells.109 Biological agents that target the integrins include the monoclonal antibodies natalizumab and ELND-004, which target α4, vedolizumab (MLN-02) and etrolizumab (RG-7413), which target α4β7, as well as rHuMab β7, which targets β7. In addition AJM-300 is a small molecule inhibitor of the α4 integrin subunit and ASP-2002 is an integrin antagonist. Other molecules that aim to prevent leucocyte infiltration include alicaforsen (ISIS-2302), an antisense oligonucleotide against ICAM-1 messenger RNA; PF-547659, an antibody against MAdCAM; and CCX282-B (GSK-1605786) and CCX-025, which both target CCR9, the key chemokine receptor in the targeting of leucocytes to the intestinal mucosa.110
The large number of these types of compounds in development is due to the success of natalizumab, which has been approved for use in CD as a second-line therapy,111 and represents the only biological therapy for CD available in the clinic that does not target TNFα. Natalizumab was initially tested in patients with CD a decade ago in a placebo controlled trial with 30 patients. The results of that first study indicated that 39% of the patients achieved remission at week 2 following treatment compared with only 8% of patients administered placebo.112 In the final phase before approval, 509 patients with moderate to severe CD were enrolled in a double-blind, placebo controlled trial that indicated that 48% of patients administered natalizumab achieved a clinical response that was sustained through 12 weeks compared with 32% of patients who received placebo. Sustained remission was achieved in 26% of patients administered natalizumab compared with only 16% of patients administered placebo.113 In addition, a recent meta-analysis of all controlled trials of natalizumab concluded that the therapy was superior to placebo in inducing remission of CD.114 Despite the success of natalizumab, it remains a second-line therapy for use in patients with treatment-refractory CD due to the increased risk of infection with the opportunistic human polyoma JC virus, which results in progressive multifocal leukoencephalopathy (PML), a risk that may be common to other adhesion molecule inhibitors. Patients undergoing treatment with natalizumab also have to enrol in a monitoring programme to ensure that the risks associated with PML are minimised.111 PML associated with natalizumab was first reported in a patient with multiple sclerosis in 2005115 followed quickly by a case in a patient with CD.116 Since then much research has been done into the risks associated with developing PML, and it is estimated that one in 1000 patients treated with natalizumab will develop PML.117 Of a group of 35 cases that have been documented in patients with multiple sclerosis treated with natalizumab, 71% survived with a shorter time from symptom onset to diagnosis, localised disease and younger age associated with improved survival suggesting the need for enhanced clinical vigilance.118
The remainder of the leucocyte infiltration inhibitors are in various stages of development. Vedolizumab is currently in phase III trials for CD and UC28 following favourable results from previous trials. In a phase II trial assessing the efficacy and safety of vedolizumab in 181 patients with active UC, remission rates were 32% and 14% for vedolizumab and placebo, respectively, at week 6, and no serious adverse events related to vedolizumab were observed.119 In contrast, a trial in a similar sized group of patients with CD over a longer time period indicated that there was no difference in response at 2 months following treatment with vedolizumab or placebo and the most common serious adverse event was worsening of CD.120 However, significant efficacy was observed in the remission rates in the 2 mg/kg group, and therefore trials of vedolizumab in patients with CD continue.
In addition, the PML risk should be absent with vedolizumab. Indeed, in a recent report carried out in experimental autoimmune encephalomyelitis in rhesus has clearly shown that natalizumab is therapeutically effective in reducing central nervous system inflammation while vedolizumab did not have any clinical effect. Overall, these results suggest that vedolizumab does not impair immune surveillance of the central nervous system and thus may have a lower risk of predisposing IBD patients to PML than natalizumab.121
Alicaforsen targets ICAM-1 by preventing the translation of the protein.122 In an initial pilot trial in 15 patients with chronically active CD alicaforsen produced a rapid and persistent clinical response and was well tolerated.123 However, in two further, larger multicentre trials the administration of alicaforsen to patients with steroid-dependent moderate CD had no greater overall effect than placebo.124 125 Interestingly, a closer examination of the patients in the second of the two multicentre trials did indicate that those patients who had a greater exposure to the drug, as calculated by the concentration of the drug in their plasma over time, did in fact show consistent improvements in clinical measures.124 Although this result showed promise further trials did not produce more useful results.126 One suggestion for the lack of efficacy for alicaforsen was that the drug was unable to penetrate the intestinal tissue from the bloodstream. With this in mind, an enema formulation was developed and administered in an open-label study in patients with UC in which it produced a 46% reduction in disease activity.127 Unfortunately, a further phase II study in patients with UC failed to show significant differences between alicaforsen and placebo after 6 weeks of treatment. Alicaforsen enema did, however, produce a prolonged reduction in disease activity relative to baseline from weeks 18 to 30 compared with placebo.128 In all of these trials no difference in adverse events was observed compared with placebo, with the exception of mild injection site reactions when alicaforsen was administered subcutaneously.122–128
