Biologic Therapy of Inflammatory Bowel Disease

Transcription

1 GASTROENTEROLOGY 2002;122: Biologic Therapy of Inflammatory Bowel Disease WILLIAM J. SANDBORN* and STEPHAN R. TARGAN *Inflammatory Bowel Disease Clinic, Division of Gastroenterology and Hepatology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota; and Inflammatory Bowel Disease Center, Cedars Sinai Medical Center, Los Angeles, California Advancing knowledge regarding the biology of chronic inflammation has led to the development of specific biologic therapies that mechanistically target individual inflammatory pathways. Many biologic therapies are being evaluated for the treatment of the chronic inflammatory bowel diseases, Crohn s disease and ulcerative colitis. Biologic compounds proven to be effective for Crohn s disease include monoclonal antibodies to tumor necrosis factor (infliximab and CDP571) and to the leukocyte adhesion molecule 4 integrin (natalizumab). Other biologic compounds for which there is insufficient evidence to judge efficacy for inflammatory bowel disease include: p55 tumor necrosis factor binding protein (onercept); interferon ; interferon -1a; anti-interferon antibody; anti-interleukin 12 antibody; p65 anti-sense oligonucleotide (blocks NF- B); granulocyte colony stimulating factor, and granulocyte macrophage colony stimulating factor; anti-interleukin 2 receptor antibody; epidermal growth factor; keratinocyte growth factor 2 (repifermin); human growth hormone; anti-cd4 antibody; and anti- 4 7 antibody. Biologic therapies that have been proven ineffective for inflammatory bowel disease include: interleukin 10; interleukin 11; antisense intercellular adhesion molecule-1; and the tumor necrosis factor receptor fusion protein etanercept. Based on the early successes of infliximab, CDP571 and natalizumab, it seems certain that biologic therapy will play an important role in the future treatment of inflammatory bowel disease. Knowledge regarding the human and rodent immune systems has evolved to include an understanding of a variety of immunologic pathways and mechanisms including T cell activation resulting from antigen presentation and costimulation (Figure 1), the differentiation of CD4 T-helper lymphocytes into T-helper-1 (Th1), T-helper-2 (Th2), T-helper-3 (Th3), and T-regulatory-1 (Tr1) subsets (Figure 2) 1 3 ; lymphocyte trafficking (Table 1) 4 ; tumor necrosis factor (TNF)-mediated inflammation (Figure 3) 5 ; and nuclear factor B (NF- B)-mediated inflammation (Figure 4). 6 Similarly, understanding of the immunologic basis of inflammatory bowel disease (IBD) has evolved from descriptive observational data to the present situation in which data from animal models and humans with IBD suggest distinct pathways of inflammation for Crohn s disease (CD) and ulcerative colitis (UC). 7 In this paradigm, the pattern of cytokines expressed by the lamina propria lymphocytes of patients with CD are consistent with a Th1 immune response; Th1 response is characterized by increased expression of interferon, interleukin (IL)-2, IL-12, and IL-18 8,9 followed by subsequent increased production of the proinflammatory cytokines TNF and IL-1, and then NF- B, 10,11 as well as a compensatory increase in the Th2 mediated anti-inflammatory cytokine IL and transforming growth factor. In UC, the pattern of cytokine expression is different, 8,9 possibly a form fruste of a Th2 immune response and an associated relatively decreased Th1 response; Th2 response is characterized by an increased expression of IL-4, IL-5, IL-6, IL-10, and IL-13. These cytokines and T-lymphocyte cell types, in addition to the pathways requisite for production or function thereof, all represent potential targets for rational, mechanism-based, biologic treatment of IBD. Some biotechnology agents have already been proven effective (or in some cases ineffective) for CD in placebo-controlled trials and a large number of additional compounds are currently under evaluation. This article reviews the biotechnology compounds that have been or are being evaluated for the treatment of patients with IBD (Table 2). Definitions of Biologic Therapy and Vehicles of Therapy Delivery Biologic therapies include: (1) native biologic preparations and isolates such as blood products and vaccines containing live, attenuated, or killed microorganisms; (2) recombinant peptides or proteins such as Abbreviations used in this paper: EGF, epidermal growth factor; IFN, interferon; IL, interleukin; JNK, c-jun N-terminal kinase; KGF-1, keratinocyte growth factor-1; Mad-CAM-1, mucosal addressin cellular adhesion molecule; MAP, mitogen-activated protein; NF- B, nuclear factor B by the American Gastroenterological Association /02/$35.00 doi: /gast

2 May 2002 BIOLOGIC THERAPY FOR IBD 1593 Figure 1. Role of CD40 in T cell activation. Antigen recognition by T cells induces the expression of CD40 ligand (CD40L). CD40L engages CD40 on the APCs and stimulates the expression of B7 molecules and the secretion of cytokines that activate T cells. Thus, CD40L on the T cells makes the APCs better APCs. T cells can express CD40L on antigen recognition even without costimulation, but sustained expression of CD40L requires B7-CD28 costimulation, as well as antigen. Thus, the B7 and CD40 pathways stimulate each other. APC, antigen-presenting cell. Reprinted with permission from Abbas AK, Lichtman AH, Pober JS, eds. Activation of T lymphocytes. In: Cellular and molecular immunology. Philadelphia, PA: Saunders, 2000:169. growth hormone, erythropoietin, and so on; (3) antibody-based therapies; (4) nucleic acid based therapies; and (5) cell and gene therapies. 13 At the present time, the biotechnology therapies that are being used in clinical practice or investigated for the treatment of IBD are predominantly proteins, usually delivered intravenously or subcutaneously. The types of therapeutic proteins in use include recombinant human proteins with immunoregulatory effects, monoclonal antibodies (chimeric, humanized, and fully human), and fusion proteins. 14 Nucleic acids in the form of anti-sense oligonucleotides have also been used. 15 Finally, as specifically immunologic pathways are better understood and relevant protein targets in aberrant mucosal inflammation are discovered Figure 2. The Th1/Th2 paradigm. Modified with permission from Romagnani S. TH1/TH2 Cells. Inflamm Bowel Dis 1999;5:

