`
`Th e
`
`Journal of
`Immunology
`
`Cytokines in Graft-versus-Host Disease
`
`Andrea S. Henden and Geoffrey R. Hill
`
`Graft-versus-host disease (GVHD) is a complication of
`allogeneic bone marrow transplantation whereby trans-
`planted naive and marrow-derived T cells damage recip-
`ient tissue through similar mechanisms to those that allow
`destruction of malignant cells, the therapeutic intent of
`bone marrow transplantation. The manifestations and
`severity of GVHD are highly variable and are influenced
`by the proportions of naive cells maturing along regula-
`tory T cell, Th1, Th2, or Th17 phenotypes. This matu-
`ration is largely influenced by local cytokines, which, in
`turn, activate transcription factors and drive development
`toward a dominant phenotype. In addition, proinflamma-
`tory cytokines exert direct effects on GVHD target tissues.
`Our knowledge of the role that cytokines play in orches-
`trating GVHD is expanding rapidly and parallels other
`infective and inflammatory conditions in which a predom-
`inant T cell signature is causative of pathology. Because
`a broad spectrum of cytokine therapies is now routinely
`used in clinical practice, they are increasingly relevant to
`transplant medicine. The Journal of Immunology, 2015,
`194: 4604–4612.
`
`G raft-versus-host disease (GVHD) is a phenomenon
`
`almost unique to allogeneic bone marrow trans-
`plantation (BMT) whereby lymphocytes are intro-
`duced and permitted to engraft and proliferate within an
`immunocompromised host. In this setting, naive (i.e., those that
`have not previously encountered Ag) donor T cells are able to
`recognize host or recipient Ags as foreign, an effect that constitutes
`the therapeutic intent of BMT, allowing destruction of leukemic
`or other malignant cells through activation of pathways of the
`adaptive immune response. This beneficial effect is termed “graft-
`versus-leukemia” (GVL). The relative contributions of memory
`T cells to GVHD and GVL were discussed elsewhere (1).
`However, the effect is not specific to malignant cells, and si-
`multaneous damage and destruction of healthy cells and tissues
`via the same or similar mechanisms give rise to GVHD. The
`morbidity and mortality of GVHD limit the clinical scenarios in
`which allogeneic hematopoietic stem cell transplantation may
`otherwise offer therapeutic benefit. Therefore, much research has
`focused on the separation of GVL and GVHD, although success
`has been limited because of the use of the same immune effector
`
`mechanisms. An example of this is T cell depletion of trans-
`plants: a reduction in GVHD is offset by attendant increases in
`the rates of relapse of primary malignancy (2, 3), in addition to
`more delayed immune reconstitution with increased morbidity
`and mortality due to opportunistic infection. An alternate focus
`has been to examine the influences on emerging innate and
`adaptive immune responses in an attempt to preserve beneficial
`GVL effects while eliminating the harmful “off-target” GVHD
`effects. In this setting, understanding the cytokine orchestration
`of the maturing immune response within allogeneic transplan-
`tation offers the opportunity to improve outcomes of this
`treatment through identification of rapidly translatable clinical
`therapeutic targets. Our understanding of events within the al-
`logeneic transplantation landscape also informs our under-
`standing of emerging innate and adaptive immune responses in
`scenarios other than BMT.
`
`Cytokines and acute GVHD
`
`The initiation of GVHD is necessarily influenced by the cy-
`tokine milieu in which it arises, and three distinct phases have
`been described (4, 5). The initial phase is triggered by tissue
`damage and associated loss of mucosal barrier function, pri-
`marily in the gastrointestinal (GI) tract, which is caused by the
`conditioning regimens needed to bring malignant disease to
`a minimal residual level suitable for subsequent immune control
`and to ablate existing immune function, allowing engraftment of
`the naive donor inoculum. Myeloablative stem cell transplan-
`tation typically uses total body irradiation or busulphan-based
`chemotherapy to achieve these dual aims; however, they also
`result in damage to the GI tract mucosa and other cells con-
`tributing to the “cytokine storm,” which is characterized by the
`release of proinflammatory cytokines: classically TNF, IL-1, and
`IL-6 (4, 6). Although less well defined, there is an appreciation
`that a similar process occurs with reduced intensity–condition-
`ing transplantation, although the dominant cytokines and
`temporal relationships may differ (7).
