signalling pathways of the tnf superfamily: a - Semantic Scholar

signalling pathways of the tnf superfamily: a - Semantic Scholar

REVIEWS SIGNALLING PATHWAYS OF THE TNF SUPERFAMILY: A DOUBLE-EDGED SWORD Bharat B. Aggarwal Two different tumour-necrosis factors (TNFs), first isola...

1MB Sizes 0 Downloads 4 Views

REVIEWS

SIGNALLING PATHWAYS OF THE TNF SUPERFAMILY: A DOUBLE-EDGED SWORD Bharat B. Aggarwal Two different tumour-necrosis factors (TNFs), first isolated in 1984, were found to be cytotoxic to tumour cells and to induce tumour regression in mice. Research during the past two decades has shown the existence of a superfamily of TNF proteins consisting of 19 members that signal through 29 receptors. These ligands, while regulating normal functions such as immune responses, haematopoiesis and morphogenesis, have also been implicated in tumorigenesis, transplant rejection, septic shock, viral replication, bone resorption, rheumatoid arthritis and diabetes; so indicating their role as ‘double-edged swords’. These cytokines either induce cellular proliferation, survival, differentiation or apoptosis. Blockers of TNF have been approved for human use in treating TNF-linked autoimmune diseases in the United States and other countries.

Cytokine Research Section, Department of Bioimmunotherapy, University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Box 143, Houston, Texas, USA. e-mail: [email protected] mdanderson.org doi:10.1038/nri1184

Similar to most discoveries in science, the discovery of tumour-necrosis factor (TNF) is marked by a series of important landmarks. More than a century ago, a German physician, P. Bruns, reported the regression of tumours in humans after bacterial infection1. This led W. Coley, an American oncologist, to use bacterial extracts (referred as Coley’s toxins) for the treatment of human cancers, with some success2. Gratia et al.3 showed tumour regression by bacterial extracts in a guinea-pig model. In 1944, Shear et al.4 isolated lipopolysaccharide (LPS) from the bacterial extracts and showed that this was responsible for the tumour regression. G. Algire, a French scientist, showed that LPS induces haemorrhagic necrosis of tumours by inducing systemic hypotension, collapse of tumour vasculature, tumour-cell anorexia and cell death5. Subsequently, O’Malley et al.6 showed that the tumour regression effects of LPS are not direct, but are mediated through the induction of a factor in the serum; which they named tumour-necrotizing factor6, renamed by L. Old’s group as tumour-necrosis factor (TNF)7. Old and coworkers originally reported7 that macrophages were the cellular source of this factor, but later they showed that a B-lymphoblastoid cell line was also a source8. This discrepancy is typical of the confusion that occurred at the

NATURE REVIEWS | IMMUNOLOGY

time due to the crudeness of the preparations. Lymphotoxin (LT) was described in 1968 by G. Granger and coworkers9 as a protein that is produced by lymphocytes that kills tumour cells. Not until 1984, when my group purified the proteins to homogeneity10,11, determined the amino-acid sequence11,12 and cloned the complementary DNA13,14, were the true chemical identities of TNF and LT, and their relationship revealed. Human LT was the first cytotoxic cytokine to be purified from a B-lymphoblastoid cell line10; its structure was determined by classic protein sequence methods11. Neutralization of LT activity by LT-specific antibodies led to the isolation of a second cytotoxic factor from a human myeloid-cell line that was known as TNF12. The determination of the amino-acid sequence of TNF indicated that the two proteins were homologous12. The binding of TNF to its receptor and its displacement by LT further confirmed the functional homology between the two proteins15. This was perhaps the earliest indication of the existence of a potential TNF superfamily. Once most of the protein sequence was known, the cDNAs for LT and TNF were isolated13,14. The sequence homology between these two proteins (30% amino-acid identity) and the existence of common cell-surface receptors led to

VOLUME 3 | SEPTEMBER 2003 | 7 4 5

© 2003 Nature Publishing Group

REVIEWS the renaming of TNF and LT to TNF-α and TNF-β, respectively. Soon after, mouse TNF-α was independently discovered as a factor that mediated LPS-induced wasting (cachexia) in mice by Beutler et al.16, and as a myeloid differentiation factor by Takeda and co-workers17. These two cytokines laid the foundation for the isolation and identification of the larger family of cytokines, now known as the TNF superfamily (FIG. 1). Considerable advances have been made during the past two decades in our understanding of the biology and the clinical role of the TNF superfamily. This review focuses on these developments. Characteristic features of ligands and receptors

At present 19 different ligands have been identified that belong to the TNF superfamily. Unlike TNF and LT, the ligands for FAS (CD95), CD27, CD30, CD40, 4-1BB, OX40 and herpes-virus entry mediator (HVEM) were identified and cloned not by protein sequencing, but by direct expression-cloning strategies18–21. This involved the use of specific antibodies, and studies of ligand− receptor interactions or biological activity. The description of the amino-acid sequences of several ligands and receptors of the superfamily led to the identification of certain regions of homology. Instead of an expressioncloning strategy, the availability of human genome sequences led scientists to use homology searches to identify several additional members of the TNF superfamily. TNF-related apoptosis-inducing ligand (TRAIL), also known as APO2L, was identified independently by two different groups22,23, followed by the identification of receptor activator of nuclear factor-κB (NF-κB) ligand (RANKL), also known as TRANCE (TNF-related activation-induced cytokine) or OPGL (osteoprotegerin ligand)24−26, vascular endothelial cell-growth inhibitor (VEGI), also known as TL1A27,28, and B-cell-activating factor (BAFF)21,30, also known as β-lymphocyte stimulator (BLYS), all by similar approaches. Glucocorticoidinduced TNF receptor family receptor ligand (GITRL) was identified and cloned by using a glucocorticoidinduced differential display technique31. Because numerous groups discovered the same cytokine independently and simultaneously, numerous names exist for each TNF family member. An intriguing feature of these ligands is that when certain ligands are shed, they inhibit the function of the ligand–receptor complex. For example, whereas the membrane-bound CD95 ligand (CD95L) kills human peripheral-blood T cells, soluble CD95L blocks this killing36. The fact that some soluble ligands act as agonists whereas others act as antagonists is intriguing. Furthermore, the transmembrane expression of most of the TNF-superfamily members indicates that they are meant to act locally. Only under non-physiological conditions, when these ligands are released, do they prove harmful. Perhaps the best example is TNF-α, which when released binds to various cell types to mediate various diseases. The 19 ligands mediate their cellular response through 29 receptors that belong to the TNF receptor (TNFR) superfamily, and they are characterized by the

746

| SEPTEMBER 2003 | VOLUME 3

presence of a cysteine-rich domain (CRD) in the extracellular portion (FIG. 1). These include four decoy receptors, known as DCR1, DCR2, DCR3 and osteoprotegerin (OPG), which reduce the ligand signalling. The ligands for death receptor 6 (DR6), receptor expressed in lymphoid tissues (RELT) and TROY have not yet been identified. Whereas germline mutations in the extracellular domain of TNFR1 suppress the secretion of the receptor leading to a family of dominantly inherited autoinflammatory syndromes37, the secretion of OPG affects the regulation of bone density38. Receptors of the TNFR superfamily are broadly divided into two groups. Those with the death domain are known as death receptors. The death domain is a region of approximately 80 amino-acid residues first identified independently by two different groups in the cytoplasmic domain of CD95 and TNFRs; their deletion abolishes ligand-induced apoptosis21,39,40. Interestingly, a TNFR that has a cysteine-rich region in its extracellular portion and serine/threonine protein kinase activity in its cytoplasmic domain has been discovered in the maize plant41. This protein, known as CRINKLY4, is a TNFR-like receptor and has kinase activity that is involved in maize epidermal differentiation. Additionally, a conserved domain in the extracellular region of TNFR1 and TNFR2 has been identified that mediates specific ligand-independent assembly of receptor trimers42. This pre-ligand-binding assembly domain (PLAD) is physically distinct from the domain that forms the main contacts with the ligand, but is required for the assembly of TNFR complexes that bind TNF and mediate signalling. Other members of the TNFR superfamily, including TRAIL receptor 1 (TRAILR1) and CD40, show similar homotypic associations. So, TNFRs and related receptors seem to function as preformed complexes rather than as individual receptor subunits that oligomerize after ligand binding. The cellular expression patterns of various members of the TNF ligands and receptors, are shown in TABLE 1. Almost all of the TNF ligands are expressed by cells of the immune system, including B cells, T cells, NK cells, monocytes and dendritic cells, the only exception is VEGI, which is expressed mainly by endothelial cells43. The TNFRs, however, are expressed by a wide variety of cells. For example, no cell type in the body has yet been found that does not express TNFR1, whereas expression of TNFR2 is mainly by immune cells and endothelial cells. Numerous activities are assigned to TNF, perhaps because its receptors are expressed by all cell types. This might also account for the non-specific toxicity of TNF. Whereas most ligands bind to a single receptor, others bind to more than one. For example, TRAIL binds to as many as five receptors (DR4, DR5, DCR1, DCR2 and OPG), whereas BAFF binds to three receptors; transmembrane activator and cyclophilin ligand interactor (TACI), B-cell maturation antigen (BCMA) and BAFFR. Why there are several receptors for the same ligand is unclear. Both homomeric and heteromeric complexes of DR4 and DR5 have been described44,45, leading to the recruitment of downstream cell-signalling intermediates. In the case of TRAIL, although DR4 and DR5 transduce most signals, DCR2 transduces partial

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group

REVIEWS

TNFR1 221

TNF-α

TNFR2 174

TNF-β

145

CD95L

DR3

193

VEGI

DR4

203

TRAIL

DR5

221

CD95 DCR3

DCR1 DCR2

154

DR6

286

?

