Basiliximab (anti-CD25) in combination with steroids may be an effective new treatment for steroid-resistant ulcerative colitis

Authors


Dr C. M. Dayan, University Research Centre for Neuroendocrinology, Bristol Royal Infirmary, Bristol, BS2 8HW, UK.
E-mail: colin.dayan@bristol.ac.uk

Summary

Background : Steroid resistance represents a major clinical problem in the treatment of ulcerative colitis. In vitro, interleukin-2 renders lymphocytes steroid resistant.

Aim : To explore the therapeutic potential of interleukin-2 receptor blockade in steroid-resistant ulcerative colitis with both in vitro measures and a pilot in vivo study.

Methods : Ten patients with steroid-resistant ulcerative colitis received a single bolus of 40 mg of intravenous basiliximab plus steroid treatment in an open-label, uncontrolled, 24-week study. The outcome was assessed using the Ulcerative Colitis Symptom Score, rectal biopsy and Inflammatory Bowel Disease Questionnaire. Lymphocyte steroid sensitivity was measured in vitro in 39 subjects in the presence or absence of basiliximab.

Results : Nine of the 10 patients achieved clinical remission within 8 weeks. At 24 weeks, seven patients were in clinical remission. Marked improvement in the Ulcerative Colitis Symptom Score was seen by 1 week (P = 0.004) and on rectal biopsy and Inflammatory Bowel Disease Questionnaire by 2 weeks (both P < 0.05). Improvements persisted to 24 weeks (Ulcerative Colitis Symptom Score, Inflammatory Bowel Disease Questionnaire, both P < 0.005). Eight of the nine responders relapsed (median, 9 weeks), but remission was re-achieved with further corticosteroids and the addition of azathioprine. At 24 weeks, seven patients were in full clinical remission, five off all steroid therapy. In vitro measurement of lymphocyte steroid sensitivity demonstrated steroid resistance in 22% of subjects. All were rendered steroid sensitive in the presence of basiliximab.

Conclusions : Basiliximab appears to be effective at inducing remission in steroid-resistant ulcerative colitis. In vitro, basiliximab also produced a dramatic increase in lymphocyte steroid sensitivity in healthy subjects. Confirmation in randomized controlled studies is required.

Introduction

Up to 30% of patients with ulcerative colitis show a poor response to steroid therapy,1, 2 and this presents a major clinical challenge. Immunosuppressant therapies have been used for steroid-resistant ulcerative colitis, notably ciclosporin, which can be effective in up to 50–80% of cases, although many patients subsequently relapse and require colectomy within 6 months.3 Ciclosporin can cause serious adverse reactions, such as renal failure and fatal immunosuppression,4, 5 and careful monitoring of patients is required, including drug level measurement. Azathioprine has been used to improve the maintenance of remission achieved with ciclosporin, and ciclosporin has been used without corticosteroids to reduce the incidence of side-effects.6 Azathioprine and 6-mercaptopurine alone are better tolerated as immunosuppressive agents, but their slow onset of action (3–4 months) precludes their use in severe disease. A therapy with comparable or greater efficacy than ciclosporin, but with a lower incidence of serious adverse reactions, would be a useful advance in the treatment of steroid-resistant ulcerative colitis and may avoid the need for emergency colectomy.

Individuals who fail to respond to steroid therapy do not necessarily have more severe disease.7 Indeed, in vitro studies have suggested that steroid resistance is an intrinsic property of an individual's T lymphocytes, independent of the presence or severity of disease.8 Lymphocyte steroid sensitivity measured in vitro has been shown to correlate well with clinical outcome in ulcerative colitis,9 and similar observations have been made in other inflammatory conditions, such as asthma,10 systemic lupus erythematosus11 and renal transplant recipients.12 In studies on healthy individuals, up to 30% were found to have lymphocyte steroid sensitivity values consistent with steroid resistance, despite the absence of disease, and lymphocyte steroid sensitivity remained fairly constant for a given individual over time.8

Preliminary evidence suggests a key paracrine role for the cytokine growth factor interleukin-2 (IL-2) in the induction of steroid resistance in T lymphocytes (T cells). IL-2 is secreted by activated T cells and acts via the high-affinity IL-2 receptor (IL-2R) on T cells themselves to promote cell survival and proliferation. The IL-2R α-chain (CD25, required for high-affinity IL-2 binding) is absent on resting T cells, but is rapidly (although transiently) expressed on activation. Lymphocytes from steroid-resistant individuals appear to produce higher levels of IL-2 than those from steroid-sensitive individuals.13 Furthermore, the addition of IL-2 (and IL-4) to steroid-sensitive lymphocytes in vitro renders them steroid resistant.14, 15 We therefore hypothesized that agents able to inhibit IL-2 action might act as steroid sensitizers.

