Basiliximab for the treatment of steroid-resistant ulcerative colitis: further experience in moderate and severe disease

Authors


Dr T. J. Creed, Henry Wellcome Laboratories for Integrative Neuroscience and Endocrinology, Dorothy Hodgkin Building, Whitson Street, Bristol, UK BS1 3NY.
E-mail: t.j.creed@bristol.ac.uk

Abstract

Summary

Background

Preliminary data have suggested that interleukin-2 receptor blockade with basiliximab may increase steroid sensitivity. We have previously reported a small case series demonstrating the potential of basiliximab as a novel agent for the treatment of steroid-resistant ulcerative colitis.

Aim

To report further experience of the efficacy and safety of treatment with the interleukin-2 receptor blocking monoclonal antibody basiliximab, in addition to steroids, for the treatment of severe and moderate steroid-resistant ulcerative colitis.

Methods

Twenty patients were enrolled – 13 patients with moderate steroid-resistant ulcerative colitis (Ulcerative Colitis Symptom Score: ≥6) and seven patients with severe steroid-resistant ulcerative colitis. All were given a single dose of 40 mg basiliximab plus standard steroid therapy in an open-label, uncontrolled trial. Primary end point was clinical remission within 8 weeks (Ulcerative Colitis Symptom Score: ≤2).

Results

Within 8 weeks, 10 of 20 (50%) patients achieved clinical remission (seven of 13 moderate, and three of seven severe). At 24 weeks, 13 of 20 (65%) patients were in clinical remission. Five patients required colectomy (four severe, one moderate ulcerative colitis) and one required rescue ciclosporin (moderate ulcerative colitis). Two patients developed herpes zoster, but treatment was generally well tolerated.

Conclusions

Basiliximab appears to promote prolonged remission after a single treatment. Taken in combination with previously reported data, basiliximab shows particular promise in moderate steroid-resistant ulcerative colitis.

Introduction

Failure to respond adequately to steroid therapy remains a major clinical problem in ulcerative colitis (UC) with 20–30% of patients showing clinically significant steroid resistance.1, 2 Ciclosporin has proved effective as second-line therapy in many such patients but its adverse side-effect profile often precludes long-term treatment, and as a result it often does not prevent the need for colectomy but merely delays it.3 More recently, biological agents have been evaluated for use in UC including anti-CD3,4 anti-α4 integrin,5 interferons,6, 7 interleukin (IL)-108 and infliximab9–13 with mixed results. Biological agents that have shown potential therapeutic value in UC have been limited by the need for repeated dosing schedules,6, 7, 11, 13, 14 or concerns with respect to side-effects.5

The T-cell autocrine growth factor IL-2 is known to be able to antagonize the actions of steroids,15 and several points of potential antagonism between the IL-2 and steroid signalling pathways have been identified.16, 17 The high-affinity receptor for IL-2 (CD25) is markedly upregulated in activated T cells. We therefore hypothesized that blockade of the receptor would improve sensitivity to steroids in T cells activated by the UC disease process, with minimal adverse effects upon the remainder of uninvolved T cells.

Basiliximab (Simulect, Novartis Pharmaceuticals, Basel, Switzerland) is a chimaeric anti-CD25 monoclonal antibody with a proven track record of efficacy combined with a low side-effect profile in the prevention of allograft rejection.18 We have previously demonstrated a marked synergistic effect of basiliximab in combination with steroids in vitro and reported early evidence of efficacy in vivo in steroid-resistant UC19 in a series of 10 patients. Basiliximab alone in vitro was relatively ineffective.16 This is consistent with the mechanistic rationale for using IL-2 receptor inhibition in combination with steroids and the disappointing results of previous studies of another anti-CD25 agent (daclizumab) in UC that was not designed to study the combination of CD25 inhibition and steroids (Protein Design Laboratories (Fremont, CA, USA) – press release DDW 2004, trial data unpublished).