EDAR

236

EDA-A1

NGFR

155

NGF

RANK

282

RANKL

LTβR

187

LTα/LTβ

FN14

28

TWEAK

OPG

HVEM 58

LIGHT

CD27

48

CD27

CD30

188

CD30

CD40

62

CD40

4-1BB 45

4-1BBL

OX40

36

OX40L

GITR

42

GITRL

BCMA 106

APRIL

106

TACI

BAFFR 82

BAFF

XEDAR 134

EDA-A2

TROY

233

RELT

243

Furin

? ? Metalloproteinase (TACE)

Cysteine-rich domain

Matrilysin

TRAF1

TRAF2

Death domain

TRAF5

TRAF6

TRAF3

Figure 1 | A diagrammatic representation of the ligands of the TNF superfamily and their receptors. All ligands, except lymphotoxin-α (LTα) and vascular endothelial cell-growth inhibitor (VEGI) which are secreted, are type II transmembrane proteins with a carboxy-terminal extracellular domain, an amino-terminal intracellular domain and a single transmembrane domain. The C-terminal extracellular domain, known as the tumour-necrosis factor (TNF) homology domain, has 20–30% amino-acid identity between the superfamily members and is responsible for binding to the receptor. TNF-related apoptosis-inducing ligand (TRAIL) and CD95 ligand (CD95L) have the highest homology among the TNF-superfamily members, in parallel with their ability to induce apoptosis. Most members of the TNF superfamily are released from the cell surface by proteolysis through distinct proteases. TNF and receptor activator of nuclear factor-κB (NF-κB) ligand (RANKL) are released by metalloproteases of the ADAM (a disintegrin and metalloprotease domain) family32,33, CD95L by matrilysin34, and B-cell-activating factor (BAFF), ectodermal dysplasin (EDA), TNF-like weak inducer of apoptosis (TWEAK) and a proliferation-inducing ligand (APRIL) by members of the furin family30,35. Mutations in the furin consensus sequence have been shown to block proteolytic release of EDA and to cause X-linked hypohydrotic ectodermal dysplasia35. Unlike the ligands, the TNF receptors (TNFRs) are characterized as type I transmembrane proteins (extracellular N-terminus and intracellular C-terminus). Because they lack a signal peptide sequence, B-cell maturation antigen (BCMA), transmembrane activator and cyclophilin ligand interactor (TACI), BAFF receptor (BAFFR) and X-linked EDA receptor (XEDAR) belong to type III transmembrane protein group. Decoy receptor 1 (DCR1) is anchored by a covalently linked C-terminal glycolipid. Osteoprotegerin (OPG) and DCR3 lack a transmembrane domain and are therefore secreted as soluble proteins. Soluble forms of several receptors have been identified, which seem to have important functions, including CD27, CD30, CD40, TNFR1, TNFR2, CD95 and 4-1BB. Soluble CD95 and 4-1BB are generated by alternative splicing18. Until now, eight different death-domain-containing receptors have been identified, death receptor 1 (DR1, also known as TNFR1), DR2 (CD95), DR3, DR4 (also known as TRAILR1), DR5 (also known as TRAILR2), DR6, EDAR and low-affinity nerve growth-factor receptor (NGFR). Although structurally NGFR belongs to the TNFR superfamily, many of its biological actions differ from other members of the family. The numbers on the left represent the number of amino acids in the cytoplasmic domain of the receptor. GITR, glucocorticoid-induced TNFR family receptor; HVEM, herpes-virus entry mediator; RELT, receptor expressed in lymphoid tissues; TRAF, TNFR-associated factor.

NATURE REVIEWS | IMMUNOLOGY

VOLUME 3 | SEPTEMBER 2003 | 7 4 7

© 2003 Nature Publishing Group

REVIEWS signals only46. Through sequestration of the ligand, DCR1, DCR2, DCR3 and OPG have been shown to inhibit cell signalling. An OPG variant protein that has two death domains, but lacks a transmembrane domain, however, has also been shown to function as a ligand through an unknown receptor47. In the case of TNF, it has been suggested that although TNFR1 is a receptor for a soluble ligand, TNFR2 mediates the signalling of the membrane-bound ligand48. Whereas some reports indicate that TNFR1 mediates apoptosis and TNFR2 mediates proliferation, others suggest that the two TNFRs transduce their signals cooperatively49,50. The genetic deletion of either receptor blocked most of the signals that are transduced by TNF50. There is also evidence of crosstalk between receptors for different ligands of the TNF superfamily. For example, CD95L18, TNF-like weak inducer of apoptosis (TWEAK)51, CD30L and CD40L52 were found to trigger through their cognate receptors the

upregulation of expression of endogenous TNF, which subsequently activates TNFR1. Gene-deletion experiments have shown that crosstalk between TNFRs and LPS receptors might occur, indicating an overlap in cell-signalling pathways between different receptors53. Cell signalling. After binding to the receptor, members of the TNF superfamily either mediate apoptosis (such as TNF, LT, CD95L, TRAIL, VEGI, TWEAK and LIGHT), survival (such as RANKL and BAFF), differentiation (such as TNF, RANKL and DR6) or proliferation (such as TNF, CD27L, CD30L, CD40L, OX40L, 4-1BBL, APRIL and BAFF) through the activation of pathways involving NF-κB, JUN N-terminal kinase (JNK), p42/p44 mitogen-activated protein kinase (MAPK) and p38 MAPK. Signalling through TNF and CD95 has been a paradigm for most other members of the TNF superfamily. None of the receptors of the mammalian TNF superfamily has any enzymatic activity. By using the yeast

Table 1 | Cellular expression of ligands and receptors of the tumour-necrosis factor superfamily Cytokine

Cells

Receptor

Cells

LTα

NK, T and B cells

TNFR1 TNFR2

Most normal and transformed cells Endothelial cells and immune cells

TNF

Macrophages, NK, T and B cells

TNFR1, TNFR2

See above

LTβ

DCs, macrophages, T, B and NK cells

LTβR

NK cells, CD4+and CD8+ T cells

CD95L

Activated splenocytes, thymocytes and nonlymphoid tissues, such as the eye and testis

CD95

Most normal and transformed cells

DCR3

Lung and colon cells

TRAIL

NK cells, T cells and DCs

DR4, DR5 DCR1, DCR2, OPG

Most normal and transformed cells Most normal and transformed cells

TWEAK

Monocytes

FN14

Endothelial cells and fibroblasts

CD27L

NK, T and B cells

CD27

CD4+ and CD8+ T cells

CD30L

T cells and monocytes

CD30

Reed–Sternberg cells

CD40L

T and B cells

CD40

Reed–Sternberg cells

4-IBBL

B cells, DCs and macrophages

4-IBB

Activated T cells, monocytes and NK cells

OX40L

T and B cells

OX40

T cells

APRIL

Macrophages, lymphoid cells and tumour cells

BCMA TACI

B cells, PBLs, spleen, thymus, lymph nodes, liver and adrenals B cells, PBLs, activated T cells, spleen, thymus and small intestine

BAFF

T cells, DCs, monocytes and macrophages

TACI BCMA BAFFR

See above See above B cells, PBLs, resting T cells, spleen and lymph nodes

LIGHT*

T cells, granulocytes, monocytes and DCs

HVEM LTβR

T cells Non-lymphoid haematopoietic cells and stromal cells

VEGI*

Endothelial cells

DR3, DCR3

Activated T cells

GITRL

N. D.

GITR

CD4+CD25+ T cells

RANKL

Activated T cells and osteoblasts

RANK OPG

Osteoclasts, osteoblasts and activated T cells Osteoclast precursors, endothelial cells and others

EDA1

Skin

EDAR

Ectodermal derivative

EDA2

Skin

XEDAR

Ectodermal derivative

N. D.

DR6

Resting T cells

N. D.

RELT

Lymphoid tissues

N. D.

TROY

Embryo skin, epithelium, hair follicles and brain

*Binds to DCR3/TR6. VEGI is also known as TNF superfamily ligand (TL1) and TL1A is a longer variant of TL1; TROY is also known as TAJ (toxicity and JNK inducer). APRIL, a proliferation-inducing ligand; BAFF; B-cell-activating factor; BCMA, B-cell maturation antigen; DC, dendritic cells; DCR, decoy receptor; DR, death receptor; EDA, ectodermal dysplasin; EDAR, ectodyplasin-A-receptor; GITR, glucocorticoid-induced TNFR family receptor; L, ligand; LT, lymphotoxin; N. D., not determined; NK, natural killer; OPG, osteoprotegerin; PBL , peripheral-blood lymphocytes; R, receptor; RANK, receptor activator of nuclear factor-κB; RELT, receptor expressed in lymphoid tissues; TACI, transmembrane activator and cyclophilin ligand interactor; TNF, tumour-necrosis factor; TRAIL, TNF-related apoptosis-inducing ligand; TWEAK, TNF-like weak inducer of apoptosis; VEGI, vascular endothelial cell-growth inhibitor; XEDAR, X-linked ectodermal dysplasia receptor.

