Bone Marrow Transplantation (2003) 32, 165–170 & 2003 Nature Publishing Group All rights reserved 0268-3369/03 $25.00 www.nature.com/bmt Graft-Versus-Host Disease Thalidomide after allogeneic haematopoietic stem cell transplantation: activity in chronic but not in acute graft-versus-host disease

S Kulkarni, R Powles, B Sirohi, J Treleaven, R Saso, C Horton, A Atra, M Ortin, C Rudin, S Goyal, S Sankpal, S Meller, CR Pinkerton, J Mehta1 and S Singhal1

Leukaemia and Myeloma Units, Royal Marsden Hospital, Surrey, UK

Summary: of morbidity and mortality after allogeneic haematopoietic stem cell transplantation, and is directly or indirectly Thalidomide was used to treat acute (n ¼ 21) or chronic responsible for death in over half of the patients who (n ¼ 59) graft-vs-host disease (GVHD) in 80 haemato- experience treatment-related mortality. Patients with acute poietic stem cell allograft recipients after failure to leukaemia who are alive and well 2 years after an allograft respond to the combination of cyclosporine and corticos- have a significantly higher risk of continued treatment- teroids with or without other agents. The median time to related mortality, and poorer survival in the presence of onset of acute GVHD was 11 days, and thalidomide was chronic GVHD requiring immunosuppressive therapy.6 started at a median of 48 days post transplant. In addition This results from a number of factors including infections to corticosteroids and cyclosporine, 13 patients had also and secondary malignancies.7,8 received other agents before thalidomide. None of the The treatment of chronic GVHD is difficult. Cyclospor- patients responded and all died of acute GVHD. For ine, prednisone, and azathioprine have been the main chronic GVHD (limited in 13, extensive in 46), thalido- agents used with variable success rates. The combination of mide was started at a median of 385 days post transplant. cyclosporine and prednisone used for high-risk chronic In addition to corticosteroids and cyclosporine, 34 patients GVHD results in a response rate of 30–40%.9 The addition received azathioprine concomitantly. In all patients, of azathioprine in standard-risk patients has been asso- thalidomide was added to the ongoing immunosuppressive ciated with adverse outcome.10 While none of the newer regimen. The median duration of therapy with thalidomide drugs has activity that is consistently superior to cyclo- was 60 days (range, 11–1210; o2 weeks in 11). In total, sporine-prednisone, alternative drugs have been found to 13 patients (22%) had complete response, eight (14%) be useful in certain situations.11 partial response and 38 (64%) no response. Response Thalidomide was developed as a sedative–hypnotic drug, rates were comparable for limited (39%) and extensive but the development of severe foetal malformation led to (33%) chronic GVHD. At a median of 53 months, 19 withdrawal of its use from clinical practice.12 Its diverse patients are alive, 13 without evidence of chronic GVHD. effects on the immune system and the cytokine cascade Survival was significantly better in patients who responded resulted in continued exploration of the drug in various to thalidomide. The principal causes of death were disorders.12 Initial murine data showed that the drug was progressive chronic GVHD (n ¼ 29) and relapsed leukae- active in acute GVHD,13 and subsequent human data mia (n ¼ 7). In conclusion, thalidomide has no activity in showed it to be active in acute14 as well as chronic acute GVHD, but has some activity in chronic GVHD in GVHD.15–17 While questions have been raised about its combination with other agents. activity in chronic GVHD in a recent blinded, placebo- Bone Marrow Transplantation (2003) 32, 165–170. controlled study,18 most phase II reports show that the doi:10.1038/sj.bmt.1704033 drug does possess some activity in GVHD.19–23 Keywords: Thalidomide; transplantation; GVHD Our preliminary impression of thalidomide was that it controlled chronic GVHD in a proportion of patients with haematological malignancies without increasing relapse24— perhaps in some way secondary to its antimalignancy Acute graft-versus-host disease (GVHD) is seen in 20–80% activity.12,25 We have now undertaken a formal retro- of marrow- or blood-derived haematopoietic stem cell spective analysis of our data to see if this impression is allograft recipients, and chronic GVHD in 30–60% of accurate. those surviving beyond day 100.1–5 GVHD is a major cause

