Bone Marrow Transplantation (2013) 48, 203–209
`& 2013 Macmillan Publishers Limited All rights reserved 0268-3369/13
`
`www.nature.com/bmt
`
`•
`
`REVIEW
`
`Chronic GVHD: Where are we? Where do we want to be?
`Will immunomodulatory drugs help?
`YPL Linhares1, S Pavletic2 and RP Gale3
`
`Chronic GVHD (cGVHD) is an important problem after allotransplants. Some risk factors for cGVHD are similar to those of acute
`GVHD (aGVHD) but others are distinct indicating sometimes overlapping but unique pathogeneses. Precise incidence and
`prevalence data of cGVHD are lacking because of diverse diagnostic criteria but a 50% risk is a reasonable estimate. Incidence and
`prevalence of cGVHD are probably growing because of increased use of unrelated donors, blood rather than bone marrow (BM)
`grafts, decreased early transplant-related mortality (TRM) and increasing frequency of allotransplants. Pathophysiology of cGVHD is
`complex and poorly understood. Notably, no reliable surrogate end point to predict mechanism(s) of cGVHD has been identified.
`Therapy of cGVHD is unsatisfactory. Corticosteroids are effective but other drugs are controversial and few are rigorously tested in
`randomized trials. Highly variable response rates are reported because of small sample sizes and inconsistencies in eligibility,
`diagnostic and response criteria. We focus on the possible role of immunomodulatory drugs (IMiDs), thalidomide lenalidomide and
`pomalidomide, in preventing and treating cGVHD. The data suggest activity of thalidomide but at doses not clinically practical in
`many instances. There are few data with lenalidomide. Trials of pomalidomide, which has immune activities like thalidomide but
`with fewer adverse effects, are beginning. Because cGVHD is not recently reviewed in Bone Marrow Transplantation, we give a brief
`background and discuss challenges in diagnosing, understanding and treating cGVHD including the recently proposed National
`Institutes of Health consensus criteria for cGVHD.
`
`Bone Marrow Transplantation (2013) 48, 203–209; doi:10.1038/bmt.2012.76; published online 14 May 2012
`
`Keywords: Chronic GVHD; IMiDs; thalidomide; lenalidomide; pomalidomide
`
`INTRODUCTION
`Chronic GVHD (cGVHD) is an important complication of allogeneic
`hematopoietic cell transplantation.1–3 It occurs in about 50% of
`persons surviving 41 year post transplant and causes substantial
`morbidity and mortality. There has been little progress over the
`past 30 years in preventing and/or treating cGVHD. Moreover,
`incidence and prevalence are increasing because of several factors
`including: (1) increased use of blood cell rather than BM grafts; (2)
`increasing use of
`incompletely HLA-matched related and
`unrelated donors;
`(3)
`increased use of donor-lymphocyte
`infusions,
`especially
`in the
`context of
`reduced-intensity
`allotransplants; (4) increased number of transplants done each
`year; and (5)
`increased proportion of
`transplants in older
`persons.4-6 The focus of our review is on the use and potential
`of IMiD-class drugs to prevent and/or treat cGVHD. These drugs
`have unique, complex immune regulatory activities. As a prelude,
`we review some relevant definitional,
`laboratory and clinical
`features of cGVHD.
`
`CHRONIC GVHD
`Reported incidences of cGVHD vary dramatically: 6–80% but 50%
`is a reasonable estimate.7–9 Several important subject-, disease-
`and transplant-related variables correlate with the risk of cGVHD
`including: (1) recipient and/or donor genetic disparity (like related
`or unrelated, degree of HLA matching); (2) graft type (blood, BM or
`umbilical cord blood): (3) graft manipulation (like T-cell depletion);
`
`and (4) whether donor-lymphocyte infusions are given post
`transplant, which are unadjusted for in these reports with diverse
`incidences.10 Some data indicate that male recipients of grafts
`from female donors, especially those who are multiparous, have
`an increased risk of cGVHD. This increased risk from multiparity
`may also apply to female recipients.
`It is unknown whether
`progeny gender of multiparous donors correlates with risk of
`cGVHD.11 Detection of antibodies to Y-chromosome encoded
`antigens in male recipients of grafts from female donors correlates
`with an increased risk of cGVHD.12 Whether this is a cause: effect
`relationship is unknown.
