Efficacy of the interleukin-2 receptor antagonist basiliximab in steroid-refractory acute graft-versus-host disease

Summary

Acute graft-versus-host disease (aGVHD) occurs in up to 80% of patients who undergo allogeneic stem cell transplantation (SCT) and contributes significantly to transplant-related mortality (TRM). We conducted a prospective phase II trial to assess the efficacy and feasibility of treating steroid-refractory aGVHD with basiliximab, a chimaeric monoclonal antibody directed against the alpha chain of the interleukin-2 (IL-2) receptor. Basiliximab was administered intravenously at a dose of 20 mg on days 1 and 4. Twenty-three patients were enrolled between October 1999 and July 2004. We found a primary overall response rate of 82·5% with four patients (17·5%) showing a complete response and 15 patients (65%) a partial response. Six patients were again treated successfully with an IL-2 receptor antagonist because of recurrence of aGVHD. The rates of infections, chronic GVHD, malignancy recurrence and 1-year TRM following immunosuppression with basiliximab were comparable with those found with other treatment modalities for aGVHD. We conclude that basiliximab is efficient and feasible for steroid-refractory aGVHD and merits further evaluation.

Acute graft-versus-host disease (aGVHD) occurs in up to 80% of patients who undergo allogeneic stem cell transplantation (SCT) and remains a major cause of treatment-related mortality (TRM) (Ferrara & Deeg, 1991; Przepiorka et al, 1996; Goker et al, 2001; Simpson, 2001). Steroid administration while continuing prophylaxis is still the first-choice treatment for aGVHD and relieves symptoms in approximately 40% of patients (Basara et al, 2000; Couriel et al, 2004). Patients with at least overall grade II aGVHD after steroid administration have a significantly worse overall survival than steroid responders (Martin et al, 1990; Weisdorf et al, 1990; Martin et al, 1991). Numerous agents, including antithymocyte globulin (ATG), monoclonal antibodies and pentostatin have been evaluated in steroid-refractory aGVHD with controversial results (Arai & Vogelsang, 2000; Deeg et al, 2001; Antin et al, 2004; Jacobsohn & Vogelsang, 2004; Bolanos-Meade & Vogelsang, 2005). However, a standard treatment for steroid-refractory aGVHD does not exist (Couriel et al, 2004).

Interleukin-2 (IL-2) plays a crucial role in the pathogenesis of aGVHD by the stimulation of activated cytotoxic donor T-lymphocytes directed against antigens of the recipient (Via & Finkelman, 1993; Goker et al, 2001). Binding of IL-2 to its receptor (CD25), a transmembrane receptor consisting of three subunits (alpha, beta, gamma), causes ras-mediated activation of antiapoptotic genes, followed by proliferation of donor T-lymphocytes (Mills et al, 1993; Gesbert et al, 1998). Thus monoclonal antibodies directed against the alpha chain of the IL-2 receptor became a focus of interest in the prevention and treatment of aGVHD (Willenbacher et al, 2001; Ji et al, 2002; Massenkeil et al, 2002; Chen et al, 2003). In contrast to pan T cell antigens, including CD2, CD3, CD5 and CD7, CD25 is mainly expressed on activated T cells and, to a lesser extent, on resting T cells (Mills et al, 1993). In relation to this, IL-2 receptor antibodies do not suppress resting T cells, which are important for engrafting and preventing infectious complications after allogeneic SCT (Couriel et al, 2004).