Patients and methods Correspondence: Professor R Powles, Royal Marsden Hospital, Downs Road, Sutton, Surrey SM2 5PT, UK. 26 1Current address: Division of Hematology/Oncology, Northwestern Comprehensive, prospectively gathered data on 80 University Medical School, Chicago, IL, USA allograft recipients treated with thalidomide for acute Received 8 May 2002; accepted 17 October 2002 (n ¼ 21) or chronic (n ¼ 59) GVHD at the Royal Marsden Thalidomide after allo-HSCT S Kulkarni et al 166 Table 1 Patient characteristics preceded by acute GVHD, progressive if it evolved from acute GVHD, and quiescent if there was a period without Parameter Acute Chronic GVHD GVHD GVHD between acute and chronic GVHD. Total 2159 Sex Treatment of GVHD Male 14 39 Female 7 20 Patients with acute GVHD were initially treated with high- Median age 24.5 y 31.5 y dose intravenous methylprednisolone (10–20 mg/kg/day) (Range) (4–51) (2–52) for at least 3 days. Further therapy was continued Diagnosis according to the clinical response. All patients also Acute leukaemia 15 34 Chronic myeloid leukaemia 5 13 continued cyclosporine during the therapy of acute GVHD. Other malignancies 112 Patients developing chronic GVHD after stopping Donor cyclosporine were recommenced on the drug. All patients HLA-matched sibling 20 47 were initially also treated with oral prednisone (1–5 mg/kg/ Matched unrelated 112 day). Other drugs such as azathioprine, tacrolimus (sub- Conditioning TBI-based 14 51 stituted for cyclosporine) and mycophenolate mofetil were Chemotherapy 7 8 added if there was no improvement. GVHD prophylaxis Cyclosporine 15 12 Cyclosporine–methotrexate 6 47 Thalidomide therapy Stem cell source Marrow 18 50 Thalidomide was started at a dose of 50–100 mg thrice a Blood 3 9 day. The dose was gradually increased to reduce the severity of the side effects with the aim of administering a total daily dose of 600–1200 mg. The dose was adjusted Hospital were reviewed. Patient characteristics are shown according to patient tolerance, and clinical signs and in Table 1. Conditioning regimens used were standard and symptoms of GVHD. Patients continued all the immuno- varied with the disease. All patients and donors provided suppressive agents they were already receiving. Four informed consent for the procedure, and all research patients with severe pulmonary involvement were treated protocols were approved by the institutional review board. with the combination of thalidomide, tacrolimus, predni- sone and azathioprine (four-drug therapy) as first-line therapy. GVHD prophylaxis Cyclosporine was started at a dose of 3 mg/kg/day Response evaluation intravenously on day À1and was switched to the oral Standard response criteria28,30 were used to evaluate route when mucositis resolved. Oral cyclosporine was response. Complete response (CR) was defined as complete started at a dose of 12.5 mg/kg/day in two divided doses. resolution of clinical signs, no new organ involvement, and The dose was gradually tapered after day 50 at the rate of normal performance status. Partial response (PR) was 5%/week with the aim of stopping the drug on day +180. defined as improvement in the performance status of at Methotrexate, when used, was given at a dose of 15 mg/m2 least two orders and at least 50% resolution of signs and on day +1and 10mg/m 2 on days +3, +6, and +11. symptoms of disease. No response (NR) was defined as lack Unless there was significant liver or kidney dysfunction, of improvement within 3 months or progression of disease methotrexate doses were not modified or omitted. on therapy.

Supportive care Statistical methods All patients were nursed in positive-pressure HEPA-filtered The probabilities of survival and mortality were determined single rooms. Patients received oral nonabsorbable anti- using the Kaplan–Meier life-table method. The percentage biotics. Prophylactic growth factors were not administered of patients responding to thalidomide was compared using routinely. Neutropenic fever was treated with broad- the Wilcoxon rank-sum test. spectrum antimicrobial agents according to the prevalent institutional guidelines. All blood products were irradiated to 2500 cGy. All patients received Pneumocystis prophy- Results laxis with oral trimethoprim-sulphamethoxazole or inhaled pentamidine. Patients with chronic GVHD received peni- Acute GVHD cillin V for pneumococcal prophylaxis.7 The median time to development of acute GVHD was 11 Grading and staging of GVHD days (range 9–40). Skin was involved in all, liver in 13 and gut in 19. In addition to high-dose methylprednisolone and Acute27 and chronic28,29 GVHD were graded and staged cyclosporine, patients also received a murine anti-IL2 according to standard previously published criteria. receptor antibody (n ¼ 5), azathioprine (n ¼ 2), Campath- Chronic GVHD was considered de novo if it was not 1G(n ¼ 2), methotrexate (n ¼ 2), total-lymphoid irradiation