Basiliximab (Simulect, Novartis Pharmaceuticals, Basle, Switzerland) is a commercially available chimeric monoclonal antibody, which binds to CD25. Basiliximab was developed as an immunosuppressive agent for use in transplantation and has been proven to be both safe and effective in this role in large randomized clinical trials.16, 17 CD25 has no cytoplasmic tail, and is therefore a good target as it is less likely to cause cytokine release with its attendant problems. Cell signalling is mediated by the β- and γ-receptor sub-units, driving subsequent T-cell proliferation. When complexed with basiliximab, IL-2Rs are unable to bind IL-2 and hence T-cell proliferation is inhibited. Studies in renal transplantation have demonstrated that the administration of basiliximab does not change the expression of IL-2Rs on circulating T cells.18 IL-2R saturations of approximately 90% have been demonstrated in vitro on cultured T lymphocytes, and this level of basiliximab binding was maintained for 36 ± 14 days when renal transplant patients received basiliximab, 20 mg, 2 h before and then 4 days after transplantation surgery.19 Although basiliximab is generally used in combination with steroids, it has not previously been used in autoimmune disease or considered to be particularly synergistic with steroid action.

In the current study, we aimed to explore the potential of basiliximab to enhance the action of steroids both in vitro and in vivo in the treatment of steroid-resistant ulcerative colitis. We report preliminary evidence of the remarkable clinical efficacy of basiliximab in steroid-resistant ulcerative colitis in an open-label, uncontrolled study. In addition, we demonstrate the in vitro effect of basiliximab on lymphocyte steroid sensitivity in a large number of healthy volunteers and patients with quiescent ulcerative colitis.

Methods

Clinical trial

This was an open-label, uncontrolled study of basiliximab for patients with steroid-resistant ulcerative colitis. The Local Research Ethics Committees gave study approval, and all patients gave written informed consent.

Patients with moderate and severe steroid-resistant ulcerative colitis were enrolled. Entry to the moderate group required an Ulcerative Colitis Symptom Score (UCSS)20 of ≥ 6 and a Baron Score21 of ≥ 2, despite treatment with prednisolone at ≥ 30 mg for a minimum of 14 days. Severe colitis was defined as six or more bloody stools per day with one or more of the following on admission: tachycardia of > 90 beats/min, temperature of > 37.8 °C, erythrocyte sedimentation rate of > 30 mm/h, plasma viscosity of > 1.85 mPa or haemoglobin level of < 10.5 g/dL (Truelove and Witts criteria22). Patients in the severe group eligible for treatment with basiliximab either fulfilled the criteria for a poor outcome after 3 days of intravenous hydrocortisone at 400 mg/day (C-reactive protein level of > 45 mg/L plus three bloody stools per day, or a stool frequency of more than eight per day),2 or had an inadequate response after 7 days of hydrocortisone treatment (more than three stools per day or visible blood in the stool, but not immediately requiring surgery). Stool samples were taken to exclude significant pathogens including Clostridium difficile.

All patients received a single intravenous infusion of basiliximab of 40 mg over 5 min. No dose adjustment for patient weight is required for basiliximab, and a single 40 mg infusion has been used previously in renal transplantation.23 In addition, all subjects continued on prednisolone 30 mg/day (moderate group) or intravenous hydrocortisone 400 mg/day (severe group) until the stool frequency fell to three or less per day with no visible blood. Patients in the severe group were then changed to oral prednisolone 30 mg/day. The prednisolone dose for all patients was subsequently reduced by 5 mg per week whilst they remained in remission. If patients relapsed after achieving remission, 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. All additional medications, including 5-aminosalicylic acid, were continued.