3 1594 SANDBORN AND TARGAN GASTROENTEROLOGY Vol. 122, No. 6 Table 1. Adhesion Molecules Involved in Lymphocyte-Endothelial Recognition Lymphocyte homing receptor Selectin-CHO L-selectin (CD62L) Predominant endothelial cell ligands PNAd (includes CD34, other protein cores) MAdCAM-1 (HEVselective sulphated CHO modification of mucin domain) Role in multistep extravasation Primary homing pathways 1 adhesion Naïve lymphocyte homing to lymph nodes; lymphocyte homing to peripheral mucosal sites of severe chronic inflammation 1 (step 1a only) Naïve lymphocyte homing to Peyer s patches CLA E-selectin 1 adhesion Memory T cell homing to skin o-linked glycans P-selectin 1 adhesion? and sulphate probably associated with PSGL-1 Integrin-Ig family 4 7 MAdCAM-1 (mucosal addressin) 1 and 2 adhesion Naïve lymphocyte homing to Peyer s patche and appendix; memory lymphocyte homing to nonpulmonary mucosal sites 4 1 VCAM-1 1 and 2 adhesion Memory lymphocyte (blast) homing to many extra-intestinal inflammatory sites L 2 (LFA-1) ICAM-1, -2,? others 2 adhesion Widespread involvement Other CD44 (HCAM) Hyaluronate 1 adhesion? Homing of activated lymphocytes (blasts) to inflammatory sites? VAP-1 1 adhesion? Unknown subset homing to (? Inflamed) HEV in human, lymph nodes, tonsils, and sites of inflammation 1, primary; 2, secondary; CHO, carbohydrate; PNAd, peripheral lymph node addressin; MAdCAM, mucosal addressin cell adhesion molecule; HEV, high endothelial venule; CLA, cutaneous lymphocyte-associated antigen; PSGL, P-selectin glycoprotein ligand; VCAM, vascular cell adhesion molecule; LFA, leukocyte function antigen; ICAM, intracellular adhesion molecule; and VAP, vascular adhesion protein. Primary adhesion involves initiation of contact through microvillous receptors and rolling. Reprinted with permission from Butcher EC, Picker LJ. Lymphocyte homing and homeostasis. Science 1996;272: by targeted studies in rodent models and genetic analyses, 16,17 small molecule libraries can be screened for compounds with activity against a specific targeted step in the pathway (conventional small molecule therapy aimed toward inhibiting a biologic target known to have a critically important function related to the initiation, perpetuation, or suppression of inflammation). Small molecule screening may be further refined by determining the structural biology of the target protein (knowledge of the amino acid sequence of the protein, the 3-dimensional structure of the protein from x-ray crystallography and/or computer modeling, and the functional importance of both the sequence and the structure to the activity of the protein). TNF Inhibitors A variety of therapeutic approaches have been used to inhibit TNF in patients with IBD including the monoclonal antibodies infliximab (formerly called ca2) and CDP571; the p 75 soluble TNF receptor fusion protein etanercept or the p55 soluble TNF receptor onercept; the MAP kinase inhibitor CNI-1493; and thalidomide. The step in the TNF production and secretion pathway that each therapy targets is shown in Figure 3. Infliximab Infliximab is a genetically engineered IgG1 murine-human chimeric monoclonal antibody containing approximately 75% human protein and 25% murine protein. 18 The murine portion of the antibody is the variable or antigenic recognition region. Proposed mechanisms of action for infliximab include neutralization of both soluble and transmembrane TNF, as well as lysis of TNF producing cells via complement fixation, antibody dependent cytotoxicity, and apoptosis of T lymphocytes caused by the IgG1 Fc portion of the antibody. 5,18,19 An initial positive experience with infliximab in a single patient with CD 20 led to an open pilot study of 10 mg/kg and 20 mg/kg doses in 10 patients, 21 and then an open dose finding study. 22 Based on the results of these preliminary studies, Targan et al. conducted a placebocontrolled trial of infliximab at doses of 5, 10, or 20

4 May 2002 BIOLOGIC THERAPY FOR IBD 1595 Figure 3. Activation pathways leading to TNF production. Each activation pathway may vary depending on the cell type, for example macrophage or T cell, and the type of extracellular stimulus. Each step in the TNF production and secretion pathway is a potential target for therapeutic manipulation using specific inhibitors (shown in circled numbers). (1) TNF gene expression can be regulated by camp (cyclic adenosine monophosphate) modulation or MAPK (mitogen-activated protein kinases) including p38, p42/44 extracellular signal-regulated kinases (ERKs), and JNK. Pentoxyfylline increases camp and blocks TNF secretion by macrophages. The MAP-kinase inhibitor CNI-1493 inhibits p38 and JNK pathways. (2) TNF expression is induced via activation of the transcription factor NF- B. NF- B can be inhibited by sulfasalazine, mesalamine, and p65 oligonucleotides. (3) Modulation of TNF expression via the transcription factors AP-1, AP-2, nuclear factor-activated T cells (NF-AT), Ets, SP-1, C/EBP, and camp response element (CRE). (4) Splicing of TNF mrna. (5) Translation of TNF mrna via adenosine-uracil multimer-rich elements (ARE). (6) Degradation of TNF mrna. Thalidomide acts to prevent degradation of TNF mrna. (7) Cleavage of pro-tnf to TNF by TNF- -converting-enzyme (TACE). This step can be inhibited by metalloproteinase inhibitors. (8) Binding of soluble and transmembrane TNF by monoclonal antibodies to TNF such as infliximab and CDP571. (9) Binding of soluble TNF by TNF receptor fusion proteins such as etanercept. (10) Binding of soluble TNF by soluble TNF receptors such as onercept. Modified with permission from Papadakis KA, Targan SR. Tumor necrosis factor: biology and therapeutic inhibitors. Gastroenterology 2000;119: mg/kg in 108 patients with active CD. 23 At 4 weeks, 48% of patients receiving infliximab 5 mg/kg, 25% of patients receiving 10 or 20 mg/kg, and only 4% of placebo-treated patients achieved complete clinical remission (remission defined as a Crohn s disease activity index [CDAI] score 150 points 24 ). Patients who did not respond in this initial study were treated with openlabel infliximab 10 mg/kg. Patients who responded (decrease in CDAI score 70 points) in either the initial blinded study or the open-label crossover study were then rerandomized at 12 weeks into a second study and re-treated with infliximab 10 mg/kg or placebo every 8 weeks through week 36, with follow-up through week During the re-treatment study, 53% of patients receiving infliximab maintained remission through week 44 compared with 20% of placebo-treated patients. Between weeks 44 and 48 (8 12 weeks after the final infliximab dose administered at week 36) the majority of infliximab-treated patients relapsed, suggesting that the duration of benefit for infliximab in many patients is approximately 8 weeks. Recently, the preliminary results of a larger re-treatment study have been presented. 26 Five