748

| SEPTEMBER 2003 | VOLUME 3

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group

REVIEWS

CD95L

TRAIL

TRAIL

CD95

DR4

DR5

DR6

TNF

VEGI

TNF

TNFR1

DR3

TNFR2

Others

TRADD

TRAF2 TRAF6

GRB2 MEKK1 FADD

SOS

IRAK1 MKK7

RIP

RAS Caspase-8

IKK RAF

JNK

MKK3

AP1

p38MAPK

Caspase-3 activation p42/44/MAPK

Apoptosis

NF-κB

Proliferation

Figure 2 | Cellular signalling pathways leading to activation of the main cellular responses by members of the TNF superfamily. Tumour-necrosis factor receptor 1 (TNFR1) binds to TNFR-associated death domain (TRADD) protein, factor associated with neutral sphingomyelinase (FAN), TNFR-associated protein 1 (TRAP1), TRAP2, mitogen-activated protein kinase (MAPK) activating death domain (MADD) and growth-factor receptor-bound protein 2 (GRB2)54,55. Through its death domain, TNFR1 sequentially recruits TRADD, TNFR-associated factor 2 (TRAF2), receptor-interacting protein (RIP) and inhibitor of nuclear factor-κB (NF-κB) kinase (IKK), leading to the activation of NF-κB; and the recruitment of TRADD, FAS-associated death domain (FADD) and caspase-8, leads to the activation of caspase-3, which in turn induces apoptosis56. A recent report indicated that FADD and caspase-8, although required, were not recruited to the TNFR1 signalling complex during TNF-induced apoptosis57. CD95, death receptor 4 (DR4) and DR5 activate apoptosis through sequential binding of FADD to caspase-8 and caspase-10 (REFS 45,58). Whether DR4/DR5 recruit FADD through TRADD is controversial. Earlier reports indicate that both DR4 and DR5 bind to TRADD44,59, but recent reports indicate that DR4/DR5 bind to FADD independently of TRADD45,57. The role of FADD in TNF-related apoptosis-inducing ligand (TRAIL) signalling is also controversial. Although an earlier report indicated that TRAIL-induced apoptosis is FADD independent59, later reports indicated that FADD is required45,57. Besides TNFR1, DR3 and DR6 are also known to bind TRADD and mediate the activation of apoptosis, NF-κB and JUN N-terminal kinase (JNK)28,60. TNFR1 activates JNK through the sequential recruitment of TRAF2, MAP/ERK kinase kinase 1 (MEKK1) and MAPK kinase 7 (MKK7). The activation of p38 MAPK by TNFR1 is less well understood and involves the recruitment of RIP by TRADD, which in turn recruits MKK3 leading to p38 MAPK activation54–56. AP1, activator protein 1; IRAK1, interleukin-1-receptor-associated kinase 1; VEGI, vascular endothelial cell-growth inhibitor.

two-hybrid system for protein−protein interaction, co-precipitation and dominant-negative approaches, several proteins that mediate TNF signalling have been identified (FIG. 2). Almost all of the members of the TNFR superfamily signal by binding to one or more TNFR-associated factors (TRAFs)61 (FIG. 3 and TABLE 2). The interaction of TRAF1, TRAF2 and TRAF5 with various cytoplasmic domains of TNFR-superfamily members requires a specific motif in the receptor (that is, Pro-Xaa-Gln-XaaThr). By contrast, TRAF6 uses a distinct motif (Gln-XaaPro-Xaa-Glu), which has been identified in CD40 and RANK62,63. However, of the known TRAFs, only TRAF2, TRAF5 and TRAF6 have been shown to mediate the activation of NF-κB and JNK. TRAF2 is known to bind to almost all of the members of the TNFR superfamily, it binds to TNFR1 through the TNFR-associated death

NATURE REVIEWS | IMMUNOLOGY

domain (TRADD) protein. Interestingly, it was recently found that the binding of TNF to TNFR2 induces inhibitor of apoptosis 1 (IAP1)-mediated ubiquitination and proteasomal degradation of TRAF2 (REF. 64). TNFR2induced degradation of TRAF2 provides further evidence of crosstalk between the two receptors. Most members of the TNF superfamily activate NF-κB through ubiquitin-mediated degradation of its inhibitor, IκBα. This process is initiated by phosphorylation of IκBα by the IκBα kinase (IKK) complex, mainly by the IKKβ catalytic subunit, and requires the regulatory subunit IKKγ (also known as NEMO). BAFF, however, has been shown to activate NF-κB by ubiquitin-mediated processing of the NF-κB2 inhibitory protein p100, which depends on the phosphorylation of p100 by IKKα. For BAFF to activate this second pathway BAFFR, NF-κBinducing kinase (NIK) and protein synthesis, but not

VOLUME 3 | SEPTEMBER 2003 | 7 4 9

© 2003 Nature Publishing Group

REVIEWS NEMO65,66 are required. This NEMO-independent cascade is physiologically relevant for the survival and maturation of splenic B cells. Similar to BAFF, LTβ receptor (LTβR) has also been shown to promote the NF-κB2 processing pathway67. IKKα has also been found to be crucial for RANK-induced NF-κB activation68. Recent studies indicate that after TNF binding, TNFR1 translocates to cholesterol- and sphingolipidenriched membrane microdomains, known as lipid rafts, in which it associates with the adaptor proteins TRADD and TRAF2, and the Ser/Thr kinase receptor-interacting protein (RIP), forming a signalling complex69 (FIG. 2). In lipid rafts, TNFR1 and RIP are ubiquitylated, which leads to their degradation by the proteasome pathway. Interfering with lipid-raft organization not only abolishes ubiquitylation, but also switches TNF signalling from NF-κB activation to apoptosis, indicating that lipid rafts are crucial for the outcome of TNF-activated signalling pathways. Refinement of cell signalling by gene deletion. As indicated earlier, the pathways through which various members of the TNFRs signal have been elucidated by investigations into protein−protein interactions, and overexpression and dominant-negative studies. These techniques, however, can produce misleading data. Gene-deletion studies have provided a different picture of the mechanism of signalling by the TNFRs. For example, deletion of the TRAF2 gene had a minimal effect on TNF-induced NF-κB activation, but resulted in defective JNK signalling70, indicating that TRAF2 is dispensable for NF-κB activation, but not for JNK activation.

TRAF1

416

Ring finger

Zinc finger

TRAF-N

TRAF-C

TRAF2

501

TRAF3

567

TRAF4

470

TRAF5

557

TRAF6

511

Figure 3 | The structures of different TNF receptor-associated factors. The tumour-necrosis factor receptor (TNFR)-associated factor (TRAF) family consists of six distinct proteins, each containing a ring finger domain (except TRAF1), a zinc-finger motif, an amino-terminal TRAF domain and a carboxy-terminal TRAF domain. TRAFs indirectly bind to the death domain of the receptors. TRAF2, TRAF5 and TRAF6 function as adaptor proteins that link the cell-surface receptors and downstream kinase cascades, which results in the activation of nuclear factor-κB (NF-κB) and activator protein 1 (AP1). Neither TRAF1 nor TRAF3 activates any kinases. The C-terminal half of the TRAF domain (TRAF-C) is highly conserved among members of the TRAF domain and it mediates homo- and heterodimerization of the TRAFs and association of TRAFs with various cell-surface receptors. The N-terminal portion of the TRAF domain (TRAF-N), predicted to form a coiled-coil structure, is less conserved. TRAF2, TRAF3, TRAF5 and TRAF6 contain five zinc-finger repeats, whereas TRAF4 has seven repeats. Apart from TRAF4, all of the TRAFs were identified by yeast two-hybrid screening using a cytoplasmic domain of various members of the TNFR family. So far, TRAF4 has no known function in the signalling of TNF or its family members.

750

| SEPTEMBER 2003 | VOLUME 3

However, TNF-induced NF-κB activation was markedly impaired in mouse embryo fibroblasts from TRAF2, TRAF5-double-knockout mice, and these cells were more susceptible to TNF-induced cytotoxicity71. These results indicated that both TRAF2 and TRAF5 are involved in TNF-induced NF-κB activation and protection from cell death. Deletion of the gene encoding RIP, however, abolished TNF-induced activation of NF-κB, but had minimal effect on JNK activation72. So RIP, which previously was not considered to be part of the TNF signalling mechanism, is now known to be a central mediator. Similarily, NIK was previously thought to be crucial for TNF-induced NF-κB activation73, but was found by gene-deletion experiments to be dispensable74. Instead, NIK was found to be required for NF-κB2 activation by LTβR and BAFFR74. Although not previously recognized, gene-deletion studies have also indicated an essential role for MAP/ERK kinase kinase 3 (MEKK3) in TNF-induced NF-κB activation75. So, our understanding of how various members of the TNFR superfamily transduce their signals is becoming increasingly refined. The fact that inhibitors of protein synthesis can enhance the cytotoxic effects of TNF and also make resistant cells sensitive to TNF has been well documented12,76. The mechanism for this, however, was unveiled only recently. Work in the past five years has indicated that TNF simultaneously activates both apoptotic and antiapoptotic or cell-survival signals (FIG. 4). The apoptotic signals do not require active protein synthesis, whereas anti-apoptotic signals do, and the synthesis is mediated through the activation of NF-κB77. It is the balance between the pro-apoptotic and anti-apoptotic signals that eventually determines whether cells will undergo apoptosis, survive or proliferate in response to TNF or other family members. Cytokines such as TRAIL and CD95 that are poor activators of NF-κB induce apoptosis more rapidly than TNF does; whereas cytokines such as RANKL mainly provide survival signals through the activation of NF-κB. Interestingly, however, NF-κB does not just regulate anti-apoptotic signals, it also regulates proapoptotic signals through the regulation of death receptors (DR1, DR2, DR4, DR5 and DR6)77−80, and death receptor ligands such as CD95L, TNF and TRAIL77,80. The pro-apoptotic effects of NF-κB could also be mediated through the upregulation of expression of p53 and c-MYC81,82. c-MYC, however, can sensitize cells to TNFmediated apoptosis by inhibiting NF-κB transactivation77. This adds additional complexity to the role of NF-κB in the regulation of apoptosis. Biological roles of the TNF-superfamily members. TNF and its family members represent a double-edged sword. Whereas physiologically they are important cytokines and required for normal responses, their inappropriate expression is harmful. Research during the past two decades has shown that TNF and its superfamily members have both beneficial and harmful activities. This is not unusual for most therapeutics (FIG. 5). First, I discuss the ‘bright side’ of TNF-superfamily members and then the ‘dark side’. Several genetic diseases are linked to the TNF/TNFR superfamily, which are discussed later.