Here, we report the use of basiliximab in combination with steroids in 20 steroid-resistant patients. This enables further analysis of the potential of this agent as a novel treatment modality for steroid-resistant UC. In combination with the previously reported data we are also able to comment upon the safety of treatment and undertake a preliminary subgroup analysis.

Materials and methods

The Local Research Ethics Committees gave study approval (reference no. E5080), and all patients gave their written informed consent.

Patients

Thirteen patients with moderate UC and seven patients with severe steroid-resistant UC were enrolled from six hospitals in the Bristol region (Bristol Royal Infirmary, Frenchay Hospital, Southmead Hospital, Gloucestershire Royal Hospital, Cheltenham General Hospital and Royal United Hospital Bath).

Resistance to steroids in inflammatory bowel disease has been variously defined,20, 21 but in this study steroid resistance was defined as an Ulcerative Colitis Symptom Score (UCSS)22 of ≥6, and a Baron Endoscopic Score23 of ≥2 despite treatment with prednisolone ≥30 mg for a minimum of 14 days. Historical controls would suggest that this group of patients would be expected to have only a 30% chance of entering remission within 8 weeks with continued steroids alone.9 Steroid resistance with severe disease activity (Truelove and Witts criteria1) was defined as either those patients that fulfilled criteria for a poor outcome after 3 days of intravenous hydrocortisone at 400 mg/day [C-reactive protein (CRP) >45 mg/L plus three bloody stools per day, or a stool frequency of >8 per day],24 or an inadequate response after 7 days of hydrocortisone treatment at 400 mg/day (>3 stools per day or visible blood in the stool, but not immediately requiring surgery). Patients with severe steroid-resistant colitis would be expected to have a colectomy rate of 85% on that admission with continued steroid therapy alone.24

Exclusion criteria included: need for urgent colectomy, pregnancy (or planned pregnancy within 6 months of receiving treatment), recent investigational drug use, prior use of infliximab, positive stool culture (including Clostridium difficile), malignancy or significant dysplasia, autoimmune disease, evidence of tuberculosis (TB) or opportunistic infection. Recent introduction of azathioprine or mercaptopurine (6-mercaptopurine) was an exclusion criterion, but stable doses for >3 months at the time of enrolment was acceptable and was continued unchanged throughout the trial. Similarly, patients on stable doses of oral 5-aminosalicylic acid (5-ASA) drugs at enrolment were not excluded, and continued on these throughout the study. Topical 5-ASA use was not permitted because of the effect upon endoscopic and histological scoring. Patient characteristics at baseline are shown in Table 1.

Table 1.   Demographics
 Moderate (n = 20)Severe (n = 10)
  1. 5-ASA, 5-aminosalicylic acid; UCSS, Ulcerative Colitis Symptom Score; IBDQ, Inflammatory Bowel Disease Questionnaire.

Age, median (range)45 years (18–66)33 years (18–64)
Male/female14/68/2
Disease extent
 Pan colitis32
 Extensive25
 Left-sided93
 Recto-sigmoid60
Median duration of continuous steroid treatment at enrolment (range)36 days (14–365)14 days (3–200)
Number on azathioprine at enrolment5 (1 previously intolerant, 2 previously ineffective) 
Number on 5-ASA at enrolment128
Median UCSS at enrolment (range)9 (6–12)10.5 (9–12)
Median IBDQ at enrolment (range)112 (63–187)105 (73–134)
Median histology score at enrolment (range)10.5 (4–15)11.5 (7–15)

Intervention and additional management

All patients received a single intravenous infusion of basiliximab 40 mg over 5 min. No dose adjustment for patient weight is required for basiliximab. This dose was selected based upon experience with its use in renal transplantation as the dose required to occupy >90% of CD25 molecules.25

All subjects continued on prednisolone 30 mg/day (moderate group) or intravenous hydrocortisone 400 mg daily (severe group) until stool frequency fell to three or less per day with no visible blood. Patients in the severe group who responded were then changed to oral prednisolone 30 mg/day. The prednisolone dose for all patients was subsequently reduced by 5 mg/week whilst they remained in remission. If patients relapsed after achieving remission, then they remained in the study but the dose of prednisolone was increased to 30 mg and azathioprine was started at a dose of 2.5 mg/kg which they continued throughout the trial.

