A pilot study of treatment of active ulcerative colitis with natalizumab, a humanized monoclonal antibody to alpha-4 integrin


Dr F. H. Gordon, Centre for Gastroenterology, 10th Floor, Royal Free and University College Medical School, Rowland Hill Street, London NW3 2PF, UK. E-mail: fionagordon@ownhomeline.freeserve.co.uk



Alpha-4 integrins facilitate leucocyte migration across vascular endothelium.


To assess the safety and efficacy of natalizumab (Antegren), a humanized antibody to alpha-4 integrin, in patients with active ulcerative colitis.


Ten patients with active ulcerative colitis, defined by a Powell-Tuck activity score > 4, received a single 3 mg/kg natalizumab infusion. The primary end-point was the change in Powell-Tuck score at 2 weeks post-infusion.


Significant decreases in the median Powell-Tuck score were observed at 2 and 4 weeks post-infusion (7.5 and 6, respectively) compared to the median baseline score (10). Five of 10 patients achieved a good clinical response at 2 weeks and one more patient by 4 weeks, defined by a Powell-Tuck score of ≤ 5. Significant improvements in quality of life scores were found at week 4. Rescue medication was required by two (20%), three (30%) and eight (80%) patients by weeks 2, 4 and 8, respectively (median, 34 days; range, 8–43 days). One patient remained in remission at 12 weeks. The median C-reactive protein at 2 weeks (6 mg/L) was lower than that pre-treatment (16 mg/L).


A single 3 mg/kg infusion of natalizumab was well tolerated by ulcerative colitis patients. The positive efficacy demonstrated in this study merits further investigation by randomized, placebo-controlled trials.


Integrins are heterodimeric glycoproteins, consisting of α and β subunits, which are important mediators of cell adhesion.1 Animal studies suggest that α4 integrin, in combination with β1 or β7 chains, is particularly important in the migration of leucocytes across vascular endothelial cells in the gut.2 Immunohistochemistry3 and flow cytometry studies4, 5 suggest that α4 integrin expression is up-regulated on the surface of lymphocytes present in inflamed areas of gut mucosa in patients with inflammatory bowel disease, and hence may represent a useful therapeutic target. Monoclonal antibodies to the α4 subunit in general, or α4β7 integrin specifically, have been shown to inhibit the migration of lymphocytes to rat mesenteric lymph nodes and Peyer's patches in vivo,6 and to ameliorate colitis in the spontaneously colitic cotton-top tamarin, an animal model of inflammatory bowel disease.7, 8 A randomized, placebo-controlled study of Act-1, a monoclonal antibody to α4β7 integrin, has demonstrated promising efficacy in patients with active ulcerative colitis, although larger confirmatory studies of this antibody are awaited.9

Natalizumab (Antegren; Elan Pharmaceuticals Inc., South San Francisco, CA, USA) is a recombinant humanized antibody to α4 integrin, which has been derived from a murine monoclonal antibody (AN100226m) raised against human α4 integrin (VLA-4).10 The complementarity-determining region of the hypervariable region of the AN100226m gene was grafted onto a human IgG4 framework to form humanized natalizumab (Antegren). In vitro, natalizumab has been shown to block the adhesion of human Jurkat α4β1-expressing cells to recombinant vascular cell adhesion molecule-1 and α4β7-expressing RPMI-8866 cells to recombinant mucosal addressin cell adhesion molecule (MAdCAM-1) (T. Yednock, personal communication, 1999). It also produces the clearance of leucocytes in the central nervous system of guinea pigs with experimental allergic encephalomyelitis,11 a disease thought to be mediated by the interaction of leucocyte α4β1 integrin and vascular cell adhesion molecule-1. A multicentre study of natalizumab in patients with active multiple sclerosis and a small phase I study in 26 healthy male volunteers have shown that a single 3 mg/kg intravenous dose is safe and well tolerated.12, 13 In patients with mild to moderately active Crohn's disease, natalizumab has been found to be safe and efficacious by randomized, double-blind, placebo-controlled trial.14, 15

Corticosteroids are the mainstay of remission induction treatment for inflammatory bowel disease, but many patients experience debilitating side-effects,16 and evidence already exists showing that, in the long term, these agents are ineffective in controlling Crohn's disease or ulcerative colitis; hence, there is a need to develop alternative methods of disease control. The aim of this open study was to assess the safety and efficacy of natalizumab in patients with active ulcerative colitis.