Patients were assessed at 0, 1, 2, 4, 6 and 8 weeks and subsequently every 4 weeks until 24 weeks. At each visit, full blood count, urea and electrolytes, liver function tests, glucose and C-reactive protein were measured. The UCSS20 was assessed, and the Inflammatory Bowel Disease Questionnaire (IBDQ) was completed at each visit.24 Rectal biopsy was performed at 0, 2 and 4 weeks. All biopsies were performed blind and were scored by an experienced histopathologist (NAS) at one sitting using the scoring system of Riley et al.25 A blood sample for lymphocyte separation and measurement of lymphocyte steroid sensitivity was taken prior to treatment with basiliximab, and 1 week after treatment.

The primary outcome was complete clinical remission within 8 weeks of treatment, as defined by UCSS ≤ 2. Secondary outcome measures included changes in UCSS, dose of prednisolone, IBDQ score, C-reactive protein, albumin and histological disease activity. Long-term outcome (24 weeks) and steroid usage were also evaluated. All adverse events were recorded, including glucocorticoid side-effects.

In vitro lymphocyte steroid sensitivity

Lymphocyte steroid sensitivity was measured according to a standard assay, as described previously.8 Briefly, peripheral blood mononuclear cells were isolated by buoyant density centrifugation and plated at 4 × 105 cells per well in 96-well plates in RPMI1640 medium with 10% foetal calf serum (Gibco, Paisley, UK). Lymphocyte steroid sensitivity was assessed by measuring the anti-proliferative effect of increasing amounts of dexamethasone (10−10−10−6 mol/L) in the presence of stimulation by phytohaemagglutinin (Sigma, Poole, UK) at a concentration of 5 µg/mL. Cultures were incubated for 48 h at 37 °C in 5% CO2, and proliferation was measured by the uptake of tritiated thymidine. Results were calculated as the counts per minute (c.p.m.) from the mean of triplicate cultures and presented as the percentage inhibition at 10−6 mol/L dexamethasone [Imax = (c.p.m. with phytohaemagglutinin alone −  c.p.m. with dexamethasone)/c.p.m. with phytohaemagglutinin alone × 100%]. Where indicated, 1 mg/mL of basiliximab was added, a concentration which achieves > 90% IL-2R occupancy in vivo.26

Statistical methods

The paired t-test was used to analyse differences in lymphocyte steroid sensitivity, IBDQ and histology scores. The Sign test was used to analyse differences in paired UCSS, serum albumin and C-reactive protein scores. As this was an open-label pilot study, no formal power calculations were performed.

Results

Clinical trial

Twenty-one patients were referred for treatment. Eleven patients were excluded for the following reasons: insufficiently active disease (six), concomitant other disease (three) (Clostridium difficile infection, alcoholism, chronic lung disease with recurrent chest infection), patient unwilling to adhere to trial follow-up arrangements (one) and colectomy prior to enrolment (one). Of the 10 patients treated, seven fulfilled the criteria for moderate disease and three for severe disease. The baseline characteristics of the patients are shown in Table 1.

Table 1.  Patient demographics at enrolment
PatientAge
(years)
SexDisease
extent
Duration of
UC (months)
Duration of continuous
steroid treatment (days)
Disease
severity

Azathioprine
  • F, female; M, male; UC, ulcerative colitis.

  • *

     Patient 9 underwent colectomy.

 135MPancolitis15635ModerateNo
 266MPancolitis 4814ModerateNo (previously
 intolerant)
 346FLeft-sided 3636ModerateNo
 456MRecto-sigmoid120365ModerateYes
 549MLeft-sided 48200SevereNo (previously
 ineffective)
 641MLeft-sided 72110ModerateYes
 745FLeft-sided13236ModerateNo
 855MLeft-sided 6015ModerateNo (previously
 ineffective)
 9*64MPancolitis  1.530SevereNo
1033FHepatic flexure 1856SevereNo