`In addition to chemotherapy and radiation-induced tissue
`damage and inflammation, recognition of pathogen-associated
`molecular patterns, such as LPS, and danger-associated mo-
`lecular patterns arising from GI microbiota have significant
`bearing on GVHD pathophysiology. The inflammatory sig-
`nals generated in the emerging adaptive immune response are
`added to by recognition of molecular motifs from both
`pathogenic and commensal organisms and subsequent acti-
`
`Bone Marrow Transplantation Laboratory, QIMR Berghofer Medical Research Institute,
`Brisbane 4006, Queensland, Australia; and The Royal Brisbane and Women’s Hospital,
`Brisbane 4029, Queensland, Australia
`
`Abbreviations used in this article: aGVHD, acute GVHD; BMT, bone marrow trans-
`plantation; cGVHD, chronic GVHD; GI, gastrointestinal; GVHD, graft-versus-host
`disease; GVL, graft-versus-leukemia; TFH, T follicular helper; Treg, regulatory T cell.
`
`Received for publication January 20, 2015. Accepted for publication March 19, 2015.
`
`Address correspondence and reprint requests to Prof. Geoffrey R. Hill, QIMR Berghofer
`Medical Research Institute, 300 Herston Road, Herston, Brisbane 4006, QLD,
`Australia. E-mail address: geoff.hill@qimrberghofer.edu.au
`
`www.jimmunol.org/cgi/doi/10.4049/jimmunol.1500117
`
`Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`
`
`The Journal of Immunology
`
`4605
`
`vation of innate lymphoid pathways. The diversity of the
`resident organisms can be affected by conditioning-associated
`inflammation and by GVHD itself; conversely, the microbiota
`present can influence the severity of GVHD (8). The quan-
`titative and qualitative contributions of this microbiota-driven
`inflammatory signal are influenced by the variety and path-
`ogenicity of organisms present, and has been demonstrated to
`affect the severity of GVHD (8–10). A reduction in the
`bacterial burden by use of antimicrobial decontamination in
`the posttransplant period also can reduce GVHD severity
`(10). Although our mechanistic understanding of this effect is
`not complete, it is apparent that GVHD mediates a loss of
`Paneth cell–derived antimicrobial peptides that play an im-
`portant role in shaping the diversity of microbiota (11), in
`addition to the use of pharmaceutical antimicrobials (9).
`Understanding these mechanisms may offer manipulable
`targets to alter this primary, inflammation-mediated, initia-
`tion phase of GVHD.
`Recognition of the range of triggers to the cytokine storm
`complements our knowledge of subsequent T cell and APC
`interactions that define the second phase of acute GVHD
`(aGVHD) pathophysiology and during which cytokines play
`a key role in driving naive T cell differentiation and expansion
`toward one maturation program or another. Type 1 or Tc1/
`Th1 maturation is recognized as the dominant pattern in
`aGVHD (12, 13), and is linked to severe GI tract pathology
`(14). Indeed, in animal models of aGVHD, Th2 and regu-
`latory T cells (Tregs) are rare (15). T cells expressing IL-17 are
`also rare, although this may reflect the plasticity of this lineage
`(16). Increased quantities of Th1-associated cytokines, TNF
`and IFN-g, in aGVHD are associated with earlier onset and
`more severe disease in preclinical models and clinical BMT (4,
`14, 17–19). Although the dominance of Th1 subsets is well
`established, Th2 and Th17 subsets are also involved in pa-
`thology, and the balance between subsets determines aGVHD
`severity (20), in addition to organ specificity (15, 20), and the
`pathogenic or protective effects of any subset cannot be
`viewed in isolation. Implicit in the known reciprocal regula-
`tion of T cell differentiation by these cytokines is the concept
`that inhibition of any one lineage may provoke unwanted and
`exaggerated differentiation down alternative differentiation
`pathways.