5 1596 SANDBORN AND TARGAN GASTROENTEROLOGY Vol. 122, No. 6 Figure 4. Schematic diagram of NF- B activation. Activation of NF- B involves the phosphorylation and subsequent proteolytic degradation of the inhibitory protein I- B by specific I- B kinases. The free NF- B (a heterodimer of p50 and p65) then passes into the nucleus, where it binds to B sites in the promoter regions of genes for inflammatory proteins such as cytokines, enzymes, and adhesion molecules. P denotes protein, and mrna messenger RNA. Modified with permission from Barnes PJ, Karin M. Nuclear factor- B a pivotal transcription factor in chronic inflammatory diseases. Copyright 1997 Massachusetts Medical Society. All rights reserved (N Engl J Med 1997;336: ). hundred seventy-three patients with active CD received an open dose of 5 mg/kg infliximab. Patients who responded to therapy ( 70 point decrease in the CDAI and 25% decrease from baseline) were then randomized into 3 groups: group I received 2 additional induction doses of placebo at weeks 2 and 6, and then maintenance doses of placebo every 8 weeks beginning at week 14; group II received 2 additional induction doses of infliximab 5 mg/kg at weeks 2 and 6, and then maintenance doses of infliximab 5 mg/kg every 8 weeks beginning at week 14; group III received 2 additional induction doses of infliximab 5 mg/kg at weeks 2 and 6, and then maintenance doses of infliximab 10 mg/kg every 8 weeks beginning at week 14. The primary endpoint was the percentage of responders at week 2 who were in remission (CDAI 150) at week 30. Three hundred thirty-five of 573 patients (59%) had clinical response at week 2. At week 30, 22% of the placebo group vs. 39% (P 0.003) of group II and 46% (P 0.001) of group III achieved clinical remission. Thus, the absolute benefit is 12% for placebo (21% of 59%) vs. 23% (39% of 59%) for group II and 27% (45% of 59%) for group III. These results show that re-treatment with infliximab every 8 weeks is more effective than placebo for maintaining remission. This study also demonstrated that a 3-dose induction regimen with infliximab (dosing at 0, 2, and 6 weeks) is more effective at inducing remission than a single induction dose (40% remission at week 10 after a 3-dose induction compared with 28% remission after a single-dose induction). 27 Finally, infliximab has been studied for the closure of CD fistulas. 28 Ninety-four patients with actively draining perianal or abdominal fistulas were randomized to treatment with infliximab 5 or 10 mg/kg or placebo administered at weeks 0, 2, and 6. Complete closure of all fistulas maintained for at least 4 weeks occurred in 55% of the infliximab 5 mg/kg group, 38% of the infliximab 10 mg/kg group, and 13% of the placebo-treated group. Pilot studies, uncontrolled studies, and secondary outcome measures from the 573-patient re-treatment trial discussed above have suggested that infliximab may also be of benefit for other indications, such as steroid sparing in CD, treatment of CD involving ileoanal pouches, 32 and refractory UC, 33,34 but these indications require further investigation. The formation of human anti-chimeric antibodies has been reported in 13% of patients with CD treated with infliximab in clinical trials. 35 Other studies using a different HACA assay have reported cumulative HACA frequencies of 36% 68%. 36,37 The development of HACA antibodies may be associated with a loss of response to infliximab. 36 Patients with HACA antibodies also develop infusion reactions more frequently. 35,37 Concomitant treatment with azathioprine, 6-mercaptopurine, or methotrexate, and either concomitant treatment or pretreatment with corticosteroids in patients treated with infliximab appears to reduce the frequency of HACA antibody formation. 35,38 Infusion reactions are

6 May 2002 BIOLOGIC THERAPY FOR IBD 1597 Table 2. Biotechnology Compounds That Have Been or Are Being Evaluated for the Treatment of Patients With Inflammatory Bowel Disease Drug target Manufacturer Indication Phase of investigation Differentiation of CD4 T-helper lymphocytes into T-helper-1 (Th1) and T-helper-2 (Th2) subsets Anti-interferon antibody Protein Design Labs Crohn s disease Phase II Anti-interleukin 2 receptor antibody Protein Design Labs Ulcerative colitis Phase IIa Anti-interleukin 12 antibody Genetics Institute Crohn s disease Phase II Interleukin 10 Schering Plough Crohn s disease, ulcerative colitis Phase III, failed Phase II, failed Lymphocyte trafficking Anti- 4 integrin antibody (natalizumab) Elan Pharmaceuticals and Biogen Crohn s disease, ulcerative colitis Phase III Phase IIa Anti- 4 7 integrin antibody (LDP-2) Millennium Pharmaceuticals Crohn s disease, ulcerative colitis Phase II Phase II Anti-sense ICAM-1 (Isis 2302) Isis Pharmaceuticals Crohn s disease, ulcerative colitis Phase III, failed Phase II Tumor necrosis factor (TNF)-mediated inflammation Chimeric anti-tumor necrosis factor antibody (infliximab) Centocor and Schering Plough Crohn s disease, ulcerative colitis Phase IV Phase II/III Humanized anti-tumor necrosis factor antibody (CDP571) Celltech Crohn s disease, ulcerative colitis Phase III Phase IIa Tumor necrosis factor receptor fusion Immunex Crohn s disease Phase II, failed protein (etanercept) p55 tumor necrosis factor binding protein Serono Crohn s disease Phase II (onercept) CNI-1493 (MAP-kinase inhibitor) Cytokine PharmaSciences Crohn s disease Phase II Thalidomide Cellgene Crohn s disease Phase IIa Miscellaneous Interferon -1a Serono Crohn s disease, ulcerative colitis Phase II Phase II Granulocyte macrophage colony Immunex Crohn s disease Phase II stimulating factor Epidermal growth factor CIBG Ulcerative colitis Phase II Keratinocyte growth factor 2 (repifermin) Human Genome Sciences Ulcerative colitis Phase II Interleukin 11 Genetics Institute Crohn s disease Phase III, failed Anti-CD40 ligand antibody IDEC Pharmaceuticals Crohn s disease Phase II defined as adverse events that occur during or within 2 hours following an infusion of infliximab. Most infusion reactions include symptoms of urticaria, dyspnea, and/or hypotension. Overall, infusion reactions occurred in 17% of infliximab-treated patients compared with 7% of placebo-treated patients. 