The three patients with severe ulcerative colitis were all being treated with intravenous hydrocortisone at enrolment. Two of these patients had received 3 days of intravenous hydrocortisone, preceded by 1 month of oral prednisolone. Both of these patients fulfilled day 3 criteria for a poor outcome (C-reactive protein >45 mg/L plus three bloody stools per day). The other patient had received 14 days of intravenous hydrocortisone, preceded by 56 days of oral prednisolone. In the moderate ulcerative colitis group, previous high-dose therapy (≥ 30 mg prednisolone per day) had been instituted for a median of 20 days (range, 14–23 days), although total continuous previous prednisolone administration at any dose was considerably longer for all patients (median, 36 days; range, 14–365 days). Six of the 10 patients were being considered for colectomy because of poor response to medical therapy at the time of entry to the study.

Initial response. Nine of the 10 (90%) patients achieved the primary outcome of complete clinical remission (UCSS ≤ 2) within 8 weeks. The remaining patient, who had severe ulcerative colitis, required a colectomy 9 days after treatment with basiliximab, and subsequently made an uneventful recovery.

All nine responders demonstrated significant improvement in the clinical activity score (UCSS) as early as week 1: at week 0, median UCSS = 9 (range, 7–11); at week 1, median UCSS = 3 (range, 2–5); P = 0.004 (Figure 1). Figure 2 shows the dose of steroids received by the study patients, which was reduced once remission was achieved according to the protocol.

Figure 1.

Median Ulcerative Colitis Symptom Score (UCSS). Filled squares represent median values and error bars denote the interquartile range. The broken line represents clinical remission (UCSS ≤ 2).

Figure 2.

Median steroid dose. Filled squares represent median values and error bars denote the interquartile range. Intravenous hydrocortisone, 400 mg, has been represented as 80 mg of prednisolone.

Quality of life scores (IBDQ) also showed a rapid and dramatic improvement (Figure 3): at week 0, mean IBDQ = 122.8; at week 2, mean IBDQ = 162.6 (the first assessment after baseline); mean difference of paired scores, 39.8 [95% confidence interval (CI), 18.2–61.4]; P < 0.005. It should be noted that the 2-week IBDQ value is comparable with the IBDQ scores reported for patients in complete remission (73% of maximum theoretically attainable score).27

Figure 3.

Median Inflammatory Bowel Disease Questionnaire (IBDQ) score. Filled squares represent median values and error bars denote the interquartile range.

In line with the rapid clinical and quality of life responses, marked changes in histological appearances were seen from week 2 (first biopsy following basiliximab), with clearing of the inflammatory infiltrate in the mucosa, resolution of cryptitis and a notable diminution in the round cell infiltrate in the lamina propria. Appearances at week 4 often showed fewer chronic inflammatory cells than would be expected in normal mucosa (Figure 4). Total histological disease activity scores25 (assessed blind) for acute and chronic inflammatory changes showed significant improvements at both week 2 (mean difference of paired samples, 3.22; 95% CI, 0.8–5.65; P = 0.015) and week 4 (mean difference of paired samples, 5.0; 95% CI, 0.81–9.19; P = 0.026) vs. baseline (Figure 5). Histological disease activity scores for individual patients were improved for nine of the 10 patients at week 2 and for seven of the 10 patients at week 4.

Figure 4.

Rectal biopsies before (week 0) (a) and after (week 4) (b) with basiliximab.

Figure 5.

Histology score. Scores based on Riley et al.25 Six histological features were graded on a four-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. One patient had an inadequate biopsy at entry and one patient had a colectomy at week 2 and therefore did not have a biopsy at week 4.

Longer term outcome. Eight patients relapsed (UCSS > 2) after initially entering remission, although only one patient had a UCSS value as high as that at enrolment (patient 4). Relapses were seen at week 4 (patients 1 and 4), week 6 (patient 6), week 8 (patient 5), week 10 (patient 7), week 12 (patient 8), week 16 (patient 2) and week 20 (patient 10). The median time to relapse in these eight patients was 9 weeks. At the time of relapse, the prednisolone doses were 0 mg (patients 2, 7, 8 and 10), 10 mg (patient 5), 15 mg (patients 1 and 4) and 20 mg (patient 6). These patients were treated with an increased dose of oral prednisolone (30 mg) and azathioprine therapy was initiated at 2.5 mg/kg daily. Two of these patients were already taking azathioprine, and one patient had previously shown intolerance to the drug (Table 1). Two patients could not tolerate azathioprine (patient 1 developed leucopenia and patient 7 developed liver function disturbance — both resolved with withdrawal of azathioprine). With this regimen, remission was re-achieved for all but one patient (week 8, patients 1 and 6; week 12, patients 4, 5 and 7; week 16, patient 8; week 24, patient 10). One relapsed again when the steroid dose was reduced to < 5 mg (patient 4). Overall, seven of the 10 patients (70%) were in remission 24 weeks after receiving basiliximab, four of whom were on azathioprine, and five of whom were receiving no steroids at all.