Assessments and end points

Patients were reviewed at weeks 0, 1, 2, 4, 6, 8, 12, 16, 20 and 24. At each visit, full blood count, urea and electrolytes, liver function tests, glucose and CRP were measured and the Inflammatory Bowel Disease Questionnaire (IBDQ)26 was completed. Rectal biopsy was performed at 0, 2 and 4 weeks. All biopsies were scored blinded by an experienced histopathologist using the scoring system of Riley et al.27 UCSS was assessed at weeks 0, 1, 2, 4, 6, 8 and 24. The primary end point was predetermined as clinical remission within 8 weeks of treatment as defined by a UCSS ≤ 2. Secondary outcome measures included reduction in UCSS, dose of steroids, IBDQ score, histological disease activity, safety and tolerability. Long-term outcomes (at 24 weeks) were also evaluated.

Statistical methods

Results are reported as median with interquartile ranges (IQR). The Sign test was used to analyse differences in histology scores, UCSS, IBDQ and steroid dose. As this was an open-label pilot study, no formal power calculations were performed. Moderate and severe groups were analysed separately and all outcomes were included in the analysis. Patients requiring colectomy or ciclosporin were given the maximal UCSS of 12 for analysis at subsequent time points. For analysis of steroid doses, intravenous hydrocortisone at 400 mg/day was considered equivalent to prednisolone 80 mg daily and patients requiring colectomy or ciclosporin were assigned a score of 40 mg of prednisolone (moderates) or 80 mg of prednisolone (severe). For IBDQ scores, patients requiring colectomy or ciclosporin had their last score carried forwards.

Results

Ulcerative Colitis Symptom Score

Within 8 weeks, 10 of 20 (50%) patients achieved clinical remission (seven of 13 moderate and three of seven severe). In addition, 13 of 13 (100%) of moderate patients and three of seven (43%) of severe patients demonstrated an improvement in UCSS over baseline within 8 weeks (see Figure 1). Median UCSS at week 8 had fallen from 10 (IQR: 8.8–11) to 5.5 (IQR: 0.8–9, P < 0.05). One patient required rescue therapy with ciclosporin (at week 2), and five patients required colectomy (four severe, one moderate). In both moderate and severe groups, improvement over baseline, if it was going to occur, became apparent in the majority of patients by the first follow-up visit (week 1, Figure 1).

Figure 1.

 Ulcerative Colitis Symptom Score (UCSS): bars denote median scores, colectomy/ciclosporin patients given score of 12. * P < 0.05, ** P = 0.03.

Figure 1 also demonstrates that after an initial rapid reduction in UCSS, a small rise in median UCSS was seen in some patients at 8 weeks. This deterioration is consistent with falling circulating basiliximab levels at this time.25 However, rechallenge with an increased steroid dose (30 mg prednisolone) as predefined in the protocol produced prompt improvement in most patients within 2 weeks, despite their initial resistance to steroids. As a consequence, at 24 weeks, 13 of 20 (65%) patients were in clinical remission (three severe, 10 moderate). Five patients required colectomy; four severe patients had colectomy within the first 2 weeks, and one moderate patient had colectomy at week 24 for persistent disease activity. One moderate patient required rescue ciclosporin at week 2 because of worsening activity of colitis and subsequently achieved full remission. Median UCSS at 24 weeks was 1.5 (IQR: 0–12, P = 0.03 compared with baseline, Figure 1).

Dose of Steroids

Significant reductions in the median steroid dose were seen despite allocating patients who underwent colectomy or received ciclosporin a dose of 40 mg or 80 mg of prednisolone (moderate and severe groups, respectively) for all subsequent statistical analysis (Figure 2). Median dose of prednisolone reduced from 30 mg at entry to 20 mg at week 8 (IQR: 8.8–32.5, P < 0.001 compared with baseline) and to 10 mg by week 24 (IQR: 0–40, P < 0.001 compared with baseline). Nine of 20 (45%) patients were in steroid-free remission at 24 weeks.

