Int J Colorectal Dis (2003) 18:1–11
`DOI 10.1007/s00384-002-0431-6
`
`R E V I E W
`
`Alan Shand
`Alastair Forbes
`
`Potential therapeutic role for cytokine
`or adhesion molecule manipulation in Crohn’s
`disease: in the shadow of infliximab?
`
`Accepted: 11 July 2002
`Published online: 14 August 2002
`© Springer-Verlag 2002
`
`A. Shand · A. Forbes (✉
`)
`St Mark’s Hospital, Watford Road,
`Harrow, HA1 3UJ, UK
`e-mail: alastair.forbes@ic.ac.uk
`Tel.: +44-20-82354016
`Fax: +44-20-82354039
`
`Introduction
`
`Abstract Background: Inappropri-
`ate activation of the intestinal im-
`mune system leading to a persistent
`inflammatory response is implicated
`in the pathogenesis of both Crohn’s
`disease and ulcerative colitis.
`Various cytokines and cell surface
`molecules expressed by immuno-
`logically active cells are involved
`and are potential targets for inter-
`vention. To date the most effective
`biological agent has proved to be
`infliximab, but many other possibili-
`ties are being considered actively.
`Studies considered: Use of the im-
`munoregulatory cytokines IL-10
`and IL-11, and of antagonists to
`IL-12, IL-15 and IL-18, to integrins,
`
`and to ICAM-1 are considered.
`Results and conclusion: IL-10 and
`the anti-integrin natalizumab are cur-
`rently best evaluated. Both are effec-
`tive, but neither is yet a direct com-
`petitor for infliximab in Crohn’s dis-
`ease, and there is no strong evidence
`in ulcerative colitis. These agents
`have considerable potential which, in
`combination with new delivery sys-
`tems (perhaps taking into account
`some form of genetic engineering),
`hold great promise.
`
`Keywords Crohn’s disease ·
`Cytokine · Intercellular adhesion
`molecule 1 · Interleukin 10 ·
`Natalizumab · Ulcerative colitis
`
`Inappropriate activation of the intestinal immune system
`leading to a persistent inflammatory response has been
`implicated in the pathogenesis of both Crohn’s disease
`and ulcerative colitis [1, 2]. The initiating event in affect-
`ed individuals remains obscure, but it may arise from ge-
`netically determined inappropriate responses to antigenic
`bacterial material within the intestinal lumen [3, 4, 5].
`The past 10 years have seen fascinating developments in
`the therapeutic approach to inflammatory bowel disease.
`Advances in biotechnology have enabled identification of
`large numbers of soluble factors or cytokines and cell
`surface molecules expressed by lymphocytes, macro-
`phages and other immunologically active cells. Complex,
`dynamic signalling between individual populations of
`cells within the immune system, mediated via binding to
`specific receptors and interactions with other soluble fac-
`tors serve to regulate the process of inflammation.
`
`Knowledge of these molecules, much of it gained
`through original observations in animal models, has al-
`lowed us to identify many of the individual steps along
`the inflammatory pathway. One example is the identifi-
`cation, through studies in mice, of at least two popula-
`tions of T-helper (Th) lymphocytes which can be defined
`according to their cytokine secretion profile. Th1 cells
`secrete (inter alia) interleukin (IL) 2 and interferon (IFN)
`γ, both of which have potent pro-inflammatory effects.
`These cytokines play an important role as part of the
`host organism’s defence against bacterial and viral infec-
`tion as well as possessing the ability to destroy certain
`malignant cells. Th2 cells, in contrast, predominantly se-
`crete IL-4, IL-5 and IL-10, which have an anti-inflam-
`matory or immunomodulatory role, but which are also
`implicated in the pathogenesis of allergic reactions and
`responses to parasitic infection. The overall balance of
`Th1 and Th2 responses within affected tissues may ulti-
`mately determine the degree of inflammation.