`Th2 differentiation is often seen as opposing Th1 differ-
`entiation; however, this subset is also recognized to cause
`aGVHD but with predominant pathology in pulmonary,
`hepatic, and cutaneous tissues (21), in contrast to the strong
`GI association with Th1. Cutaneous pathology also may be
`generated by Th17 cells; although they are more commonly
`associated with chronic GVHD (cGVHD), they also have
`been associated with acute pathology (22–24). Th17 differ-
`entiation is initiated by IL-6 (25), and RORgt is the defining
`transcription factor (26), whereas maintenance and amplifi-
`cation relies on IL-23 and IL-21, respectively (27). The use of
`RORC-deficient donor T cells results in attenuated aGVHD
`severity and lethality (26). Further studies are needed to better
`define the role of this subset in late aGVHD versus early
`cGVHD, as well as the relative contribution of IL-17 from
`CD4 and CD8 T cells to end-organ pathology.
`The third and final effector phase of aGVHD is charac-
`terized by target tissue damage, with the hallmark histological
`finding being apoptosis, most commonly in the GI tract, liver,
`
`and skin. This tissue damage is mediated by more than one
`immunological mechanism. First, cognate T cell–MHC
`interactions are required for effector, usually CD8, T cells that
`are able to evoke cytolytic machinery, including perforins and
`granzymes that induce target cell death via apoptosis (5, 6,
`28). Interestingly, granzyme B–deficient donor T cells me-
`diate less severe GVHD but may still generate GVL (29), via
`reduced activation-induced death of CD8 T cells (30). In
`a complementary pathway in which cognate T cell–MHC
`interactions are not required, myeloid cells, in addition to
`lymphoid cells, are primed during aGVHD to release cyto-
`pathic quantities of inflammatory cytokines (e.g., TNF, IL-6)
`that directly invoke apoptosis (31). Importantly, TNF is also
`involved in GVL effects, and inhibition can compromise
`antitumor immunity (32). Damage to the primary target
`organs of GVHD is driven by chemokine expression that
`results in tissue homing of lymphocyte populations. LPAM-1
`(a4b7 integrin) and L-selectin (CD62L) are associated with
`homing to GI and GALTs and cutaneous lymphoid Ag to
`skin (33) and are necessary for induction of GVHD tissue
`damage at these sites (34). Cytokines, including IFN-g, are
`known to induce upregulation of chemokines and recep-
`tors (35), and these mechanisms were shown to be important
`in determining the severity of GVHD within an inflam-
`matory environment (36–38). The expression of mole-
`cules associated with lymphocyte exhaustion (e.g., PD1) and
`their ligands (e.g., PD-L1) on nonlymphoid tissue also was
`shown to be cytokine (IFN) dependent and contribute to the
`constraint of
`lymphocyte-mediated tissue damage late in
`the aGVHD setting (39, 40). Therefore, the inflamma-
`tory cytokines present in this setting participate in positive-
`and negative-feedback loops in both lymphocyte and non-
`lymphocyte populations.
`
`Cytokines and cGVHD
`
`cGVHD represents a distinct pathophysiological entity from
`aGVHD that traditionally is separated by time of onset; how-
`ever, it is now recognized by its distinct end-organ pathology.
`Although the cardinal feature of aGVHD is apoptosis, fibrosis
`is the predominant mechanism of tissue damage in cGVHD.
`Additionally, primary target organs also differ, with lung and
`skin being the primary target organs in cGVHD, manifesting
`as bronchiolitis obliterans and scleroderma (41, 42). Sicca
`symptoms secondary to salivary and lacrimal gland destruc-
`tion and oral lichenoid GVHD are also prominent. Despite
`these disparate pathophysiological manifestations, clear roles
`for cytokine control of this phase of disease also were dem-
`onstrated, unaccompanied by large-scale conditioning-related
`tissue damage and the “cytokine storm” that initiates aGVHD.
`IL-17 and subsequent T cell differentiation along the Th17
`pathway are becoming more strongly associated with cGVHD.
`Initially identified with the use of G-CSF in stem cell mo-
`bilization of donors and prominent Th17 differentiation (43),
`IL-17 was shown more recently to result in CSF1-dependent
`macrophage accumulation in skin and lung, which drives tis-
`sue fibrosis (44). We demonstrated recently that, consistent
`with this, systemic IL-17 levels increase late after clinical
`BMT, at a time when cGVHD develops (45). It is also clear
`that T follicular helper (TFH) cells and IL-21 play important
`roles in the development of cGVHD via the stimulation of
`germinal center B cells and alloantibody generation (46).
