Antibody deficiency associated with an inherited autosomal dominant mutation in TWEAK

Abstract

Mutations in the TNF family of proteins have been associated with inherited forms of immune deficiency. Using an array-based sequencing assay, we identified an autosomal-dominant deficiency in TNF-like weak inducer of apoptosis (TWEAK; TNFSF12) in a kindred with recurrent infection and impaired antibody responses to protein and polysaccharide vaccines. This mutation occurs in the sixth exon of TWEAK and results in the amino acid substitution R145C within the conserved TNF-homology domain of the full-length protein. TWEAK mutant protein formed high molecular weight aggregates under nonreducing conditions, suggesting an increased propensity for intermolecular interactions. As a result, mutant TWEAK associated with B-cell–activating factor (BAFF) protein and down-regulated the BAFF-mediated activation of the noncanonical NF-κB pathway through inhibition of p100 processing to p52, resulting in inhibition of BAFF-dependent B-cell survival and proliferation. As BAFF mediates T-cell–independent isotype switching and B-cell survival, our data implicate TWEAK as a disease-susceptibility gene for a humoral immunodeficiency.

Mutations in the TNF superfamily (TNFSF) and their cognate receptors have been identified in humans with inherited forms of autoimmune or immune deficiency disorders (1), for example, autoimmune lymphoproliferative syndrome [FAS (CD95 or TNFRSF6) mutations], TNF receptor-associated periodic syndrome (TNFR1 mutations), hyper IgM syndrome (CD154/CD40 mutations), and EBV sensitivity (CD27 mutations). Common variable immunodeficiency (CVID) represents a group of heterogenous primary antibody deficiency disorders characterized by a marked reduction in serum Ig levels with impaired antibody responses to common bacteria and viruses that result in recurrent infections (2). Single-gene defects in B-cell–activating factor receptor BAFF-R (TNFRSF13C) (3) and TACI (transmembrane activator and calcium-modulating cyclophilin ligand interactor, CD267, TNFRSF13B) (4, 5) are associated with some cases of CVID.

Both BAFF-R and TACI bind to BAFF (B-cell activating factor, BLYS, TNFSF13B). BAFF is expressed primarily by monocytes and dendritic cells (6) and may induce T-cell–independent isotype switching in naive B cells in the absence of CD40 engagement and exogenous cytokines (7). However, BAFF seems to play a major role in B-cell survival through interaction with BAFF-R (8) based on studies of mice deficient in either BAFF or BAFF-R, which both display severe defects in peripheral B-cell maturation and decreased levels of immunoglobulins (9, 10). BAFF can form active heterotrimeric molecules when coexpressed with the TNF ligand APRIL (a proliferation-inducing ligand, TNFSF13), and circulating BAFF/APRIL heterotrimers are present in serum samples from patients with autoimmune disorders (11, 12). These findings raise the possibility that BAFF may form heteromeric complexes with other members of the TNF ligand superfamily and play a role in other autoimmune or immunodeficient disease states.

In a recent array-based resequencing study, we screened 148 candidate genes implicated in B-cell development, class switch recombination, and NF-κB signaling in 35 patients with defective antibody production (13). In one kindred with impaired humoral immunity and recurrent infections we identified a heterozygous mutation in the sixth exon of TWEAK (TNF-like weak inducer of apoptosis, APO3L, TNFSF12) that leads to a single amino acid substitution of arginine 145 to cysteine (position 52 in the secreted form). Similar to BAFF and many other TNFSFs, TWEAK can function as a type II transmembrane protein or as a secreted soluble cytokine that is cleaved by furin proteases (14). In general, TWEAK is widely expressed in many tissues and cell types, including monocytes/macrophages, dendritic cells, natural killer (NK) cells, and T cells, and its expression is increased during inflammation (15). Although the precise physiologic function of TWEAK awaits further delineation, it has been proposed that TWEAK may promote proliferation in endothelial cells (16) and modulate innate immunity transition to adaptive Th1 immunity (17) through its receptor Fn14 (TWEAKR, TNFRSF12A) (18). We found that mutant R145C TWEAK has impaired signaling function in the Fn14 pathway and may also affect B-cell development through interaction with BAFF.