Bone Marrow Transplantation Thalidomide after allo-HSCT S Kulkarni et al 167 Chronic GVHD N=59

CR PR NR N=13 N=8 N=38

12 alive 1 dead 2 alive 6 dead 5 alive 33 dead

25 chronic GVHD 3 on Second 2 on 4 chronic GVHD 2 on 5 relapsed immunosupression immunosupression 2 Relapsed leukaemia malignancy leukaemia immunosupression 3 other

1 on 1 on thalidomide 0 on thalidomide thalidomide

Figure 1 Thalidomide in chronic graft-versus-host disease – patient flow.

(n ¼ 1) and antithymocyte globulin (n ¼ 1). All patients had Table 2 Organ involvement in 59 patients with chronic GVHD progressive GVHD, with no response to the first- or Organ n % second-line therapy. The overall Glucksberg grade at the time of starting thalidomide was III in 11 and IV in 10. Skin 49 83 Thalidomide was started at a median of 48 days (range Liver 35 60 Oral mucosa 3152 13–84) after the transplant, and the median duration of Eyes 2136 treatment was 16.5 days (range 1–70). No patient GI tract 9 15 responded, and all died of progressive GVHD at a median Lungs 14 24 of 69 days (range 23–118 days) after the allograft (Figure 1). Muscles 12 Thrombocytopenia 22 38

Chronic GVHD

In all, 10 of the patients (Table 1) were treated for chronic Table 3 Response by extent of chronic GVHD (P=0.49) GVHD that had developed following cell-mediated im- Extent CR PR NR Alive Dead munotherapy of relapsed disease with donor leucocytes.31,32 The remainder were treated for chronic GVHD developing Limited (n=13) 3 (23%) 3 (23%) 7 (54%) 5 (38%) 8 (62%) after the transplant. Table 2 shows the organs involved, Extensive (n=46) 10 (22%) 5 (11%) 31 (67%) 14 (30%) 32 (70%) Table 3 the severity of the process and Table 4 the mode of Total (n=59) 13 (22%) 8 (14%) 38 (64%) 19 (32%) 40 (68%) onset. The initial therapy of chronic GVHD consisted of prednisolone and cyclosporine in all the patients. In total, 39 patients also received azathioprine at some stage. Of Table 4 Response by the mode of onset of chronic GVHD which, 14 patients responded to first-line therapy partly (P=0.8) before experiencing progressive GVHD. Onset CR PR NR Alive Dead Thalidomide was started at a median of 385 days (range, 92–3432) post-transplant. The median duration of therapy Progressive (n=20) 4 (20%) 3 (15%) 13 (65%) 6 (30%) 14 (70%) Quiescent (n=30) 8 (27%) 3 (10%) 19 (63%) 10 (33%) 20 (67%) was 61 days (range, 1–1210), and the median daily dose De novo (n=9) 1 (11%) 2 (22%) 6 (67%) 3 (33%) 6 (67%) 400 mg (range, 50–1200). The 11 patients receiving therapy for o2 weeks, most because of excessive sedation, were Total 13 (22%) 8 (14%) 38 (64%) 19 (32%) 40 (68%) considered non-responders. As expected, the duration of therapy for CR patients (median 196 days; range 30–721) was longer than that for PR (median 120 days; range no correlation between any type of organ involvement and 14–337) and NR (median 39 days; range 1–1210) (P ¼ 0.005). response to thalidomide. Five of the 14 patients with lung GVHD responded to therapy (three CR and two PR). Response in chronic GVHD Figure 1summarizes the outcome. A total of 19patients are alive at a median of 53 months (range, 5–159) post Figure 1shows response. CR was seen in 13(22%), eight transplant. The median Karnofsky performance score is had PR (14%), and 38 (64%) did not respond. The overall 80% (range, 70–100%). The actuarial 5-year probability response rate was 36%. As Tables 3 and 4 show, the extent of survival is 37% (95% CI: 23–51%) (Figure 2). In all, 12 of the disease or the mode of onset did not affect response patients are alive without chronic GVHD, and seven with. rates. There was no correlation between response to The causes of death were chronic GVHD (n ¼ 29), thalidomide and response to previous therapy. There was relapsed leukaemia (n ¼ 7), secondary malignancy (n ¼ 2),