`
`
`
`4606
`
`BRIEF REVIEWS: CYTOKINES IN GVHD
`
`This is particularly relevant to bronchiolitis obliterans, be-
`cause preliminary evidence suggests that Th17 differentia-
`tion and CSF1 dysregulation are also involved in this
`aberrant immunological pathway (44). Thus, inhibition of
`Th17 differentiation and CSF1 appear highly relevant to
`the prevention and treatment of cGVHD. Inhibition of ter-
`minal cytokines involved in fibrosis, such as TGF-b and
`IL-13, represent additional targets; however, TGF-b inhi-
`bition is problematic given its important role in Treg ho-
`meostasis (47).
`
`Cytokines and T cell–differentiation programs
`
`The ability of cytokines to drive cellular differentiation is
`recognized in situations other than GVHD, with the accep-
`tance that derivation of phenotypically distinct erythroid,
`myeloid, and lymphoid populations from a common long-
`term hematopoietic stem cell is dependent upon binding of
`cytokines to their cognate receptors (48, 49). Similarly, the
`maturation of the naive T cell population in the context of
`BMT and GVHD is also driven by cytokines and subsequent
`transcriptional pathways elicited thereafter (50). A summary
`of these effects is shown in Fig. 1 and outlined further below.
`Cytokines and Th1 differentiation. IFN-g, IL-2, and TNF are
`the key cytokines generated during Th1 differentiation (14),
`and phenotypic differentiation is initiated by IL-12 and
`controlled by the transcription factor T-bet (25, 48). IFN-g
`in this setting participates in positive feedback to reinforce Th1
`responses, in addition to exerting effects on nonlymphocytes,
`as well as on nonhematopoietic cells (18). Initial attempts
`to define the role of IFN-g in determining aGVHD severity
`were hampered by conflicting data supporting the exacerba-
`tion and amelioration of pathology; however, subsequent
`work demonstrated this to be related to differing effects on
`donor and host tissues, in addition to tissue-specific effects
`
`on nonhematopoietic tissues. Donor lymphocyte IFN-g
`signaling enhanced GVHD via the promotion of Th1 dif-
`ferentiation, and it also is directly cytotoxic to gut mucosa
`(18). Tissue-specific effects are also seen in pulmonary pa-
`renchyma in which a protective role for IFN-g was dem-
`onstrated and these effects have also been described by
`other groups (15). IFN-g provides evidence for a paradigm
`where cytokines may exert effects in nonhematopoietic tissue,
`in addition to specific effects on lymphocytes and other he-
`matopoietic cells. Evidence of similar patterns for other cy-
`tokines is continually being defined and allows selection of
`appropriate targets for inhibition in the clinic.
`With regard to other Th1-associated cytokines, a similarly
`complex effect is seen for IL-2, both mechanistically and in
`therapeutic outcomes (51). Initially used at a high dose in an
`attempt to augment proliferation of lymphocytes as “immu-
`notherapy” for solid malignancies (52, 53),
`it was subse-
`quently found, paradoxically, to have a critical role in sup-
`porting Treg populations and in controlling GVHD (51,
`54). The promotion of regulatory pathways in GVHD was
`demonstrated in small numbers of patients when used in
`a “low dose” (55). These apparently dose-dependent effects
`are likely explained by competition for consumption of this
`cytokine by maturing Treg and T effector cell populations
`(56); in the clinical transplantation scenario, they are further
`complicated by the use of calcineurin inhibitors, which also
`target this pathway, when used as GVHD prophylaxis post-
`transplant.
`Cytokines and Th2 differentiation. The presence of IL-25 and,
`subsequently, IL-4 supports the development of T cells of
`the Th2 lineage that traditionally have been described as
`being involved in allergy and host defense against parasites
`and helminths. Th2 cells produce IL-4, IL-5, IL-10, and
`IL-13, with transcriptional control exerted by GATA3 (57).
`
`.. Tissue
`
`damage
`
`Acute
`GVHD
`
`Chronic
`GVHD
`
`Cytokine driven
`cellular differentiation
`
`Naive
`cells
`
`chemotherapy -,,
`conditioning
`damage
`
`+ Radiation and _.,.