Results

Case Report.

The proband patient P2 developed pneumococcal meningitis complicated by multiple infarcts resulting in left-sided hemiparesis at 3 y of age. A review of the family medical history revealed that both siblings (P1 and P2 in the family pedigree; Fig. 1A) and their father (B1) had a history of frequent ear or sinopulmonary infections since infancy. The father had been hospitalized several times in infancy with pneumonia and had a prolonged hospitalization for osteomyelitis, although the frequency and severity of his infections had decreased in adulthood. The father had chronic thrombocytopenia, and both children had intermittent neutropenia in the peripheral blood. The three affected family members had multiple warts. Subsequent immunologic evaluation in both siblings and the father revealed the absence of antibody responses to T-cell–dependent and polysaccharide antigens (Table 1). Serum Ig profiles and lymphocyte phenotypes were further analyzed for the father and the two proband patients (Tables 1–3). All three affected members had low levels of IgM and IgA. Low levels of IgG were revealed in patients P1 and P2, especially P2. Although the absolute IgG level of the father was considered normal, both IgG2 and IgG4 subclasses were far below normal range (Table 1). The father and patient P1 had reduced number of B cells, and nearly all the B cells (>94%) in the three affected patients were IgM⁺ IgD⁺ naive cells (Table 2, column 6). Furthermore, all three affected members had a slight increase in the percentage of TCRαβ⁺ CD4⁻CD8⁻ T cells [i.e., double-negative T (DNT) cells] and an increase in total T cell numbers, especially CD8⁺ T cells, whereas the level and percentage of CD4⁺ T cells were generally within the normal range (Table 3). The maternal medical history was unrevealing. The paternal grandmother died of lymphoma, and the paternal grandfather was not available for detailed analysis.

Fig. 1.

Identification of a rare heterozygous mutation in the TWEAK gene in a family diagnosed with CVID. (A) Pedigree of the affected family members. Highlighted, mutation carrier; square, male; circle, female. (B) Representative image of resequencing array signals showing the heterozygous mutation in the TWEAK coding region. (C) Confirmation of the heterozygous mutation by Sanger sequencing. (D) Schematic representation of TWEAK structure showing full-length and secreted forms and location of the mutation site. (E) Sequence conservation of the TWEAK gene among mammals. Highlighted box indicates the mutation site.

Table 1.

Serum Ig characterization of three patients carrying the heterozygous mutation in TWEAK

Patient	Age at diagnosis, y	Sex	IgM, mg/dL	IgG, mg/dL	IgA, mg/dL	Pneumococcal antibodies, μg/mL	Diphtheria antibodies, IU/mL	Tetanus antibodies, IU/mL
Normal	—	—	34–342	642–1,730	91–499	>0.1	>0.1	>0.6
B1	31	M	61	672*	34†	<0.1†	0.04†	<0.6†
P1‡	6	F	<21†	562†	<10†	<0.03†	<0.1†	0.02†
P2‡	4	M	<21†	16†	<10†	<0.03†	<0.1†	<0.10†

*Low IgG2 (85 mg/dL; normal range, 171–632 mg/dL) and IgG4 (0.3 mg/dL; normal range, 2.4–121 mg/dL).

^†Abnormal values.

^‡Blood samples for subjects P1 and P2 were drawn before IVIG replacement therapy.

Table 3.

T lymphocyte phenotype report for the three patients

Patient	Total CD3+ T cells	CD3+ T cells, %	Total CD4+ T cells	CD4+ T cells, %	Total CD8+ T cells	CD8+ T cells, %	Total α/β DNT cells	α/β DNT cells, %
Normal	650–2,108	57.3–86.4	362–1,275	29.0–57.3	344–911	25.2–50.8	2–17	<1.4
B1	3,036*	88.8*	1,294	37.8	1,795*	52.5*	33*	1.4*
P1	2,061	92.9*	698	31.5	1,302*	58.7*	31*	1.7*
P2	2,235*	82.5	810	30.0	1,346*	49.7	63*	3.2*

*Abnormal values.

Table 2.

