Patients

This randomised, double blinded, placebo controlled, multicentre trial was conducted in Belgium, Canada, and the USA at 40 sites between April 2003 and May 2004. The institutional review board or ethics committee at each site approved the protocol. All patients gave written informed consent.

Criteria for eligibility included patients 12 years of age or older with UC of at least four months’ duration which was moderately active, as defined by a Mayo Clinic score¹² of 5–10 (inclusive). Men and women of childbearing potential consented to the use of double barrier contraception for the duration of the study. Concurrent therapies for UC, including stable doses of 5-aminosalicylates, methylprednisolone 32 mg/day or less (or equivalent dose of other corticosteroid), azathioprine, and 6-mercaptopurine were permitted. Concomitant therapy with methotrexate, ciclosporin, tacrolimus, antibiotics, or rectally administered corticosteroids or 5-aminosalicyate containing medications was not permitted. Patients were eligible if they had a haemoglobin level ⩾8.5 g/dl, white blood cell count ⩾3.0×10⁹/l, lymphocytes ⩾0.5×10⁹/l, platelets ⩾100×10⁹/l, neutrophils ⩾1.5×10⁹/l, aspartate aminotransferase and alanine aminotransferase within three times the upper limit of normal, and creatinine ⩽1.6 mg/dl. Patients were excluded if they had Crohn’s disease or indeterminate colitis, a positive pregnancy test or breast feeding in female patients, imminent need for surgery for UC or toxic megacolon, positive stool culture for enteric pathogens (including C difficile), prior treatment with daclizumab, known human immunodeficiency virus, hepatitis B or C infection, prior malignancy within five years or current malignancy (excluding adequately treated non-melanoma skin carcinoma or in situ carcinoma of the cervix), use of any investigational drug within 30 days prior to screening, use of a monoclonal antibody within 12 weeks of screening (including infliximab), infection with varicella, herpes zoster, or severe viral infection within six weeks of screening, exposure to varicella within 21 days of screening, or vaccination with live virus within four weeks prior to screening.

Study design

Eligible patients were randomly assigned in a 1:1:1 ratio to receive either an intravenous infusion of daclizumab (Protein Design Labs, Inc., Fremont, California, USA) at a dose of: 1 mg/kg at weeks 0 and 4 alternating with placebo at weeks 2 and 6 (the dosing regimen used in the uncontrolled pilot study in patients with UC); or 2 mg/kg at weeks 0, 2, 4, and 6, or placebo at weeks 0, 2, 4 and 6, and followed through to week 20. Randomisation was performed at each site by an unblinded study pharmacist and treatment allocation was stratified according to: (1) investigational site; and (2) baseline use of oral corticosteroid therapy. Both patients and investigative staff (except for the study pharmacist at each site) were blinded to treatment assignment. The dose of all concurrent medications remained constant through to week 8. After week 8, the daily dose for patients receiving oral corticosteroids could be tapered at the investigator’s discretion.

Patient schedule and efficacy and safety evaluations

Patients were assessed at weeks 0, 2, 4, 6, 8, 12, 16, and 20. The Mayo score¹² was determined at weeks 0 and 8. The components of the scoring system are: stool frequency, rectal bleeding, endoscopic findings, and the physician’s global assessment. Each subscore is rated on a scale from 0 to 3, indicating normal to severe activity. Remission was defined as a score of 0 on the rectal bleeding and endoscopy subscores and a score of 0 or 1 on the stool frequency and physician’s global assessment subscores. Response was defined as a decrease from baseline in the total Mayo score by at least 3 points. Clinical response was defined as a decrease from baseline in the combined rectal bleeding, stool frequency, and physician’s global assessment subscores by at least 2 points. Endoscopic response was defined as a decrease from baseline in the endoscopic subscore by at least one point. All randomised patients were required to have mucosal biopsies for histological assessment during endoscopy performed at baseline and at week 8. Histological assessment of inflammation was performed with the validated grading scale of Geboes, an instrument with six domains (structural (architectural change), chronic inflammatory infiltrate, lamina propria neutrophils and eosinophils, neutrophils in epithelium, crypt destruction, and erosions or ulceration). Scores can range from 0 to 5.4, with higher scores indicating more severe histological inflammation.¹³

Data for all 159 randomised patients were included in the safety analysis. At each visit, adverse events and concomitant medications were recorded and, at some visits, samples were collected for laboratory evaluations. Safety evaluations included vital signs, physical examination, haematology, serum biochemistry, and urinalysis. Samples for determination of antibodies to daclizumab were collected at weeks 0, 4, 8, 12, 16, and 20.

Immunohistochemical analysis of lamina propria T cells

At 15 selected clinical sites, additional mucosal biopsies were obtained to determine lamina propria lymphocyte CD25 expression and saturation with daclizumab, T cell activation (Ki67 staining), and T cell apoptosis (caspase 3 staining). Immunohistochemisty was performed on 5 µm thick cryostat sections, using a three step indirect immunoperoxidase technique. Expression of CD25 in mucosal biopsies was investigated using antibody 2A3 (347640; Becton Dickinson, San Jose, California, USA) (anti-CD25 daclizumab competing antibody) and 143-13 (Neomarker Lab, Fremont, California, USA; Vision MS-785-P) (anti-CD25 daclizumab non-competing antibody) in a dilution of 1/.5 and 1/10, respectively.