`(bO
`
`IL-4 -IL-12 -IL-1
`
`IL-6 -
`
`IFNy
`TNF
`
`Tcells
`
`Oo O
`
`NK/NKTI
`Innate
`Lymphoid
`Cells(ILC)
`
`Monocytes
`and
`macrophages
`
`Microbiota
`derived DAMPs
`and PAMPs
`
`Tc1/Th1; IFNy, TNF, IL-2
`
`NK/NKT/ILC activation
`
`-
`
`FIBROSIS
`
`APOPTOSIS
`
`y
`
`granzyme
`mediated (cid:173)
`cytolysis
`
`TNF,
`IL-1
`and IL-6
`
`Activated monocytes
`and macrophages
`
`......_ TGF-~
`
`FIGURE 1. Cytokine drivers in the three phases of aGVHD initiation and end-organ pathology. Initial inflammatory signals are elicited by cellular damage from
`chemo- and radiotherapy, in addition to those derived from gut microbiota following GI tract damage and loss of integrity. Cytokines act on naive T cell, ILC, and
`myeloid cell populations, resulting in differentiation to Th1, Th2, and Th17 cell subsets, activated ILC subsets, and activated myeloid cells. End-organ tissue damage
`in aGVHD is caused by apoptosis elicited by Th1/Tc1 cytokines and cytolytic machinery, including perforin and granzyme, following cognate TCR–MHC
`interactions. Additional inflammatory pathways that are not dependent on cognate T cell pathways, including IL-6– and TNF–mediated apoptosis, following
`release of these cytokines from activated monocyte and macrophage populations. End-organ damage in cGVHD classically follows aGVHD and is mediated by
`Th2/Th17 cells and monocyte/macrophage populations secreting TGF-b that result in tissue fibrosis. The influence of ILCs on GVHD requires further de-
`lineation but they may be regulatory, at least early after BMT.
`
`
`
`The Journal of Immunology
`
`4607
`
`In aGVHD, the Th2 program appears to mediate skin and
`lung pathology (15, 18), as opposed to the strong association
`of Th1 cells with gut and liver damage. Recent work dem-
`onstrated a role for an IL-25–dependent immature non-B
`non-T cell innate immune effector cell population that is re-
`sponsible for propagation of Th2 responses and production
`of IL-4, IL-5, and IL-13; where lacking, this results in an
`impaired ability to expel helminths (58–60). This cytokine
`may be seen as having protective effects on GI tissues, an
`outcome that is clearly attractive in the context of aGVHD
`pathology.
`Other Th2 cytokines were shown to have protective roles in
`GVHD, including IL-10. Despite mixed results when initially
`given as therapy to patients, its production by B lymphocytes
`in animal models of transplantation reduces the severity of
`GVHD (61). These outcomes were mediated by effects on
`donor T cell expansion; similarly, reductions in IL-4 and IL-
`10 were demonstrated in patients with cGVHD (62, 63).
`Cytokines and Th17 differentiation. Th17 cells are a more recent
`addition to the Th1/Th2 paradigm (25), and roles for IL-17–
`producing cells of both Th17 and Tc17 varieties are still being
`defined in the GVHD setting and in other immune pathologies
`(64). Initiation of Th17 development is triggered by IL-6 and
`TGF-b and is associated with transcriptional activation of
`RORgt after phosphorylation of STAT3, as well as with the
`secretion of IL-17, IL-21, and IL-22. There is an increasing
`appreciation for the role that Th17 cells play in determining
`the severity of GVHD (65), with a particular role for IL-6
`becoming apparent (66, 67). Recently, our group demonstrated
`the importance of this effect in cohorts in which IL-6 inhibition
`represents a potentially effective therapeutic strategy to reduce the
`severity of aGVHD in clinical stem cell transplantation (45). IL-
`22 may be secreted by Th17 cells and, in this setting, it appears to
`be pathogenic (68); conversely, it may be secreted by innate
`lymphoid cells where, in the GI tract at least, it appears to be
`an important protective cytokine (69). IL-21 can be produced by
`both Th17 and TFH cells, and it promotes aGVHD by impairing
`Treg homeostasis (70). Given that it also has an important role in
`inducing aberrant, allospecific germinal center B cell responses
`and cGVHD, inhibition of this Th17-associated cytokine rep-
`resents another attractive therapeutic target for GVHD control
`after BMT.