Figure 2.

 Steriod dose: patients requiring colectomy or ciclosporin were assigned a score of 40 mg (moderate) or 80 mg (severe). Intravenous (IV) hydrocortisone is represented as equivalent to 80 mg prednisolone. Bars denote median values (* P < 0.01).

Quality of life – IBDQ

This validated 32-part questionnaire of quality of life and disease activity showed significant improvements commensurate with the observed reduction in clinical activity (Figure 3). The range of possible scores is 32–224 with a higher score denoting quiescent disease. Baseline IBDQ was 107 (IQR: 87–126) and rose to 163 by week 8 (IQR: 109–180, P = 0.01 compared with baseline). By week 24 the median IBDQ had risen further to 179 (IQR: 134–194, P < 0.001 compared with baseline).

Figure 3.

 Inflammatory Bowel Disease Questionnaire (IBDQ) score: bars denote median score. Patients undergoing colectomy or receiving ciclosporin had their last score carried forwards. Bars denote median values (*P = 0.01, ** P = 0.001).

Histology

Biopsies were scored by an experienced blinded histopathologist. Patients undergoing colectomy or receiving ciclosporin had their previous histology score carried forward. Because of severe activity of the colitis, some biopsies were inadequate for scoring (pus only) and excluded from statistical analysis (subsequent scores could thus not be carried forward in those undergoing colectomy). One patient declined a biopsy, and one patient did not have a biopsy because of excessive bleeding following a previous biopsy (on two antiplatelet agents). Forty-seven of 60 (78%) biopsies were available for scoring. Baseline median Riley score was 10 (IQR: 8–12) and fell to 7.5 (IQR: 3–11.3, P = 0.02) at week 2 and to 6 (IQR: 1.3–10.3, P = 0.03) at week 4 (Figure 4).

Figure 4.

 Histology: six histological features were graded on a 4-point scale: 1, acute inflammatory cell infiltrate; 2, crypt abscesses; 3, mucin depletion; 4, surface epithelial integrity; 5, chronic inflammatory cell infiltrate; 6, crypt architectural irregularities. Scores shown are total scores for all features. Patients undergoing colectomy or receiving ciclosporin had previous score carried forward. Bars denote median score (* P = 0.02, ** P = 0.03).

Safety

Basiliximab was very well tolerated. There were no infusion reactions. Adverse events are noted in Table 2, but in the absence of a placebo group, it is not possible to determine to what degree these are attributable to the basiliximab infusion. Both individuals who developed herpes zoster were taking azathioprine and steroids. One developed signs of zoster following a colectomy at week 2. The second patient developed zoster 16 weeks post-basiliximab infusion. Both episodes resolved rapidly with no significant complications and did not require anti-viral therapy.

Table 2.   Adverse events
Adverse eventsNumber of patients
Herpes zoster2
Rash2
Pyrexia1
Bleeding following rectal biopsy1
Dyspnoea1
Lethargy1
Paraesthesiae1
Loin pain1

One patient required rescue ciclosporin (moderate group) after receiving basiliximab, but had no adverse events from this combination and rapidly achieved remission.

The patient who developed rectal bleeding was on two antiplatelet agents for prevention of stroke. He required admission to hospital for observation, but the bleeding stopped with conservative management. The remaining reported adverse events were mild and required no medical intervention.

Discussion

In our initial clinical report of the first 10 steroid-resistant patients in this series (seven moderate, three severe),19 we observed a significant overall improvement in the patients by both weeks 8 and 24. This further series of 20 patients demonstrates comparable improvements. The data from all 30 patients over 6 months of follow-up supports the safety and good tolerability profile of basiliximab in combination with steroids.

In addition, by combining the two series we are able to undertake some preliminary subgroup analysis with data from 20 moderate and 10 severe patients.