35 In addition, a syndrome of delayed hypersensitivity (myalgia and/or arthralgia with fever and/or rash, and in some patients pruritus, facial, hand, or lip edema, dysphagia, urticaria, sore throat, and headache) was reported in up to 25% of patients retreated with infliximab after a drug holiday of 2 4 years. 35,39 The formation of autoantibodies has been reported in patients with CD treated with infliximab in clinical trials including antinuclear antibodies (ANA) in 34% of patients and anti double-stranded DNA antibodies in 9%. 35 Another study reported that the cumulative frequency for developing a positive ANA was 50%. 40 Rarely, these patients may develop features of drug-induced lupus. 35 The occurrence of non-hodgkin s lymphoma appears to be rare but has been reported in 3 patients with CD and at least 4 patients with rheumatoid arthritis. 35,41,42 Long-term follow-up is needed before it is known with certainty whether or not there is an increased risk of lymphoma and other malignancies in patients treated with infliximab, but the paucity of reports to date is relatively reassuring. Infections occur more frequently in patients treated with infliximab compared with placebo (32% vs. 22%). 35 Serious infections may occur after treatment with infliximab including sepsis, tuberculosis, histoplasmosis, listeriosis, and aspergillosis. 35,43,44 Tuberculin skin testing, and chest radiograph in patients with a positive skin test, before treatment with infliximab is recommended. 35,43,45 Patients with evidence of prior tuberculosis infection must receive appropriate antitubercular therapy before treatment with infliximab. 46 There is preliminary informa-

7 1598 SANDBORN AND TARGAN GASTROENTEROLOGY Vol. 122, No. 6 tion on the outcome of pregnancies in women receiving infliximab, but the data are insufficient to draw conclusions at the present time. 47 CDP571 CDP571 is a genetically engineered IgG4 humanized monoclonal antibody constructed by grafting the complementarity-determining region of a mouse antihuman TNF monoclonal antibody into a human IgG4 antibody framework. 18 CDP571 contains approximately 95% human protein and 5% murine protein, and thus theoretically is less immunogenic than infliximab. Proposed mechanisms of action for CDP571 are limited to neutralization of both soluble and transmembrane TNF caused by the IgG4 Fc portion of the antibody. 18 Three placebo-controlled trials with CDP571 have been conducted in patients with CD. The first study of 31 patients with active CD showed that CDP571 5 mg/kg resulted in a greater decrease in the mean CDAI score at week 2 compared with placebo. 48 In a second larger study, Sandborn et al. reported on 167 patients with active CD treated with CDP571 at doses of 10 or 20 mg/kg or placebo. 49 At 2 weeks, 54% of patients receiving CDP mg/kg and 37% of patients receiving CDP mg/kg had a clinical response (decrease in CDAI score 70 points) compared with only 27% of placebo-treated patients. The differences were statistically significant for CDP mg/kg vs. placebo. Subsequently, patients in this study were retreated every 8 (for 2 doses) or 12 weeks (for 1 dose) with CDP mg/kg or placebo, and then followed through week 24. There were trends toward a greater rate of clinical remission at week 24 in patients receiving re-treatment (maintenance therapy) with CDP571 but the differences compared with placebo were not statistically significant. In this study, there was also a trend toward a greater rate of fistula closure in patients with draining perianal fistulas in patients treated with CDP571 as compared with placebo (50% vs. 15% closure rate). In the third study, 71 patients with steroid-dependent CD were treated with CDP mg/kg at baseline followed by 10 mg/kg at week 8 or placebo at baseline and again at week 8, and then followed through week Corticosteroids were tapered over 10 weeks. At week 16, 44% of CDP571 treated patients had successfully discontinued steroids while maintaining clinical remission compared with only 22% of placebo-treated patients. The Kaplan Meier survival estimate of time to relapse was statistically greater for the CDP571 treated patients. In this study, there was a trend toward a greater rate of fistula closure in patients with draining perianal fistulas among patients treated with CDP571 as compared with placebo (25% vs. 0%). Large placebo-controlled phase III studies are underway to confirm the efficacy of CDP571 for the indications of clinical response (reduction in signs and symptoms), steroid sparing, and re-treatment (maintenance of improvement). An uncontrolled pilot study suggested that CDP571 may also be of benefit in patients with active UC. 51 Anti-idiotype antibodies occurred overall at a rate of 5.3% in patients receiving treatment with CDP ,50 Infusion reactions occurred in 12.7% of CDP571-treated patients compared with 7.7% of placebo-treated patients. 49,50 Anti double-stranded DNA antibodies occurred in 5.3% of CDP571 treated patients and 0% of placebo-treated patients (concomitant immunosuppressive therapy was not related to the development of anti double-stranded DNA antibodies). 49,50 To date, there have been no cases of delayed hypersensitivity reactions, drug-induced lupus, or non-hodgkin s lymphoma. Etanercept There are 2 distinct cell-surface TNF receptors, designated p55 and p75. Soluble, truncated versions of membrane TNF receptors (consisting of only the extracellular, ligand-binding domain) are present in body fluids and are thought to be involved in regulating TNF activity in some chronic inflammatory diseases such as rheumatoid arthritis. 