`
`Heterogeneity of effect conferred by cellular target
`
`Increasingly, the effects mediated by a particular cytokine are
`being defined as dependent on the cells in which it transduces
`a signal and may be considered to regulate effector cell pop-
`ulations, as well as to confer susceptibility or protection to other
`inflammatory signals in target organs and tissues. Modes of
`signaling also influence these responses, with a recent appreci-
`ation for the differential pathology induced by the binding of
`cytokines by membrane-bound receptors as opposed to soluble
`receptors.
`Hematopoietic and nonhematopoietic cells. Cytokines may exert
`effects on cells of hematopoietic origin, in addition to non-
`hematopoietic target tissues directly. IFN-g (18) and IL-22
`(68, 69) are clear examples that were already discussed.
`Further examples are seen in cGVHD: IL-2, IL-10, and
`TGF-b may act directly on tissue fibroblasts in affected
`organs to mediate pathology (71),
`in addition to known
`effects of IL-2 in supporting Treg populations (51, 55) and
`
`B cell–derived IL-10 being protective in the initiation of
`aGVHD (61). A role for innate lymphoid cells as cytokine-
`responsive mediators of protection from GVHD is emerging
`(72), with ILC1, ILC2, and ILC3 subtypes demonstrating
`similar transcriptional control and cytokine profiles to Th1,
`Th2, and Th17 cells (73).
`Donor and host cells. The effect of a single cytokine can
`be dependent upon its roles in hematopoietic and non-
`hematopoietic tissue; however, in BMT, donor or host origin of
`the cell transducing the signal is as an additional factor influ-
`encing outcomes. Type I IFN is an example of a cytokine for
`which signaling through recipient APCs results in less severe class
`II–dependent GVHD in the colon, whereas signaling through
`donor APCs may amplify GVHD responses. The former is
`mediated through decreased donor CD4 proliferation, and the
`latter is mediated through more effective cross-presentation of
`alloantigens to CD8 T cells (19). IL-4 is another example. A
`subpopulation of recipient NKT cells secretes high levels of IL-4
`and indirectly expands donor Treg populations to promote
`tolerance after BMT (74). In contrast, IL-4 may drive donor
`Th2 differentiation directly and enhance GVHD that likely
`usually represents chronic disease. Appreciation of these
`mechanisms is important, because treatment of a recipient or
`graft can be temporally separated and offers the opportunity to
`select desirable effects while avoiding potentially deleterious
`outcomes.
`Receptor disposition. Additional complicating factors exist when
`considering cytokine-mediated effects in immune-mediated and
`inflammatory conditions. IL-6 is an example of a cytokine for
`which signaling via “classical” or membrane-bound receptor–
`ligand interactions produces differing pathology than does
`signaling mediated through soluble or trans receptor binding.
`These effects were described originally in mouse models of
`rheumatoid arthritis, in which trans signaling (by the IL-6–
`soluble IL-6R complex) recapitulated inflammatory joint
`disease in IL-6–deficient mice, whereas injection of the
`native IL-6 cytokine itself did not (75). Subsequently, IL-6
`signaling through the trans pathway has been thought to be
`more inflammatory in nature than classical signaling, in part
`relating to the ability of IL-6 to signal through cells that ex-
`press the gp130 receptor complex but do not basally express
`IL-6R (76). A similar paradigm was demonstrated in al-
`lergic asthma: Th2 expansion appears to be driven by trans
`signaling whereby expansion of Tregs was limited by classical
`IL-6 signaling, and inhibition with anti–IL-6R mAb resulted
`in increased numbers of Tregs (77), and an increase in asthma
`risk was associated with a single nucleotide polymorphism
`that results in an increase in soluble IL-6R and trans signaling
`(78). Appreciation of the mechanisms by which a cytokine can
`mediate differential effects is critical to understanding both
`disease pathophysiology and effective clinical translation of
`therapeutics. The availability of mAbs to cytokine receptors,
`such as tocilizumab for IL-6R, which inhibits all IL-6 signaling,
`in addition to more specific inhibitors of signalling pathway
`components, such as soluble gp130:Fc, which inhibits IL-6
`trans signaling only, is a clear example.