Despite the early relapses in four of the nine patients who initially responded, the median UCSS demonstrated sustained improvements compared with baseline at all time points: week 8, 0.5 (range, 0–5; P < 0.008); week 24, 0 (range, 0–7; P = 0.004). Sustained improvements in the mean difference of the paired IBDQ scores were also seen throughout the study: week 4, 50.38 (95% CI, 25.8–74.9); week 6, 51.88 (95% CI, 27.2–76.6); week 8, 55.6 (95% CI, 33.2–77.9); week 24, 62.6 (95% CI, 38.1–87.1); all P < 0.005 vs. baseline. In addition, the mean albumin levels improved significantly: mean albumin at baseline, 35.9 g/dL; mean albumin at week 8, 41.1 g/dL (P < 0.05); mean albumin at week 24, 43.3 g/dL (P < 0.05). Changes in C-reactive protein were not statistically significant: mean C-reactive protein at baseline, 23.1 mg/L; mean C-reactive protein at week 24, 12.3 mg/L (P = 0.69). However, C-reactive protein values in many patients were undetectable (< 10 mg/L) at the end of the study, and a significant improvement might have been seen if a highly sensitive assay had been used.

Adverse events. One patient developed pyrexia 1 week after treatment with basiliximab which was treated with antibiotics. Subsequently, all cultures were negative, there was no evidence of bowel perforation and the patient recovered completely. The following five adverse events in five patients were recorded: mild paraesthesia of the feet, transient photosensitivity, transient lethargy, loin pain and upper respiratory tract infection. Many patients had steroid side-effects (moon face, weight increase, etc.) at entry from previous steroid use. No increase in steroid side-effects was seen after treatment with basiliximab and, as steroid requirements fell, the resolution of some steroid side-effects was noted.

In vitro lymphocyte steroid sensitivity

Lymphocyte steroid sensitivity was measured in 39 healthy subjects (32 normal volunteers and seven with quiescent ulcerative colitis) on 45 occasions. Eleven (24%) measurements (22% of subjects, n = 10) demonstrated Imax < 60%, a value associated with an intermediate or poor response to steroids in vivo in our previous study (‘steroid resistant’; Figure 6).9 The percentage of subjects appearing as ‘steroid resistant’in vitro (24%) was consistent with previous observations in healthy individuals.8 Basiliximab alone was less potent as an anti-proliferative agent than dexamethasone alone. However, with basiliximab in combination with dexamethasone, all measurements showed Imax >60% (‘steroid sensitive’; Figure 6). Of note, all measurements from ‘steroid-resistant’ individuals (Imax <60%) were modulated to ‘steroid sensitive’. The median percentage inhibition values were: dexamethasone (10−6 mol/L) alone, 77.2%[interquartile range (IQR), 60.9–85.9%]; basiliximab alone, 29.4% (IQR, 10.35–44.2%); dexamethasone (10−6 mol/L) + basiliximab, 92.2% (IQR, 89.2–95.6%). Thus, the addition of basiliximab increased the median inhibition of lymphocyte steroid sensitivity from 77.2% to 92.2% (P < 0.0001) (Figure 6). Furthermore, at sub-maximal doses of steroid (10−10− 10−8 mol/L), it was possible to confirm that the interaction between basiliximab and dexamethasone was synergistic. At 10−8 mol/L, for example, a difference was seen between the median suppression with dexamethasone in the presence and absence of basiliximab even after the suppression with basiliximab alone had been subtracted from the former (32.6% vs. 20.5%, n = 45, P < 0.001).