5,18 Etanercept is a genetically engineered fusion protein consisting of 2 identical chains of the recombinant human TNF-receptor p75 monomer fused with the Fc domain of human IgG1. 18 Etanercept is a fully human protein, and thus theoretically is less immunogenic than infliximab. The proposed mechanism of action for etanercept is neutralization of soluble TNF. 18 A small pilot study of etanercept 25 mg administered subcutaneously twice weekly for 12 weeks in 10 patients with active CD reported a clinical response (decrease in the CDAI 70 points) at week 2 in 6 patients. 52 Subsequently, a placebo-controlled trial in 43 patients with active CD failed to demonstrate efficacy for etanercept. 53 These negative results confirm those of another unpublished placebo-controlled dose-finding study of etanercept in 49 patients with active CD that had a similar negative outcome (Immunex Corporation, Seattle, WA, data on file). Onercept As discussed previously, the soluble p55 receptor is a soluble, truncated version of the p55 membrane TNF receptor (consisting of only the extracellular, ligandbinding domain). Onercept is a genetically engineered recombinant human TNF-receptor p55 monomer. Oner-

8 May 2002 BIOLOGIC THERAPY FOR IBD 1599 cept is a fully human protein, and thus theoretically is less immunogenic than infliximab. The mechanism of action for onercept is unknown but may be neutralization of soluble TNF. A small randomized dose-finding study of onercept 11.7 mg or 50 mg administered 3 times per week for 2 weeks in 12 patients with active CD was reported by Rutgeerts et al. 54 At 2 weeks, 18% of patients receiving onercept 11.7 mg and 67% of patients receiving CDP mg had a clinical remission (CDAI score 150 points). A large placebo-controlled phase II study is underway to confirm the efficacy of onercept for the indication of induction of remission. CNI-1493 CNI-1493 is a small molecule (guanylhydrazone compound) that acts to inhibit mitogen-activated protein kinase (MAP-kinase) pathways. CNI-1493 appears to block TNF gene expression by inhibiting the p38 and c-jun N-terminal kinase ( JNK) MAP kinase signaling pathways that lead to translational activation of the TNF gene. 5,55,56 Although CNI-1493 is a small molecule, not a biologic therapy, we have elected to include it in this review because the molecular target of the drug is a specific step in the pathway that leads to the production of TNF, and this compound is a small molecule alternative to inhibition of TNF with biotechnology therapies. A small randomized dose-finding study of CNI mg/m 2 or 25 mg/m 2 administered intravenously daily for 12 days in 10 patients with active CD was reported by Hommes et al. 57 At 2 weeks, 8 of 9 evaluable patients had a clinical response (decrease in CDAI score 70 points) and 4 of 9 evaluable patients were in remission (CDAI score 150 points). Thalidomide Thalidomide is a small molecule that acts to inhibit TNF biosynthesis by enhancing degradation of TNF mrna in macrophages, as well as inhibiting the TH1 polarizing cytokine IL Again, although thalidomide is a small molecule, not a biologic therapy, we have elected to include it in this review because the molecular target of the drug is a specific step in the pathway that leads to the production of TNF, and this compound is a small molecule alternative to inhibition of TNF with biotechnology therapies. Two uncontrolled pilot studies have been conducted using thalidomide to treat active inflammatory and fistulizing CD. Vasiliauskas et al. 58 used thalidomide mg/day for 12 weeks in patients with active inflammatory CD and reported response rates of 67% and remission rates of 0% 33%. Ehrenpreis et al. 59 used thalidomide mg/day for 12 weeks in patients with active inflammatory and fistulizing CD. Eighty percent of patients with fistulizing disease had fistula closure, and 50% of patients with active inflammatory disease had a clinical response. Inhibitors of Lymphocyte Trafficking A variety of therapeutic approaches have been used to inhibit lymphocyte trafficking in patients with IBD including monoclonal antibodies to 4 integrin (natalizumab) and 4 7 integrin (LDP-02); and antisense to intercellular adhesion molecule-1 (ICAM-1). The targets for these therapies are shown in Table 1. The 4 integrin is expressed at a moderate or high level on almost all lymphocytes Integrin usually exists in combination with either a 1or 7 subunit and interacts predominantly with the endothelial ligands vascular cellular adhesion molecule 1 and mucosal addressin cellular adhesion molecule (Mad-CAM-1), respectively. The interaction between 4 7 integrin and Mad-CAM-1 is important in mediating leukocyte homing to gut mucosa. 61 Natalizumab Natalizumab is a recombinant IgG4 humanized monoclonal antibody constructed by grafting the complementarity-determining region of a mouse antihuman 4 integrin monoclonal antibody into a human IgG4 antibody framework. 62 Studies in the cotton-top tamarin monkey demonstrated that monoclonal antibody blockade of 4 integrin was an effective therapy for the spontaneous colitis that these animals develop. 63 Two placebo-controlled trials with natalizumab have been conducted in patients with CD. The first study in 30 patients with active CD demonstrated that natalizumab 3 mg/kg resulted in a greater decrease in the mean CDAI score at week 2 compared with placebo. 64 Clinical remission (CDAI score 150 points) occurred more frequently in the natalizumab group than in the placebo group (39% vs. 8%). In a second larger study, Ghosh et al. reported on 244 patients with active CD treated with natalizumab at a single 3 mg/kg dose, two 3 mg/kg doses at a 4-week interval, two 6 mg/kg doses at a 4-week interval, or placebo. 65 At 6 weeks, 29% of patients receiving a single dose of natalizumab 3 mg/kg, 46% of patients receiving 2 doses of natalizumab 3 mg/kg, and 31% of patients receiving 2 doses of natalizumab 6 mg/kg had a clinical remission (CDAI score 150 points) compared with 27% of placebo-treated patients. The differences were statistically significant for treatment with 2 doses of natalizumab 3 mg/kg vs.