B lymphocyte and platelet phenotype report for the three patients

Patient	Total CD19+ B cells	CD19+ B cells, %	CD27+ B cells, %	CD19+IgD−CD27+ cells, %	IgM+IgD+ B cells/total B cells, %	Platelets, × 103/μL
Normal	47–409	3.5–7.1	0.7–6.3	0.4–2.3	63–94	161–347
B1	74	2.2*	0.9	0.4*	97*	55*
P1	19*	0.9*	ND*	ND*	95*	226
P2	126	4.6	0.8	0.1*	98*	206

ND, not determined (i.e., too low to be determined accurately).

*Abnormal values.

Identification of Rare Heterozygous Mutation in TWEAK.

By using a custom resequencing array (13), we identified a single-base heterozygous mutation, c.433C>T, in the sixth exon of TWEAK in the two siblings P1 and P2 (Fig. 1 B and C). This mutation results in the amino acid substitution from arginine (R) to cysteine (C) at position 145 in full-length TWEAK protein or position 52 in the secreted protein. The p.R145C mutation is located in the conserved TNF-homology domain (THD) of TWEAK and is close to the only N-glycosylation site (N125), which may be responsible for trimerization of the TWEAK monomer and receptor binding (Fig. 1 D and E); in addition, PolyPhen prediction (http://genetics.bwh.harvard.edu/pph/) indicated that the R-to-C change may be deleterious (score of 1.992). Other than this mutation, the two siblings had no unusual nucleotide changes in known CVID disease-susceptibility genes such as BAFF, BAFF-R, or TACI (19), or in the related genes FAS, FASLG (FAS ligand, TNFSF6), CASP3 (caspase 3, apoptosis-related cysteine peptidase), and CASP8 (caspase 8) (20). The same heterozygous mutation was confirmed in their father, B1, but was not detected in more than 200 internally sequenced control samples, or in the latest build of the National Center for Biotechnology Information Single Nucleotide Polymorphism database.

Mutant TWEAK Fails to Induce Cell Death and Activate NF-κB and MAPK Pathways.

As a TNF-like weak inducer of apoptosis, TWEAK is able to induce apoptosis in certain cell lines through induction of TNF-α secretion (21) or through other independent mechanisms (22). We therefore tested whether the R145C mutation affected the ability of TWEAK to induce cell death in TWEAK-sensitive tumor cell lines. We first purified Flag-tagged soluble WT or mutant TWEAK protein from stable-transfected HEK293 cells, and an equal amount of each protein was used to treat HT-29 cells, an adenocarcinoma cell line that is sensitized to TNF-induced cell death. Only WT TWEAK induced significant apoptosis in HT-29 cells in the presence of IFN-γ; mutant TWEAK showed little apoptotic effect (Fig. 2A, compare columns 2 and 3). The proapoptotic effect of TWEAK was clearly blocked by preincubation of the protein with recombinant Fn14:Fc, the TWEAK receptor (Fig. 2A, column 5), but not by nonspecific human IgG (column 4). Moreover, this proapoptotic effect of the WT TWEAK protein was dose-dependent, whereas the mutant protein had little effect at any dose tested (Fig. 2B).

Fig. 2.

Mutant TWEAK fails to induce cell death and activate NF-κB pathways. (A) Mutant TWEAK fails to induce cell death in HT-29 cells. Flag-tag recombinant WT and mutant TWEAK soluble proteins were purified by M2 anti-Flag agarose gel and quantified by ELISA. HT-29 cells were cultured in the presence of WT TWEAK, mutant soluble TWEAK, or control BSA (3 ng/mL) together with IFN-γ (10 ng/mL). In some experiments, the WT soluble TWEAK protein was preincubated with human IgG or recombinant Fn14:Fc for 2 h before addition to HT-29 cells. Cell viability was determined in triplicate after 3 d. EV, empty vector; Tmu, TWEAK mutant; Twt, TWEAK WT (*P < 0.05 vs. BSA control). (B) Purified mutant TWEAK protein showed minimal capability to induce cell death in HT-29 cells. Equal amounts of WT TWEAK, mutant TWEAK, or control BSA were added in triplicate to HT-29 cell cultures, and cell viability was determined after 3 d (*P < 0.05 vs. BSA control). (C) Mutant TWEAK forms high molecular weight aggregates. Lysates of Escherichia coli transformed with WT or mutant soluble TWEAK were subjected to SDS/PAGE and Western blotting under reducing and nonreducing conditions. (D) The R145C mutation does not reduce binding of TWEAK to its receptor Fn14. 293T cells were transfected with RFP fusion constructs expressing full-length WT TWEAK or mutant TWEAK. Cells transfected with RFP empty vector served as controls. Transfection efficiency of each construct was monitored by RFP expression. Surface expression of TWEAK was measured by FACS analysis using five different concentrations of FITC-conjugated Fn14:Fc and gating on RFP⁺ cells. MFI, mean fluorescence intensity. Data in the plots were based on four independent experiments. (E) Mutant TWEAK protein was unable to induce NF-κB activation in THP1 cells. THP1 cells were treated with 6 ng/mL purified WT or mutant TWEAK soluble protein at 37 °C for the indicated time periods. To observe the effect on IκB degradation, 50 μg/mL cycloheximide (CHX) was added to the cell culture. At the indicated times, cells were lysed and subjected to Western blotting with the indicated antibodies. (F) Western blot analysis of phospho-JNK expression was performed as in A but without CHX treatment.