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group

REVIEWS

Table 2 | Interaction of TRAFs with the TNFR-superfamily members Receptor

TRAF1

TRAF2

TRAF3

TRAF5

TNFR2

+

+





TRAF6 –

LTβR

+

+

+

+

N. D.

CD40

+

+

+

+

+

CD30

+

+

+

+

N. D.

CD27

N. D.

+



+



OX40



+

+



N. D.

4-1BB

+

+





N. D.

GITR

+

+

+





HVEM

+

+

+

+



RANK

+

+

+

+

+

TACI



+



+

+

BCMA

+

+

+

+

+

TROY



+



+

+

DR6

+









XEDAR

+









FN14

+

+

+

+



BAFFR





+





NGFR









+

BAFFR, B-cell-activating factor receptor; BCMA, B-cell maturation antigen; DR6, death receptor 6; GITR, glucocorticoid-induced TNFR family receptor; HVEM, herpes-virus entry mediator; LTβR, lymphotoxin-β receptor; N. D., not determined; NGFR, nerve growth-factor receptor; RANK, receptor activator of nuclear factor-κB; TACI, transmembrane activator and cyclophilin ligand interactor; TNFR, tumour-necrosis factor receptor; TRAF, TNFR-associated factor; XEDAR, X-linked ectodermal dysplasia receptor.

The bright side of the TNF superfamily

SYSTEMIC LUPUS ERYTHEMATOSUS

(SLE). A disease of unknown origin in which tissues and cells are damaged by the deposition of pathogenic antibodies and immune complexes. Patients generally have abnormal B- and T-cell functions. SJOGREN’S SYNDROME

(SS). An autoimmune disease characterized by diffuse lymphoid-cell infiltrates in the salivary and lacrimal glands, resulting in dry eyes and mouth due to insufficient secretion. GRAFT-VERSUS-HOST DISEASE

(GVHD). Tissue damage in a recipient of allogeneic transplanted tissue (usually a bone-marrow transplant) that results from the activity of donor cytotoxic T cells that recognize the recipient’s tissue as foreign. GVHD varies markedly in severity, but can be life threatening in severe cases. Typically, damage to the skin and gut mucosa leads to clinical manifestations.

Anticancer potential. Although initially thought to be a potent anticancer agent, it is now generally believed that TNF has limited activity in the suppression of cancer, mainly because of its systemic toxicity83. If this systemic toxicity could be suppressed, the therapeutic usefulness of TNF might increase. For example, recent studies showed that the inhibition of matrix metalloproteinases blocks the lethal hepatitis and apoptosis induced by TNF and allows safe antitumor therapy84. Also, TNF is a successful treatment for locally advanced limb softtissue sarcomas and other large tumours when administered by isolated limb perfusion — this approach can avoid the need for amputation85. TNF was approved in Europe after a multicentre trial in patients with locally advanced soft-tissue sarcomas considered to be unresectable by an independent review committee; the response rate to isolated limb perfusion with TNF plus melphalan was 76% and the limb was saved in 71% of the patients. Moreover, the trial showed the efficacy of isolated limb perfusion of TNF and melphalan against various other limb-threatening tumours, such as skin cancers and drug-resistant bony sarcomas. Laboratory models of isolated limb perfusion have helped to elucidate the mechanisms of action and to develop new treatment methods. These studies have identified TNF-mediated vasculotoxic effects on the tumour vasculature and have shown that addition of TNF to the perfusate results in an increase of three to six times in uptake of the vasoactive drugs melphalan or doxorubicin by tumours. Moreover, isolated limb perfusion is

NATURE REVIEWS | IMMUNOLOGY

an effective procedure for gene therapy mediated by an adenovirus vector. Various clinical phase I and phase II studies can be expected in the next few years. A multicenter, random-assignment trial of isolated limb perfusion that compares melphalan alone and melphalan with TNF in patients with melanoma, is also being carried out by the National Cancer Institute in the United States86. Alternative approaches to target TNF involve the use of agonistic antibodies specific for TNFR1 or gene therapy using TNFR1, which can also induce apoptosis of tumour cells when overexpressed. In contrast to TNF, TRAIL has been found to specifically kill tumour cells without harming normal cells, and so it is being explored for the treatment of cancer. Animal studies investigating the anticancer potential of TRAIL have been promising46,87. In addition to TRAIL, agonistic antibodies specific for its receptors DR4 (REF. 88) and DR5 (REF. 89) are also being examined for their antitumour activities. The initial results with animal studies have been promising. The role of TRAIL and other members of the TNF superfamily in cancer immunotherapy has also been documented90,91. Regulation of the immune system. The immune system is regulated not only by cell proliferation and differentiation, but also by apoptosis. Loss-of-function mutations in the signalling molecules that are involved in apoptosis can cause hyperproliferation of cells. By contrast, amplification of the death cascade can cause the destruction of various tissues. Ligands of the TNF superfamily control and orchestrate the immune response at several levels19. TNF, LTα, LTβ and RANKL provide crucial signals for the morphogenesis of secondary lymphoid organs. Pro-apoptotic members, such as TNF, CD95L and TRAIL, contribute to the function of cytotoxic effector cells in the recognition and destruction of virus-infected cells. CD95−CD95L interactions are important for regulating the immune system in several ways. Activated T cells express CD95 and CD95L, and they are sensitive to CD95-induced apoptosis, indicating that the activated T cells undergo suicide or kill each other (fratricide) to downregulate the immune response92. CD95-induced apoptosis might also be involved in clonal deletion of thymocytes that are highly reactive to antigens expressed in the thymus93. Several autoimmune diseases have been identified in which CD95− CD95L interactions are non-functional94. Elevated serum levels of CD95L have been seen in patients with natural killer (NK) cell large granular lymphoma leukaemia, SYSTEMIC LUPUS ERYTHEMATOSUS (SLE), rheumatoid arthritis, SJOGREN’S SYNDROME, lymphohistocytosis, myocarditis and acute GRAFT-VERSUS-HOST DISEASE58. Studies of the naturally occurring CD95-mutant mice (lpr, lymphoproliferation) and CD95L-mutant mice (gld, generalized lymphoproliferative disease) have enhanced our understanding of the role of these TNF-superfamily members in the regulation of the immune system. Mature T cells from lpr or gld mice do not die after activation. In addition to T-cell defects, CD95-deficient mice accumulate B cells and have

VOLUME 3 | SEPTEMBER 2003 | 7 5 1

© 2003 Nature Publishing Group

REVIEWS

IMMUNE-PRIVILEGED SITE

Immune-privileged sites are areas in the body with a decreased immune response to foreign antigens, including tissue grafts. These sites include the brain, eye, testis and uterus.

elevated levels of immunoglobulin of various classes95. It has been shown that CD95L-expressing T cells kill CD95-expressing activated B cells. The eye is an IMMUNEPRIVILEGED SITE, which ensures that inflammation does not destroy the tissue. This is mediated through the expression of CD95L by the cornea cells, which kills the inflammatory cells and is, therefore, responsible for protection of the eye from inflammation96. TRAIL has also been shown to have an important role in regulation of the immune system. TRAIL can regulate lymphocyte functions in the periphery by inhibiting cell-cycle progression by T cells97. Apoptosis of thymocytes is defective in mice that are deficient for TRAIL, such that thymic deletion induced by T-cell receptor ligation is markedly impaired. TRAIL-deficient mice are also hypersensitive to collagen-induced arthritis and streptozotocin-induced diabetes, and develop heightened autoimmune responses98,99. So, TRAIL mediates thymocyte apoptosis and is important in the induction of autoimmune diseases99. The potential role of TRAIL and other members of the TNF superfamily in organ transplantation have also been documented90.

Haematopoiesis. Almost all of the TNF ligands are produced by haematopoietic cells (TABLE 1) and interactions mediated by these ligands between haematopoietic cells, such as T cells−B cells, T cells−monocytes and T cells− T cells have a crucial role in regulating their proliferation. For example, CD95L is involved in peripheral T-cell homeostasis; 4-1BBL preferentially co-stimulates CD8+ T cells; CD27 co-stimulates T cells; CD30L is important in processes mediated by T helper 2 (TH2) cells; CD40L is essential for T-cell-dependent antibody responses and OX40 co-stimulates T cells. All B cells also express receptors for CD30, CD40, TNF, LT, APRIL, BAFF and CD27L. Three different receptors for BAFF are expressed by B cells and control the maturation and survival of these cells. B-cell development is completely blocked in mice that are deficient for BAFF or BAFFR. Additionally, both TNF and LT have been implicated in B-cell proliferation90,91. BCMA deficiency does not lead to any severe B-cell phenotypes, indicating that BCMA is dispensable. TACI deficiency resulted in increased numbers of B cells, but the ratio of mature B-cell subsets in the spleen was normal. The activation of immune precursor cells to fully competent effector cells also depends on CD40L for B cells; 4-1BBL, OX40L and CD27L for T cells; and CD40 and RANKL for dendritic cells.

Balance between life and death

Protection against microbial infection. An effective host response against infection with bacteria depends partly on the ability to produce an appropriate TH1-type cytokine profile, but TNFR signalling is also involved. It has been shown that deletion of either TNF or TNFR1 in mice leads to increased susceptibility to Listeria monocytogenes, rapid death from infection and resistance to LPS-mediated septic shock, indicating a crucial role of this cytokine in protection from infection. Antibodies specific for TNF (such as infliximab) and soluble TNFRs (such as enbrel) have been approved for human use in the treatment of Crohn’s disease and rheumatoid arthritis, respectively. Anti-TNF therapies, however, have considerably increased the risk of development of certain diseases, such as tuberculosis100. This is not surprising considering the important role of TNF in killing Mycobacterium tuberculosis by macrophages.

TNF superfamily ligands (TNF, CD95L,TRAIL and VEGI) Cycloheximide/ actinomycin D

TNFR FADD Caspase-8 Other caspases Die (apoptosis)

The dark side of the TNF superfamily

NF-κB TRAF1 XIAP

DCR BCL-XL

FLIP

Survivin

IAP

Live (Survive/proliferate)

Figure 4 | The balance between life and death mediated by TNF-superfamily members . Apoptotic signals do not require active protein synthesis, whereas anti-apoptotic signals do and the synthesis is mediated through the activation of nuclear factor-κB (NF-κB)77. Several NF-κB-regulated proteins that can suppress apoptosis have been identified. These include tumour-necrosis factor receptor (TNFR)-associated factor 1 (TRAF1), inhibitor of apoptosis 1 (IAP1) and IAP2, survivin, FAS-associated death domain-like interleukin-1 converting enzyme inhibitory protein (FLIP), X-chromosome linked inhibitor of apoptosis protein (XIAP), decoy receptor (DCR) and BCL-XL. These proteins inhibit different steps in the apoptotic pathways. For example, DCR sequesters the receptors, whereas FLIP blocks the activation of caspase-8, survivin inhibits caspase-3 activation, IAP suppresses caspase-9 activation, XIAP inhibits caspase-3 activation, and BCL-XL prevents the release of cytochrome-c from the mitochondria and so blocks caspase-9 activation.