Bone Marrow Transplantation Thalidomide after allo-HSCT S Kulkarni et al 168 1.0 peripheral neuropathy (n ¼ 2) and deep vein thrombosis (n ¼ 2) were uncommon. One patient with a normal blood count at the start of therapy developed thrombocytopenia 0.8 that recovered spontaneously on discontinuation of thali- domide. 0.6

Chronic GVHD 0.4 (19 of 59 alive) Probability Discussion

0.2 Acute GVHD (0 of 21 alive) Our data show that thalidomide is ineffective in steroid- refractory acute GVHD, but is effective in a proportion of 0.0 patients with chronic GVHD when used in combination 0 20 40 60 80 100 120 with cyclosporine and prednisone (with or without addi- Time since starting Thalidomide (mo) tional immunosuppressive agents). Response of chronic GVHD to thalidomide therapy is associated with better Figure 2 Overall survival of patients with acute and chronic GVHD from survival. the time of starting thalidomide (Po0.0001). The efficacy and potential of thalidomide as a single agent in the treatment of GVHD is undefined. Indeed, 1.0 because of the difficulty in deviating from established immunosuppressive therapy (making thalidomide mono- CR/PR therapy difficult), the length of time often required to treat 0.8 (14 of 20 alive) chronic GVHD, and the often-fluctuating clinical course of chronic GVHD make the determination of its activity 0.6 challenging. Indeed, the only controlled study evaluating its use in chronic GVHD found the drug to be inactive.18 Thalidomide possesses a number of anti-inflammatory 0.4 Probability and immunomodulatory properties, which may underlie its NR activity in GVHD.12 These have been discussed in detail (5 of 39 alive) 0.2 elsewhere,12,33 and include inhibition of tumour necrosis factor-a (TNF-a), IL-6, IL-10, and IL-12 production, 0.0 increase in IL-2, IL-4 and IL-5 production, modulation 0 20 40 60 80 100 120 of interferon-g production, increase in total lymphocyte counts as well as CD4+ and CD8+ T cells, and Time since starting Thalidomide (mo) costimulation of T lymphocytes. It also inhibits angiogen- Figure 3 Effect of response to therapy on overall survival of chronic esis and alters adhesion molecule expression.12 GVHD from the time of starting thalidomide (Po0.0001). Whether the lack of effect of thalidomide in acute GVHD is related to the speed of development of acute idiopathic pneumonitis (n ¼ 1), and myocardial infarction GVHD or the difference in the immunological mechanism (n ¼ 1). A significantly higher proportion of responding involved in its pathogenesis (compared to chronic GVHD) patients survived compared to non-responders (14 of 21 vs is not known. It may be the latter in view of some of the 5 of 38; Po0.0001; Figure 3). paradoxical effects the drug has (vide supra) because, when Of the 10 patients who received thalidomide for chronic used for GVHD prophylaxis, it has been shown to increase GVHD following cell-mediated immunotherapy, two had acute GVHD.34 It is also possible that optimum duration of CR, two PR and six NR. Two patients who achieved CR therapy necessary for any potential benefit will not be are alive one off immunosuppression and one with active achieved due to the rapidity of progression of severe acute GVHD. Both with PR died of relapsed leukaemia. One GVHD. with NR is alive with extensive chronic GVHD on There is some suggestion that thalidomide may have immunosuppression (including thalidomide), three died of particular activity in pulmonary chronic GVHD.35 This is GVHD, and two died of relapsed leukaemia. Leukaemia certainly plausible because angiogenesis underlies fibrosis, relapsed in two out of 21patients who responded to and thalidomide inhibits angiogenesis. We did not find any thalidomide as compared to three out of 38 who did not strong relation between response to thalidomide and the respond. Two relapses occurred among patients who were nature of organ involvement; five of 14 patients with lung treated successfully for chronic GVHD that had appeared involvement in our series responded to thalidomide with primarily because of the allograft (as opposed to immu- disappearance of symptoms and improvement in pulmon- notherapy-induced GVHD). ary function studies. The extent (severity) or mode of onset of chronic GVHD did not influence the response rate. et al17 Toxicity Vogelsang reported higher response rate with thalidomide for refractory disease as compared to primary Thalidomide was generally well tolerated. Sedation, dizzi- therapy of high-risk disease, whereas Parker et al20 found ness and constipation were common. Symptoms of similar response rate in high risk and standard risk GVHD.