The loss of apoptotic function of mutant TWEAK protein may result from structural changes induced by the R145C mutation. As this mutation removes a positive charge from the extracellular domain but leaves a free thiol group in the cysteine residue, an increase in intermolecular binding to itself or to other proteins is expected. Indeed, SDS/PAGE under nonreducing conditions revealed high molecular weight aggregates in lysates of cells expressing the secreted form of mutant TWEAK protein (Fig. 2C).

As previously reported, high molecular weight aggregates of TWEAK appear to be nonfunctional under some conditions (23), possibly because of changes in protein conformation or loss of receptor binding activity. As the proapoptotic function of TWEAK is dependent on its receptor Fn14, we tested whether the loss of function in this mutant was a result of loss of binding capability to Fn14. To our surprise, 293T cells transfected with mutant TWEAK seemed to have slightly increased binding to FITC-labeled Fn14 compared with cells transfected with WT TWEAK (Fig. 2D). This indicates that the failure of R145C TWEAK to induce cell death in TWEAK-sensitive cell line is not a result of diminished interaction with Fn14.

TWEAK has been proposed to activate cell death pathways through multiple pathways involving activation of NF-κB and MAP kinases (22). To determine whether mutant TWEAK affects IκB degradation, which leads to NF-κB release and activation, we treated THP1 cells (a human acute monocytic leukemia cell line) with purified Flag-tagged WT or mutant TWEAK soluble protein. As expected, WT TWEAK induced IκB-α degradation in the presence of cycloheximide (CHX) whereas the mutant protein failed to induce IκB-α degradation (Fig. 2E). Similarly, mutant TWEAK lost the ability to induce MAPK pathways, as evidenced by minimal induction of JNK phosphorylation in THP1 cells treated with mutant TWEAK protein (Fig. 2F). Taken together, these data suggest that mutant TWEAK is a partially functional ligand that can still bind to its receptor but has lost the capability to stimulate NF-κB activation and induce cell death.

Mutant TWEAK Associates with BAFF.

The free thiol group at the mutant position C145 may allow TWEAK protein to readily bind to other TNFSFs with structurally similar THDs through formation of an additional disulfide bridge. BAFF, a critical B-cell activation and survival factor, contains many cysteine residues and is prone to form homotrimers, oligomers, or heterotrimers with other proteins such as APRIL (11). In fact, active BAFF/APRIL heterotrimers were detected in serum samples from patients with systemic immune-based rheumatic diseases (11). As the TWEAK receptor Fn14 resembles BCMA (B cell maturation antigen, TNFRSF17), one of the BAFF receptors, we reasoned that the THD of TWEAK, which shares structural similarity with that of BAFF, might interact with BAFF, and thus the TWEAK mutation might affect this binding. To test this hypothesis, we cotransfected 293T cells with full-length BAFF together with either WT or mutant TWEAK. After 48 h, cells were lysed and immunoprecipitated with anti-BAFF or anti-TWEAK antibodies and then immunoblotted with antibodies against TWEAK or BAFF, respectively. Both experiments clearly demonstrated that mutant TWEAK has stronger binding to BAFF than WT TWEAK (Fig. 3 A and B). In contrast, mutant TWEAK does not bind to CD40L (Fig. S1).