There are two chimaeric monoclonal IL-2 receptor antagonists with different pharmacokinetic features: basiliximab, with a terminal half-life of 7 d, and daclizumab, with a terminal half-life of 20 d (Kovarik et al, 1997; Vincenti et al, 1998). Both have been used in recent years to prevent and treat rejection after solid organ transplantation and have shown high efficacy and a favourable toxicity profile (Henry & Rajab, 2002; Van Gelder et al, 2004; Webster et al, 2004). To the best of our knowledge, there has only been one prospective trial investigating the chimaeric IL-2 receptor antagonist daclizumab in steroid-refractory aGVHD (Przepiorka et al, 2000). This trial, and some retrospective analyses investigating the IL-2 receptor antagonist basiliximab (Pasquini et al, 2000; Massenkeil et al, 2002; Couriel et al, 2004), demonstrated the efficacy and feasibility of both chimaeric IL-2 receptor antagonists for treating steroid-refractory aGVHD. However, a recently published phase II trial analysing the combination of daclizumab and steroids as initial therapy for aGVHD was terminated early because of reduced overall survival and increased recurrence of the underlying disease, compared with the administration of steroids alone (Lee et al, 2004).

We conducted a prospective, single centre phase II trial to assess the efficacy and feasibility of basiliximab for steroid-refractory aGVHD.

Patients and methods

Patients

Patients’ characteristics are summarised in Table I. All patients underwent haematopoietic cell transplantation at our centre and were of white-Caucasian origin. Once aGVHD developed, methylprednisolone was administered as first-choice treatment (2 mg/kg body-weight daily). Patients were eligible to enter the trial if aGVHD was progressive or unchanged (minimum of overall grade II) for at least 3 d after steroid administration. Further important eligibility criteria included: no severe infections, aged over 18 years, and informed written consent. Exclusion criteria were: chronic graft-versus-host disease (cGVHD) (i.e. GVHD for more than 100 d after SCT), and prior administration of basiliximab or any IL-2 receptor antagonist. The trial was approved by the local ethics committee and conducted in accordance with the declaration of Helsinki.

Table I.  Patients’ characteristics.

	Number of patients (%)
*One patient with severe aplastic anemia (SAA), one patient with refractory anaemia with excess of blasts (RAEB II), and one patient with hairy cell leukaemia. †Fludarabine/busulphan (one patient), TBI alone (one patient, retransplanted), TBI/etoposide (one patient), and cyclophosphamide/busulphan (one patient). CSA, ciclosporin; MTX, methotrexate; MMF, mycophenolate mofetil; ATG, antithymocyte globulin; TBI, total body irradiation.
Female/male	10 (43)/13 (57)
Median age (range), years	51 (31– 63)
Disease underlying SCT
Acute myeloid leukaemia	8 (35)
Acute lymphoblastic leukaemia	3 (13)
B cell-chronic lymphocytic leukaemia	3 (13)
Chronic myeloid leukaemia	2 (9)
Multiple myeloma	4 (17)
Other*	3 (13)
GVHD prophylaxis
CSA	1 (4)
CSA + MTX	6 (26)
CSA + MMF	3 (13)
CSA + MTX + ATG	10 (44)
CSA + MMF + ATG	3 (13)
Conditioning regimen
Treosulphan/fludarabine	9 (39)
TBI/cyclophosphamide	8 (35)
Cyclophosphamide/fludarabine	2 (9)
Others†	4 (17)
Donor characteristics
Matched unrelated donor	12 (52)
Mismatched unrelated donor	3 (13)
HLA-matched sibling	7 (31)
HLA-mismatched sibling	1 (4)
Stem cell source
Bone marrow	2 (9)
Peripheral blood stem cells	21 (91)

Basiliximab (Simulect^®; Novartis Pharma AG, Basel, Switzerland) was administered intravenously at a dose of 20 mg on days 1 and 4. The drug was dissolved in 5 ml of distilled water, followed by dilution in 250 ml of isotonic sodium chloride. The solution was administered over a period of 30 min without premedication. If a patient showed remission of aGVHD after receiving basiliximab, steroids were slowly tapered. In cases of recurrent aGVHD, basiliximab was administered again.

Assessment of efficacy

aGVHD was graded according to the revised Glucksberg scale (Przepiorka et al, 1995). Evaluation of efficacy included daily skin inspection, measurement of the daily amount of stool in cases of intestinal aGVHD, and documentation of nausea or vomiting. Alanine transaminase, aspartate transaminase, alkaline phosphatase and bilirubin were measured at least three times weekly to assess liver involvement of aGVHD. If the GVHD origin was in doubt for any target organ, biopsies were taken of the skin, intestinal tract and liver.