Expression of Ki67 in mucosal biopsies was assessed using the primary antibody MiB1 (Ki67, IgG1, kappa; Dako, Glostrup, Denmark) in a dilution 1/100. Expression of T cells was assessed using the CD5 (leu1) antibody (Becton Dickinson) in a dilution 1/2.5. Expression of caspase 3 in mucosal biopsies was assessed using a polyclonal antibody against caspase which detects the large fragment of activated caspase-3 (17–20 kDa) (ASP175; Cell Signaling Technology, Beverly, Massachusetts, USA) in a dilution 1/20. Detection was performed using swine antirabbit/biotin (1/500; Dako) followed by a wash in phosphate buffered saline for five minutes and incubation for 30 minutes with the avidin-biotin complex/horseradish peroxidase (Dako). Subsequently, sections were washed for five minutes in phosphate buffered saline and treated for 10 minutes with DAB. These slides were faintly counterstained with Harris haematoxylin, rinsed with distilled water, and coverslipped with glycerol. Controls that were negative consisted of omission of the primary or secondary antibody and the use of the chromogen alone. In addition, staining with an irrelevant, isotype matched, antibody directed against cytokeratin 10 was performed. Positive controls for CD5 consisted of lymph nodes; for Ki67, the positive controls were the proliferative compartment of intestinal crypts. For CD25, skin biopsies from patients with psoriasis were used, and for caspase 3, samples from colorectal cancers were used. The number of positive cells was counted in 10 fields of 144 µm² using a calibrated eyepiece. The percentage of T cells staining for CD25 was counted by dividing the number of CD25 positive cells by the number of CD5 positive cells. To confirm that CD25 staining actually stained T cells, double stainings for CD25 and CD5 were performed in two cases.

Evaluation of CD25 expression on peripheral lymphocytes and peripheral T cell counts

At 15 selected clinical sites, anticoagulated peripheral blood samples were collected at screening and at weeks 0, 4, 8, 12, 16, and 20 for determination of the relative proportions of lymphocytes, T cells (CD3), T cell subsets (CD4 and CD8), and CD25 expression on all lymphocytes, T cells, CD4, and CD8 T cells using fluorescence activated cell sorter (FACS) analysis (FACSCalibur; Becton Dickinson). At the same time points, the percentage of CD25⁺ cells and mean fluorescence intensity were determined with two murine IgG1 antibodies specific for CD25-anti-CD25 antibody 2A3 that competes with daclizumab for binding to CD25, and non-competitive anti-CD25 antibody MA251. The percentage of CD25 (2A3)⁺ cells indicates the percentage of CD25⁺ cells that have no DAC bound to CD25, and the percentage of CD25 (MA251)⁺ cells indicates the percentage of cells expressing CD25 on the cell surface. We considered CD4 and CD8 cells saturated with DAC when the level of CD4⁺/CD25 (2A3)⁺ was at or below 1%.

Absolute CD3, CD4, and CD8 T cell counts were calculated at screening and at weeks 0, 4, 8, and 20 based on lymphocyte counts measured on a haematology analyser, and the percentage of CD3⁺, CD3⁺CD4⁺, and CD3⁺CD8⁺ T cells determined by FACS.

Statistical analysis

The primary efficacy end point was remission at week 8. The major secondary end points were response at week 8, clinical response at week 8, endoscopic response at week 8, and Mayo scores and total histopathology disease severity scores at weeks 0 and 8. All randomised patients who had valid baseline assessments and initiated infusions of study drug were analysed, regardless of the number of administrations of study drug that they completed (“intention to treat” approach). Patients who prematurely withdrew from the study for efficacy reasons were included in the analyses using “last observation carried forward” methods. Early terminations due to safety or issues that were not treatment related, such as relocation, were included in the analyses up to the last valid visit prior to withdrawal.

To compare the proportion of patients achieving a specified end point (that is, remission, response, clinical response, endoscopic response) between treatment groups, a two sided Fisher’s exact test was used. Continuous response parameters were compared using a t test with unequal variance.

We estimated that 50 patients were needed in each of the daclizumab groups and 50 patients in the placebo group in order to have 80% power to detect a true difference in the proportion of patients with moderate UC who achieved remission at week 8, assuming the proportion in either of the two active groups was 35% and that placebo was 10%. We planned to recruit a total of 150 patients.