`
`Translational application
`
`Accepted murine models of transplantation and rapid and
`reproducible multiplexed techniques to measure cytokines in
`serum or cell culture supernatants or intracellular cytokine
`
`
`
`4608
`
`BRIEF REVIEWS: CYTOKINES IN GVHD
`
`production by flow cytometry have allowed identification of,
`and will continue to define, the key cytokines in aGVHD
`(45, 79), as well as facilitate clinical translation of findings.
`However, a number of factors must be considered when ex-
`trapolating laboratory observations
`into clinical cohorts.
`Variation exists in transplantation protocols, patient pop-
`ulations, modes of conditioning, and posttransplant immune-
`suppression strategies. The last factor is of particular im-
`portance when considering the translation of observations
`made in animal models to the clinical setting, where immune
`suppression with cyclosporin or tacrolimus, combined with
`methotrexate or mycophenolate, is considered standard of
`care to avoid life-threatening acute and severe GVHD. How-
`ever, most animal models of transplantation rely solely on
`radiation-based conditioning. Therefore, effective transla-
`
`tion will require validation of observations made in animal
`models with clinical cohorts, because standard immune-
`suppressing agents were shown to affect cytokine levels pro-
`duced by T cells and NK cells, and profiles vary with stem cell
`source (80, 81). Importantly, IFN-g, TNF, and IL-1 are not
`systemically dysregulated in clinical subjects after BMT who
`receive immune suppression in the same way as seen in rodent
`models (45). With this in mind, it should be noted that no
`cytokine-inhibition strategy or cytokine administration has
`proved efficacious in randomized studies. In general, en-
`couraging results seen in preclinical studies and early-phase
`clinical trials have either not progressed into phase III studies,
`or effects have not been robust within this context [e.g., IL-1
`(82) and TNF inhibition (83, 84)]. Table I provides a sum-
`mary of cytokines and inhibitors that have been explored for
`
`Table I.
`
`Summary of relevant cytokine-targeted therapeutic studies
`
`Cytokine Inhibitors
`
`Phase I and/or II Clinical Trial Data
`
`Randomized, Double-Blind
`Controlled and/or Phase III
`Clinical Trial Data
`
`TNF-aR2 (etanercept)
`Prophylaxis
`Treatment
`TNF-a binding mAb (infliximab)
`Prophylaxis
`Treatment
`IL-1Ra (anakinra)
`Prophylaxis
`Treatment
`IL-2Ra/anti-CD25 (basiliximab/daclizumaba)
`Prophylaxis
`Treatment
`IL-6R (tocilizumab)
`Prophylaxis
`Treatment
`Keratinocyte growth factor (palifermin)
`Prophylaxis
`
`Cytokines
`
`IL-2 (aldesleukin)
`Prophylaxis
`Treatment
`IL-11 (oprelvekin)
`Prophylaxis
`
`Cytokines with Non-GVHD Benefits
`
`IFN-a (INTRON A, Roferon-A)—promotion of GVL with concomitant
`promotion of GVHD
`Prophylaxis
`Treatment
`Keratinocyte growth factor (palifermin)—for reduction of oral mucositis
`Prophylaxis
`
`Potential GVHD Therapies
`
`IL-17
`IL-17A mAb (secukinumab, ixekizumab, perakizumab)
`IL-17RA mAb (brodalumab)
`IL-17A/TNF (ABT122)
`IL-22 (fezakinumab)
`IL-12p40/23 mAb (ustekinumab)
`IL-23p19 (guselkumab, tildrakizumab)
`
`IL-13 (lebrikizumab, tralokinumab)
`
`+ (83)
`+ (102–105)
`
`2 (84)
`+ (99, 100)
`
`+ (106)
`
`+ (107)
`+ (108–110)
`
`+ (45)
`+ (113, 114)
`
`2 (115–119)
`
`+ (55)
`+ (51)
`
`2 (121)
`
`+ (122–126)
`+ (127, 128)
`
`+ (116–118)
`
`•
`•
`
`•
`2 (101)
`
`2 (82)
`•
`
`•
`2 (111, 112)
`
`•
`•
`
`2 (120)
`
`•
`•
`
`•
`
`•
`•
`
`+ (120)
`
`Phase I and/or II Clinical
`Data Outside GVHD
`
`Phase III Clinical Trial
`Data Outside GVHD
`
`+ Ixekizumab (129)
`+ (130)
`Ongoing
`Ongoing
`+ (97, 132)
`+ Guselkumab (133)
`+ Tildrakizumab (134)
`+ Tralokinumab (135)
`
`+ Secukinumab (131)
`•
`•
`•
`•
`
`Tildrakizumab (ongoing)
`+ Lebrikizumab (136)
`Tralokinumab (ongoing)
`
`Cytokines and their antagonists are included that have been tested within a trial setting to prevent or treat GVHD or other complications of allogeneic BMT. Also included is a list
`of newer therapeutics with potential application to GVHD that are undergoing testing in other disease settings.