`Some studies report prior acute GVHD (aGVHD) strongly
`correlated with cGVHD.13 A recent multivariate analysis by
`Flowers et al.14 in which National
`Institutes of Health (NIH)
`definitions for aGVHD and cGVHD were used showed grade-3/-4
`aGVHD was a risk factor for cGVHD. However, there were also
`some unique risk factors correlated solely with cGVHD including:
`(1) blood cell grafts and (2) older donor age. Also, use of female
`donors for male recipients had a greater effect on risk of cGVHD
`than on aGVHD. There was little change in hazard ratios with other
`variables after adjusting the hazard ratio of cGVHD for aGVHD.
`This suggests that the association of certain risk factors with
`cGVHD is independent of aGVHD. These data support the notion
`cGVHD is not merely a time-dependent expression of aGVHD.
`aGVHD and cGVHD were previously considered to have distinct
`clinical and temporal features. By convention, the temporal cutoff
`for classifying a clinical syndrome as aGVHD or cGVHD was 100
`
`1Department of Medicine, Division of Hematology/Medical Oncology, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA; 2GVHD and Autoimmunity Unit,
`Experimental Transplantation and Immunology Branch, National Cancer Institute, Bethesda, MD, USA and 3Department of Medicine, Section of Haematology, Division of
`Experimental Medicine, Imperial College, London, UK. Correspondence: Dr RP Gale, Department of Medicine, Section of Haematology, Division of Experimental Medicine, Imperial
`College, 11693 San Vicente Boulevard, Suite 335, Los Angeles, CA 90049-5105, USA.
`E-mail: robertpetergale@aol.com
`Received 29 March 2012; accepted 29 March 2012; published online 14 May 2012
`
`

`

`•
`
`204
`
`Immune modulators in chronic GVHD
`YPL Linhares et al
`
`this binary classification is
`transplant. However,
`days post
`obscured when aGVHD occurs after day 100, when there
`is delayed development of hematopoietic chimerism, when
`there are flares after tapering post transplant immune suppression
`or following post transplant infusions of donor lymphocytes. In
`2005, the NIH consensus project proposed two main GVHD
`categories each with two subcategories.10,15 For aGVHD there is
`(1) classic aGVHD occurring within 100 days post transplant and
`(2) persistent, recurrent or late onset aGVHD, a clinical syndrome
`resembling aGVHD but persisting beyond or developing after 100
`days post transplant commonly referred to just as ‘late’ aGVHD.
`For cGVHD there is (1) classic cGVHD subtype (without features
`characteristic of aGVHD) and (2) overlap syndrome with
`synchronous clinical
`features of cGVHD and aGVHD. The old
`‘limited’ and ‘extensive’ cGVHD staging nomenclature is replaced
`with the more informative individual organ severity scoring
`(grades 0–3) and global cGVHD (mild–moderate–severe) stage
`which in addition to cGVHD manifestations also account for
`subjects symptoms and quality of life.10,16 Feasibility of use and
`clinical utility of the NIH consensus cGVHD classification and
`severity scoring is validated in several prospective studies.8,16,17
`Several
`retrospective
`and prospective
`analyses
`identified
`additional prognostic variables which complement
`the NIH
`consensus staging system including low platelet levels, poor per-
`formance score when cGVHD is diagnosed, and gastrointestinal
`tract
`involvement which significantly influence outcome and
`progressive onset
`type of
`cGVHD as adversely affecting
`survival.18,19 These correlations need confirmation.