Figure 6.

In vitro modulation of lymphocyte steroid sensitivity: dexamethasone alone, basiliximab alone and dexamethasone plus basiliximab. Dex, dexamethasone (10−6 mol/L); Bas, basiliximab (1 mg/L); Imax = 60%.

At trial entry, the mean inhibition of lymphocyte steroid sensitivity with dexamethasone in the study subjects with active disease was 58%, and 50% fulfilled the in vitro criteria for steroid resistance. Hence, the frequency of steroid-resistant individuals was approximately twice that seen in the healthy population. This is consistent with previous observations as, although the lymphocyte steroid sensitivity measured in vitro correlates with the failure to respond adequately to steroids in vivo, it is not an absolute predictor.9 After the addition of basiliximab in vitro, Imax increased in all subjects and the mean lymphocyte suppression increased to 79.1%. Two of the 10 subjects remained steroid resistant (Imax of 41.6% and 59%). Both of these patients achieved remission. The subject who was marginally steroid resistant (patient 6, Imax = 59%) had only a brief relapse at 6 weeks and remained in remission for the remainder of the study. However, the patient with the lowest Imax value (41.6%, patient 4) had an early relapse at week 4, achieved remission again at 12 weeks, but relapsed again before the end of the study. In vivo treatment with basiliximab did not significantly alter the in vitro lymphocyte steroid sensitivity levels (P = 0.19). This might be expected, as less than 10% of lymphocytes in the peripheral blood spontaneously express IL-2R in vivo (although 100% do so after in vitro activation by phytohaemagglutinin). Hence, the majority of blood T lymphocytes would not have bound basiliximab from the serum.

Discussion

Treatment with basiliximab plus glucocorticoids in vivo led nine of the 10 patients with steroid-resistant ulcerative colitis to achieve complete clinical remission within 8 weeks, and in vitro led 100% of healthy individuals and 80% of patients to achieve the criteria for steroid sensitivity. Seven of the 10 patients were in remission at 24 weeks after treatment with basiliximab, including two of the three patients who initially had severe ulcerative colitis. Five of the patients in remission at 24 weeks were not taking any steroids, with the other two patients taking 5 and 10 mg of prednisolone, respectively. The patient on 10 mg was able to withdraw steroids altogether 2 weeks after finishing the trial follow-up, and the other patient was maintained in remission on 5 mg daily. Four of the patients in remission were on a maintenance dose of azathioprine. Importantly, many of these patients had experienced active disease for up to 1 year prior to study entry, requiring prolonged therapy with high-dose steroids. Six of the 10 patients overall were being considered for a colectomy because of a poor response to medical therapy. In contrast with these results, data from previous studies have indicated that patients with active moderate ulcerative colitis, despite 14 days of prednisolone (30 mg/day), would be expected to have a response rate of, at most, 30% after 8 weeks with continued prednisolone therapy alone.28 In severe ulcerative colitis, patients with poor predictors of outcome similar to those described in this study, have been shown to have a chance of only 15% of avoiding a colectomy on admission.2

In addition to the primary outcome of clinical remission at 8 weeks, a single treatment with basiliximab resulted in significant clinical improvement in all patients within 1 week (UCSS), confirmed symptomatic improvement within 2 weeks (IBDQ score) and marked histological improvement within 4 weeks. The blind histological review clearly demonstrated a statistically significant reduction in acute inflammatory scores. However, there was also a notable diminution in the round cell infiltrate in the lamina propria. Indeed, in most cases, in the week 4 biopsy, this was less than would be expected for normal colorectal mucosa. Thus, basiliximab in combination with steroids not only reduced the active inflammation, but also resulted in a striking reduction in the chronic inflammation characteristically associated with chronic ulcerative colitis. Significant improvements in albumin levels were also seen. C-reactive protein levels showed a trend towards improvement, but this did not reach statistical significance.