`
`

`

`2
`
`Crohn’s disease is a classic example of a disease me-
`diated predominantly by Th1 cytokines, while a pattern
`more similar to the Th2 profile is seen in ulcerative coli-
`tis [6, 7, 8, 9]. However, it is clear that the Th1/Th2 par-
`adigm, based on a murine model, is an oversimplifica-
`tion of the inflammatory process and does not precisely
`mirror the situation encountered in humans. Nevertheless
`animal studies do allow the generation of useful hypoth-
`eses. The ability to synthesise individual cytokines
`through the use of recombinant DNA technology, along
`with the capability to interfere specifically with many of
`their actions, facilitates the testing of such hypotheses.
`This is rapidly leading to the development of new, poten-
`tially therapeutic strategies aimed at selective enhance-
`ment of the effects of immunoregulatory cytokines or
`blockade of the pro-inflammatory. In Crohn’s disease the
`goal of therapy has been to up-regulate the Th2 response,
`thus attenuating the predominantly Th1-driven inflam-
`matory process. Several cytokine-orientated strategies
`have been tested in animal models and now also in a
`number of chronic conditions in human subjects, includ-
`ing rheumatoid arthritis, psoriasis and multiple sclerosis
`[10, 11]. The multi-system approach has led to the inter-
`esting, if often frustrating, finding that although the basic
`mechanisms of the inflammatory process may not be tis-
`sue specific, the effects of highly selective blockade or
`antagonism of various aspects of the process may be. For
`example, a biological agent which is highly effective in
`the treatment of rheumatoid disease, or psoriasis may be
`of limited or no use in Crohn’s disease. This may reflect
`the differential importance of individual cytokine path-
`ways in a number of disease processes and may also in-
`dicate the involvement of other important pathways such
`as altered apoptosis, for which there is separate evi-
`dence.
`To date the most effective biological agent developed
`for the treatment of intestinal inflammation is infliximab
`(Remicade, Schering-Plough) a chimeric monoclonal an-
`tibody which blocks the binding of tumour necrosis fac-
`tor (TNF) α to its receptors on cells or in solution. A
`great deal has been written about the dramatic clinical
`effects of this compound in the treatment of Crohn’s dis-
`ease, and many patients have benefited from its use [12,
`13, 14, 15]. Other anti-TNF-α agents are also at ad-
`vanced stages of therapeutic development, with similar
`early results. Although a considerable experience has
`built up worldwide in the use of infliximab and other an-
`ti-TNF agents, there still remains some uncertainty re-
`garding their optimal usage and potential long-term side
`effects. An international panel of gastroenterologists re-
`cently produced guidelines on the indications for the use
`of this agent [16, 17]
`However, a number of other biological therapies
`which focus on other specific points in the inflammatory
`process are currently under scrutiny. These include en-
`hancing the immunoregulatory effects of the Th2 cyto-
`
`kines, IL-10 and IL-11, and blocking the pro-inflamma-
`tory effects of IL-12, IL-15 and IL-18. Another pro-
`inflammatory cytokine, IL-8, is also over-expressed in
`Crohn’s disease [18, 19, 20] and may potentially become
`a therapeutic target in the future. However, although ful-
`ly human anti-IL-8 neutralising antibodies are available
`[21], and their use has been described in the treatment of
`other inflammatory conditions [22, 23], studies of their
`use in inflammatory bowel disease have yet to be pub-
`lished. To date only two of these cytokine-directed bio-
`logical agents, IL-10 and IL-11, have come to trials in
`human subjects with intestinal inflammation [24, 25, 26,
`27].
`The release of pro-inflammatory cytokines also trig-
`gers the activation of inflammatory cells and promotes
`their recruitment to sites of inflammation. Central to
`these processes are the up-regulation of a host of cellular
`adhesion molecules, including intercellular adhesion
`molecule (ICAM) 1, vascular adhesion molecule 1, se-
`lectins and integrins, both on vascular endothelium and
`on lamina propria mononuclear cells [28, 29]. Up-regu-
`lation of ICAM-1 expression has been demonstrated in a
`number of inflammatory conditions including rejection
`of hepatic and renal allografts, hepatitis and inflammato-
`ry bowel disease [30, 31, 32, 33]. Neutralising antibodies
`to ICAM-1 have been shown to inhibit dextran sulphate
`sodium (DSS) induced colitis in mice [34] when admin-
`istered topically. The novel approach of interfering with
`messenger RNA (mRNA) expression for ICAM-1 by us-
`ing antisense oligonucleotide technology has also been
`described in humans [35]. Antisense oligonucleotides
`can be produced to target mRNA coding for the protein
`of interest. The oligonucleotide becomes incorporated
`with mRNA, producing a hybrid messenger signal and
`resulting in reduced expression and reduced quantities of
`the protein in question.