Bone Marrow Transplantation Thalidomide after allo-HSCT S Kulkarni et al 169 We conclude that thalidomide holds some promise for 14 Lim SH, McWhannell A, Vora AJ et al. Successful treatment the treatment of refractory chronic GVHD, and if with thalidomide of acute graft-versus-host disease after bone- successful, may improve survival. Since drug intolerance marrow transplantation. Lancet 1988; 1: 117. is an important factor resulting in early discontinuation of 15 Vogelsang GB, Hess AD, Santos GW. Thalidomide for therapy, its analogues such as CC501336 may improve treatment of graft-versus-host disease. Bone Marrow Trans- plant response rates through better tolerance. 1988; 3: 393–398. 16 McCarthy DM, Kanfer EJ, Barrett AJ. Thalidomide for the therapy of graft-versus-host disease following allogeneic bone marrow transplantation. Biomed Pharmacother 1989; 43: Acknowledgements 693–697. 17 Vogelsang GB, Farmer ER, Hess AD et al. Thalidomide for This study was supported by the Bud Flanagan Leukaemia the treatment of chronic graft-versus-host disease. N Engl J Fund, the David Adams Leukaemia Fund, the Cancer Research Med 1992; 326: 1055–1058. Campaign and the Institute of Cancer Research. 18 Koc S, Leisenring W, Flowers ME et al. Thalidomide for treatment of patients with chronic graft-versus-host disease. Blood 2000; 96: 3995–3996. 19 Cole CH, Rogers PC, Pritchard S et al. Thalidomide in the References management of chronic graft-versus-host disease in children following bone marrow transplantation. Bone Marrow Trans- 1Powles RL, Clink HM, Spence D et al. Cyclosporine A to plant 1994; 14: 937–942. prevent graft-versus-host disease in man after allogeneic bone- 20 Parker PM, Chao N, Nademanee A et al. Thalidomide as marrow transplantation. Lancet 1980; 1: 327–329. salvage therapy for chronic graft-versus-host disease. Blood 2 Sullivan KM, Shulman HM, Storb R et al. Chronic graft- 1995; 86: 3604–3609. versus-host disease in 52 patients: adverse natural course and 21Rovelli A, Arrigo C, Nesi F et al. The role of thalidomide in successful treatment with combination immunosuppression. the treatment of refractory chronic graft-versus-host disease Blood 1981; 57: 267–276. following bone marrow transplantation in children. Bone 3 Powles RL, Morgenstern GR, Kay HE et al. Mismatched Marrow Transplant 1998; 21: 577–581. family donors for bone-marrow transplantation as treatment 22 Mehta P, Kedar A, Graham-Pole J et al. Thalidomide in for acute leukaemia. Lancet 1983; 1: 612–615. children undergoing bone marrow transplantation: series at a 4 Storb R, Deeg HJ, Whitehead J et al. Methotrexate single institution and review of the literature. Pediatrics 1999; and cyclosporine compared with cyclosporine alone for 103: e44. prophylaxis of acute graft vs host disease after marrow 23 Browne PV, Weisdorf DJ, DeFor T et al. Response to transplantation for leukaemia. N Engl J Med 1986; 314: thalidomide therapy in refractory chronic graft-versus-host 729–735. disease. Bone Marrow Transplant 2000; 26: 865–869. 5 Powles R, Mehta J, Kulkarni S et al. Allogeneic blood and 24 Kulkarni S, Powles R, Mehta J et al. Thalidomide in GVHD– bone-marrow stem-cell transplantation in haematological is anti-GVHD effect separable from the antiangiogenesis? malignant diseases: a randomised trial. Lancet 2000; 355: Blood 1998; 92 (Suppl. 1): 344b. 1231–1237. 