`cGVHD is a multiorgan alloimmune and autoimmune disorder
`characterized by immune dysregulation,
`immune deficiency,
`impaired end-organ function and decreased survival.20 Although
`these features of cGVHD were noted 430 years ago there has
`been little progress sorting out the precise mechanism(s). Therapy
`of cGVHD, whatever the cause(s), is typically prolonged immune
`suppression which may
`further
`aggravate
`the
`immune
`suppression intrinsic to cGVHD.21
`
`PATHOPHYSIOLOGY
`The pathophysiology of cGVHD is poorly understood (to say the
`least) despite several decades of research. T cells (Th1 and Th2) and
`B cells are important suggesting a global loss of immune tolerance
`including abnormalities in the function of regulatory T cells (Tregs;
`CD4 þ , CD25 þ , FoxP3 þ ) which maintain self-tolerance.22
`Studies in mice indicate that normal Tregs can suppress aGVHD
`and cGVHD and that deficient and/or defective Tregs worsens
`aGVHD and cGVHD.23 There are conflicting data in humans
`concerning the role of Tregs in the development of cGVHD.24 A
`recent study suggests that therapy with low-dose IL-2 can increase
`Tregs and improve severe cGVHD.25
`Several
`studies
`suggest
`that aGVHD is associated with
`predominant Th1-type immune response and cGVHD with a
`predominant Th2-type immune response.26 Th1 cells produce
`IFN-g that mediates cell-mediated immunity whereas Th2 cells
`produce IL-4, IL-5 and IL-13 that mediate humoral immunity. T-cell
`dysregulation results in cytokine abnormalities which may be
`important in cGVHD. High levels of TNF-a, IL-6, TGF-b and IL-1b
`and low levels of INF-g and IL-10 are reported in persons with
`cGVHD.27–31 A recent study by Imanguli et al.31 challenges this
`paradigm of cGVHD as type-II cytokine-mediated disorder. Data
`from this study suggest that activation of the type-I IFN axis is
`important in oral cGVHD.
`Autoimmunity and autoreactive T cells are important in cGVHD.
`Some studies report a functional host thymus is not required to
`induce cGVHD and that quiescent autoreactive T and B cells in
`transplants from non-autoimmune donors can be activated and
`expanded to induce cGVHD.32 In contrast, involvement of thymus-
`dependent T-cell pathways in cGVHD development begins with
`
`injury to the thymus from chemotherapy and/or radiation and/or
`from aGVHD leading to loss of self-tolerance.26,33,34 Some data in
`mice suggest that during aGVHD graft-donor CD4 þ T cells that
`can recognize and react against host tissues develop in the
`thymus in and mediate cGVHD.34
`Some data suggest an important role for B cells in cGVHD.
`Because B cells produce antibodies and can, in some instances,
`expose or present antigen to T cells, it may enhance development
`of cGVHD. Anti-nuclear, -mitochondrial, -parietal, -smooth muscle
`and -parotid autoantibodies are present in some persons with
`cGVHD.35,36 Also, persons with autoantibodies have more cGVHD-
`associated symptoms than persons without autoantibodies.35
`Autoantibodies against platelet-derived growth factor receptor
`may have a role in cGVHD.37 These platelet-derived growth factor
`receptor-a autoantibodies stimulate thyrosine phosphorylation,
`resulting in a cascade of events that may contribute to inflamma-
`tion and fibrosis.
`Some studies report elevated levels of BAFF (B-cell activating
`factor of the TNF family) in persons with cGVHD. BAFF is produced
`by T cells and granulocytes and supports differentiation and
`survival of normal B cells in persons with cGVHD and autoimmune
`diseases.23,38 Fujii and coworkers reported persons with early
`onset cGVHD have elevated levels of sBAFF, sIL-2Ra, sCD13 and
`anti-dsDNA autoantibodies. sBAFF, anti-dsDNA and antinuclear
`antibody are elevated in late onset cGVHD, suggesting B-cell
`activation.39–41 These observations may explain occasional reports
`of benefit of therapy of cGVHD with anti-B-cell antibodies like anti-
`CD20 (rituximab for instance).40
`These T- and B-cell pathways and others are potential targets
`for treating cGVHD. However, it is important to note no immune
`parameter(s)
`is a reliable biomarker of diagnosis,
`severity,
`therapy outcome of cGVHD.42 Consequently,
`prognosis or
`clinical
`trials,
`ideally
`randomized, blinded and placebo
`controlled, are the sole way to know whether a therapy
`intervention in cGVHD is safe and effective.