Assessment of response

A complete response was defined as the resolution of all clinical signs of aGVHD and normalisation of the bilirubin level following basiliximab administration. Relief of aGVHD by at least one grade in one target organ without progression in another was regarded as a partial response. Basiliximab was considered to be ineffective if a patient showed unchanged aGVHD in all target organ systems or progression of aGVHD in any target organ at a maximum of 7 d after administration of the first dose.

Assessment of toxicity, infections and cGVHD

Toxicity was evaluated in accordance with Common Toxicity Criteria (CTC) of the National Cancer Institute (version 3·0). Fungal infections were classified as ‘proven’, ‘probable’ or ‘possible’ according to the European Organisation for Research and Treatment of Cancer (EORTC) criteria (Subira et al, 2003). Cytomegalovirus (CMV) antigenin was calculated by determining serum pp65 antigen at least three times weekly. cGVHD was scored according to the revised Seattle criteria (Lee et al, 2003).

Study design

The primary endpoint of the study was to determine the number of patients who achieved complete or partial remission of aGVHD. Secondary endpoints included assessing the onset and duration of remission induction, the infusion-related toxicity, the frequency of cGVHD, the recurrence rate of the underlying malignancy, the infection rate and the 1-year TRM.

Statistical analysis

We used Simon's two-stage minimax design (Simon, 1989). In the present trial, a rate of ≥70% of patients with a partial or complete response (objective response) was thought to justify the evaluation of basiliximab in further trials. If ≤50% of patients showed an objective response, basiliximab would not be considered effective enough to be tested in further trials. Applying an α-error of 5% and a β-error of 20%, the calculated total number of patients required was 37. To avoid treating too many patients with an ineffective regimen the trial had to be terminated early if ≤11 of the first 23 patients showed a complete or partial response (first stage). If at least 12 patients showed an objective response, the trial was to be continued until 37 patients were evaluable (second stage). Basiliximab was to be investigated further if at least 23 of 37 patients showed a partial or complete response.

Results

Patients’ characteristics

Twenty-three patients entered the trial between October 1999 and July 2004. Patients’ characteristics are depicted in Table I. The conditioning regimen was reduced-intensity in 13 patients (57%) and conventional in 10 patients (43%). Eighty-three per cent of related, as well as unrelated donors, were 6/6 (A, B, DRB1) HLA-full matched. In-vitro purging was not carried out for any of the donor stem cells. The donor sex was female in 10 patients (44%), male in 12 patients (52%) and unknown in one patient (4%). One patient developed aGVHD after the second allogeneic SCT, performed because of graft failure (patient 12, Table II). Grade 2–4 aGVHD of the skin, liver and intestinal tract was observed in 20 (87%), 4 (17%) and 10 (43%) of the patients, respectively. Eleven patients (48%) exhibited overall grade II aGVHD, while overall grade III or IV aGVHD was observed in 12 patients (52%) (Table II). Grade III and IV aGVHD were less frequent in patients who received reduced-intensity conditioning (5/13 patients, 38%) than in those who underwent conventional conditioning (7/10 patients, 70%).

Table II.  Efficacy of basiliximab.