Patients with active ulcerative colitis, whose Powell-Tuck score was > 4,17 were assessed for trial eligibility. The disease was required to have been diagnosed at least 3 months prior to trial entry and to have been confirmed by at least two of the following diagnostic criteria: clinical history, histology, radiology or endoscopy.18 Female participants were required to have a negative pregnancy test at study entry and to use effective contraception throughout the study. Patients receiving oral corticosteroids (dose equivalent to ≤ 40 mg prednisolone daily) or 5-aminosalicylic acid-derived drugs (dose ≤ 2.4 g daily) were included, provided that the dose had not been increased within 2 weeks of study entry. Likewise, patients receiving azathioprine (dose ≤ 2 mg/kg daily) were eligible, provided that this had not been commenced or the dose increased within 4 months of study entry.

Exclusion criteria

The following groups were not included in the study: in-patients; patients who had undergone or who were thought likely to require intestinal surgery up to 3 months prior to, or 3 months after, study entry, respectively; patients with proctitis only; patients receiving ciclosporin, methotrexate or tacrolimus therapy; patients with an ileostomy, colostomy or ileal pouch; patients with laboratory-confirmed intestinal infection; and patients known to have had malignant neoplasia at any site. Patients weighing more than 100 kg, and pregnant or breast-feeding women, were also excluded.


Patients were assessed for trial eligibility at an initial screening visit, at which the following screening investigations were performed: full blood count, erythrocyte sedimentation rate, biochemical screen including immunoglobulins and C-reactive protein, urinalysis, resting electrocardiogram and sigmoidoscopic examination including rectal biopsy. The proportions of peripheral blood T cells (TCRαβ+) and B cells (CD19+) were also measured using fluorescence-activated cell-sorter analysis (Becton Dickinson, Abingdon, UK).19 The study was approved by the ethics committee of the Royal Free Hampstead NHS Trust and all patients gave written informed consent prior to participating in the study.

Patients received a single 3 mg/kg intravenous infusion of natalizumab over 25–45 min and were observed for at least 6 h post-infusion. The Powell-Tuck score was calculated, based on patient symptoms and examination findings, at 1, 2, 4, 8 and 12 weeks after treatment (Table 1). Sigmoidoscopic examination and rectal biopsies for histology were taken at each follow-up visit. The appearance of the rectal mucosa at sigmoidoscopy was scored as follows: 0, normal; 1, granular mucosa; 2, contact bleeding/spontaneous haemorrhage; 3, severe changes; ulcers/severely haemorrhagic mucosa and pus present.20 Biopsies were scored histologically as described by Saverymuttu et al.21 Quality of life was assessed by the Inflammatory Bowel Disease Questionnaire22 at screening and 4 weeks after treatment. Venous blood was taken at each visit for the haematological, biochemical and fluorescence-activated cell-sorter analyses described above, and for measurement of serum natalizumab and anti-natalizumab antibody levels.

Table 1.  . Sample Powell-Tuck colitis activity index score sheet. Adapted from Powell-Tuck et al.17Thumbnail image of

Statistical analysis

Differences between pre-treatment and week 2 Powell-Tuck scores, C-reactive protein, erythrocyte sedimentation rate, platelet counts and albumin concentrations were analysed by the Wilcoxon signed rank test. The study end-point was the change in Powell-Tuck score at 2 weeks. The number of patients who required `rescue' medication at 2 weeks was used as a secondary end-point. `Rescue' medication was defined as the initiation of or increase in daily dose of any of the following drugs by any route of administration: corticosteroids, 5-aminosalicylates or azathioprine/6-mercaptopurine. The `last-observation-carried-forward' principle was applied to analyses of the Powell-Tuck scores for those patients who had received rescue medication or who had been withdrawn.


Five of the 15 screened patients did not enter the study because of a failure to meet disease activity criteria on the day of treatment (one patient) or personal reasons (four patients). The clinical characteristics of the 10 patients who received natalizumab are shown in Table 2, nine of whom completed the post-infusion follow-up to 12 weeks. One patient was withdrawn from the study at 2 weeks post-infusion due to severe disease requiring urgent colectomy.