752

| SEPTEMBER 2003 | VOLUME 3

Several of the harmful activities of TNF and its superfamily members are thought to be mediated through the activation of NF-κB. All members of the TNF ligand superfamily can activate NF-κB, although some members are more potent and ubiquitous than others. So far, over 200 genes have been shown to be regulated by NF-κB, and their products have been implicated in a wide variety of diseases. In general, the pathology that is linked to TNF and its family members is similar to that associated with inappropriate NF-κB activation. Cancer. As mentioned earlier, TNF was discovered as a cytokine that could kill tumour cells, however, it is now clear that TNF can contribute to tumorigenesis by mediating the proliferation, invasion and metastasis of tumour cells. TNF has been shown to be an autocrine

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group

REVIEWS

Haematopoiesis

Protection from bacterial infection

Innate immunity

Tumour regression

Immune surveillance

AIDS Systemic lupus erythematosus

Crohn’s disease

Osteoporosis/ Bone resorption

Alzheimer’s disease Multiple sclerosis

Septic shock

TNF

Lymphoproliferative diseases

Transplant rejection

Pulmonary fibrosis

Diabetes (type II)

Tumorigenesis Rheumatoid arthritis

Tumour metastasis

Heart failure Atherosclerosis

Liver disease

Allergic asthma

Fever

Figure 5 | The main physiological and pathological effects linked to members of the TNF superfamily. Different studies indicate that although tumour-necrosis factor (TNF) and its superfamily members are essential for haematopoiesis, protection from bacterial infection, immune surveillance and tumour regression (indicated in green); its dysregulation leads to various diseases (indicated in blue). This type of role as a ‘double-edged sword’ is typical of most members of the TNF superfamily.

growth factor for a wide variety of tumours. Through the activation of NF-κB, TNF induces the expression of various genes that are involved in invasion and metastasis, including adhesion molecules, urokinase plasminogen activator (UPA), matrix metalloproteinase 9 (MMP9), cyclo-oxygenase 2 (COX2) and vascular endothelial growth factor (VEGF). In addition, activation of NF-κB can suppress apoptosis, which further contributes to tumorigenesis. Several reports indicate that TNF is highly carcinogenic and mice deficient for TNF are resistant to skin carcinogenesis101. However, by contrast, a recent report showed that NF-κB blockade can trigger invasive epidermal neoplasia102.

OSTEOPETROSIS

A hereditary bone disease with intense positive balance of body calcium. Autosomal recessive osteopetrosis is a rare, fatal disease characterized by the accumulation of excessive bone mass due to defective bone resorption. The pathogenesis of osteopetrosis is controversial. Defects in osteoblast−osteoclast interactions, incorrect differentiation of osteoclasts, abnormal contact between osteoclasts and the extracellular matrix, and abolished signalling can occur in this disease. OSTEOPOROSIS

A condition that involves loss of bone due to an increase in the number of osteoclasts.

Autoimmunity. Several TNF ligands have been implicated in the development of autoimmunity. For example, transgenic overexpression of BAFF results in many autoimmune symptoms in mice91. These mice have an increased number of mature B cells, splenomegaly, DNA-specific antibodies, proteinuria and glomerulonephritis91. These phenotypes mimic those of SLE-like symptoms in humans. So, BAFF inhibitors might prove useful in regulating autoimmunity. It is also now becoming apparent that TNF has an important role in the pathogenesis of type II diabetes mellitus. TNF has been shown to interfere with an insulin-signalling mechanism by inhibiting the tyrosine kinase activities of the insulin receptor and serine phosphorylation of the insulin receptor substrate 1 (REF. 103). These inhibitory effects of TNF on insulin signalling seem to be mediated through the upregulation of expression of suppressor of cytokine signalling 3 (SOCS3)104. The role of TNF has also been explored using mice in which the TNF or TNFR gene has been

NATURE REVIEWS | IMMUNOLOGY

deleted105. The absence of TNF resulted in markedly improved insulin sensitivity in diet-induced obesity and in the ob/ob model of obesity. The TNF-deficient obese mice had lower levels of circulating free fatty acids, which are associated with the development of insulin resistance, and were protected from the obesity-related reduction in insulin receptor signalling in muscle and fat. This indicates that TNF is an important mediator of insulin resistance in obesity through its effects on several important sites of insulin action. Other diseases. Besides the diseases indicated earlier, ligands of the TNF superfamily have been linked with various other diseases including chronic heart failure, bone resorption, AIDS, Alzheimer’s disease, transplant rejection, atherosolerosis and hepatotoxicity. The development of chronic heart failure involves phenotypic changes in several homeostatic systems so that, as the disease advances, chronic heart failure might be seen as a multi-system disorder with its origins in the heart extending to many extra-cardiac manifestations. Higher levels of TNF are found in the blood and myocardium of patients with chronic heart failure than in controls, and TNF has been implicated in several of the pathophysiological processes that are thought to be important in the progression of chronic heart failure106,107. Therapies directed against this cytokine therefore provide an approach to control heart failure. TNFR fusion proteins have been developed that target circulating TNF itself. Both TNF and RANKL have been shown to have a crucial role in bone destruction108. These ligands induce osteoclastic differentiation of macrophages and myeloid progenitor cells. OPG protects bone integrity by downregulating osteoclastogenesis, therefore supporting the hypothesis that RANKL mediates bone loss. Further evidence comes from gene-deletion studies, which showed marked OSTEOPETROSIS after the deletion of either RANK or RANKL109,110. Conversely, the deletion of OPG leads to OSTEOPOROSIS111. Several of these activities are mediated by TNF ligand-induced activation of NF-κB. Genetic mutations in the TNF/TNFR superfamily

Mutations in both ligands and receptors of the TNF superfamily have been found in humans. These include mutations in TNFR1, CD95/CD95L, RANK, ectodermal dysplasin (EDA) and CD40L. Missense mutations of TNFR1 lead to autosomal dominant periodic fever syndromes that are characterized by unexplained episodes of fever and severe localized inflammation37. The mutation disrupts the conserved extracellular disulphide bonds, causing a decrease in the level of soluble TNFR1 in the plasma of patients. Leukocytes from individuals with a Cys52Phe mutation in TNFR1 express increased levels of membrane TNFR1 and showed reduced receptor cleavage after stimulation. The autoinflammatory phenotype results from impaired downregulation of membrane TNFR1 and diminished shedding of potentially antagonistic soluble receptor. The existence of TNFR1-associated periodic syndromes (TRAPS) establishes an important class of TNFR mutations112.

VOLUME 3 | SEPTEMBER 2003 | 7 5 3

© 2003 Nature Publishing Group

REVIEWS

TNF Melanoma

Fc–TNFR2 (Enbrel)

TRAIL Anti-DR4 Anti-DR5

Rheumatoid arthritis

Cancer TNF/TNFR superfamily

Anti-BAFF Autoimmune diseases

Anti-TNF antibodies (Infliximab)

OPG Inhibitors of TNF production (Thalidomide)

Crohn’s disease

Osteoporosis

Leprosy

Figure 6 | Present and future therapeutics based on members of the TNF superfamily and their receptors. Based on tumour-necrosis factor (TNF) and its receptors, several therapeutic agents have been approved for the treatment of inflammatory diseases (such as rheumatoid arthritis and Crohn’s disease), cancer (melanoma) and others (leprosy, autoimmunity and osteoporosis). Therapeutics based on TNF-related apoptosis-inducing ligand (TRAIL) and its receptors are presently being tested for the treatment of cancer. BAFF, B-cell-activating factor; DR, death receptor; OPG, osteoprotegerin; R, receptor.

Individuals with

CANALE−SMITH SYNDROME

or

AUTO-

IMMUNE LYMPHOPROLIFERATIVE SYNDROME (ALPS) have pheno-

CANALE-SMITH SYNDROME

(CSS). An inherited disease characterized by massive lymphadenopathy, hepatosplenomegaly and systemic autoimmunity to erythrocytes and platelets. AUTOIMMUNE LYMPHOPROLIFERATIVE SYNDROME

(ALPS). ALPS is characterized clinically by chronic nonmalignant lymphoproliferation and autoimmunity, and is caused by a genetic defect in apoptosis. Most patients with ALPS have heterozygous mutations in the CD95 gene. OSTEOBLASTS

Cells that are responsible for the formation of bone. OSTEOCLASTS

Cells that are responsible for bone resorption. They are rare cells with only 2−3 cells seen per 1 mm3 of bone. However, the loss of function in osteoclasts, problems with their differentiation and decrease in their number lead to bone osteosclerosis/osteopetrosis. Conversely, an increase in their number or function induces bone osteoporosis, indicating that osteoclasts have a pivotal role in bone homeostasis.