Patient number	aGVHD* prior to basiliximab administration				aGVHD* after basiliximab administration				Response
	Skin	Liver	Intestine	Overall aGVHD	Skin	Liver	Intestine	Overall aGVHD
*Graded according to the revised Glucksberg scale (Przepiorka et al, 1995). CR, complete response of aGVHD; PR, partial response of aGVHD.
1	3	2	1	III	4	3	1	IV	Refractory
2	3	0	0	II	0	0	0	0	CR
3	1	0	2	III	0	0	1	II	PR
4	3	0	0	II	0	0	0	0	CR
5	3	0	0	II	1	0	0	I	PR
6	3	1	0	II	1	0	0	I	PR
7	3	0	0	II	1	0	0	I	PR
8	0	0	3	III	0	0	0	0	CR
9	2	0	0	II	1	0	0	I	PR
10	3	0	0	II	1	0	0	I	PR
11	3	2	0	III	1	2	0	III	PR
12	2	0	3	III	1	0	1	II	PR
13	3	0	0	II	1	0	0	I	PR
14	2	0	4	III	1	0	1	II	PR
15	2	2	3	III	2	2	3	III	Refractory
16	3	2	3	III	4	3	3	IV	Refractory
17	1	1	4	III	1	1	1	II	PR
18	2	0	3	III	1	0	0	I	PR
19	2	0	0	II	1	0	0	I	PR
20	2	1	3	III	1	1	1	II	PR
21	3	0	0	II	0	0	0	0	CR
22	3	0	3	III	4	3	4	IV	Refractory
23	3	0	1	II	1	0	0	I	PR

Biopsies were taken from the skin, the upper and lower intestinal tract in two, five and six patients, respectively (9%, 22% and 26%). All biopsies confirmed clinically staged aGVHD. A liver biopsy was not taken in any of the patients. The first dose of basiliximab was administered at an average of 32 d (range 12–95 d) after allogeneic SCT .

Response evaluation

A response was evaluable in all patients enrolled in the trial. Basiliximab administration led to a complete response of aGVHD in four patients (17·5%) and a partial response in 15 patients (65%). Four patients (17·5%) showed refractory aGVHD (Tables II and III). Remission of aGVHD was induced 24–120 h after the first dose of basiliximab. Remission of aGVHD was sustained in 47% (9/19 patients), while 53% (10/19 patients) showed transient relief of aGVHD with a median duration of 4·4 weeks (range 2–8 weeks). Six of these 10 patients again received basiliximab or the chimaeric IL-2 receptor antagonist daclizumab, and they all achieved another partial or complete response (data not shown).

Table III.  Response to basiliximab and follow-up.

Patient number	Response	Duration of remission of aGVHD	cGVHD	Recurrence of underlying malignancy	Causes of death	Survival (days after transplantation)
*Bacterial sepsis; †fungal sepsis. LD, limited cGVHD; ED, extensive cGVHD; + patient still alive.
1	Refractory	–	–	No	Sepsis*	45
2	CR	Sustained	No	No	–	898+
3	PR	Sustained	–	No	Sepsis*	99
4	CR	7 weeks	LD	No	–	830+
5	PR	Sustained	LD	Yes	–	841+
6	PR	2 weeks	No	Yes	Leukaemia	123
7	PR	Sustained	LD	Yes	–	772+
8	CR	Sustained	LD	Yes	–	708+
9	PR	Sustained	No	No	–	604+
10	PR	Sustained	No	Yes	Leukaemia	126
11	PR	3 weeks	–	No	aGVHD	55
12	PR	4 weeks	LD	No	–	204+
13	PR	4 weeks	No	No	–	183+
14	PR	8 weeks	–	No	–	99+
15	Refractory	–	ED	No	cGVHD	144
16	Refractory	–	–	No	Sepsis*	27
17	PR	2 weeks	–	No	aGVHD	50
18	PR	2 weeks	ED	No	–	1437+
19	PR	Sustained	LD	Yes	Leukaemia	572
20	PR	4 weeks	ED	No	–	1633+
21	CR	Sustained	No	No	Sepsis*	109
22	Refractory	–	–	No	Sepsis†	67
23	PR	8 weeks	ED	No	cGVHD	308

To determine basiliximab responsiveness of each target organ system, we analysed the number of patients with a minimum one-grade aGVHD reduction for each target organ system. This was achieved by 17 of 22 patients (77%) with skin aGVHD, one of seven (14%) with liver involvement and eight of 12 (67%) with intestinal involvement.