Our in vitro data suggest that basiliximab can act as a steroid sensitizer. Basiliximab with dexamethasone demonstrated a greater anti-proliferative effect than either agent alone and, at sub-maximal concentrations of dexamethasone, it was possible to see that the addition of basiliximab had an effect that was greater than the sum of the steroid and basiliximab effects alone. Basiliximab has been given alone in some clinical settings with mixed results,29–33 and our data do not completely exclude an action of basiliximab alone. However, our data suggest that, as a sole agent, it is less effective than in combination with steroids. Basiliximab was generally very well tolerated, as seen in previous trials in transplantation. However, hypersensitivity has been reported with the use of basiliximab, notably after repeated administration.34 Hence, patients must be observed for 2 h after administration. Repeated use may result in sensitization, and cannot be recommended at present. Furthermore, repeated use of basiliximab may theoretically result in the formation of human anti-chimeric antibodies, with a resultant decrease in efficacy, as has been seen with the use of anti-tumour necrosis factor treatments.

A particularly attractive feature of anti-CD25 (anti-IL-2R) therapy is that its steroid-sensitizing action is likely to be restricted to the immune system, as only leucocytes express CD25 and are IL-2 dependent. No increase in steroid side-effects is expected in other tissues (e.g. impaired glucose tolerance, osteoporosis, hypertension, weight gain or skin change), and this was reflected in the findings of this study. In effect, basiliximab is expected to make the action of steroids more ‘lymphocyte specific’ and, as such, might benefit steroid-sensitive as well as steroid-resistant individuals by permitting a reduction in steroid dose.

Our data demonstrate a promising short-term (8-week) response to basiliximab, but relapses occurred later in all but one patient. This is not unexpected given the pharmacokinetics of basiliximab (T1/2 = 13.4 days; duration of IL-2R saturation in adults, 4–6 weeks).26 The relapse after approximately 8 weeks in some patients is consistent with the expected time course of IL-2R blockade by basiliximab after a single dose. However, in the patients who relapsed, remission was easily re-achieved with an increase in steroid dose to 30 mg of prednisolone and/or the addition of azathioprine in all but one patient. This response was striking in a group of patients who had previously failed prolonged high-dose steroid treatment. After 24 weeks of follow-up, seven of the 10 patients remained in remission, with the majority requiring no steroid treatment. Thus, we envisage that basiliximab may have a role in the rapid induction of remission and the avoidance of colectomy, but alternative therapies may be required to maintain remission, analogous to the emerging role of infliximab in Crohn's disease.35 An ‘8-week window’ of remission following treatment with basiliximab may be sufficient to allow 6-mercaptopurine or azathioprine to take effect. Sub-group analysis may reveal which patient characteristics are associated with the most favourable outcome from basiliximab treatment, but this study was too small to draw such conclusions.

During the preparation of the manuscript, the results of a pilot study of daclizumab (Zenepax, Roche Pharmaceuticals, Basel, Switzerland), an alternative IL-2R monoclonal antibody, for the treatment of medically resistant ulcerative colitis were published.36 In this study, no overall histological improvement was seen up to 8 weeks (although there was a reduction in CD25-positive cells in the mucosa), and clinical remission was achieved in only five of 10 patients. No data on the clinical progress beyond 8 weeks were given. Importantly, however, only four patients were treated with oral steroids during this study. Thus, daclizumab was not used as a steroid-sensitizing agent for steroid-resistant patients, but as an alternative immunosuppressant. We believe that this may explain the greater efficacy seen in our study (50% initial remission rate for daclizumab, compared with 90% remission with basiliximab plus steroids). If true, this supports the view that, although CD25 inhibition has immunosuppressant properties alone, it works best in this context as a steroid-sensitizing agent. However, differences between the two monoclonal antibodies, as seen with the use of different anti-tumour necrosis factor agents in Crohn's disease, may also be responsible for the different efficacies seen.

The in vitro data in this paper relate to the effects of basiliximab and steroids on peripheral blood mononuclear cells, rather than colonic mucosal lymphocytes. However, there are increasing data indicating that steroid resistance in lymphocytes is a generalized phenomenon, affecting all lymphocytes, and that the degree of ‘generalized’ steroid resistance correlates with the response to steroids in disease. Data in ulcerative colitis,9 systemic lupus erythematosus11 and asthma10 support this concept. This correlation between peripheral blood lymphocyte steroid sensitivity and clinical response to steroid therapy suggests that it may be meaningful to measure generalized steroid resistance in lymphocytes as an important predictor of outcome in response to steroid therapy. However, further studies should be performed to investigate steroid resistance in mucosal lymphocytes, which may indeed differ in some aspects.