`Unfortunately, in vivo human studies have so far
`largely failed to live up to the expectations placed upon
`them by the prior animal and in vitro studies. This illus-
`trates a potentially generalisable problem, whereby the
`efficacy of a highly targetted biological therapy may be
`lost in the translation from experimental benchwork to
`the more complex clinical scenario where several differ-
`ent systems may be in operation at any given time.
`In this review we propose to highlight the current
`state of those cytokine and adhesion molecule-directed
`therapies for Crohn’s disease which do not specifically
`block receptors to TNF-α.
`
`Enhancing immunoregulatory cytokines
`
`Interleukin-10
`
`IL-10, one of the family of Th2 or contrainflammatory
`cytokines, was first described by Fiorentino et al. [36] in
`
`

`

`3
`
`1989 as cytokine synthesis inhibitory factor. It is capable
`of inhibiting the effector functions of activated macro-
`phages and monocytes in vitro and down-regulating the
`production of inflammatory cytokines such as IFN-γ and
`IL-2 by Th1 lymphocytes. This made it a theoretically
`attractive agent to use in the treatment of inflammatory
`conditions where a Th1 cytokine pattern predominates,
`such as Crohn’s disease. Genetically engineered IL-10–/–
`“knock-out” mice, incapable of IL-10 production, spon-
`taneously develop a Crohn’s disease-like enterocolitis
`[37, 38] if kept in standard, non-sterile laboratory condi-
`tions. However, similar IL-10–/– mice kept in germ-free
`conditions do not develop intestinal inflammation, sug-
`gesting that the presence of luminal bacteria is required
`for the development of intestinal inflammation, and that
`correcting IL-10 deficiency does may not fully reverse
`inflammation. Indeed, in the IL-10–/– mouse model the
`prophylactic administration of exogenous IL-10 did pre-
`vent the development of enterocolitis but was less effec-
`tive at ameliorating or reversing the condition once it
`was established [39].
`One study of the safety of multiple doses of recombi-
`nant human IL-10 (rhuIL-10; Schering-Plough, Kenil-
`worth, N.J., USA) [40] and three randomised, placebo-
`controlled trials [24, 25, 26, 26] have now been conduct-
`ed, involving a total of 535 human subjects with Crohn’s
`disease. Pharmacokinetic studies following subcutaneous
`administration of rhuIL-10 demonstrate a dose-depen-
`dent absorption with peak concentrations in serum at
`6–8 h. The mean elimination half-life was constant re-
`gardless of dose [24]. Unfortunately, the clinical results
`of the trials have been disappointing, without a dramatic
`therapeutic benefit being demonstrated in those with
`chronically active Crohn’s disease refractory to conven-
`tional therapy [25] or in those with mild to moderately
`active disease [24]. A third trial, looking at the preven-
`tion of recurrent Crohn’s disease lesions in the neo-ter-
`minal ileum following first ileal resection failed to dem-
`onstrate any reduction in the development of such le-
`sions 3 months after surgery [26].