25 Singhal S, Mehta J, Desikan R et al. Antitumor activity of 6 Singhal S, Powles R, Treleaven J et al. Long-term outcome of thalidomide in refractory multiple myeloma. N Engl J Med adult acute leukaemia patients who are alive and well two 1999; 341: 1565–1571. years after allogeneic bone marrow transplantation from 26 Powles R, Milan S, Horton C et al. The Royal Marsden an HLA-identical sibling. Leuk Lymphoma 1999; 34: 287–294. Hospital leukaemia–myeloma database: an ‘operations re- 7 Kulkarni S, Powles R, Treleaven J et al. Chronic GVHD search’ resource for assessing clinical outcomes and planning is associated with long term risk of pneumococcal infections new drug trials. Blood 2001; 98 (Suppl. 1): 426a. in bone marrow transplant patients. Blood 2000; 95: 27 Glucksberg H, Storb R, Fefer A et al. Clinical manifestations 3683–3686. of graft-versus-host disease in human recipients of marrow 8 Kulkarni S, Powles R, Treleaven J et al. Melphalan/TBI is not from HL-A-matched sibling donors. Transplantation 1974; 18: more carcinogenic than cyclophosphamide/TBI for transplant 295–304. conditioning: follow-up of 725 patients from a single centre 28 Sullivan KM, Shulman HM, Storb R et al. Chronic graft- over a period of 26 years. Bone Marrow Transplant 2000; 25: versus-host disease in 52 patients: adverse natural course and 365–370. successful treatment with combination immunosuppression. 9 Sullivan KM, Witherspoon RP, Storb R et al. Alternating-day Blood 1981; 57: 267–276. cyclosporine and prednisone for treatment of high-risk chronic 29 Akpek G, Zahurak M, Piantadosi S et al. Development of a graft-v-host disease. Blood 1988; 72: 555–561. prognostic model for grading chronic graft-versus-host disease. 10 Sullivan KM, Witherspoon RP, Storb R et al. Prednisone and Blood 2001; 97: 1219–1226. azathioprine compared with prednisone and placebo for 30 Martin PJ, Schoch G, Fisher L et al. A retrospective analysis treatment of chronic graft-v-host disease: prognostic influence of therapy for acute graft-versus-host disease: initial treatment. of prolonged thrombocytopenia after allogeneic marrow Blood 1990; 76: 1464–1472. transplantation. Blood 1988; 72: 546–554. 31Mehta J, Powles R, Treleaven J et al. Induction of graft- 11 Vogelsang GB. How I treat chronic graft-versus-host disease. versus-host disease as immunotherapy of leukaemia relapsing Blood 2001; 97: 1196–1201. after allogeneic transplantation: single-center experience of 32 12 Singhal S, Mehta J. Thalidomide in cancer: potential uses and adult patients. Bone Marrow Transplant 1997; 20: 129–135. limitations. BioDrugs 2001; 15: 163–172. 32 Singhal S, Powles R, Kulkarni S et al. Long-term follow-up of 13 Vogelsang GB, Hess AD, Gordon G, Santos GW. Treat- relapsed acute leukaemia treated with immunotherapy after ment and prevention of acute graft-versus-host disease allogeneic transplantation: the inseparability of graft-versus-host with thalidomide in a rat model. Transplantation 1986; 41: disease and graft-versus-leukaemia, and the problem of extra- 644–647. medullary relapse. Leuk Lymphoma 1999; 32: 505–512.

Bone Marrow Transplantation Thalidomide after allo-HSCT S Kulkarni et al 170 33 Corral LG, Kaplan G. Immunomodulation by thalidomide and 35 Forsyth CJ, Cremer PD, Torzillo P et al. Thalidomide thalidomide analogues. Ann Rheum Dis 1999; 58 (Suppl. 1): responsive chronic pulmonary GVHD. Bone Marrow Trans- I107–I113. plant 1996; 17: 291–293. 34 Chao NJ, Parker PM, Niland JC et al. Paradoxical effect of 36 Vacca A, Singhal S, Ribatti D, Dammacco F. Angiogenesis in thalidomide prophylaxis on chronic graft-versus-host disease. plasma cell disorders. In: Mehta J, Singhal S (eds). Myeloma. Biol Blood Marrow Transplant 1996; 2: 86–92. Martin Dunitz: London, 2002.

Bone Marrow Transplantation