754

types that are similar to those of lpr/gld mice, which have mutations in CD95 and CD95L, respectively. This is a good example of a case in which the identification of a mouse mutation has led to the understanding of a human disease113. These patients have an increased risk of developing T- and B-cell lymphomas114. Their risk of non-Hodgkin’s lymphoma and Hodgkin’s disease, respectively, was 14 and 51 times greater than expected, due to defective lymphocyte apoptosis, implicating that CD95-mediated apoptosis is important for preventing B-cell and T-cell lymphomas. CD95 mediates apoptosis through the activation of caspase-8. Human family members with an inherited genetic deficiency of caspase-8 have been described115. Homozygous individuals have defective lymphocyte apoptosis and homeostasis but, unlike individuals with ALPS, they also have defects in the activation of T cells, B cells and NK cells, which leads to immunodeficiency. So, caspase-8 deficiency in humans is compatible with normal development and shows that caspase-8 has a postnatal role in immune activation of naive lymphocytes. Interactions between the B-cell surface antigen CD40 and its ligand expressed by activated T cells have a crucial role in isotype switching. This is illustrated by the failure of isotype switching in patients with X-linked hyper-IgM syndrome in who the CD40L gene is mutated116. Familial expansile osteolysis is another rare, autosomal dominant bone disorder that is characterized by focal areas of increased bone remodelling. The osteolytic lesions, which usually develop in the long bones during early adulthood, show increased activity of OSTEOBLASTS and OSTEOCLASTS. Two heterozygous insertion mutations in exon 1 of RANK were identified in patients with the disease. One was a duplication of 18 bases and the other

| SEPTEMBER 2003 | VOLUME 3

a duplication of 27 bases, both of which affected the signal-peptide region of RANK. Expression of recombinant forms of the mutant RANK proteins indicated disruptions to the expression levels and lack of normal cleavage of the signal peptide. Both mutations caused an increase in RANK-mediated NF-κB signalling in vitro, consistent with the presence of an activating mutation117. Hypohydrotic ectodermal dysplasia is a genetic disease seen in humans and mice, characterized by the loss of hair, sweat glands and teeth. The main X-linked form results from mutations in EDA — a TNF-like ligand. A phenotypically indistinguishable autosomal form of the disease results from mutations in the receptor for EDA (EDAR). EDAR signalling is mediated by the activation of NF-κB. EDA−EDAR signalling mediates cell interactions in the ectoderm and regulates the initiation and morphogenesis of hair and teeth. It is also required for the development of fish scales, indicating that this pathway and its function have been conserved during the evolution of ectodermal organs118. Additionally, mutations in death receptors have been found in various cancers. For example, mutations in CD95 have been found in thyroid lymphoma, prostatic intraepithelial neoplasia and testicular germ-cell tumours115−121. Nucleotide substitution in the ectodomain of DR4 has been associated with lung cancer, and head and neck cancer122. Therapeutic implications

From this description, it is clear that members of the TNF superfamily and their receptors have a wide-ranging role in many cellular and physiological functions. These roles have been translated into new generation therapies (FIG. 6). TNF has been approved, in Europe, for human use in the treatment of sarcomas and melanomas85,86. Antibodies specific for TNF have been approved for the treatment of Crohn’s disease123 and are also being exploited for the treatment of various cancers124. Anti-TNF therapy using soluble TNFR2 has been approved for rheumatoid arthritis125. Several other therapies, including treatment with TRAIL or agonistic antibodies specific for DR4 and DR5 are being tested for anticancer properties at present. Due to the important role of BAFF in various autoimmune diseases, inhibitors of this cytokine are also under development. Conclusion

Almost two decades ago, TNF was thought to be a single molecule, now it is clear that it belongs to a superfamily of cytokines that have a wide variety of activities. Although over 50,000 papers have been published on TNF since its initial isolation in 1984, there is little information about the other members of this superfamily. As all of the TNF-superfamily members, except LT, are transmembrane proteins, how these proteins are released from the cell surface remains to be elucidated. The ligands for receptors such as DR6, RELT and TROY remain to be identified. Future studies will indicate the differences in the signal transduction mechanisms between the various members of the TNF superfamily, which might allow for the development of more specific

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group

REVIEWS therapeutics. Several therapeutics based on TNF have been approved for use in humans, and more are likely to be found in the future based on other members of this family. Experience indicates that all of the

1. 2. 3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

21.

22.

23.

Bruns, P. Die Heilwirkung des Erysipels auf Geschwulste. Beitr. Klin. Chir. 3, 443–446 (1868). Coley, W. B. Contribution to the knowledge of Sarcoma. Ann. Surg. 14, 199–220 (1891). Gratia, A. & Linz, R. C. R. Seances Soc. Biol. Ses Fil. 108, 421–428 (1931). Shear, M. J. & Turner, F. C. Chemical treatment of tumors. V. Isolation of the hemorrhage-producing fraction from Serratia marcescens (Bacillus prodigiosus) culture filtrate. J. Natl Cancer Inst. 4, 81–97 (1943). Algire, G. H., Legallaies, F. Y. & Anderson, B. F., Vascular reactions of normal and malignant tissues in vivo. V. The role of hypotension in the action of a bacterial polysaccharide on tumors. J. Natl Cancer Inst. 12, 1279–1295 (1952). O’Malley, W. E., Achinstein, B. & Shear, M. J. Action of bacterial polysaccharide on tumors. II. Damage of Sarcoma 37 by serum of mice treated with Serratia marcescens polysaccharide, and induced tolerance, J. Natl Cancer Inst. 29, 1169–1175 (1962). Carswell, E. A. et al. An endotoxin-induced serum factor that causes necrosis of tumors. Proc. Natl Acad. Sci. USA 72, 3666–3670 (1975). Williamson, B. D. et al. Human tumor necrosis factor produced by human B-cell lines: synergistic cytotoxic interaction with human interferon. Proc. Natl Acad. Sci. USA 80, 5397–5401 (1983). Williams, T. W. et al. Lymphocyte in vitro cytotoxicity: lymphotoxins of several mammalian species. Nature 7, 1076–1077 (1968). Aggarwal, B. B. et al. Human lymphotoxin: production by a lymphoblastoid cell line, purification and initial characterization. J. Biol. Chem. 259, 686–691 (1984). This is the first report on the isolation and characterization of lymphotoxin (LT) or tumournecrosis factor-β (TNF-β). Aggarwal, B. B. et al. Primary structure of human lymphotoxin derived from 1788 lymphoblastoid cell line. J. Biol. Chem. 260, 2334–2344 (1985). Aggarwal, B. B. et al. Human tumor necrosis factor: production, purification and characterization. J. Biol. Chem. 260, 2345–2354 (1985). This is the first report on the isolation and characterization of TNF-α. Gray, P. W. et al. Cloning and expression of cDNA for human lymphotoxin, a lymphokine with tumor necrosis activity. Nature 312, 721–724 (1984). Pennica, D. et al. Human tumor necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 312, 724–729 (1984). Aggarwal, B. B. et al. Characterization of receptors for human tumor necrosis factor and their regulation by γ-interferon. Nature 318, 665–667 (1985). This is the first report to describe the characterization of TNF receptors (TNFRs) and indicates that LT and TNF have a common receptor. Beutler, B. et al. Identity of tumour necrosis factor and the macrophage-secreted factor cachectin. Nature 316, 552–554 (1985). Takeda, K. et al. Identity of differentiation inducing factor and tumour necrosis factor. Nature 323, 338–340 (1986). Smith, C. A. et al. The TNF receptor superfamily of cellular and viral proteins: activation, co–stimulation, and death. Cell 76, 959–962 (1994). Locksley, R. M. et al. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 104, 487–501 (2001). Yonehara, S. et al. A cell-killing monoclonal antibody (antiFas) to a cell surface antigen co-downregulated with the receptor of tumor necrosis factor. J. Exp. Med. 169, 1747–1756 (1989). Itoh, N. et al. The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 66, 233–243 (1991). Wiley, S. R. et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 3, 673–682 (1995). Pitti, R. M. et al. Induction of apoptosis by Apo-2 ligand, a new member of the tumor necrosis factor cytokine family. J. Biol. Chem. 271, 12687–12690 (1996).

TNF-superfamily members are double-edge swords and more studies are required to determine whether it is their agonist or antagonist feature that should be targeted for therapy.

24. Anderson, D. M. et al. A homologue of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell function. Nature 390, 175–179 (1997). 25. Wong, B. R. et al. TRANCE (tumor necrosis factor (TNF)related activation-induced cytokine), a new TNF family member predominantly expressed in T cells, is a dendritic cellspecific survival factor. J. Exp. Med. 186, 2075–2080 (1997). 26. Lacey, D. L. et al. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93, 165–176 (1998). 27. Haridas, V. et al. VEGI, a new member of the TNF family activates nuclear factor-κB and c-Jun N-terminal kinase and modulates cell growth. Oncogene 18, 6496–6504 (1999). 28. Migone, T. S. et al. TL1A is a TNF-like ligand for DR3 and TR6/DcR3 and functions as a T cell co-stimulator. Immunity 16, 479–492 (2002). 29. Mukhopadhyay, A. et al. Identification and characterization of a novel cytokine, THANK, a TNF homologue that activates apoptosis, nuclear factor-κB, and c-Jun NH2terminal kinase. J. Biol. Chem. 274, 15978–15981 (1999). 30. Schneider, P. et al. BAFF, a novel ligand of the tumor necrosis factor family, stimulates B cell growth. J. Exp. Med. 189, 1747–1756 (1999). 31. Nocentini, G. et al. A new member of the tumor necrosis factor/nerve growth factor receptor family inhibits T cell receptor-induced apoptosis. Proc. Natl Acad. Sci. USA 94, 6216–6221 (1997). 32. Black, R. A. et al. A metalloproteinase disintegrin that releases tumour-necrosis factor-α from cells. Nature 385, 729–733 (1997). 33. Lum, L. et al. Evidence for a role of a tumor necrosis factor-α (TNF-α)-converting enzyme–like protease in shedding of TRANCE, a TNF family member involved in osteoclastogenesis and dendritic cell survival. J. Biol. Chem. 274, 13613–13618 (1999). 34. Powell, W. C. et al. The metalloproteinase matrilysin proteolytically generates active soluble Fas ligand and potentiates epithelial cell apoptosis. Curr. Biol. 9, 1441–1447 (1999). 35. Chen, Y. et al. Mutations within a furin consensus sequence block proteolytic release of ectodysplasin-A and cause X-linked hypohidrotic ectodermal dysplasia. Proc. Natl Acad. Sci. USA 98, 7218–7223 (2001). 36. Suda, T. et al. Membrane Fas ligand kills human peripheral blood T lymphocytes, and soluble Fas ligand blocks the killing. J. Exp. Med. 186, 2045–2050 (1997). 37. McDermott, M. F. et al. Germline mutations in the extracellular domains of the 55 kDa TNF receptor, TNFR1, define a family of dominantly inherited autoinflammatory syndromes. Cell 97, 133–144 (1999). 38. Simonet, W. S. et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 89, 309–319 (1997). 39. Itoh, N. et al. A novel protein domain required for apoptosis. Mutational analysis of human Fas antigen. J. Biol. Chem. 268, 10932–10937 (1993). This report and reference 40 were the first to describe the presence of a death domain in TNFR and CD95. 40. Tartaglia, L. A. et al. A novel domain within the 55 kd TNF receptor signals cell death. Cell 74, 845–853 (1993). 41. Becraft, P. W. et al. CRINKLY4: a TNFR-like receptor kinase involved in maize epidermal differentiation. Science 273, 1406–1409 (1996). 42. Chan, F. K. et al. A domain in TNF receptors that mediates ligand-independent receptor assembly and signaling. Science 288, 2351–2354 (2000). This report refutes the previous theory proposing that the ligand induces oligomerization of the receptor. The authors identified a domain in the receptor that mediates ligand-independent assembly of the receptor. 43. Zhai, Y. et al. VEGI, a novel cytokine of the tumor necrosis factor family, is an angiogenesis inhibitor that suppresses the growth of colon carcinomas in vivo. FASEB J. 13, 181–189 (1999). 44. Schneider, P. et al. TRAIL receptors 1 (DR4) and 2 (DR5) signal FADD-dependent apoptosis and activate NF-κB. Immunity 7, 831–836 (1997). 45. Kischkel, F. C. et al. Apo2L/TRAIL-dependent recruitment of endogenous FADD and caspase-8 to death receptors 4 and 5. Immunity 12, 611–620 (2000).