Fig. 3.

Mutant TWEAK associates with BAFF. (A and B) In vitro association of mutant TWEAK with BAFF. Full-length BAFF and WT (Twt) or mutant TWEAK (Tmu) constructs in pIRES vector were cotransfected into 293T cells. After 48 h, cells were lysed and immunoprecipitated with anti-BAFF (A) or anti-TWEAK (B) antibodies, and the resulting precipitates were subjected to Western blot analysis with the opposing antibodies. Numbers under the bands indicate relative intensity of BAFF:TWEAK complex against BAFF-IP or TWEAK-IP input. (C) Association of BAFF and TWEAK in dendritic cells from patients and normal controls. Monocytes from patients and control subjects were separated and cultured in the presence of GM-CSF and IL-4 for 6 d. The resulting dendritic cells were activated with lipopolysaccharide and subjected to IP and immunoblotting with the indicated antibodies. (D) Detection of BAFF–TWEAK heteromers in serum samples of patients. Serum samples from affected family members and normal control subjects were added to an ELISA plate coated with anti-TWEAK monoclonal antibodies and detected by HRP-conjugated monoclonal antibody against BAFF. Results from the two siblings (P1 and P2) and their father (B1) are shown (*P < 0.05 vs. normal controls).

To confirm the in vivo association between BAFF and TWEAK, we differentiated monocytes of the two siblings into dendritic cells that express TWEAK and BAFF and performed similar coimmunoprecipitation (IP) experiments. Using a monoclonal antibody against BAFF, BAFF–TWEAK association was observed in activated dendritic cells from patient samples but not in those from normal controls (Fig. 3C). IP experiments with antibody against TWEAK gave a similar result (Fig. S2). We also conducted hybrid ELISA tests on serum samples of the two patients (P1 and P2) and their father (B1). Serum samples from 20 normal individuals served as controls. The ELISA plates were coated with anti-TWEAK antibody and detected by using a specific HRP-conjugated anti-BAFF antibody. A low level of BAFF–TWEAK complex was generally detected in normal control serum samples, whereas serum from the two patients displayed significantly higher levels of BAFF–TWEAK protein complex, with patient P2 showing the highest levels (Fig. 3D). Interestingly, the BAFF–TWEAK complex level was lower in the father, B1, than in the two siblings, which may explain the reduced severity of immunodeficient symptoms in the father. The incomplete penetrance of the phenotype in the kindred could relate to differences in expression of the mutant protein as well as differences in genetic background of the affected individuals.

Mutant TWEAK Inhibits BAFF Signaling Function in B Cells.

Association of mutant TWEAK with BAFF may affect binding of BAFF to its receptors. However, neither WT nor mutant TWEAK binds BAFF-R (Fig. S3). BAFF has three receptors: a high-affinity receptor (BAFF-R), a moderate-affinity receptor (TACI), and a low-affinity receptor (BCMA). Cotransfection of full-length BAFF with WT or mutant full-length TWEAK slightly enhanced binding of BAFF to BAFF-R (Fig. 4A) but not to TACI (Fig. 4B); however, the enhancement effect was indistinguishable between WT and mutant TWEAK (Fig. 4A). Because an important function of BAFF is to provide continuous survival signaling to antigen-activated B cells through BAFF-R, we reasoned that mutant TWEAK might affect downstream signaling of BAFF-R. Activation of BAFF-R induces the noncanonical NF-κB pathway by promoting processing of a precursor protein p100 (NF-κB2) to an active p52 subunit (24). Indeed, membrane fraction preparations from BAFF stably expressing HEK293 cells transfected with mutant TWEAK showed decreased BAFF-induced processing of p100 in naive B cells as evidenced by an increase in p100 levels and a decrease in p52 levels (Fig. 4C, lane 4). In contrast, BAFF-expressing stable cell lines transfected with WT TWEAK showed enhanced p100 processing with increased p52 levels compared with BAFF-expressing stable transfectants alone (Fig. 4C, lane 3).

Fig. 4.