`However, rhuIL-10 was remarkably well tolerated by
`the majority of patients, in spite of the need for adminis-
`tration by subcutaneous injection. Daily doses ranging
`from 1 to 20 µg/kg were given over a 28-day period, and
`43 patients were treated for 12 weeks with either 4 µg/kg
`per daily or 8 µg/kg rhuIL-10 twice weekly. In the larg-
`est study, involving a total of 329 patients, 95% of
`rhuIL-10 treated patients complained of at least one ad-
`verse event. This figure was remarkably similar in the
`group who received placebo (94%). The most common
`side effects cited by both groups were headache, abdom-
`inal pain and fever. All of these were described as mild
`to moderate in severity. Dose dependent, asymptomatic
`haematological abnormalities, with reduction in haemo-
`globin concentration, haematocrit, white cell indices or
`platelet counts were seen in up to 33% of cases. Lym-
`
`phopenia was seen in 9.5% of those receiving active
`treatment and in 9% of placebo treated patients, while a
`reduction in haemoglobin concentration occurred in 11%
`of those on the lowest doses of therapy (1, 4 and 8 µg/kg
`rhuIL-10) and 8% on placebo. In contrast, those patients
`who received the highest dose (20 µg/kg rhuIL-10) expe-
`rienced a significantly greater frequency of haematologi-
`cal side effects, with 33% showing a decrease in haemo-
`globin (P=0.0003 vs. placebo). No patient experienced
`any adverse symptoms in relation to these laboratory ab-
`normalities. In contrast to infliximab [41, 42], there have
`been no reported cases of lymphoproliferative disease
`associated with treatment with rhuIL-10. One patient
`died of pre-existing severe coronary heart disease while
`being treated with rhuIL-10, a death considered by the
`investigators to be unrelated to the therapy.
`Best et al. [43] assessed the efficacy of rhuIL-10 by
`means of the Crohn’s Disease Activity Index (CDAI).
`Defining remission as a CDAI score less than 150, with
`a drop of at least 100 points compared to baseline scores,
`no significant differences were seen between individual
`doses of rhuIL-10 and placebo. Of 66 patients 11 (17%)
`receiving placebo achieved remission, compared to be-
`tween 10% (95% confidence interval, 4.3–20.4%) for
`those receiving 20 µg/kg and 22% (95% CI, 12.5–34%)
`for those receiving 4 µg/kg. A lesser degree of benefit,
`defined as “clinical improvement”, or a reduction in
`CDAI of at least 100 points was seen more frequently in
`rhuIL-10 treated patients than in the placebo group, al-
`though this difference did not achieve statistical signifi-
`cance. A subgroup of patients with the highest CDAI
`scores (greater than the median CDAI) appears to have
`derived the most benefit from treatment with rhuIL-10 in
`terms of clinical well being and of parameters such as
`white blood cell count and C-reactive protein levels. A
`greater response to placebo was also seen in those with a
`CDAI above the median value, but in the absence of
`changes in white cell counts or C-reactive protein. Simi-
`lar remission rates were reported by Fedorak et al. [24]
`although induction of remission was a secondary end-
`point in their study. No placebo-treated patients achieved
`remission, and 5 µg/kg rhuIL-10 was found to be the
`most efficacious dose, inducing remission in 29.4% of
`patients (95% CI, 10.3–56.0%).
`Mary and Modigliana [44] also report a trend towards
`improvement in endoscopic appearances as measured by
`the Crohn’s Disease Endoscopic Index of Severity
`(CDEIS) in a greater proportion of those receiving
`rhuIL-10 than placebo, with 50 of 81 (62%, 50.5–72.3%)
`showing improvement compared to 10 of 23 (43%,
`23.3–65.5%; P=0.15). Fedorak et al. [24] confirmed this
`finding, using a global physician assessment score, with
`improvement seen endoscopically in 51% of the treat-
`ment group and 33% of the placebo group. In the special
`situation of the neo-terminal ileum following a first re-
`section for Crohn’s disease, Colombel et al. [26] report-
`
`

`

`4
`
`ed that after 12 weeks treatment with rhuIL-10 46% of
`the treatment group had evidence of endoscopic lesions
`vs. 52% of the placebo group. Even more disappointing
`was the observation that three patients developed clinical
`evidence of disease recurrence, sufficiently severe to
`warrant therapeutic intervention within the study period,
`all of whom had received rhuIL-10. It should be noted,
`however, that the criteria for response were tighter than
`those in many other studies of new agents in Crohn’s
`disease.