The molecular mechanism of steroid resistance in inflammatory disease remains unclear. The glucocorticoid receptor (GR) exists in different isoforms. GR-α is the predominant form and is physiologically active. GR-β is a truncated splice variant and does not bind glucocorticoid. GR-β mRNA is detectable in many patients with steroid-resistant ulcerative colitis (83%), compared with patients with steroid-sensitive ulcerative colitis (9%), and the presence of GR-β has been suggested as a mechanism of steroid resistance.37 However, GR-β mRNA was only found in small amounts (< 0.2% compared with GR-α), suggesting that it may be a marker for steroid-resistant disease rather than the cause.38 Our data suggest a role for IL-2, and are consistent with previous studies.14 Binding of IL-2 to its receptor results in the activation of Janus kinases (JAK1 and JAK3), which then leads to the activation of signal transducer and activator of transcription 5 (STAT5).39 Other signalling pathways activated by the receptor include the mitogen-activated protein kinases, extracellular signal-regulated kinase (ERK), p38 and c-Jun N-terminal kinase (JNK)(Figure 7).

Figure 7.

Interleukin-2 (IL-2) and glucocorticoid signalling in activated T cells. Cross-talk between IL-2 and glucocorticoid signalling pathways. Signalling molecules extracellular signal-regulated kinase (ERK), p38, c-Jun N-terminal kinase (JNK), signal transducer and activator of transcription 5 (STAT5), nuclear factor kappa B (NFκB) and activator protein-1 (AP1) are activated by the IL-2 receptor, and all have been shown to interfere with glucocorticoid signalling (conversely, the glucocorticoid receptor has also been shown to interfere with STAT5, NFκB and AP1). Akt, AKT virus protein kinase; JAK, Janus kinase; PI3K, phosphoinositide 3-kinase.

Each of these signalling molecules has been shown to interfere with the activity of GR and, in several cases, the inhibition is mutual (Figure 7). Thus, STAT5 inhibits transactivation by GR,40, 41 as do ERK,42, 43 JNK43 and p38.15 It seems reasonable to speculate, therefore, that the resistance of lymphocytes to glucocorticoids can result from the activation of one or more signalling pathways by IL-2R, and that basiliximab reverses this process by inhibiting IL-2 signalling. Our data do not exclude the possibility that basiliximab acts independently of steroids to suppress ulcerative colitis. However, the presence of multiple proven interactions between the IL-2 and glucocorticoid signalling pathways, the limited effect of basiliximab alone on lymphocyte steroid sensitivity in vitro (Figure 6), the pharmacological synergy seen with glucocorticoids at intermediate doses in vitro and the limited effects of anti-CD25 blockade when not directly combined with steroids in vivo36 make this unlikely. Indeed, the impressive synergistic effect of basiliximab with steroids on lymphocyte steroid sensitivity in vitro, even in healthy volunteers, supports the idea that it may be effective in T-cell-dependent inflammatory conditions other than ulcerative colitis.

In conclusion, the data presented here provide preliminary evidence that basiliximab in combination with glucocorticoids is a very effective and well-tolerated treatment for the induction of remission in steroid-resistant ulcerative colitis. When used in combination with other slower acting treatments (such as azathioprine or 6-mercaptopurine), we envisage that anti-CD25 therapy may avoid the need for surgery in the majority of patients with steroid-resistant disease, and may also reduce the incidence of steroid-related side-effects. Further controlled trials are now required to confirm these promising findings.

Acknowledgements

We are grateful to two patients who supplied personal donations to provide initial funding for the in vitro studies, and to Marcus Arden-Jones and David Youds (Novartis Pharmaceuticals, Frimley, UK) for providing support to the study.

A project grant to fund a research salary (TJC) and basiliximab for the in vitro study and the clinical trial were supplied by Novartis Pharmaceuticals. Three authors (SDH, MNN, CMD) have lodged a patent application for the use of basiliximab as a steroid sensitizer.

Ancillary