`
`CLINICAL FEATURES
`lungs,
`liver,
`cGVHD usually targets the skin, eyes, mouth, gut,
`joints and genitourinary system. Typical skin manifestations are
`sclerosis and poikiloderma and lichen-type lesions. There are
`often hyperkeratotic oral plaques. A lung biopsy may show
`bronchiolitis obliterans. These clinical
`features resemble auto-
`immune diseases like progressive systemic sclerosis, systemic
`lupus erythematosis and Sjo¨ gren’s syndrome.10
`cGVHD is categorized in the NIH global scoring system as mild,
`moderate or severe depending on the number of organs involved
`and the severity of the abnormality(ies).10,15 Systemic immune
`suppression is usually advised for persons with moderate or
`severe cGVHD. Systemic therapy is also considered in persons with
`thrombocytopenia (platelets o100  10e9/L) or progression while
`prednisone.43
`receiving
`cGVHD eventuates
`in
`impaired
`performance score, poor quality-of-life and death. 8-10,21,44-46
`
`TREATMENT
`interventions that prevent aGVHD are not
`Pharmacological
`effective in preventing cGVHD. Strategies using anti-thymocyte
`globulins for in-vivo T-cell depletion show promise but no benefit
`on survival.47 The standard initial therapy of cGVHD is prednisone
`with or without a calcineurin inhibitor. However, only about 50%
`of persons have a durable response.22,43 There is no standard next
`therapy. Recommended interventions include about 40 drugs, all
`studied in poorly standardized, phase-2 trials or reported in
`retrospective case analyses, including sirolimus, tacrolimus, myco-
`phenolate, MTX, MoAbs, pentostatin,
`imatinib, extracorporeal
`low-dose IL-2 and many others.25,48-49 Choice
`photopheresis,
`
`Bone Marrow Transplantation (2013) 203 – 209
`
`& 2013 Macmillan Publishers Limited
`
`

`

`Immune modulators in chronic GVHD
`YPL Linhares et al
`
`Thalidomide
`
`Lenalidomide
`
`Pomalidomide
`
`•
`
`205
`
`~
`
`0
`
`NH2
`
`0
`
`o QOYso
`8:$
`-µ=o
`
`NH
`
`O 0
`
`++++
`
`++++
`
`+++
`
`+++
`
`+++
`
`+++++
`
`+++++
`
`++++
`
`++++
`
`++++
`
`+ +
`
`Anti-inflammation
`
`T-cell stimulation
`
`Anti-angiogenesis
`
`++++
`
`+ +
`
`Anti-proliferation
`
`Pro-erythroid
`
`Figure 1. Biological activities of IMiDs.
`
`Survival was 64% with 55 months follow-up and 76% in persons
`failing prior therapy and 48% in those with previously untreated
`high-risk cGVHD. Confidence intervals were not reported and
`there was no comparator cohort. Main adverse effects were
`sedation, neuropathy and constipation.58 Heney et al.61 reported
`responses in five of six persons receiving thalidomide, 100–200 mg/day.
`Response was best in skin cGVHD. Two subjects developed
`neuropathy. Cole et al.62 reported five children with advanced
`cGVHD treated with thalidomide, 12–25 mg/kg/day. There was
`one complete response and four partial responses. Adverse effects
`were minimal and there was no neuropathy. Parker et al.63 treated
`80
`subjects with
`advanced
`cGVHD with
`thalidomide,
`400–1200 mg/day. There were nine complete and seven partial
`responses. Twenty-nine subjects discontinued because of adverse
`effects, including sedation, constipation, neuritis, neutropenia and
`rash. Rovelli et al.64 used thalidomide, 3–12 mg/kg/day in 14
`children with cGVHD. Six complete responses and four partial
`responses were reported. Browne et al.65 treated 37 subjects with
`advanced cGVHD with thalidomide, 200–800 mg/day.
`In all, 1
`subject had a complete and 13, partial responses. Responses were
`more common in children than in adults. Kulkarni et al.66 reported
`data on 59 subjects with advanced cGVHD using thalidomide,
`600–1200 mg/day. Thirteen subjects had a complete and eight, a
`partial response. Two subjects developed poly-neuropathy, two,
`deep vein thromboses and one, thrombocytopenia. There are a
`case series and two case reports not included in Table 1. Mehta
`treated two children with cGVHD. One had a complete and the
`other a PR. Adverse effects were sedation and constipation.67
`Forsyth reported a response to thalidomide, 400 mg/day, in a case
`of bronchiolitis obliterans from cGVHD.68 Staumont-Salle reported
`a response to thalidomide, 100 mg/day, in a subject with lichenoid
`vulvar lesions from cGVHD.69
`Table 2 summarizes data from three randomized trials. Two
`it.70–72
`were attempts to treat cGVHD and one,
`to prevent
`Koc et al.70 reported a placebo-controlled trial of thalidomide,
`200–800 mg/day,
`in adults and 3–12 mg/kg/day in children.