`Quality of life as assessed by previously validated
`questionnaires (Medical Outcomes Study Short-Form
`36-Item) [45] and Inflammatory Bowel Disease Ques-
`tionnaire (IBDQ) [46]) tended to parallel the measure-
`ments of clinical disease activity. Therapy with rhuIL-10
`appears to have had a beneficial effect on quality of life
`although this does not always occur in tandem with clini-
`cal improvement. Patients with mild to moderately active
`Crohn’s disease treated with 5 or 10 µg/kg rhuIL-10
`showed an improvement in IBDQ score compared to
`those on placebo. In patients with chronic active disease,
`whose baseline quality of life might be expected to be
`poor, the greatest improvements in IBDQ scores were
`seen in those who achieved remission (with a mean im-
`provement of 35.1 points), followed by those who im-
`proved clinically (mean improvement 28.5 points). How-
`ever, even those 215 patients whose CDAI scores failed
`to improve showed a mean 10.3 point improvement in
`IBDQ scores. It should be noted, however, that the Mary
`and Modigliana [44] study was not sufficiently powered
`to detect minimal important changes in either quality of
`life scale.
`Treatment of Crohn’s disease with rhuIL-10 was not
`associated with the formation of anti-rhuIL-10 antibodies
`in any patient tested.
`
`Interleukin-11
`
`IL-11, is a pleiotropic cytokine of mesenchymal cell ori-
`gin with potent thrombocytopoietic effects in vivo [47].
`This property has been utilised in the treatment of
`thrombocytopenic patients undergoing chemotherapy for
`malignant disease [48]. In addition to its stimulatory ef-
`fects on platelet production, it also appears to have pro-
`tective effects on intestinal morphology and barrier func-
`tion in animal models of radiotherapy, stimulating recov-
`ery of enterocytes [49] and protecting against radiation
`damage [50]. Furthermore, IL-11 can downregulate ex-
`pression of the pro-inflammatory cytokines IL-1, TNF-α
`and IFN-γ.
`Initial studies of IL-11 in animal models of experi-
`mental colitis were encouraging in terms of improving
`histological appearances and reducing levels of myelo-
`peroxidase in the colonic mucosa [51, 52]. A multi-cen-
`tre trial in 76 human subjects with Crohn’s disease [27]
`
`did show that the agent was superior to placebo in terms
`of achieving a clinical response (42% response rate in
`active treatment group vs. 7% in placebo group). How-
`ever, remission as defined by a CDAI of less than 150 or
`a drop of more than 70 points from baseline was only
`seen in 33% (compared to 17% of the placebo group).
`As might have been expected, there was a significantly
`higher mean proportional elevation of platelet counts in
`the IL-11 treated group than with placebo (37–43% vs.
`9%, P<0.05), and this effect increased with higher doses
`of IL-11 and with longer duration of therapy. There was
`no excess of adverse events in the treated group, and no
`serious events were reported. In particular, there were no
`cases of thrombotic disease, although patients with a past
`history of venous thrombosis were excluded from the
`study. A number of serious side effects, including pe-
`ripheral oedema, tachycardia, atrial arrythmias and
`stroke had been reported in earlier studies in cancer pa-
`tients and warnings are included on the package insert of
`IL-11. No anti-IL-11 antibodies were detected in any pa-
`tient following 21 days of therapy.
`
`Blockade of pro-inflammatory cytokines
`
`Interleukin-12
`
`IL-12 is a 70-kDa heterodimer consisting of two cova-
`lently liked polypeptide subunits (p35 and p40) coded
`for by separate genes [53, 54]. Produced by monocytes
`and macrophages in response to exposure to bacteria or
`bacterial products, it was originally discovered in 1989
`and named natural killer stimulatory factor (NKSF) [55].
`IL-12 plays a pivotal role in determining the outcome of
`the effector T-cell response, inducing the synthesis of
`IFN-γ by naive T-cells and driving a Th1 response. It ex-
`hibits tumoricidal activity [56], partly through induction
`of IFN-γ, but also by inhibiting angiogenesis in growing
`tumours via effects of the interferon inducible protein 10
`[57]. Homodimers of p40 subunit IL-12 bind to the IL-
`12 receptor and are capable of inhibiting the IL-12 medi-
`ated inflammatory response [58].