Toxicity and infectious complications

Basiliximab did not cause any infusion-related toxicity. Its administration was followed by infections in 15 of 23 patients (65%). Bacterial infections occurred in 10 patients, fungal infections in two, viral infections in five and cerebral toxoplasmosis in one patient. Common bacterial infections included pneumonia, bronchitis and central-venous catheter-associated coccal sepsis. Four of 10 patients (40%) with bacterial infections died from sepsis, while six (60%) showed reduction of bacterial infection after standard antibiotic therapy was initiated (Table III). A fungal infection was rated as proven in one patient and as possible in another according to EORTC criteria. Viral infections included herpes simplex virus and herpes zoster virus.

CMV disease did not occur in any of the patients. However, 29% (5/17 patients) showed CMV antigenin between 10 d and 6 weeks after administration of the first dose of basiliximab. One patient required foscavir therapy because of ganciclovir resistance of CMV, but all others responded to standard intravenous ganciclovir treatment. Six cases of CMV antigenin were not evaluated because CMV prophylaxis had been carried out with ganciclovir (two patients) or because CMV antigenin coincided with the first dose of basiliximab (four patients).

One-year mortality

One-year mortality could be assessed in all but three patients who are still alive less than 1 year after allogeneic SCT. Eleven of these 20 patients (55%) died within the first year after allogeneic SCT, nine patients (45%) from TRM and two patients (10%) from non-TRM because of recurrence of the underlying malignancy.

Assessment of cGVHD and follow-up

cGVHD was evaluable in 16 of 23 patients but could not be assessed in seven patients because of survival for ≤100 d after SCT (six patients) or follow-up for less than 100 d after SCT (one patient). cGVHD occurred in 10 of 16 patients (63%); six of them presented with limited cGVHD and four with extensive cGVHD.

Twelve of 23 patients (52%) died at an average of 144 d (range 27–572 d) after allogeneic SCT, while 11 (48%) were still alive after a mean follow-up of 746 d (range 99–1633 d) (Table III). Patients with refractory aGVHD died at an average of 71 d after SCT, while those with a complete or partial response survived for an average of 508 d after allogeneic SCT, with 11 patients still alive. Causes of death included sepsis (five patients), acute or chronic GVHD (four patients) and recurrence of the underlying malignancy (three patients). The underlying malignancy recurred after basiliximab administration in 26% (6/23 patients).

Discussion

Management of steroid-refractory aGVHD in the allogeneic SCT setting still remains a major challenge that calls for novel treatment modalities. The phase II trial performed to investigate the chimaeric IL-2 receptor antagonist basiliximab in steroid-refractory aGVHD revealed an overall response rate of 82·5% with a complete response in 17·5% and a partial response in 65% of the patients. These results are in accordance with previous trials showing response rates of 56–83% when the chimaeric IL-2 receptor antagonist daclizumab or basiliximab was used for treating steroid-refractory aGVHD (Pasquini et al, 2000; Przepiorka et al, 2000; Willenbacher et al, 2001; Massenkeil et al, 2002). However, while reporting a lower overall response rate of 71%, Massenkeil et al (2002) recently found a complete response rate of 53% after basiliximab administration, which was higher than the 17·5% demonstrated in our trial. This discrepancy might be because of the physicians’ influence in aGVHD assessment, particularly in patients with multiple co-morbidity, which is not uncommon after allogeneic SCT. Moreover, we cannot exclude the underestimation of minor changes in aGVHD assessment when comparing retrospective trials to prospective trials like the one presented here.

The prompt and sustained remission of aGVHD to basiliximab administration favours the hypothesis that aGVHD relief was mainly due the antibody and not steroids or ciclosporin (CSA). However, we cannot exclude concomitant effects of steroids and CSA in some of our patients.