`Thalidomide was discontinued in 23 of 25 subjects in the
`thalidomide cohort and 17 of 26 subjects in the placebo cohort
`because of
`intolerance, mostly neutropenia,
`sedation and
`neuropathy. This high discontinuation rate in the placebo cohort
`underscores the variable natural course of cGVHD and subsequent
`difficulty of performing clinical trials in persons with advanced
`cGVHD.
`It also underscores the need for placebo-controlled
`double-bind studies.
`It is also possible that thalidomide was
`ineffective in this population. Arora et al.71 reported a randomized,
`placebo-controlled trial of
`initial
`treatment of cGVHD with
`thalidomide. There was no benefit of adding thalidomide,
`200–800 mg/day, vs placebo to cyclosporine and prednisone.
`Chao et al.72 reported thalidomide increased the incidence and
`severity of cGVHD when given in a prevention study.
`
`between drugs is based on logistics; cost, failed prior treatments,
`toxicity profile and subject or clinician preference.
`Therapy of cGVHD is a difficult problem to address because of
`the logistical challenges in conducting such trials and the lack of
`standardized criteria for study design. To improve conduct and
`interpretation of clinical
`trials,
`the NIH-Sponsored Consensus
`Development Project published guidelines addressing diagnosis
`and staging, histopathology, biomarkers, assessment of therapy
`response, ancillary therapy and supportive care and the design of
`cGVHD clinical trials.10,50–54
`cGVHD typically is diagnosed within 6 months post transplant
`and lasts 2–5 years. About 85% of survivors can discontinue
`systemic immune suppression by that time. Five-year survival of
`persons with cGVHD is 30–40% for those with high-risk disease
`and for persons failing corticosteroids. There is better 5-year
`survival, about 70–80%,
`in persons with lower risk cGVHD and
`those responding to corticosteroids.9,55,56
`Treatment goals for cGVHD include reversing symptoms and
`signs, preventing recurrence, disability or death. A goal could be
`also correcting associated immune abnormalities. This is, of
`course, challenging, as therapy of cGVHD typically involves
`immune-suppressive drugs that have multiple acute and cumu-
`lative toxicities. The therapy of cGVHD is largely unsatisfactory and
`most persons, especially those who fail corticosteroids, should be
`treated on investigational protocols whenever possible.45
`
`IMID-CLASS DRUGS
`Thalidomide
`Thalidomide is effective in modulating cGVHD in rodents and is
`studied as therapy and prevention of cGVHD in humans.57
`Vogelsang et al.58 reported that
`thalidomide is a safe and
`effective treatment
`for severe cGVHD. Several other studies
`reproduced these findings. Thalidomide is
`the third most
`commonly used drug in phase-2 trials of therapy of cGVHD in
`persons failing corticosteroids.59
`Thalidomide has diverse immune-modulating effects including:
`(1) reduced levels of TNF-a; (2) co-stimulation of T cells to produce
`IL-2 and IFN-g; (3) inhibition of other cytokines like IL-1b,
`IL-6,
`IL-12;
`(4) downregulation of cell surface adhesion molecules
`involved in leukocyte migration; and (5) anti-angiogenesis.48,57,60
`Its biological activities are contrasted with other IMiDs in Figure 1.
`There are seven phase-2 and three phase-3 trials of thalidomide in
`cGVHD. Analysis of the trials is complex for several reasons:
`(1) different definitions of cGVHD; (2) different response criteria;
`(3) inclusion of children in some studies; (4) different doses of
`thalidomide;
`(5) different prior
`therapy (ies);
`(6) different
`objectives
`(for example,
`therapy vs prevention and initial
`therapy vs second-line therapy after corticosteroid failure); and
`(7) thalidomide alone vs combinations and other variables. This
`heterogeneity, not uncommon in cGVHD trials, makes it difficult/
`impossible to draw precise conclusions regarding safety and
`efficacy of thalidomide in cGVHD.
`Table 1 summarizes data from 7 trials in 245 subjects with
`cGVHD receiving thalidomide after failing initial therapy at dosesof
`100–1600 mg/day. These data include children and adults in
`diverse therapy settings and using diverse response criteria. In all,
`46 subjects are reported to have had a complete (19%) and 51 a
`partial response (21%) for an overall response in 97 subjects (40%).