Within 8 weeks, 19 of 20 (95%) of patients in the moderate group had improved (reduced) UCSS and 14 of 20 (70%) had achieved full clinical remission; 10 of 20 (50%) were in remission at the 8-week time point. Median UCSS at week 8 had fallen from 9 (IQR: 8–10) to 2.5 (IQR: 0–6, P < 0.001). Median UCSS in the severe group as a whole (allocating colectomy/ciclosporin patients a maximum score of 12) fell from 10.5 (IQR: 10–11) to 8.5 (IQR: 0.5–12) that did not reach statistical significance.

Baseline median histology score was 10.5 (IQR: 8.3–11) in the moderate group and fell to 6 (IQR: 3–9, P = 0.001) at week 2, and 4 (IQR: 2–7.3, P = 0.021) at week 4. Median histology score in the severe group began higher at 11.5 (IQR: 8–13.3); the fall at week 2, 7.5 (IQR: 5–12) and week 4, 7.5 (IQR: 3.3–12, P = 0.63) did not reach statistical significance but in the five non-colectomy patients, histology scores were all ≤4 by week 4.

In the moderate group, median dose of prednisolone had reduced from 30 to 15 mg (IQR: 3.8–21.3, P = 0.001) at week 8 and to 0 mg (IQR: 0–10, P < 0.001) by week 24, with 12 of 20 patients completing steroid therapy altogether at this stage. In the severe group, the median dose of prednisolone fell from 80 to 50 mg at week 8 (IQR: 50–80, P = 0.06) and 55 mg at week 24 (IQR: 6.3–80, P = 0.06), although much of this was accounted for by the score of ‘80 mg’ allocated to the colectomy patients in the severe group. Of the patients who did not require colectomy, only one of five was taking more than 10 mg of prednisolone at either week 8 or week 24.

This provides an early indication that the combination of steroids and basiliximab appears to work particularly well in moderate steroid-resistant UC, with 70% of patients in remission within 8 weeks, rising to 75% at 24 weeks (n = 20). These improvements were achieved at the same time as a reduction in steroid dose so that 13 of 20 moderate patients were on no steroids at the end of the study. The improvements were also directly paralleled by significant improvements in IBDQ and blinded histology. In the severe patients (n = 10), the results divided into two groups: 50% did not respond rapidly and required colectomy in the first week while 50% responded well with four of five in remission at 8 weeks and five of five in remission at 24 weeks. In addition, basiliximab appeared to act very quickly, enhancing the effect of steroids as early as 1 week after infusion.

This was an open-label uncontrolled study and caution should be exercised in comparing results with controlled studies. Nonetheless the response rates for steroid-resistant cases of UC were striking. Remission rates were significantly greater than reported in all recent studies of biological agents, namely 10–53%.4–10 Data from previous studies indicate that patients with active moderate UC despite 7 days prednisolone (30 mg/day) would be expected to have response rates of at most 30% after 6 weeks with continued prednisolone therapy alone.9 Thus, it appears that a delayed response to steroids is likely to explain <50% of the response seen after the addition of anti-CD25 in the present study. In addition, patients enrolled in the present study had a minimum of 14 days of high-dose steroids at enrolment, and so one might expect the remission rate with steroids alone in this cohort to be even lower. In severe UC, patients with poor predictors of outcome similar to those in this study have been shown to have only a 15% chance of avoiding colectomy on that admission.24 Whilst there were only 10 patients in the severe arm of this study, a remission rate of 50% might be considered very encouraging. Data from the ACT 1 and ACT 2 trials13 of infliximab in a group of patients similar to our moderate UC group demonstrated a clinical response rate to infliximab of 65% and 69%, compared with our response rate of 95% in the moderate group (n = 20). Furthermore, the proportion of patients who were able to discontinue steroids in the ACT 1 trial was 22% (week 30) when compared with 60% (week 24) of moderate patients in this study.