NATURE REVIEWS | IMMUNOLOGY

46. Ashkenazi, A. et al. Targeting death and decoy receptors of the tumour-necrosis factor superfamily. Nature Rev. Cancer. 2, 420–430 (2002). 47. Kwon, B. S. et al. TR1, a new member of the tumor necrosis factor receptor superfamily, induces fibroblast proliferation and inhibits osteoclastogenesis and bone resorption. FASEB J. 12, 845–854 (1998). 48. Grell, M. et al. The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell 83, 793–802 (1995). This report shows that soluble TNF activates TNFR1 and membrane TNF activates TNFR2. 49. Weiss, T. et al. TNFR80-dependent enhancement of TNFR60-induced cell death is mediated by TNFRassociated factor 2 and is specific for TNFR60. J. Immunol. 161, 3136–3142 (1998). 50. Mukhopadhyay, A. et al. Genetic deletion of the tumor necrosis factor receptor p60 or p80 abrogates ligandmediated activation of nuclear factor-κB and of mitogenactivated protein kinases in macrophages. J. Biol. Chem. 276, 31906–31912 (2001). 51. Schneider, P. et al. TWEAK can induce cell death via endogenous TNF and TNF receptor 1. Eur. J. Immunol. 29, 1785–1792 (1999). 52. Grell, M. et al. Induction of cell death by tumour necrosis factor (TNF) receptor 2, CD40 and CD30: a role for TNF-R1 activation by endogenous membrane-anchored TNF. EMBO J. 18, 3034–3043 (1999). 53. Takada, Y. et al. Genetic deletion of the TNF receptor p60 or p80 sensitizes macrophages to lipopolysaccharide-induced nuclear factor-κB, mitogen-activated protein kinases and apoptosis. J. Biol. Chem. 278, 23390–23397 (2003). 54. Wajant, H. et al. Tumor necrosis factor signaling. Cell Death Differ. 10, 45–65 (2003). 55. Baud, V. et al. Signal transduction by tumor necrosis factor and its relatives. Trends Cell. Biol. 11, 372–377 (2001). 56. Chen, G. et al. TNF-R1 signaling: a beautiful pathway. Science 296, 1634–1635 (2002). 57. Harper, N. et al. Fas-associated death domain protein and caspase-8 are not recruited to the TNF-R1 signaling complex during TNF-induced apoptosis. J. Biol. Chem. 278, 25534–25541 (2003). 58. Nagata, S. et al. Fas ligand-induced apoptosis. Annu. Rev. Genet. 33, 29–55 (1999). 59. Pan, G. et al. The receptor for the cytotoxic ligand TRAIL. Science 276, 111–113 (1997). 60. Pan, G. et al. Identification and functional characterization of DR6, a novel death domain-containing TNF receptor. FEBS Lett. 431, 351–356 (1998). 61. Arch, R. H. et al. Tumor necrosis factor receptor-associated factors (TRAFs) — a family of adapter proteins that regulates life and death. Genes Dev. 12, 2821–2830 (1998). 62. Darnay, B. G. et al. Activation of NF-κB by RANK requires tumor necrosis factor receptor-associated factor (TRAF) 6 and NF-κB-inducing kinase. Identification of a novel TRAF6 interaction motif. J. Biol. Chem. 274, 7724–7731 (1999). 63. McWhirter, S. M. et al. Structural and biochemical analysis of signal transduction by the TRAF family of adapter proteins. Cold Spring Harb. Symp. Quant. Biol. 6, 551–562 (1999). 64. Li, X. et al. TNF-RII and c-IAP1 mediate ubiquitination and degradation of TRAF2. Nature 416, 345–347 (2002). 65. Claudio, E., Brown, K., Park, S., Wang, H. & Siebenlist, U. BAFF-induced NEMO-independent processing of NF-κB2 in maturing B cells. Nature Immunol. 3, 958–965 (2002). 66. Kayagaki, N. et al. BAFF/BLyS receptor 3 binds the B cell survival factor BAFF ligand through a discrete surface loop and promotes processing of NF-κB2. Immunity 17, 515–524 (2002). 67. Dejardin, E. et al. The lymphotoxin-β receptor induces different patterns of gene expression via two NF-κB pathways. Immunity 17, 525–535 (2002). 68. Cao, Y. et al. IKKα provides an essential link between RANK signaling and cyclin D1 expression during mammary gland development. Cell 107, 763–775 (2001). 69. Legler, D. F., Micheau, O., Doucey, M. A., Tschopp, J. & Bron, C. Recruitment of TNF receptor 1 to lipid rafts is essential for TNFα-mediated NF-κB activation. Immunity 18, 655–664 (2003).

VOLUME 3 | SEPTEMBER 2003 | 7 5 5

© 2003 Nature Publishing Group

REVIEWS

70.

71.

72.

73.

74.

75.

76.

77. 78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

756

This paper shows that after binding of TNF, TNFR1 translocates to cholesterol- and sphingolipidenriched membrane microdomains, known as lipid rafts, in which it associates with receptor-interacting protein (RIP), TNFR-associated death domain (TRADD) and TNFR-associated factor 2 (TRAF2), forming a signalling complex. Yeh, W. C. et al. Early lethality, functional NF-κB activation, and increased sensitivity to TNF-induced cell death in TRAF2-deficient mice. Immunity 7, 715–725 (1997). Tada, K. et al. Critical roles of TRAF2 and TRAF5 in tumor necrosis factor-induced NF–κB activation and protection from cell death. J. Biol. Chem. 276, 36530–36534 (2001). Although a dominant-negative mutant of TRAF2 blocks TNF-induced nuclear factor-κB (NF-κB) activation, the deletion of TRAF2 has no marked effect on TNF-induced NF-κB activation. Here, the authors show that cells from TRAF2- and TRAF5double-knockout mice are impaired in TNF-induced NF-κB activation. Kelliher, M. A. et al. The death domain kinase RIP mediates the TNF-induced NF-κB signal. Immunity 8, 297–303 (1998). Malinin, N. L., Boldin, M. P., Kovalenko, A. V. & Wallach, D. MAP3K-related kinase involved in NF-κB induction by TNF, CD95 and IL-1. Nature 385, 540–544 (1997). Yin, L. et al. Defective lymphotoxin-βreceptor-induced NF-κB transcriptional activity in NIK-deficient mice. Science 291, 2162–2165 (2001). Yang, J. et al. Mekk3 is essential for early embryonic cardiovascular development. Nature Genet. 24, 309–313 (2000). Sugarman, B. J. et al. Recombinant human tumor necrosis factor-α: effects on proliferation of normal and transformed cells in vitro. Science 230, 943–945 (1985). Karin, M. et al. NF-κB at the crossroads of life and death. Nature Immunol. 3, 221–227 (2002). Kasof, G. M. et al. Tumor necrosis factor-α induces the expression of DR6, a member of the TNF receptor family, through activation of NF-κB. Oncogene 20, 7965–7795 (2001). Zheng, Y. et al. NF-κB RelA (p65) is essential for TNF-αinduced Fas expression but dispensable for both TCRinduced expression and activation-induced cell death. J. Immunol. 166, 4949–4957 (2001). Ravi, R. et al. Regulation of death receptor expression and TRAIL/Apo2L-induced apoptosis by NF-κB. Nature Cell Biol. 3, 409–416 (2001). You, Z. et al. c-Myc sensitizes cells to tumor necrosis factormediated apoptosis by inhibiting nuclear factor-κB transactivation. J. Biol. Chem. 277, 36671–36677 (2002). Klefstrom, J. et al. Induction of TNF-sensitive cellular phenotype by c-Myc involves p53 and impaired NF-κB activation. EMBO J. 16, 7382–7392 (1997). Feinberg, B. et al. A phase I trial of intravenouslyadministered recombinant tumor necrosis factor-α in cancer patients. J. Clin. Oncol. 6, 1328–1334 (1988). Wielockx, B. et al. Inhibition of matrix metalloproteinases blocks lethal hepatitis and apoptosis induced by tumor necrosis factor and allows safe antitumor therapy. Nature Med. 7, 1202–1208 (2001). Eggermont, A. M., de Wilt, J. H. & ten Hagen, T. L. Current uses of isolated limb perfusion in the clinic and a model system for new strategies. Lancet Oncol. 4, 429–437 (2003). Lans, T. E. et al. Role of tumor necrosis factor on toxicity and cytokine production after isolated hepatic perfusion. Clin. Cancer Res. 7, 784–790 (2001). Kelley, S. K. et al. Preclinical studies to predict the disposition of Apo2L/tumor necrosis factor-related apoptosis-inducing ligand in humans: characterization of in vivo efficacy, pharmacokinetics, and safety. J. Pharmacol. Exp. Ther. 299, 31–38 (2001). Chuntharapai, A. et al. Isotype-dependent inhibition of tumor growth in vivo by monoclonal antibodies to death receptor 4. J. Immunol. 166, 4891–4898 (2001). Ichikawa, K. et al. Tumoricidal activity of a novel anti-human DR5 monoclonal antibody without hepatocyte cytotoxicity. Nature Med. 7, 954–960 (2001).