`This result is encouraging but there is the undoubtedly important
`issue like possible of selective reporting of favorable outcomes.
`These trials are reviewed in below.
`Vogelsang et al.58 used thalidomide, 800–1600 mg/day, as initial
`therapy of 21 subjects with high-risk cGVHD and as salvage
`therapy for 23 subjects with cGVHD failing initial therapy. Initial
`dose in children was 3 mg/kg given four times daily. Complete
`response was reported in 14 subjects and partial response in 12.
`
`& 2013 Macmillan Publishers Limited
`
`Bone Marrow Transplantation (2013) 203 – 209
`
`

`

`•
`
`206
`
`Immune modulators in chronic GVHD
`YPL Linhares et al
`
`Table 1. Phase-2 trials of thalidomide in advanced cGVHD
`
`Vogelsang et al.59
`Heney et al.62
`Cole et al.63
`Parker et al.64
`Rovelli et al.65
`Browne et al.66
`Kulkarni et al.67
`
`Abbreviation: cGVHD ¼ chronic GVHD.
`
`Dose
`
`800–1600 mg/day (X3 mg/kg/day children)
`100–200 mg/day
`12–25 mg/kg/day
`400–1200 mg/day
`3–12 mg/kg/day
`200–800 mg/day
`600–1200 mg/day
`
`Table 2. Phase-3 trials of thalidomide in initial therapy or prevention of cGVHD
`
`N
`
`44
`6
`5
`80
`14
`37
`59
`
`Response
`
`CR-14; PR-12
`CR-2; PR-3
`1 CR; PR-4
`CR-9; PR-7
`CR-6; PR-4
`CR-1; PR-13
`CR-13; PR-8
`
`First therapy
`Koc et al.71
`Arora et al.72
`
`Prevention
`Chao et al.73
`
`Abbreviation: cGVHD ¼ chronic GVHD.
`
`Dose
`
`200–800 mg/day
`200–800 mg/day
`
`400 mg/day
`
`N
`
`51
`54
`
`59
`
`Response
`
`Early discontinuation due to toxicity
`No difference
`
`Increased cGVHD
`
`In summary, several phase-2 trials report about 50% responses
`to thalidomide doses of 100–1600 mg/day. Much of the variability
`between trials reflects small sample sizes, heterogeneous subjects
`and diverse, poorly defined response criteria among other
`complexities discussed above. Thalidomide doses of 4200 mg/
`day were poorly tolerated. Phase-3 trials show no convincing
`benefit of thalidomide for prevention or initial therapy of cGVHD.
`Although thalidomide may be active against cGVHD at high doses
`in rodents,
`these doses cannot be reproducibly and safely
`achieved in humans in most
`instances. An alternative, not
`mutually exclusive problem is the variable course of cGVHD with
`exacerbations and improvements unrelated to therapy interven-
`tions, which may mimic drug adverse effects and/or therapy
`response.
`
`Lenalidomide
`Given the many reports and widespread use of thalidomide in
`corticosteroid-resistant cGVHD, there are remarkably few data on
`use of lenalidomide in this setting. A recent Boolean PUBMED
`search of the English-language literature 1966-present using the
`search terms lenalidomide AND chronic graft-versus-host disease
`identified fewer than 10 reports most of which were anecdotes.
`One phase-2 study of lenalidomide maintenance for myeloma
`after allogeneic transplantation was discontinued because of a
`claimed increased risk of aGVHD.73 However,
`there was no
`concurrent
`control
`arm. Another
`report
`suggested that
`lenalidomide induced a syndrome resembling aGVHD in
`autotransplant recipients.74 It is difficult to interpret these few
`data. One possibility is that concerns about BM toxicity of
`lenalidomide preclude widespread use. Another is publication
`bias: trials or treatments may have been done but were not
`reported because of unfavorable results. The bottom line is
`efficacy of lenalidomide in corticosteroid-resistant cGVHD is not
`known because it appears not extensively studied.
`
`Pomalidomide
`Pomalidomide is a novel immune-modulating drug with 4000-fold
`greater inhibition of TNF-a production compared with thalido-
`mide.75 A comparison of biological activities of pomalidomide
`with thalidomide and lenalidomide is included in Figure 1.