Despite the substantial rate of aGVHD remission induced by basiliximab, some of our patients had a relapse that required further aGVHD treatment after an initial complete or partial response. We found a median response duration of 4·4 weeks, which is comparable with the 5·5 weeks documented by Pasquini et al (2000) after basiliximab administration. Interestingly, the majority of our patients with aGVHD recurrence again achieved remission of aGVHD after administration of basiliximab or the monoclonal IL-2 receptor antagonist, daclizumab. This observation might be because of basiliximab causing transient inhibition of the IL-2 receptors, since it blocks them for only 30–45 d (Kovarik et al, 1997). However, some of our patients had a relapse of aGVHD 2–4 weeks after the first dose of basiliximab, by which time, a significant number of IL-2 receptors should have been inhibited by the antibody. Therefore, further mechanisms might account for aGVHD recurrence, including steroid tapering or IL-2 bypassing processes of aGVHD pathogenesis. Moreover, we cannot exclude the possibility that basiliximab may inhibit not only cytotoxic but also regulatory T-cells, a T-cell subset that also expresses CD25 on the cell surface (Dieckmann et al, 2002; Couriel et al, 2004). It has recently been shown that regulatory T-cells might be important for the prevention and resolution of aGVHD (Taylor et al, 2002; Edinger et al, 2003).

Basiliximab was well tolerated without any infusion-associated side effects. This confirmed the favourable toxicity profile of chimaeric IL-2 receptor antibodies demonstrated by many controlled trials in the solid organ transplantation setting (Henry & Rajab, 2002; Van Gelder et al, 2004; Webster et al, 2004). In this regard, basiliximab seems to be superior to other agents frequently used for treating aGVHD, including ATG and mycophenolate mofetil (Couriel et al, 2004).

Determination of infectious complications following basiliximab administration revealed rates of bacterial, fungal, viral and other infections comparable with those found in severe aGVHD treated by modalities other than IL-2 receptor antagonists (Marr & Bowden, 1999; Van Burik & Weisdorf, 1999; Boeckh et al, 2003; Ding et al, 2003; Fukuda et al, 2003). Similar infection rates have been described in other studies that investigated chimaeric IL-2 receptor antibodies for aGVHD treatment or prophylaxis (Przepiorka et al, 2000; Massenkeil et al, 2002; Chen et al, 2003). However, one trial demonstrated an increased rate of CMV reactivation after administration of the chimaeric IL-2 receptor antibody daclizumab for treating aGVHD (Willenbacher et al, 2001).

As suggested by others, IL-2 receptor inhibition by basiliximab or daclizumab might diminish the graft-versus-leukaemia effect, thus increasing the recurrence rate of the underlying malignancy (Blaise et al, 1995; Lee et al, 2004). We found a recurrence rate of 26% (for a mean follow-up of 432 d including all patients after SCT), which was comparable with that found in large series after peripheral blood SCT and aGVHD treatment modalities other than direct inhibition of the IL-2 receptor (Tanimoto et al, 2004). It is notable that a recently published phase II trial analysing the combination of daclizumab and steroids as the initial aGVHD therapy was terminated early because it led to a lower overall survival and a higher recurrence of the underlying malignancy than administration of steroids alone (Lee et al, 2004). These discrepancies might be explained by different strategies of IL-2 receptor antagonist initiation, e.g. upfront management in the trial performed by Lee et al (2004) as opposed to our use of basiliximab in steroid-refractory aGVHD. Furthermore, pharmacokinetic differences between the two chimaeric IL-2 receptor antibodies as the cause of the different results cannot be excluded, considering the significantly longer terminal half-life of daclizumab compared with basiliximab (20 and 7 d, respectively) (Kovarik et al, 1997; Vincenti et al, 1998).

Since aGVHD represents a major risk factor for developing cGVHD (Przepiorka et al, 2001; Lee et al, 2003), we analysed the occurrence of cGVHD. The rate of 63% of patients developing extensive or limited cGVHD is consistent with that found in other large trials (Weisdorf et al, 1990; Lee et al, 2003). Determination of the responsiveness of each target organ to basiliximab administration revealed the highest response rate for skin involvement, followed by intestinal tract and liver involvement. This sequence of responsiveness has been described elsewhere (Willenbacher et al, 2001).

This first prospective phase II study using basiliximab in steroid-refractory aGVHD underlines its therapeutic potential in a way that seems to justify a randomised trial with this antibody. We therefore stopped patient recruitment and initiated a prospectively randomised phase III trial.