`+, positive data; •, lack of data in this setting; 2, negative data.
`
`
`
`The Journal of Immunology
`
`4609
`
`efficacy in the treatment or prevention of GVHD, in addition
`to agents with potential efficacy in GVHD that are being
`explored in other disease settings. Importantly, most studies
`in GVHD examine the usefulness of cytokine antagonists as
`an adjunct to standard modalities of immune suppression
`rather than their efficacy in isolation, as is usually the case in
`preclinical testing. Thus, it will be important to follow some
`recently defined general principles, taking into consideration
`concurrent immune suppression, when planning to translate
`findings from mice to patients (28).
`The effect of a particular cytokine in any one individual is
`also affected by human genetic heterogeneity, and data already
`demonstrated a clear impact of single nucleotide polymorphisms
`in cytokine loci on GVHD outcomes (85, 86). Despite these
`difficulties in directly translating laboratory observations to the
`clinic, cytokine therapy for GVHD remains fertile ground for
`new and effective therapeutics, because a number of agents that
`augment or antagonize cytokine pathways are already available,
`having been explored and validated in other autoimmune and
`inflammatory disease settings. Cytokine inhibition,
`initially
`with TNF and subsequently with IL-6, is considered routine
`care for rheumatoid arthritis patients whose disease is not
`controlled by more nonspecific immune suppression with
`corticosteroids, methotrexate, and calcineurin inhibitors (87–
`90). Imperfect disease control, when used as monotherapy, has
`paved the way for the use of combination cytokine inhibition,
`and the rational combination of TNF and IL-17 showed efficacy
`in preclinical models of disease (91). Novel targets, such as IL-
`32, IL-34, and IL-35, are also being explored (92). The success
`of cytokine inhibition in rheumatoid arthritis is also paralleled in
`other inflammatory diseases, including psoriatic dermatitis and
`arthritis (93), with TNF inhibition having a demonstrated role,
`in addition to promising newer targets, such as IL-22 and IL-23
`(94, 95). Evidence for the value of cytokine inhibition exists in
`diseases other than inflammatory arthropathies, with efficacy
`demonstrated for TNF (96), as well as IL-12/23 (97, 98), in
`inflammatory bowel disease. In this setting of proven efficacy for
`cytokine inhibition in other diseases of dysregulated immunity,
`further definition of the role of cytokines in the determination of
`GVHD severity is likely to translate rapidly into efficacious
`therapies for the transplant patient population.
`
`Conclusions
`Cytokines are a defining influence on evolving immune
`responses in BMT and in the generation of GVHD, the major
`pathology limiting the wider application of transplantation.
`The classical appreciation of a naive T cell being influenced by
`a cytokine to mature into a more differentiated phenotype is
`made more complex in the GVHD setting as the impact of
`cytokines acting in different tissue compartments (e.g., lym-
`phoid and nonlymphoid or donor and host) and the use of
`classical and trans signaling pathways become better appreci-
`ated. Although these factors require further work to better
`define these complex interactions,
`they go some way in
`explaining the previously mixed and conflicting results asso-
`ciated with some cytokine therapies (99–101). Appreciation
`of this complexity will allow for the development of more
`log