9 1600 SANDBORN AND TARGAN GASTROENTEROLOGY Vol. 122, No. 6 placebo. Large phase III induction and maintenance of remission trials in patients with CD are underway. An uncontrolled pilot study suggested that natalizumab may also be of benefit in patients with active UC. 66 LDP-02 LDP-02 is a recombinant IgG1 humanized monoclonal antibody constructed by grafting the complementarity-determining regions of a mouse anti-human 4 7 integrin monoclonal antibody into a human IgG1 antibody framework. Studies in the cotton-top tamarin monkey reported that monoclonal antibody blockade of 4 7 integrin was an effective therapy for the spontaneous colitis that these animals develop. 67 A phase IIa study in 29 patients with active UC demonstrated endoscopic improvement with LDP-02 in 1 of 5 (20%) patients who received 0.15 mg/kg subcutaneous, in 0 of 5 (0%) patients who received 0.15 mg/kg intravenously, in 3 of 5 (60%) patients who received 0.5 mg/kg intravenously, in 1 of 5 (20%) patients who received 2.0 mg/kg intravenously, and in 2 of 8 (25%) patients who received placebo. 68 Large, placebo-controlled, dose-finding trials are ongoing in patients with active CD and UC. ISIS 2302 (Antisense to ICAM-1) Isis 2302 is a 20 base phosphorothioate oligodeoxy-nucleotide designed to hybridize to a sequence in the 3 untranslated region of the human ICAM-1 message. The oligonucleotide-rna heterodimer so formed serves as a substrate for the ubiquitous nuclease RNase-H with subsequent cleavage and reduction in cellular specific message content and consequent reduction in ICAM-1 expression. A phase IIa placebo-controlled dose finding study reported that Isis 2301 had short-term efficacy in patients with moderately to severely active CD despite treatment with corticosteroids. 69 This study used placebo (5 patients) and 3 doses of Isis 2302: 0.5, 1.0, and 2.0 mg kg 1 d 1 intravenously for 13 of 26 days. Seven of 15 patients treated with Isis 2302 improved compared with 1 of 5 patients treated with placebo. Unfortunately, 2 subsequent trials involving 75 patients with active CD treated with subcutaneous Isis 2302 administered at a dose of 0.5 mg/kg for 2 days, 1 week, 2 weeks, 4 weeks, or placebo 70 and 299 patients with active CD treated with intravenous Isis 2302 administered at a dose of 2 mg/kg 3 days per week for 2 or 4 weeks, or placebo 71 failed to demonstrate efficacy. In the 299 patient study, a small subgroup of 9 obese female patients who received Isis 2302 and who achieved high serum concentrations of Isis 2302 that were identified post hoc had an increased rate of remission. 71 An additional randomized controlled trial, in which sufficient Isis 2302 is administered to ensure high serum concentration, is underway in patients with active CD. Inhibitors Of TH1 Polarization The therapeutic approaches that have been or may be used to inhibit TH1 polarization in patients with IBD include monoclonal antibodies to IL-12, interferon (IFN)-, IL-18, and IL-2 receptor (daclizumab and basiliximab); and the immunomodulatory recombinant human protein IL-10. The targets for these therapies are showed in Figure 2. Anti interleukin-12 IL-12 plays a central role in promoting Th1 responses and is critical in the regulation of differentiation and activation of helper T lymphocytes IL-12 interacts with a heterodimeric receptor composed of a 1 chain, which is instrumental in initial binding of IL-12 to the cell, and a 2 chain, which transduces the IL-12 signal. The latter consists in part of the phosphorylation of the STAT4 transcription factor, which translocates to the nucleus after phosphorylation where it plays an essential role in the transcription of the IFN- gene. IL-12 is secreted by antigen-presenting cells (dendritic cells and macrophages) following the interaction of the IFN- with activated T cells expressing CD40 ligand. IL-12 then acts on activated naive T cells via the IL-12 receptor as described previously to induce their differentiation into Th1 cells producing IFN-, and on activated-memory T cells to turn on their already-established capacity to produce IFN-. Studies in animal colitis models have demonstrated that antibodies to IL-12 can prevent or treat intestinal inflammation A phase II dose-finding trial with a human anti IL-12 antibody that has been genetically modified in its variable region so that it has a high affinity for human IL-12 in patients with active Crohn disease is currently underway. Interferon and Anti interferon As discussed previously, increased production of IFN- by Th1 cells is part of the process of polarization toward a Th1 immunologic response. Clinically, recombinant IFN- is used to enhance the killing of phagocytosed bacteria in chronic granulomatous disease. 78 Theoretically, the administration of IFN- to patients with CD could result in disease exacerbation. However, small pilot studies with low-dose recombinant IFN- at doses of 12,000,000 IU daily by IV infusion or 1.5 g/kg (45,000 U/kg) 3 times weekly by subcutaneous injection did not result in disease exacerbation. 79,80 Alternatively, the administration of monoclonal antibodies to IFN-

10 May 2002 BIOLOGIC THERAPY FOR IBD 1601 could theoretically be efficacious for the treatment of CD. In animal model studies, anti IFN- antibodies were less effective than anti IL-12 antibodies for the reversal of 2,4,6-trinitrobenzene induced colitis. 77 A phase II study of anti IFN- antibody therapy in patients with CD is underway. IL-10 IL-10, a T-helper type 2 cytokine, suppresses the production of IL-2 and IFN- by T-helper type 1 cells 81 and decreases IL-12 production by macrophages. Mice deficient in IL-10 have been shown to develop chronic enterocolitis. 82 Administration of IL-10 to IL-10 deficient mice weanlings completely prevented IBD, whereas administration of IL-10 to IL-10 deficient mice with established IBD resulted in some improvement but did not cure their disease. 83 A phase IIa dose-response study reported that therapy with recombinant IL-10 (rhuil- 10) at doses of 0.5, 1, 5, 10, or 25 g/kg intravenously for 7 days was beneficial as compared with placebo in 46 patients with medically refractory CD. 84 Unfortunately, these encouraging preliminary results were not confirmed by 3 large III placebo-controlled dose-response trials in 95 patients with mild to moderately active CD, patients with chronically active CD refractory to corticosteroids, 86 and 373 patients with steroid-dependent CD. 87 In another phase II study, there was no difference in the rates of endoscopic recurrence in 62 patients with CD undergoing an initial ileal or ileocolonic resection who were treated with rhuil-10 or placebo for 12 weeks. 88 Finally, a phase II placebo-controlled dose response trial of rhuil-10 failed to show a beneficial effect in 94 patients with mild to moderately active UC. 89 Development of systemic administration of rhuil-10 by intravenous or subcutaneous routes for the indications of UC and CD has been discontinued because of a lack of efficacy in these controlled trials. Recent animal studies have suggested that local administration of IL-10 to the colon via genetically engineered lactococcus bacteria that are orally administered may result in high colonic mucosal concentrations of IL-10 and potentially enhanced efficacy. 90 Cong generated bacteria specific Tr1 regulatory cells. 