`Immunohistochemical studies have demonstrated a
`higher density of cells (usually macrophages) staining
`positive for IL-12 within the lamina propria of gut biop-
`sy specimens from patients with Crohn’s disease than in
`non-inflamed control specimens [59]. Berrebi et al. [60]
`also describe focally enhanced expression of IL-12 in
`lamina propria macrophages in gastric biopsy tissue
`from children with Crohn’s disease when compared to
`those with Helicobacter pylori
`induced gastritis and
`those with histologically normal stomachs. Lamina pro-
`pria mononuclear cells from patients with Crohn’s dis-
`ease express transcripts for both subunits of the IL-12
`molecule [61]. Studies of T-cell clones generated from
`inflamed intestinal tissues of patients with Crohn’s dis-
`
`

`

`5
`
`ease have demonstrated high levels of IFN-γ but not IL-4
`production in response to stimulation with phorbol 12-
`myristate 13-acetate. This contrasts with T-cell clones
`derived from those with ulcerative colitis or non-inflam-
`matory intestinal disorders, which tend to produce IL-4
`in addition to IFN-γ. Immunohistochemical analysis of
`intestinal specimens from Crohn’s disease patients re-
`vealed a similar picture with large numbers of CD4+ T-
`cells showing IFN-γ but not IL-4 reactivity. These cells
`were rarely seen in control specimens. Furthermore, cul-
`ture of T-cells from Crohn’s disease tissue in the pres-
`ence of neutralising IL-12 antibodies leads to a down-
`regulation of expression of IFN-γ producing CD4+ cells
`[59]. The production of IL-12 is thought to be a localised
`phenomenon, restricted to the lamina propria mononu-
`clear cells, as no IL-12 was detected in unstimulated pe-
`ripheral blood mononuclear cells taken from Crohn’s
`disease patients [61]. That enhanced IL-12 expression
`plays a role in Crohn’s disease is established, but just
`what initiates this enhanced expression and then perpetu-
`ates it remains unclear. There may be a factor specific to
`Crohn’s disease responsible for stimulating lamina pro-
`pria mononuclear cells to produce IL-12, or it may arise
`as a result of activation of macrophages by luminal bac-
`terial antigens or by IFN-γ produced by activated Th1
`cells.
`Animal models support a therapeutic role for block-
`ade of the biological effects of IL-12 in inflammatory
`bowel disease. Neurath et al. [62] studied the effects of
`intraperitoneal systemic administration of rat anti-mouse
`IL-12 neutralising antibodies to mice in whom an experi-
`mental colitis had been induced by rectal administration
`of
`the hapten 2,4,6-trinitrobenzene sulphonic acid
`(TNBS). Administration of neutralising anti-IL-12 anti-
`bodies either 5 or 20 days after the induction of TNBS
`colitis led to a striking improvement in the clinical con-
`dition of affected mice, with reversal of weight loss and
`a complete histological abrogation of the TNBS-induced
`inflammation in the majority of cases. Despite the publi-
`cation of this study almost 7 years ago these observa-
`tions have not yet led to any therapeutic trials of either
`neutralising IL-12 antibodies or the use of IL-12 p40 ho-
`modimers in human subjects. However, interest in this
`pivotal cytokine may be revived by a report from Bau-
`ditz et al. [63] who demonstrated significant reductions
`in both TNF-α and IL-12 production by lamina propria
`and peripheral blood mononuclear cells from nine
`Crohn’s patients treated with thalidomide for 12 weeks.
`A similar, dose-dependent reduction in both these cyto-
`kines was obtained when lamina propria mononuclear
`cells were cultured in the presence of increasing concen-
`trations of thalidomide. Interestingly, the authors noted
`no corresponding alterations in production of IL-1β or
`IL-6 and conclude that the clinical improvements ob-
`served with thalidomide therapy may be partially medi-
`ated by inhibition of IL-12 as well as of TNF-α.
`
`Interleukin-18
`
`In 1995 Okamura et al. [64] first described IL-18 as a
`new cytokine, capable of inducing IFN-γ production by
`T-cells. This 24-kDa precursor protein is mainly pro-
`duced by macrophages and requires cleavage by IL-1β
`converting enzyme to produce the biologically active
`18.3-kDa protein [65, 66]. The effects of IL-18 are medi-
`ated via its receptor, which consists of two subunits. The
`IL-1 receptor related protein (IL-1Rrp) functions as a
`binding domain, while the Il-1R accessory protein-like
`(IL-1RacPL) subunit is required for signalling [67, 68].