As a small molecule inhibitor of CCR9129 CCX282-B has recently been shown to be well tolerated and safe with no drug-related adverse events in a 36-week maintenance study in patients with moderate to severe CD. In particular, no serious or opportunistic infections were observed.130 The results of the double-blind, placebo controlled study indicated that following the induction of remission by CCX282-B, remission was maintained for 36 weeks in approximately 50% of patients, whereas it decreased progressively in patients maintained on placebo. In addition, 19% of subjects administered CCX282-B had normal levels of CRP compared with only 9% of subjects given placebo, and fewer patients given CCX282-B required steroid rescue therapy.130 This suggests that CCX282-B may be a useful adjunct therapy for the maintenance of remission and the molecule has just been moved into phase III clinical trials.131
The potential of PF-547659, an anti-MAdCAM antibody, has only just begun to be tested with the results from a first in human study published recently and longer term studies still on going.28 In the placebo controlled first in human study 80 patients with UC were administered PF-547659 over 12 weeks with no obvious drug-related side effects. The overall response rates at week 12 were 42% for patients treated with PF-547659 and 21% for patients administered placebo. Remission was achieved in 22% of patients treated with PF-547659 and none of the patients given placebo.132 Similarly, only one trial of AMJ300 has been completed as has one trial of rHuMab β7. The results of the AMJ300 study indicated that this was well tolerated in patients with CD but that there was no significant difference in clinical response compared with placebo. However, among a subset of patients with a high initial CD activity index, AJM300 did produce a significant response suggesting that a larger study is needed to examine the effects of the drug.133 Like PF-547659, the initial trial of rHuMab β7 was carried out in patients with UC and the results showed that rHuMab β7 was well tolerated and that at day 71 10 out of 15 and three out 15 patients administered rHuMab β7 achieved clinical response or remission, respectively. This was compared with two out of three and none out of three patients who achieved clinical response or remission when administered placebo.134 The remaining biological agents, ELND-004, etrolizumab, ASP-2002 and CCX-025 are either still in preclinical development or are undergoing initial phase I trials with no clinical data available.
Cell-based therapies for IBD
Besides targeting specific molecular components of the adaptive and innate immune systems, a final method of addressing the pathological inflammation present in IBD is the use of cell-based therapies to stimulate or replace aberrant immune cells. To this end, several therapies involving stem cells or autologous immune cells are in development for the treatment of IBD. These therapies include remestemcel-L (Prochymal) and multistem, which are mesenchymal/haematopoietic stem cell based; PDA-001, which uses placenta-derived stem cells; and OvaSave, an autologous T-cell therapy. All of these approaches are currently in phase I–III trials in either CD, UC, or both.28
The results from previous trials of cell-based therapies in small groups of patients with IBD are promising. For example, the use of autologous non-myeloablative haematopoietic stem cell transplantation in 24 patients with severe CD resulted in remission in all the patients treated. Over a 5-year follow-up the percentage of patients who remained relapse free was 91% after 1 year, 57% after 3 years and 19% at the end of 5 years.135 Another study using mesenchymal stem cells had more mixed results, with only three out of nine patients showing a clinical response and three patients requiring surgery due to disease worsening.136 Stem cells have also been used specifically to address the problem of fistulas. In a small trial of 10 patients, intrafistular injections of mesenchymal stem cells resulted in sustained complete closure in seven patients as well as a reduction in CD activity.137 Similarly, the use of adipose-derived stem cells with fibrin glue in perianal fistulas resulted in healing in 71% of patients compared with only 16% healing in patients treated with fibrin glue alone.138 Despite the success of cell-based therapies, the complex, time consuming and expensive process needed to harvest, expand and transplant the cells makes it difficult to treat large numbers of patients. It is likely that cell-based therapy will be reserved for specific situations such as fistulas and as a last resort when all other treatment avenues have been tried.
Speculations for the future
To conclude the description of the current and potential therapies for IBD it is sufficient to say that although there are many avenues to explore, as yet no single ‘magic bullet’ has been discovered. Many paths have been followed only to lead to dead ends. What is particularly perplexing is that while basic science is important, it did not play a large role in the development of the anti-TNF agents, the most potent IBD treatments developed so far. Many ideas born at the bench do not translate to the bedside. This issue is most likely to stem from the complexity of the pathogenesis and progression of IBD, processes that have not been fully characterised. In addition, while most of the drugs are investigated for luminal disease, there is also an enormous need for further investigation of approaches to treat complications of CD such as fibrotic disease and fistulas.