91 It is possible that this approach to IL-10 therapy will be applied to patients with CD or UC in the future. Anti IL-18 IL-18 is a recently described cytokine that mainly exists in activated macrophages and epithelial cells and shares biological activities with IL-12 in driving the development of Th1 by inducing IFN-. 92,93 IL-12 and IL-18 may act synergistically in this role. 94 IL-18 is up-regulated in the intestinal mucosal cells of patients with CD. 93 In animal model studies, anti IL-18 antibodies were effective in the reduction of severity of dextran sulfate sodium-induced colitis in mice. 94 A monoclonal antibody to human IL-18 has been created. 95 The efficacy of anti IL-18 antibody for patients with CD should be evaluated. Anti IL-2 Receptor Antibodies (Daclizumab and Basiliximab) IL-2 is produced by Th1 cells after IL-12, IFN-, and IL-18 induce differentiation of naive T helper cells to Th1 cells. Daclizumab, a human anti IL-2 receptor antibody, acts to block the binding of IL-2 to the IL-2 receptor. A pilot study of daclizumab in patients with refractory UC reported a beneficial effect in 4 of 5 patients. 96 Use of the chimeric monoclonal antibody to the IL-2 receptor, basiliximab, has not been reported in patients with IBD. Controlled trials of daclizumab and basiliximab in patients with IBD would be of interest. Inhibitors of NF- B NF- B designates a group of transcription factors with important functions in the intestinal immune system including the transcriptional control of various promoters of pro-inflammatory cytokines (including IL-1, IL-2, IL-12, and TNF- ), cell-surface receptors, transcription factors, and adhesion molecules (including ICAM-1). 11 The NF- B family consists of NF- B1 (p50; precursor protein: p105), NF- B2 (p52; precursor protein: p100), p65 (RelA), c-rel (Rel), and RelB which share the Rel homology domain (Figure 4). 6,11 Sulfasalazine and mesalamine are nonselective inhibitors of NF- B. 97,98 Selective inhibition of NF- B is an attractive therapeutic target for IBD. 99 An animal study of local (enema) administration of antisense phosphorothioate oligonucleotides to the p65 subunit of NF- B inmice with 2,4,6-trinitrobenzene sulfonic acid colitis and colitis in IL-10 deficient mice demonstrated therapeutic benefit. 100 Human studies of selective NF- B inhibitors in patients with IBD are needed. Inhibitors of T-Cell Activation The activation of T cells is a complex process that requires the presentation of an antigen by MHC class II on an antigen-presenting cell to the T-cell receptor on a T cell. 101 However, this antigen presentation in isolation leads to anergy (tolerance) or apoptosis and additional signals from costimulatory molecules are required for T-cell activation. Antigen recognition by T cells induces the expression of the costimulatory molecule CD40 li-

11 1602 SANDBORN AND TARGAN GASTROENTEROLOGY Vol. 122, No. 6 gand (CD40L) on the T-cell surface. CD40L then engages another costimulatory molecule CD40 on the antigen-presenting cell and stimulates the expression of the costimulatory molecule B7 (also on the surface of the antigen-presenting cell) and the secretion of cytokines that activate T cells. B7 then engages yet another costimulatory molecule CD28 on the surface of the T cell. T cells can express CD40L on antigen recognition even without costimulation, but sustained expression of CD40L requires B7-CD28 costimulation, as well as antigen. Thus, the B7 and CD40 pathways stimulate each other. Several other costimulatory molecules have been discovered, including ICOS, B7RP, 105 B7-1, 104,106 B7-2, 104,106 and B7H 107 that may be relevant targets for inhibition of T-cell activation. To date, there is little experimental data regarding the utility therapies targeting the T-cell receptor, MHC class II, or the various costimulatory molecules. However, a phase II trial of a monoclonal antibody to CD40L in patients with CD is underway. The targets for these potential therapies are shown in Figure 1. Anti-CD4 Antibodies T cells with the CD4 marker, T helper cells, play a central role in modulating cellular immunity through the secretion of multiple cytokines, which in turn modulate numerous effector functions including: immunoglobulin secretion, complement activation, neutrophil chemotaxis, and macrophage activation. In patients with IBD, T suppressor (CD8) and T helper (CD4) cells in the intestinal lamina propria and the epithelium are present in the usual proportions 108,109 but they appear to have an increased level of activation. 110 The importance of CD4 cells in IBD was highlighted by a report of complete remission of CD in a patient who became infected with the human immunodeficiency virus. 111 cm-t412 cm-t412 is a genetically engineered IgG1 murine-human chimeric monoclonal antibody to CD4 containing approximately 75% human protein and 25% murine protein. In a small phase I study, 8 patients with CD and 4 patients with UC were treated with cm-t412 with a dose of 20 mg/d intravenously (IV) for 7 days and then re-treated with a dose of 40 mg/d IV for 4 days at variable times between weeks 2 and ,113 Seven of 8 patients with CD and 3 of 4 patients with UC had endoscopic and histologic remission with a mean remission duration of 11 months and 12 months respectively. Similarly, in another small phase I dose-finding study, 12 patients with CD were treated with cm-t412 at doses of 10 mg/d, 30 mg/d, or 100 mg/d IV for 7 days (4 patients per group). 114 Improvement in the CDAI scores occurred in 9 of 12 patients. In the first study, CD4 counts remained persistently low (below 550 cell/ L) in all patients, but none experienced opportunistic infections. 112,113 Similarly, in the second study CD4 counts dropped significantly compared with baseline with inversion of the CD4/CD8 ratio, but there were no opportunistic infections. 114 Because of concern about toxicity from sustained CD4 lymphopenia, further studies have not been performed. MAX.16H5 MAX.16H5 is a monoclonal antibody to CD4. In a phase I study, 3 patients with CD and 7 patients with UC were treated with MAX.16H5 at a dose of 0.3 mg kg 1 d 1 IV for 7 days Two patients with CD and 2 patients with UC were re-treated 4 weeks later. All 3 patients with CD and 4 of 7 patients with UC responded. Sustained response was seen in 5 patients, particularly in patients who were re-treated. Circulating CD4 cells were significantly depleted but the effect was of short duration (24 hours). BF-5 BF-5 is a murine monoclonal antibody obtained by immunizing mice with activated T cells, of IgG1 isotype and CD4 specific. In a phase I dose-finding study, 12 patients with CD were treated with BF-5: 8 patients received 0.5 mg kg 1 d 1 IV for 7 days and 4 patients received 0.5 mg kg 1 d 1 IV for 1 day and then 1.0 mg kg 1 d 1 IV for 6 days. 118 Four of 11 patients improved. Significant decline of CD4 counts did not occur. Growth Factors A variety of growth factors may play an important role in IBD including transforming growth factor (which is secreted by Th3 cells and provides negative feedback on the differentiation of naive T helper cells to Th1 or Th2 subtypes), epidermal growth factor (EGF), and keratinocyte growth factor-1 (KGF-1). 119 Epidermal Growth Factor In the gastrointestinal tract, EGF stimulates cell proliferation. The role of EGF in the pathogenesis of IBD is unclear. 119 A placebo-controlled trial of EGF enemas in 23 patients with active left-sided UC demonstrated remission in 82% of EGF-treated patients and 8% of placebo-treated patients. 120 The benefit appeared to last for at least 2 weeks after the discontinuation of treatment.