`Binding of IL-18 is followed by recruitment of IL-1R as-
`sociated kinase and activation of nuclear factor (NF) κB
`via an interaction with the TNF receptor associated fac-
`tor 6 [69].
`In humans, four different isoforms (a–d) of a naturally
`occurring IL-18 inhibitor, IL-18 binding protein (IL-
`18BP) are coded for by the IL-18BP gene [70, 71]. Only
`isoforms a and c possess any neutralising capacity.
`As with IL-12, IL-18 is a key player in the develop-
`ment of Th1 responses. Serum levels of IL-18 are signif-
`icantly higher in patients with Crohn’s disease than in
`normal or ulcerative colitis controls, and the expression
`of IL-18 is up-regulated in lamina propria mononuclear
`cells from inflamed mucosa in Crohn’s disease patients
`[72]. When administered together, IL-12 and IL-18 act in
`a synergistic manner to promote induction of high con-
`centrations of IFN-γ leading to severe intestinal inflam-
`mation in mice [73]. A similar effect was observed in an-
`imal models of therapy for malignant disease when lethal
`levels of IFN-γ were generated following gene transfer
`of both IL-12 and IL-18 [74]. This synergy is mediated
`in part by IFN-γ dependent mechanisms, with up-regula-
`tion of expression of the IL-18 receptor by IL-12 [75].
`However, IL-18 has other pro-inflammatory properties
`which are not dependent on IFN-γ. For example, in the
`presence of lipopolysaccharide it can also directly stimu-
`late cultures of peripheral blood mononuclear cells to
`produce IL-1β and TNF-α [76]. Furthermore, IL-18 acti-
`vation of NF-κB induces production of the pro-inflam-
`matory cytokines IL-8 and macrophage inflammatory
`protein 1α [77]
`Two trials reporting the effects of blockade of IL-18 in
`experimental colitis have been reported recently [78, 79].
`The first used recombinant human IL-18BP-6his-tagged,
`isoform a (rhuIL-18BPa) to treat mice with TNBS-in-
`duced colitis [79]. The severity of disease, measured clin-
`ically by degree of weight loss and also by histological
`examination of the colon, was significantly reduced in
`those mice in whom endogenous IL-18 activity had been
`neutralised. There were also significant reductions in the
`levels of TNF-α, IL-1β and IL-6 in colonic homogenates
`from IL-18BPa treated mice compared with controls.
`There were no such reductions in IFN-γ levels, suggest-
`ing an effect of IL-18 which is independent of IFN-γ. The
`
`

`

`6
`
`authors speculate that it may be IL-18 itself, and not its
`effects on IFN-γ production which mediates the inflam-
`mation seen in TNBS and DSS experimental colitis. One
`very powerful argument in favour of this is the fact that
`colitis can be induced by instillation of TNBS in IFN-γ
`receptor “knock-out” mice [80]. The second study, re-
`ported in abstract form only at present, demonstrated in-
`creased levels of IL-18 mRNA in colonic tissue of mice
`in whom an experimental colitis had been induced by in-
`stillation of DSS. Administration of IL-18BPa led to at-
`tenuation of the clinical effects of experimental colitis,
`such as weight loss, and significantly reduced histological
`evidence of colonic inflammation [78].
`The results of IL-18BPa therapy in these animal mod-
`els of Th-1 predominant colitis are encouraging, but
`there are no reports of the use of this compound in hu-
`mans. One potential advantage of IL-18BPa over inflixi-
`mab is that it is a naturally occurring compound rather
`than a chimeric, humanised antibody. As such, the risks
`of adverse events during administration of the compound
`may be lessened. Although the relevance of the human
`anti-chimeric antibodies seen in a minority of patients
`who have received infliximab is not clear [42], such re-
`actions would be unlikely following administration of
`IL-18BPa.