During the course of IBD many molecules are upregulated or downregulated in patients with IBD, but this may have nothing to do with their biological significance and their potential as clinical targets. Similarly, while murine models of IBD have been useful in pointing towards crucial targets they do not fully describe the situation in a patient. Often many of the observations made in mice do not translate and we have very limited information on how the disease changes over time in humans. Animal models have several limitations, including reproducibility, and may sometimes provide contrasting results according to the model. In addition, although animal models of IBD have been extensively used to investigate drugs for translational purposes, some key drugs routinely used in the clinic such as corticosteroids or 5-aminosalicylic acid are totally ineffective in animal models. In addition, several drugs that have been incredibly effective in experimental colitis have been lost in translation. These observations should serve to remind investigators that animal models are not fully able to reproduce the complex spectrum of human IBD.
A patient who develops IBD is probably born with one or more of a range of genetic susceptibilities. The resultant mutated protein may then form the basis of an immunoregulatory imbalance. At some point in the patient's life there is a priming event that results in abnormal permeability and is eventually followed by a triggering event that results in microscopic disease. As the symptoms worsen a diagnosis of IBD is made and clinical disease becomes evident. At this point therapy is attempted, in some patients there will be a response, in others no response. There is then a lifetime of remission, recurrences, different medications, hospitalisations, operations, increased risk of cancer and so on. At each of these stages, the characteristics of the underlying inflammation are likely to be changing as the disease progresses. For example, the mucosal levels of IFN-γ, IL-12p40 and IL-12Rβ2 chain mRNA vary between early and late CD. IFN-γ is significantly higher in early stage disease compared with late stage disease. Equally, both IL-12p40 and IL-12Rβ2 levels are significantly higher in early stage CD.139 This has been clearly shown in paediatric IBD but data on adults are urgently needed, because knowing what could be the right cytokine to be targeted in the right disease stage could be the key for clinical success.
The huge potential for variation in inflammatory response from patient to patient and within a single patient over time suggests the need for a personal and integrated approach depending on the status of each patient. It is possible that the future of a personalised approach to IBD therapy would include computing gene and protein function. As this future is still far off, what is the best approach to treating patients at the current time to improve outcomes? It is also important to consider the fact that more than one therapeutic option will probably be necessary during the course of a patient's disease/lifetime, particularly as a host of cells and molecules are involved in IBD. Currently, such options are used in succession as a step-up approach rather than in combination. To date there has been limited investigation of a combination approach but a recent study looked at the effect of combining infliximab and azathioprine in comparison with either drug as monotherapy. The results of the Study of Biologic and Immunomodulator Naive patients in Crohn's disease (SONIC) indicated that patients with moderate to severe CD were more likely to achieve corticosteroid-free remission if they were treated with a combination of infliximab plus azathioprine rather than infliximab or azathioprine monotherapies. Consistent with this result, the SUCCESS trial in patients with UC also demonstrated that two drugs in combination, infliximab plus azathioprine, is better than infliximab or azathioprine alone,140 suggesting that inflammation can be more efficiently suppressed by simultaneously tackling more than one pathway.
Pending the development of new compounds, the current treatment strategies need to be optimised, as suggested by the results of the SONIC and SUCCESS trials. In particular, new outcome measures and endpoints need to be developed so that effective treatments can be delivered appropriately in order to alter the biological processes underlying IBD and not just control the symptoms. Even with the development of biological therapies for IBD, the long-term outlook for patients has not changed, with little evidence to suggest that current therapies alter disease progression and significantly reduce the long-term need for surgery.15 141 It is important, for example, to define what constitutes early CD so that early intervention can be attempted to prevent the progression from inflammatory to structuring/penetrating disease.142 Similarly, the recent development of an IBD disability index will give researchers a more structured way of evaluating the long-term effects of both IBD and treatment on patient functional status.143 In terms of disease progression, a scoring system to measure structural damage is also being developed so that disease progression can be plotted over time and monitored rather than just captured at one specific time point.144
Overall, much progress has been made in understanding the pathogenesis of IBD, and many potentially useful treatments are under development, but few are already available. What is required now is a more targeted, patient-based approach that makes use of genetic information and biomarkers both to direct the development of and deliver tailor-made therapy.
Time will tell us if the above directions of investigation will be successful for IBD patients.
The author is grateful to Yolande Chvatchko, Amanda Proudfoot, Denise Carte and Michele Laddaga for critical reading and illustration suggestions, and to Sarah. The author would also like to thank Sarah A De La Rue of Readable Science for her assistance with this manuscript.
Competing interests Silvio Danese has served as a speaker, a consultant and an advisory board member for Schering-Plough, Abbott Laboratories, UCB, Ferring, Cellerix, Millenium Takeda, Nycomed, Actelion, Astra Zeneca, Danone, Chiesi, Novo Nordisk, Merk-Serono and Cosmo Pharmaceuticals.
Provenance and peer review Commissioned; externally peer reviewed.