Characteristics of patients

A summary of patient disposition is provided in fig 1⇓. Fifty six patients with moderate UC were randomised to receive intravenous therapy with placebo at weeks 0, 2, 4, and 6; 56 were randomised to receive daclizumab 1 mg/kg intravenously at weeks 0 and 4; and 47 were randomised to receive daclizumab 2 mg/kg intravenous at weeks 0, 2, 4, and 6. There were 150 patients included in the primary efficacy analysis. Of the nine patients not included in the efficacy analysis, eight terminated early due to safety or issues that were not treatment related, and one patient had an incomplete Mayo score at week 8.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Overview of patient enrolment, treatment assignments, and subjects available for final evaluation.

The 150 patients had either completed treatment or had withdrawn prior to week 8 due to inefficacy. The safety population comprised all 159 randomised patients who received at least one dose of study medication. Baseline characteristics were similar in the three treatment groups (table 1⇓).

Efficacy

Overall, 12 patients (21%) in the placebo group withdrew from the study prematurely compared with 14 patients (25%) and 14 patients (30%) in the daclizumab 1 mg/kg and 2 mg/kg groups, respectively, primarily due to treatment failure.

At week 8, 2% of patients in the daclizumab 1 mg/kg group (1/54; p = 0.11 v placebo) and 7% of patients in the daclizumab 2 mg/kg group (3/43; p = 0.73 v placebo) were in remission, compared with 10% of patients in the placebo group (5/53) (fig 2A⇓). Similarly, 25% of patients in the daclizumab 1 mg/kg group (13/53; p = 0.04 v placebo) and 33% of patients in the daclizumab 2 mg/kg group (14/43; p = 0.29 v placebo) were in response at week 8, compared with 44% of patients in the placebo group (23/52) (fig 2B⇓). Also, for clinical and endoscopic Mayo score response and for mean clinical at week 8, no differences between the three groups were observed (table 2⇓)

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

 Proportions of patients with remission (A) and response (B) at week 8.

At baseline, 57 of 159 patients (36%) were receiving corticosteroids. An exploratory analysis in this patient subgroup demonstrated that 0% of patients in the daclizumab 1 mg/kg group (0/21; p = 0.45. placebo) and 6% of patients in the daclizumab 2 mg/kg group (1/16; p = 1.00 v placebo) were in remission, compared with 6% of patients in the placebo group (1/17).

At baseline, most patients had histological signs of active disease: placebo, 92% (n = 12); daclizumab 1 mg/kg, 100% (n = 12); daclizumab 2 mg/kg, 92% (n = 13). Histological severity scores at week 8 were similar in all three treatment groups. Also, the majority of patients had moderate to severe histological disease regardless of treatment assignment (placebo, 75%; daclizumab 1 mg/kg, 75%; daclizumab 2 mg/kg, 79%).

Safety

The incidence of adverse events was similar in the placebo (75%) and two daclizumab groups (76.6% in the 2 mg/kg group and 83.9% in the 1 mg/kg group). A similar number of patients discontinued treatment because of an adverse event in the placebo group (0%) as in the 1 mg/kg and 2 mg/kg daclizumab groups (0% and 2%, respectively). The most frequently reported adverse events in the three groups were similar (table 3⇓). One patient receiving 2 mg/kg daclizumab experienced a moderate infusion reaction after the first dose of study drug. The patient continued in the trial and had no further events. Serious adverse events occurred in 3.6% of patients in the placebo group, and in 12.5% and 4.3% in the 1 mg/kg and 2 mg/kg daclizumab treatment groups, respectively (table 3⇓). Most of the serious adverse events involved exacerbation of underlying UC. No deaths occurred during the study and no patient developed solid tumour or haematological malignancies. There were no clinically significant changes in laboratory values in either treatment group.

The incidence of infections was similar in the placebo and in the two daclizumab groups. Pneumonia occurred in two patients in the daclizumab 1 mg/kg group. None of the patients developed opportunistic infections.

Pharmacokinetics

Immunohistochemical analysis of lamina propria T cells

Immunohistochemical analysis showed, on average, a low number of CD5⁺ T cells in the mucosa of patients and a low percentage of CD25⁺ T cells (29–39%) prior to treatment. The numbers of both 143-13 positive cells and 2A-3 positive T cells were not significantly changed at day 56 in comparison with pretreatment values in any of the treatment groups, revealing CD25 expression on T cells and lack of CD25 saturation with daclizumab (fig 3⇓). In only four patients, no CD25 staining was detected with 2A3 anti-CD25 antibody on day 56 although T cells were stained with 143-13 antibody. These data indicate that in these subjects CD25 was expressed on T cells and saturated with daclizumab. Average counts of Ki67⁺ and caspase-3⁺ T cells demonstrate no effect of treatment when pre- and post-treatment samples were compared.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

 CD25 expression in colonic biopsies before and after daclizumab treatment. Immunohistochemistry using a non-competing antibody (143-13) and showing (brown colour) CD25 positive lymphocytes before (A) and after (B) treatment with high dose daclizumab. CD25 staining remained unchanged using both a competing (2A-3) and non-competing anti- CD25 antibody (143-13). Note that at baseline only a minority of mononuclear cells were CD25⁺. Pictures are representative of all patients evaluated.