`
`Interleukin-15
`
`IL-15 and IL-2 have similar biological properties in vit-
`ro, probably arising from their shared receptor signalling
`components (IL-2/15Rβγc) [81]. IL-15 was identified in
`1994 by two independent groups as a soluble factor
`which was capable of stimulating cells of the IL-2 de-
`pendent CTLL-2 T-cell line, in the presence of neutrali-
`sing IL-2 antibodies [82, 83]. The mature IL-15 protein
`is a 14- to 15-kDa protein coded for by exons 5–8 of the
`IL-15 gene, and secreted by monocytes, macrophages
`and other antigen presenting cells, as well as by placenta,
`skeletal muscle, kidney lung and heart. It plays a vital
`role in the development of natural killer T-lymphocytes,
`activating their proliferation, and regulating their cyto-
`toxicity, cytokine production and interaction with macro-
`phages.
`The role of IL-15 in inflammatory bowel disease has
`yet to be fully described. However, current evidence
`shows that it is detectable in the sera of patients with ul-
`cerative colitis but not in patients with Crohn’s disease
`[84]. These serum levels may correlate with disease ac-
`tivity in ulcerative colitis [85] but do not appear to be
`able to predict relapse. Elevated numbers of peripheral
`blood mononuclear cells expressing IL-15 are seen in
`both ulcerative colitis and severe Crohn’s [84]. Sakai et
`al. [86] describe elevated levels of IL-15 in the superna-
`tant fluid of cultured colonic biopsies from patients with
`active
`inflammatory bowel disease (both UC and
`
`Crohn’s disease) compared to non-inflamed controls.
`However, a similar increase was seen in biopsy speci-
`mens from patients with inactive ulcerative colitis. Ex-
`pression of IL-15 mRNA was also significantly in-
`creased in inflamed rectal mucosa of IBD patients. Liu et
`al. [87] also found localised IL-15 production by macro-
`phages in inflamed mucosa from IBD patients.
`The anti-inflammatory effects of attenuating IL-15
`over-expression by its administering an excess of its sol-
`uble receptor (IL-15Rα) have been proven in an animal
`collagen-induced arthritis [88]. To date, no studies of the
`use of IL-15 blocking strategies in the treatment of ex-
`perimental IBD have been published, although several
`such strategies exist, including use of an antagonist IL-
`15mutant/Fcγ2a fusion protein [89] and a polyclonal
`neutralising antiserum [90].
`
`Attenuating recruitment of leucocytes to sites
`of inflammation
`
`Anti-integrin antibody therapy
`
`Integrins are glycoprotein molecules, widely expressed
`on leucocytes, and which function as mediators of leuko-
`cyte adhesion to vascular endothelium [91]. These het-
`erodimeric molecules consist of an α- and a β-subunit.
`The α4 integrin is expressed on almost all lymphocytes
`and to a lesser degree on monocytes and eosinophils
`[92]. The α4 integrin subunits usually exist in combina-
`tion with a β1 or β7 subunit, permitting interaction with
`vascular cellular adhesion molecule 1 or mucosal cellu-
`lar adhesion molecule 1, respectively. The interaction be-
`tween α4β7 and mucosal cellular adhesion molecule 1 is
`important in mediating the recruitment of leucocytes to
`intestinal mucosa in intestinal inflammation [93].
`In a recent double-blind, placebo-controlled study of
`30 patients with active Crohn’s disease, a single intrave-
`nous infusion of 3 mg/kg of the recombinant, humanised
`α4β7 antibody natalizumab (Antegren, Elan Pharma,
`Letchworth, UK), resulted in a significant fall in CDAI
`and improvement in quality of life [94]. However, the
`improvement in CDAI was not statistically different
`from that observed in placebo-treated patients. By
`4 weeks post-infusion 50% of the actively treated group
`had required rescue therapy with corticosteroids, com-
`pared to 59% of the placebo group. There was no signifi-
`cant difference in the number of patients entering remis-
`sion between the two groups, and only 10% of the treat-
`ed group remained in remission at 12 weeks follow-up.
`Treated patients required rescue therapy a median of
`22 days post infusion (range 17–89 days). The single in-
`fusion was generally well tolerated with 80% of patients
`reporting no adverse events throughout the follow-up pe-
`riod. In those who did experience adverse events, head-
`ache, abdominal pain and return of Crohn’s symptoms
`
`

`

`7
`
`wer