Mutant TWEAK affects BAFF downstream signaling. (A) Association of TWEAK with BAFF slightly increases binding of BAFF to BAFF-R. 293T cell lines were cotransfected with full-length DsRed2–BAFF fusion vector together with empty vector (EV) or WT or mutant TWEAK in pEGFP-C1 vector. Surface expression of BAFF was measured by FACS analysis using recombinant Fc-BAFF-R and APC-conjugated secondary IgG. (B) TWEAK does not affect binding of BAFF to TACI. Experiments were conducted as in A except that surface expression of BAFF was measured by FACS analysis using recombinant TACI:Fc. (C) Membrane fraction of HEK293 stable BAFF-expressing cell line cotransfected with mutant TWEAK showed decreased p100 processing in human B cells. Membrane fractions were prepared from HEK293 BAFF stable cells transfected with WT or mutant TWEAK or empty vectors and used to stimulate purified human B cells. After 42 h, B cells were lysed and subjected to Western blotting using anti-p52 antibodies. B, BAFF; Tmu, TWEAK mutant; Twt, TWEAK WT. Numbers under the bands indicate relative intensity of p100 and p52 bands. (D) Mutant TWEAK down-regulates the BAFF-induced B-cell proliferation response. B cells were purified from PBMCs and cocultured in the presence of anti-human IgM F(ab′)2 fragment and irradiated HEK293 cells that stably expressed full-length BAFF and were transfected with WT or mutant TWEAK construct or empty vector. B-cell proliferation response was measured as described in SI Materials and Methods (*P < 0.05 vs. 293-BAFF plus EV control).

To further test whether down-regulation of BAFF-R signaling by mutant TWEAK is associated with a decreased proliferation response in activated B cells, we performed an in vitro B-cell proliferation assay using [³H]thymidine incorporation. WT TWEAK, mutant TWEAK, or empty vector were transfected into BAFF-expressing stable HEK293 cell lines. After 36 h, cells were irradiated and cocultured with purified human B cells that were stimulated with anti-IgM F(ab′)2 fragment. B cells cultured with WT TWEAK transfectants showed a slightly higher proliferation response than B cells cultured with control transfectants, whereas B cells cultured with mutant TWEAK transfectants showed a decreased proliferation response (Fig. 4D). BAFF-mediated B-cell survival is primarily mediated by BAFF-R (25, 26). To test whether mutant TWEAK down-regulates BAFF-dependent B-cell survival, human B cells were treated with recombinant BAFF with or without purified soluble TWEAK WT or mutant, or with purified BAFF/mutant TWEAK heteromers (Fig. S4). As expected, addition of WT or mutant soluble TWEAK did not affect BAFF-induced B-cell survival, whereas addition of BAFF/mutant TWEAK heteromers decreased cell viability to the levels seen in medium alone at all time points tested. Finally, to analyze whether mutant TWEAK affects the BAFF-induced B-cell isotype switching response, human B cells were treated with recombinant BAFF with or without purified WT or mutant soluble TWEAK, or with purified BAFF/mutant TWEAK heteromers. BAFF alone promoted secretion of IgG and IgA, whereas addition of BAFF/mutant TWEAK heteromers decreased BAFF-mediated IgG and IgA secretion in cultured B cells (Fig. S5).

Taken together, these data indicate that mutant TWEAK may dominantly inhibit B-cell survival and proliferation in addition to Ig class switching through association with BAFF and down-regulation of the noncanonical BAFF-induced NF-κB pathway.

Discussion

The combination of candidate-gene sequencing and protein functional analysis used in this study provides an effective strategy for identifying genes responsible for rare monogenic immunodeficiency disorders (27). Here, we describe an autosomal dominant TWEAK mutation that is associated with impaired antibody responses, reduced IgM and IgA levels, and an increased number of DNT cells (i.e., TCRαβ⁺ CD4⁻CD8⁻ T cells). The TWEAK p.R145C mutation changes a charged arginine residue to a cysteine at a position close to the receptor binding sites in the THD. Although this mutation does not affect binding of TWEAK to its receptor, it appears to impair its ability to induce apoptosis in TWEAK-sensitive cell lines by decreasing activation of NF-κB and MAPK pathways. The demonstration that mutant TWEAK associates with BAFF indicates that the mutant protein might also dominantly inhibit B-cell function by forming noneffective ligand trimers or oligomers, thereby blocking effective receptor binding and downstream signaling.