`
`Pomalidomide is extensively used in humans in the setting of
`clinical trials primarily for the treatment of multiple myeloma and
`myeloproliferative neoplasm-associated myelofibrosis.76,77 It offers
`high potency without the dose-limiting toxicities of neuropathy
`and sedation. In persons with multiple myeloma, the dose-limiting
`toxicity is BM suppression with a maximum-tolerated dose of
`2 mg/day.78 Several features of pomalidomide suggest it may be
`useful in treating cGVHD including: (1) in-vitro suppression of TNF-
`a (human monocytes);79 (2) increasing Th1 (mouse cancer vaccine,
`human CD4 þ T cells in vitro);80,81 (3) suppression of Th2 (mouse
`cancer vaccine);80 and (4) stimulation of
`IL-12 and sIL-2Ra
`(humans).78,80 However, other effects of pomalidomide have
`potential
`adverse
`effects
`in
`treating
`cGVHD including:
`(1)
`increased CD45RO þ (memory) CD4 and CD8 T cells
`(humans);78 (2) decreased Tregs;82 (3)
`increased Th2 (polarized
`human CD4 þ T cells in vitro);81 and (4) increased B cells (in-vitro
`human CD19 þ cells).83 Recently, cereblon was identified as an
`essential mediator of Ienalidomide and pomalidomide anticancer
`activity in multiple myeloma. These drugs react with cereblon to
`mediate IFN-regulator factor downregulation. This may affect
`development of Th-17 cells
`sometimes
`implicated in the
`development of cGVHD.84,85 Whether this is a possible target of
`activity of pomalidomide in cGVHD is unknown. Pomalidomide is
`effective in treating experimental scleroderma in mice and
`in a model of bleomycin-induced skin fibrosis (Celgene Corp;
`unpublished data). Because, as discussed, the precise patho-
`genesis of cGVHD is unknown (and may differ
`in different
`persons), it is impossible to predict the impact of therapy with
`pomalidomide outside the context of a controlled clinical trial.
`There is one report of a small phase-2 study of pomalidomide in
`cGVHD. Pusic et al.86 treated eight subjects failing corticosteroids.
`Subjects received 3 mg/day with dose reductions to 2 mg, 1 mg
`and 0.5 mg/day. Seven subjects had dose reductions because of
`muscle cramps, tremor and fatigue. Five subjects discontinued
`therapy for worsening cGVHD of the skin and mouth (N ¼ 1), pain
`(N ¼ 1) and no response (N ¼ 3). There was no BM suppression,
`somnolence, constipation or
`thromboembolic events. Three
`persons reached the primary evaluation end point at 6 months
`at the 2 mg (N ¼ 2) or 1 mg dose (N ¼ 1). These three had global
`PRs per NIH criteria (erythema and gastrointestinal) and oPR
`ongoing improvements (skin, mouth and eyes). This study shows
`
`Bone Marrow Transplantation (2013) 203 – 209
`
`& 2013 Macmillan Publishers Limited
`
`

`

`•
`
`207
`
`feasibility of giving pomalidomide to humans with cGVHD and
`absence of serious side effects at doses of p2 mg/day. A
`randomized phase-2 trial of pomalidomide in persons with
`corticosteroid-resistant cGVHD is planned. There are also large,
`ongoing studies of pomalidomide in multiple myeloma and
`myeloproliferative neoplasm-associated myelofibrosis.
`
`CONCLUSIONS
`thalidomide
`Considerable preclinical data support efficacy of
`therapy for cGVHD. Data from uncontrolled clinical
`trials of
`therapy after failure of other drugs, mostly corticosteroids, support
`this notion. Randomized trials have not been performed
`for the second-line cGVHD treatment. Data from randomized trials
`of cGVHD prevention or
`initial
`therapy are less convincing.
`One randomized therapy trial could not be completed because
`few subjects could tolerate the prescribed dose (nor could
`subjects receiving placebo) whereas another study showed
`no benefit when thalidomide was added to standard post
`transplant immune suppression. The one prevention trial showed
`no benefit.
`Disparate results of intervention to prevent and/or treat aGVHD
`or cGVHD are common. Examples include daclizumab that seems
`effective in corticosteroid-resistant aGVHD but detrimental when
`added to corticosteroids as initial therapy or mycophenolate that
`is ineffective when added to steroids for the initial treatment of
`cGVHD.87,88 Reasons for
`this are complex and incompletely
`understood.