90. Adams, A. B. et al. The role of TNF receptor and TNF superfamily molecules in organ transplantation. Am. J. Transplant 2, 12–18 (2002). 91. Mckay, F. et al. BAFF AND APRIL: a tutorial on B cell survival. Annu. Rev. Immunol. 21, 231–264 (2003). 92. Nagata, S. et al. The Fas death factor. Science 267, 1449–1456 (1995). 93. Kishimoto, H. et al. A role for Fas in negative selection of thymocytes in vivo. J. Exp. Med. 187, 1427–1438 (1998). 94. Hayashi, T. & Faustman, D. L. Role of defective apoptosis in type 1 diabetes and other autoimmune diseases. Recent Prog. Horm. Res. 58, 131–153 (2003). 95. Cohen, P. L. et al. Lpr and gld: single gene models of systemic autoimmunity and lymphoproliferative disease. Annu. Rev. Immunol. 9, 243–269 (1991). 96. Griffith, T. S. et al. Fas ligand-induced apoptosis as a mechanism of immune privilege. Science 270, 1189–1192 (1995). 97. Song, K. et al. Tumor necrosis factor-related apoptosisinducing ligand (TRAIL) is an inhibitor of autoimmune inflammation and cell cycle progression. J. Exp. Med. 191, 1095–1104 (2000). 98. Cretney, E. et al. Increased susceptibility to tumor initiation and metastasis in TNF-related apoptosis-inducing liganddeficient mice. J. Immunol. 168, 1356–1361 (2002). 99. Lamhamedi-Cherradi, S. E. et al. Defective thymocyte apoptosis and accelerated autoimmune diseases in TRAIL–/– mice. Nature Immunol. 4, 255–260 (2003). 100. Gardam, M. A. et al. Anti-tumour necrosis factor agents and tuberculosis risk: mechanisms of action and clinical management. Lancet Infect. Dis. 3, 148–155 (2003). 101. Moore, R. J. et al. Mice deficient in tumor necrosis factor-α are resistant to skin carcinogenesis. Nature Med. 5, 828–831 (1999). Given the associations between chronic inflammation and epithelial cell cancer, the authors studied susceptibility to skin carcinogenesis in mice deficient for TNF. TNF–/– mice were resistant to the development of benign and malignant skin tumours, whether induced by initiation with DMBA (12dimethylbenz[a]-anthracene) and promotion with TPA (12-0-tetradecanoyl-phorbol-13-acetate) or by repeated dosing with DMBA. 102. Dajee, M. et al. NF-κB blockade and oncogenic Ras trigger invasive human epidermal neoplasia. Nature 421, 639–643 (2003). 103. Hotamisligil, G. S. et al. Free in PMC tumor necrosis factor-α inhibits signaling from the insulin receptor. Proc. Natl Acad. Sci. USA 91, 4854–4858 (1994). 104. Emanuelli, B. et al. SOCS-3 inhibits insulin signaling and is upregulated in response to tumor necrosis factor-α in the adipose tissue of obese mice. J. Biol. Chem. 276, 47944–47949 (2001). 105. Uysal, K. T. et al. Protection from obesity-induced insulin resistance in mice lacking TNF-α function. Nature 389, 610–614 (1997). The authors investigated the role of TNF in obesity and insulin resistance, by generating obese mice with a targeted null mutation in the gene encoding TNF and those encoding the two receptors for TNF. The absence of TNF improved insulin sensitivity in both diet-induced obesity and that resulting for the ob/ob model of obesity. 106. Bolger, A. P. et al. Tumour necrosis factor in chronic heart failure: a peripheral view on pathogenesis, clinical manifestations and therapeutic implications. Drugs 60, 1245–1257 (2000). 107. Diwan, A. et al. Inflammatory mediators and the failing heart: a translational approach. Curr. Mol. Med. 3, 161–182 (2003). 108. Romas, E. et al. Involvement of receptor activator of NF-κB ligand and tumor necrosis factor-α in bone destruction in rheumatoid arthritis. Bone 30, 340–346 (2002). 109. Kong, Y. Y. et al. OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymphnode organogenesis. Nature 397, 315–323 (1999). Bone remodelling and bone loss are controlled by a balance between osteoprotegerin ligand (OPGL) and OPG. In addition, OPGL regulates lymph-node organogenesis, lymphocyte development and

| SEPTEMBER 2003 | VOLUME 3

110.

111.

112.

113.

114.

115.

116.

117.

118. 119. 120.

121.

122.

123.

124.

125.

interactions between T cells and dendritic cells. Here, the authors report that activated T cells can directly trigger osteoclastogenesis through the expression of OPGL. Fata, J. E. et al. The osteoclast differentiation factor osteoprotegerin-ligand is essential for mammary gland development. Cell 103, 41–50 (2000). Bucay, N. et al. Osteoprotegerin-deficient mice develop early onset osteoporosis and arterial calcification. Genes Dev. 12, 1260–1268 (1998). Aganna, E. et al. Tumor necrosis factor receptor-associated periodic syndrome (TRAPS) in a Dutch family: evidence for a TNFRSF1A mutation with reduced penetrance. Eur. J. Hum. Genet. 9, 63–66 (2001). Nagata, S. et al. Human autoimmune lymphoproliferative syndrome, a defect in the apoptosis-inducing Fas receptor: a lesson from the mouse model. J. Hum. Genet. 43, 2–8 (1998). Straus, S. E. et al. The development of lymphomas in families with autoimmune lymphoproliferative syndrome with germline Fas mutations and defective lymphocyte apoptosis. Blood 98, 194–200 (2001). Chun, H. J. et al. Pleiotropic defects in lymphocyte activation caused by caspase-8 mutations lead to human immunodeficiency. Nature 419, 395–399 (2002). Ramesh, N. et al. CD40–CD40 ligand (CD40L) interactions and X-linked hyperIgM syndrome (HIGMX–1). Clin. Immunol. Immunopathol. 76, S208–S213 (1995). Hughes, A. E. et al. Mutations in TNFRSF11A, affecting the signal peptide of RANK, cause familial expansile osteolysis. Nature Genet. 24, 45–48 (2000). Thesleff, I. & Mikkola, M. L. Death receptor signaling giving life to ectodermal organs. Sci. STKE 131, PE22 (2002). Takakuwa, T. et al. Frequent mutations of Fas gene in thyroid lymphoma. Cancer Res. 61, 1382–1385 (2001). Takayama, H. et al. Fas gene mutations in prostatic intraepithelial neoplasia and concurrent carcinoma: analysis of laser capture microdissected specimens. Lab. Invest. 81, 283–288 (2001). Takayama, H. et al. Frequent Fas gene mutations in testicular germ cell tumors. Am. J. Pathol. 161, 635–641 (2002). Fisher, M. J. et al. Nucleotide substitution in the ectodomain of trail receptor DR4 is associated with lung cancer and head and neck cancer. Clin. Cancer Res. 7, 1688–1697 (2001). Suenaert, P. et al. Anti-tumor necrosis factor treatment restores the gut barrier in Crohn’s disease. Am. J. Gastroenterol. 97, 2000–2004 (2002). Couriel, D. R. et al. Role of tumor necrosis factor-α inhibition with inflixiMAB in cancer therapy and hematopoietic stem cell transplantation. Curr. Opin. Oncol. 12, 582–587 (2000). Feldmann, M. et al. Anti-TNFα therapy of rheumatoid arthritis: what have we learned? Annu. Rev. Immunol. 19, 163–196 (2001).

Acknowledgements This research was supported by The Clayton Foundation for Research, by the Department of Defense of the US Army Breast Cancer Research Program and by the National Institutes of Health. I would like to thank Y. Takeda, A. Bharti, U. Bhardwaj and S. Shishodia for assistance with the graphics, U. Gaur and L. Ford for help in preparation of the manuscript and W. Pagel for a careful review of the manuscript.

Online links DATABASES The following terms in this article are linked online to: LocusLink: http://www.ncbi.nlm.nih.gov/LocusLink/ 4-1BB | APRIL | BAFF | BAFFR | BCMA | CD27 | CD30 | CD30L | CD40 | CD40L | CD95 | CD95L | DCR1 | DCR2 | DCR3 | DR4 | DR5 | DR6 | EDA | EDAR | GITRL | HVEM | IAP1 | IKKα | IKKβ | IKKγ | LT | LTβR | MEKK3 | NF-κB | OPG | OX40 | RANKL | RELT | RIP | SOCS3 | TACI | TNF | TNFR1 | TNFR2 | TRADD | TRAF1 | TRAF2 | TRAF5 | TRAF6 | TRAIL | TRAILR1 | TROY | TWEAK | VEGI FURTHER INFORMATION Bharat Aggarwal’s lab: http://www.mdanderson.org/departments/ bioimmunother/

www.nature.com/reviews/immunol

© 2003 Nature Publishing Group