Table 2.  . Demographic and treatment characteristics of 10 patients with active ulcerative colitis prior to receiving a single 3 mg/kg natalizumab infusion Thumbnail image of

Disease activity

The median Powell-Tuck score of 6 at 2 weeks post-infusion was significantly less than the baseline median score of 10 (P=0.004). Significant improvement in the Powell-Tuck score was also found at weeks 1 (P=0.02) and 4 (P=0.008) post-treatment. True remission is defined by a Powell-Tuck score of zero, which was achieved by two patients. During the study (i.e. post hoc), the investigators felt that a good clinical response should be defined as a Powell-Tuck score of ≤ 5, as the patients' scores were in the range 8–12 at study entry. This was achieved in five of the 10 patients at 2 weeks post-infusion, and in one further patient by week 4 (Figure 1). The mean decreases in the Powell-Tuck score were 4.0 and 4.8 points at weeks 2 and 4, respectively, i.e. scores fell from 9.7 points at baseline to 5.7 and 4.9 points at weeks 2 and 4, respectively.

Figure 1.

. Individual Powell-Tuck colitis activity scores of 10 ulcerative colitis patients prior to and 1, 2, 4 and 8 weeks after 3 mg/kg natalizumab infusion. Boxes indicate time point at which rescue therapy commenced.

The median Inflammatory Bowel Disease Questionnaire quality of life score improved significantly from 101 at baseline to 162 at 4 weeks post-infusion (P=0.02) in the eight patients for whom both data sets were available (Figure 2; range of possible scores, 32–224).

Figure 2.

. Inflammatory Bowel Disease Questionnaire (IBDQ) quality of life scores of eight ulcerative colitis patients prior to and 4 weeks after a single 3 mg/kg natalizumab infusion. Higher values indicate improved quality of life.

Rescue medication was required by two (20%), three (30%) and eight (80%) patients by weeks 2, 4 and 8, respectively (median, 34 days; range, 8–43 days). Two of the eight patients rescued at week 8 initially demonstrated a good clinical response (Powell-Tuck score ≤ 5), but experienced a subsequent return of their symptoms.

Adverse events

Adverse events were infrequent and were generally thought to be unrelated to natalizumab. The most commonly reported events were `aggravated ulcerative colitis', representing either a worsening or return of disease symptoms (six patients), headache, vomiting, lethargy and sore throat (each of which was reported by two patients). Three patients required hospital admission during the study, one of whom had failed to respond to natalizumab and required colectomy at 2 weeks despite rescue with oral and intravenous corticosteroids (patient 10). A second patient (patient 4) was admitted with Campylobacter jejunii enteritis, caused by a food-related infection, at 28 days post-infusion, and a third patient (patient 5) developed nausea, vomiting and rigors after commencing oral azathioprine at 53 days post-infusion. None of these admissions were assessed as being related to natalizumab.

Haematological and biochemical indices

The median C-reactive protein value of 6 mg/L at 2 weeks post-infusion was significantly less than that at baseline (16 mg/L) (Figure 3), but serum albumin and erythrocyte sedimentation rate values were unchanged at 1 and 2 weeks post-infusion. No significant changes occurred in haemoglobin, platelet counts or serum biochemistry (except C-reactive protein) post-infusion of natalizumab.

Figure 3.

. Median serum C-reactive protein (CRP, ▪) levels and erythrocyte sedimentation rate (ESR, □) values of 10 ulcerative colitis patients prior to and following a single 3 mg/kg natalizumab infusion. Bars are ± s.d. and asterisk denotes significant difference from baseline value (< 0.05).

Drug levels and antibodies

The mean serum half-life of natalizumab in the 10 patients treated was 3.8 days. The mean serum concentrations of natalizumab achieved during the first 4 weeks post-infusion are shown in Figure 4. At 8 weeks, the serum concentration of natalizumab was negligible (< 0.05 μg/mL) in all patients. Low-titre (3.4–16 μg/mL) anti-idiotype antibodies to natalizumab developed transiently in one patient (patient 6) and were detected at weeks 4, 8 and 12. The patient had a good clinical response to natalizumab and did not require rescue therapy until week 8 (Figure 1).

Figure 4.

. Mean serum natalizumab levels of 10 ulcerative colitis patients following a single 3 mg/kg natalizumab infusion. Bars are ± s.d. Mean t½=3.8 days (s.d. 1.5).