Of particular interest among the observations in these patients is the increased number of DNT cells and presence of cutaneous papillomatosis. Previous reports suggest that TWEAK works with other proapoptotic TNFSF ligands such as FASLG, TRAIL (TNF-related apoptosis inducing ligand, TNFSF10), and TNF-α to facilitate cytotoxicity in many cell types, including activated monocytes (28), dendritic cells (29), NK cells (30), and T cells (31). Autoimmune lymphoproliferative syndrome caused by impaired FAS-mediated cell death is characterized by an accumulation of DNT cells and autoimmunity (32). It seems that the loss of apoptotic function of TWEAK protein is correlated to the increase in peripheral DNT cells and CD8⁺ T cells in patients carrying the mutant R145C allele; however, the exact link and underlying apoptotic mechanism awaits further study. Because the patients have papillomatosis, we were intrigued by the fact that TWEAK protein can be up-regulated by IFN-γ or phorbol myristate acetate in cultured human peripheral NK cells (17, 28). Although further investigation is warranted, TWEAK expression by NK cells and its subsequent engagement of Fn14 on the surface of epithelial cells may be important for controlling local immune responses to papilloma virus.

Mutations in BAFF-R in humans have been associated with reduced antibody production. The findings in our patients of absent antibody responses to T-cell–dependent and polysaccharide antigens, as well as reduced B cell numbers, have also been reported in BAFF-R–deficient patients (3). However, in contrast to the patients presented here, BAFF-R–deficient patients have normal or even elevated IgA serum concentrations. Differences in genetic background among patient groups may influence the development of some of these phenotypes. It is possible that R145C TWEAK/BAFF heterotrimers or oligomers also bind to TACI in a dominant-negative manner and limit the receptor’s association with another TNF ligand, APRIL. Consistent with this hypothesis are the observations of reduced serum IgA levels in TACI-deficient humans and mice deficient in APRIL (4, 5, 33). Thus, the phenotype that results from the R145C TWEAK mutation may not only be influenced by its interaction with BAFF, but also by association of the mutant heteromeric complex with multiple TNF family receptors.

TWEAK is located on human chromosome 17p13.1 approximately 878 bp upstream of APRIL, and these two genes are tightly linked in human and mouse with similar intron–exon organization. APRIL is also an important regulator of T-cell–independent class-switch pathways in B cells. An intergene splicing transcript TWE-PRIL, which combines the first six exons of TWEAK and the last five exons of APRIL, was identified in activated T cells and monocytes (34). TWE-PRIL is a membrane-anchored protein that possesses the intracellular transmembrane stalk region of TWEAK but the receptor-binding domain of APRIL, thus signaling through the APRIL binding domain may be linked to downstream pathways shared with the TWEAK intracellular domain. As the R145 residue is also present in TWE-PRIL, the possibility that the R145C mutation may interfere with the signaling pathways of APRIL cannot be ruled out. However, we were not able to detect a band corresponding to the TWE-PRIL protein in activated monocytes or T cells using monoclonal antibodies against APRIL or TWEAK in normal subjects. Furthermore, we prepared single-cell suspensions from lymph nodes of normal subjects and did not detect TWE-PRIL expression on the surface of T cells and monocytes. These results call into question whether TWE-PRIL exists as a functional protein in human leukocytes.

These findings lead us to consider the physiologic function of naturally occurring heteromeric TNF ligands. Further study of these complexes could lead to the rational design of immunosuppressive agents that target specific aspects of immunity. In addition, further analysis of cysteine mutations in other TNF family members and their contribution to the formation of heteromeric complexes might provide a mechanism for some of the phenotypic variability associated with immune deficiency disorders.

Materials and Methods

Patients and Protocols.

All patients and healthy controls were studied with informed consent at the NIH clinical center under clinical protocol (00-I-006), which was approved by the NIAID institutional review board. A detailed of summary of research methods (7, 13, 23, 35), reagents, and statistical analysis is available in the SI Materials and Methods.