The overall (moderate and severe cases, n = 30) induction of remission rate within 8 weeks of 63% seen in the present study is comparable with response rates of between 60% and 90% reported with ciclosporin,3, 28–32 but there was no evidence of the significant associated adverse effects that are seen with the latter agent.33 Importantly, the use of basiliximab does not exclude the use of ciclosporin, indeed basiliximab has previously been extensively used in combination with ciclosporin in transplantation.34

Many of the other emerging biological agents for medical use are associated with significant adverse events such as leucopoenia, or cytokine release syndrome. The target for basiliximab is CD25, which has no cytoplasmic tail, and therefore it is not associated with cytokine release syndrome. Furthermore, CD25 is only expressed on activated T cells, and therefore it is only the active clones of T cells that are targeted. This therefore does not result in widespread T-cell depletion of resting cells, or increased sensitivity to steroids at other target organs, which might precipitate worsening steroid side-effects.

A surprising finding was that basiliximab appeared to induce a prolonged phase of steroid sensitization, such that relapses that occurred around 6–8 weeks (when antibody levels would be expected to be declining in the plasma) responded very promptly to moderate doses of steroid without the need for further IL-2 receptor blockade (Figure 1). This resulted in remission rates at 24 weeks that were greater than at 8 weeks. A possible mechanism for this effect is the induction of apoptosis in large numbers of steroid-resistant T-cell clones in the bowel following anti-CD25 treatment. Preliminary data from our own laboratory, tracking proliferation in the presence of steroids in vitro using carboxy-fluorescein diacetate succinimidyl ester staining and flow cytometry, suggests that steroid resistance relates to a subpopulation of T cells that continue to proliferate even in the presence of high-dose steroids. If so, such clones would be expected to predominate in the gut after 2 weeks of steroid therapy. CD25 blockade appears to restore the sensitivity of these cells to steroids, resulting in rapid depletion in the presence of steroid. It may then take many months for this population to recover, leaving the patient clinically very much more steroid-sensitive over this period of time even though no more anti-CD25 is administered. The steroid sensitization seen following administration of single-dose basiliximab raises the possibility of employing it in conjunction with reduced dose steroid therapy in steroid-sensitive individuals. This could potentially minimize steroid side-effects in those patients with significant steroid-related problems such as osteoporosis or diabetes.

Preliminary data are available on the use of an alternative anti-CD25 monoclonal antibody, Daclizumab (Zenepax, Protein Design Laboratories), a humanized rather than chimaeric antibody. This agent initially showed promise in an uncontrolled pilot study by van Assche et al.,35 but a subsequent RCT with 159 patients (unpublished) was reported at Digestive Diseases Week 2004 to have negative results. However, although not all the study details are available yet, there appear to have been important differences in study design compared with the current study. Many of the patients in the van Assche study35 and subsequent RCT were not truly steroid-resistant, not all received concomitant steroid therapy, many received submaximal doses of daclizumab and different end points were used. We have previously demonstrated in vitro that the action of basiliximab upon T cells is as a steroid sensitizer, and that its effects alone are inferior to steroids alone.19 We therefore believe that CD25 blockade, unlike other monoclonal antibody therapies, is only likely to be beneficial when used in combination with steroid treatment.

In conclusion, we have confirmed in this extended pilot study that basiliximab in combination with steroids is potentially very effective for the induction of remission in both moderate and severe steroid refractory UC. In particular, patients with moderately severe steroid-resistant colitis appear to benefit. Furthermore, the treatment is well tolerated and appears to generate a state of increased sensitivity that persists for many months after disappearance of the monoclonal antibody from the circulation. Anti-CD25 therapy may thus possess many features that make it very different from other biological agents used in inflammatory bowel disease. If these findings are confirmed in a randomized-controlled trial (RCT), anti-CD25 may prove to be a much needed addition to our limited range of treatment options in steroid resistant UC.

Acknowledgements

Tom Creed was supported by a research grant from Novartis Pharmaceuticals (UK).

The authors would like to thank the following members of Novartis Pharmaceuticals who provided encouragement and technical support: Marcus Arden Jones, David Youds, Cory Ingliss, George Gooding (UK), Michael Cooreman (USA).

Ancillary