Fluorescence-activated cell-sorter analysis of peripheral blood lymphocytes

A significant increase in total lymphocyte count occurred post-infusion of natalizumab (Figure 5) from 1.66 ± 1.03 (mean ± s.d.) pre-infusion to 3.32 ± 1.07, 2.88 ± 0.78 and 2.01 ± 0.62 at 1, 2 and 4 weeks, respectively (Figure 5; < 0.05 compared to baseline). Fluorescence-activated cell-sorter analyses revealed that this increase was due to a rise in both B-cell (CD19+) and T-cell (TCRαβ+) counts.

Figure 5.

. Total, B and T lymphocyte counts of 10 patients with active ulcerative colitis before and after a single 3 mg/kg infusion of natalizumab. Asterisks denote significant changes from baseline values (**< 0.005; *< 0.05).

Other findings

There were no significant differences in sigmoidoscopic appearance scores following natalizumab treatment, although six of the 10 (60%) patients demonstrated an improvement in their sigmoidoscopic score at 2 and 4 weeks post-infusion. There were also no significant histological changes observed between pre-treatment rectal biopsies (mean score, 6.0 ± 3.0) and those performed at 1, 2 and 4 weeks post-infusion (mean scores, 6.2 ± 2.2, 6.3 ± 2.0 and 5.9 ± 3.5, respectively).


The results of this study show that natalizumab is safe and well tolerated by patients with active ulcerative colitis. The numbers of patients are too small in this open study to assess the evidence of efficacy reliably, but the overall significant improvement in Powell-Tuck score found at weeks 2 and 4 post-treatment is encouraging. Significant improvements in quality of life scores at 4 weeks and C-reactive protein at 2 weeks also support these findings.

Interestingly, natalizumab's mean serum half-life of 3.8 days in this study was considerably shorter than the 8.9 days predicted by healthy volunteer studies.13 A similar disparity in half-life between healthy controls and patients with active inflammatory bowel disease has also been noted in a trial of monoclonal antibody to tumour necrosis factor-α,23 and may represent increased inflammatory cell turnover and antibody consumption in patients with active inflammatory bowel disease. Despite natalizumab's short half-life in ulcerative colitis patients, a significant increase in the number of circulating peripheral blood lymphocytes was sustained to at least 4 weeks post-infusion, suggesting that lymphocyte trafficking had been affected.

Since this study has been completed, further in vitro work has suggested that a minimum serum natalizumab concentration of 5 μg/mL is required to saturate at least 80% of circulating α4 integrins (Elan Pharmaceuticals Inc., unpublished data, 1999), a concentration which was sustained to 2 weeks in only three of the nine patients for whom these data were available. Thus, although lymphocyte dynamics were altered by natalizumab, heightened blockade of α4 integrins may be necessary to improve efficacy, and further studies are required to assess the optimal dose and dose interval in ulcerative colitis patients.

Although this small study demonstrates a positive efficacy for natalizumab, some patients did not respond, notably patient 10, in spite of experiencing an increase in circulating lymphocytes. Furthermore, the percentage increase in peripheral blood lymphocytes at each time point did not correlate with the patients' individual clinical responses to natalizumab overall (data not shown), suggesting that factors other than α4 integrins may be important in perpetuating mucosal inflammation in the human gut. The lack of a demonstrable change in histological parameters post-infusion is disappointing, but perhaps predictable in the light of the pharmacokinetic findings.

In summary, we have demonstrated that natalizumab (Antegren), an anti-α4 integrin antibody, is well tolerated in patients with active ulcerative colitis. There is a positive efficacy, but the present dosing regimen is suboptimal. A larger randomized, placebo-controlled trial is necessary to establish the efficacy of natalizumab and hence its potential role in the treatment of ulcerative colitis.


We are grateful to Dr Bruce MacFarlane (Watford General Hospital, UK) for his help in recruiting patients to the study, and to Shelley Rana and Judy Sercombe for technical assistance. The protocol was designed in collaboration between the Inflammatory Bowel Disease Study Group of the Royal Free Hospital and Elan Pharmaceuticals (Europe) Ltd. The study was funded by a grant from Elan Pharmaceuticals (Europe) Ltd.