`Its possible corticoisteroid-resistant cases of GVHD
`are biologically different than untreated cases. We conclude
`thalidomide is likely to be effective as second-line therapy of
`cGVHD therapy but that it is difficult to give doses compatible
`with those effective in preclinical models.
`There are few data regarding lenalidomide therapy or preven-
`tion of cGVHD.
`It is unclear whether this represent publication
`bias, limited use or other factors. Persons with cGVHD typically
`have various degrees of BM dysfunction and the BM toxicity of
`lenalidomide poses a substantial challenge. Although low-dose
`lenalidomide has not been extensively evaluated, it seems unlikely
`this will be a useful approach to therapy or prevention of cGVHD.
`There are few data of pomalidomide in cGVHD. The single
`phase-1/-2 trial
`is difficult to interpret. A randomized phase-2
`study will start soon. Lack of neurotoxicity and BM toxicity is
`attractive but additional clinical data are needed.
`There are major challenges to developing therapy for cortico-
`steroid-resistant cGVHD. Prominent among these are (1) cGVHD is
`complex and there are no convincing surrogate in-vitro or in-vivo
`parameters to predict benefit. We are left with clinical trials and
`ultimately, placebo-controlled randomized trials, which are
`difficult and costly to perform; (2) although there are considerable
`efforts to define cGVHD and therapy response, none is prospec-
`tively validated. This confounds design and execution of clinical
`trials using end points other than survival; (3) one of the major
`consequences of cGVHD is immune suppression. However,
`immune-suppressive drugs are our dominant therapy interven-
`tion. This may aggravate rather than help cGVHD outcomes;
`(4) cGVHD has multiple impacts confounding outcomes analyses.
`cGVHD is correlated with decreased survival and with disability,
`but preventing or decreasing cGVHD is correlated with increased
`graft failure, infections and cancer recurrence. Consequently, the
`most convincing outcome of a trial of cGVHD intervention is a
`survival benefit. This is difficult to show in a chronic disease and is
`confounded by competing, unrelated causes of therapy failure.
`Progress in treating and/or preventing cGVHD is a substantial
`challenge in improving survival of recipients of blood cell or BM
`allotransplants. Whether IMiD-class drugs will be useful
`in this
`setting remains to be determined. Current focus is on pomalido-
`mide; trials are progressing.
`
`Immune modulators in chronic GVHD
`YPL Linhares et al
`
`FUTURE CONSIDERATIONS
`Progress in diagnosing, staging and treating cGVHD is ideally
`based on an accurate and reliable understanding of pathogenesis.
`Unfortunately, this is unlikely and we remain with empirical clinical
`trials of drugs that seem promising. We can make progress in
`cGVHD by standardizing diagnosis, staging and evaluation of
`response using the proposed NIH consensus criteria, perhaps with
`some added variables. Use of validated biomarkers may also help
`but none are currently available. There is progress in developing
`collaborations and further
`testing the NIH criteria in large
`prospective observational and interventional trials.89–91 The end
`point of any therapy intervention in cGVHD must be a clinically
`important
`improvement
`in transplant outcomes, especially
`survival and quality of life needs to be confirmed in a double-
`blind randomized clinical trial. Substantial progress in preventing
`and treating aGVHD was made 30 years ago without a precise
`understanding of etiology or pathogenesis. We hope similar
`progress may be made in preventing and treating cGVHD which is
`an even more complex and challenging problem because of
`multiple confounded outcomes of therapy interventions. The goal
`is to prevent cGVHD-associated morbidity and mortality without
`losing graft-vs-cancer effects in diseases where it exists. Lack of
`specificity of current cGVHD therapies makes achieving this goal
`challenging and difficult. Currently, we need to focus on
`developing better treatment strategies.
`
`CONFLICT OF INTEREST
`SZP is an employee of the Center for Cancer Research, National Cancer Institute and
`National Institutes of Health. RPG is a parttime employee of Celgene Corp.
`
`ACKNOWLEDGEMENTS
`RPG acknowledges support from the NIHR Biomedical Research Centre funding
`scheme. Statements included in this article do not represent the official position of
`the NCI, NIH or the US government. SZP receives partial research funding support
`through the Cooperative Research and Development Agreement for intramural-PHS
`clinical re