12 May 2002 BIOLOGIC THERAPY FOR IBD 1603 KGF Fibroblast factor 7, also known as KGF-1, is abundantly expressed in the gastrointestinal tract where it is a potent stimulant of intestinal epithelia cells. 119,121 The amount and distribution of KGF-1 in the gastrointestinal tract is altered in patients with IBD. 119 Animal studies with the recombinant human KGF-1 (rhukgf) demonstrated a beneficial effect in the dextran sodium sulfate mouse and the CD45RB Hi mouse colitis models. 122 Animal studies with a homolog of KGF-1, keratinocyte growth factor-2 (KGF2, repifermin) showed a beneficial effect in the dextrans sodium sulfate mouse colitis model and in the indomethacin small intestinal injury rat model. 123,124 A phase II study of repifermin in patients with active UC has recently been completed, and the results are pending. Miscellaneous Biotechnology Therapies IFN- Interferon is produced naturally by virally infected cells to induce resistance of the cells to viral infection. Recombinant IFN- -2a, IFN- -2b, and IFN- -n are used clinically to treat HIV-related Kaposi s sarcoma, melanoma, chronic hepatitis B infection, and chronic hepatitis C infection. 78 The cellular responses to interferon appear to be mediated through the JAK- STAT pathway. 125 Uncontrolled studies with IFN- -2a reported response rates of up to 50% 126,127 in patients with CD and up to 93% in patients with UC. 128 An open-label, randomized comparison of IFN- -2a and prednisolone enemas in patients with left-sided UC reported similar efficacy. 129 Uncontrolled studies of IFN- -2b reported response rates ranging from 33% 50% in patients with CD IFN- -1a Interferon is produced naturally by virally infected cells to induce resistance of the cells to viral infection. Recombinant interferon is used clinically to treat multiple sclerosis. 78 The cellular response to interferon appears to be mediated through the JAK-STAT pathway. 125 A pilot study of interferon in 5 patients with CD reported improvement in 80%. 133 A phase IIa placebo-controlled trial of interferon up to 88 g 3 times per week for up to 12 weeks in 18 patients with UC demonstrated response rates of 50% in the interferon group and 14% in the placebo group. 134 A large phase IIb trial in patients with CD is being initiated. Granulocyte Colony Stimulating Factor (Filgrastim) and Granulocyte-Macrophage Colony Stimulating Factor (Sargramostim) Genetic syndromes associated with intestinal inflammation that is phenotypically similar to CD include chronic granulomatous disease, glycogen storage disease, and Chediak Higashi syndrome. 135 A case report suggested that administration of filgrastim (human granulocyte colony-stimulating factor, also known as recombinant methionyl human granulocyte colony-stimulating factor or r-methug-csf) was beneficial in closing CD perianal fistulas. 136 Subsequently, 2 pilot studies of filgrastim (18 patients) and sargramostim (human granulocyte-macrophage colony-stimulating factor, also known as recombinant human granulocyte-macrophage colonystimulating factor or RhuGM-CSF) (11 patients) have reported that filgrastim and sargramostim may be effective for treating chronically active and fistulizing CD. 137,138 A phase II trial with sargramostim in patients with active CD is underway. Human Growth Hormone Growth hormone is a regulatory peptide that increases amino acid and electrolyte uptake by the intestines, decreases intestinal permeability, and induces the expression of insulin-like growth factor I, which in turn stimulates collagen synthesis. The rationale for the use of growth hormone in CD is to reverse the catabolic process associated with inflammation. A placebo-controlled trial of recombinant human growth hormone (somatropin) in 37 patients with active CD demonstrated a greater decrease in the mean CDAI score for somatropin-treated patients than for placebo-treated patients. 139 The proportion of patients entering remission was not reported. Additional controlled trials using the more conventional endpoints of improvement and remission are planned. IL-11 IL-11, a cytokine produced by cells of mesenchymal origin, has multiple biologic effects including a potent thrombocytopoietic effect, as well as enhancement of barrier function of the intestinal mucosa. 140,141 Administration of IL-11 in a trinitrobenzene sulfonic acid rat model of colitis demonstrated beneficial effects. 142 Based on this rationale, recombinant human IL-11 (rhil11) has been evaluated for the treatment of CD. The overall results of a phase IIa dose-response study with placebo or rhuil11 administered subcutaneously in 76 patients with active CD were negative, but there was a trend toward a beneficial effect with rhuil11 16 g/kg given either in a 2 or 5 day per week regimen. 143 A second phase IIb/III dose response study with rhuil11 administered subcu-