Genetic susceptibility underlies the pathogenesis of cancer. We and others have previously identified a novel susceptibility gene TNFRSF19, which encodes an orphan member of the TNF receptor superfamily known to be associated with nasopharyngeal carcinoma (NPC) and lung cancer risk. Here, we show that TNFRSF19 is highly expressed in NPC and is required for cell proliferation and NPC development. However, unlike most of the TNF receptors, TNFRSF19 was not involved in NFκB activation or associated with TRAF proteins. We identified TGFβ receptor type I (TβRI) as a specific binding partner for TNFRSF19. TNFRSF19 bound the kinase domain of TβRI in the cytoplasm, thereby blocking Smad2/3 association with TβRI and subsequent signal transduction. Ectopic expression of TNFRSF19 in normal epithelial cells conferred resistance to the cell-cycle block induced by TGFβ, whereas knockout of TNFRSF19 in NPC cells unleashed a potent TGFβ response characterized by upregulation of Smad2/3 phosphorylation and TGFβ target gene transcription. Furthermore, elevated TNFRSF19 expression correlated with reduced TGFβ activity and poor prognosis in patients with NPC. Our data reveal that gain of function of TNFRSF19 in NPC represents a mechanism by which tumor cells evade the growth-inhibitory action of TGFβ.

Significance:

TNFRSF19, a susceptibility gene for nasopharyngeal carcinoma and other cancers, functions as a potent inhibitor of the TGFβ signaling pathway.

Graphical Abstract:http://cancerres.aacrjournals.org/content/canres/78/13/3469/F1.large.jpg. Cancer Res; 78(13); 3469–83. ©2018 AACR.

Nasopharyngeal carcinoma (NPC) is a malignant tumor that originates in the nasopharynx epithelium. Multiple factors, including genetic susceptibility, Epstein–Barr virus (EBV) infection, and environmental factors, contribute to NPC development. NPC exhibits a striking geographic and ethnic distribution; the incidence of NPC is unusually high in Southeast Asia, southern China, and North Africa. In addition, familiar aggregation of NPC and the occurrence of multiple cases of the disease in first-degree relatives have been reported in endemic regions, strongly indicating that genetic susceptibility plays a key role in NPC (1, 2). However, little is known about the precise genetic changes attributable to the pathogenesis of NPC (3–5).

The fundamental abnormality resulting in the development of cancer is the uncontrolled cell proliferation. Cytokine TGFβ is one of the few classes of endogenous inhibitors of cell growth. TGFβ signals through a complex of membrane-bound type I (TβRI) and type II (TβRII) receptors, both of which are serine/threonine kinases. TβRII activates TβRI upon formation of the ligand–receptor complex by phosphorylating the cytoplasmic GS domain of TβRI, which turns on the kinase activity of TβRI, followed by the phosphorylation of receptor-regulated Smads (R-Smads) including Smad2 and Smad3 at the C-terminal SSXS motif. Phosphorylated R-Smads then form a trimeric complex with the common mediator Smad4 (Co-Smad), which is translocated from the cytoplasm to nucleus to cooperate with other transcriptional modulators to initiate the transcriptional regulation of target genes. Several direct target genes of the TGFβ pathway include plasminogen activator inhibitor 1 (PAI-1), the CDK inhibitors p15INK4b and p21Cip1, and TGFβ itself. In addition to canonical Smad-dependent TGFβ signaling, the TGFβ receptor complex also mediates non-Smad signaling, including activation of the MAPK, Erk, p38, and JNK kinases (6). Under physiologic conditions, TGFβ arrests the cell cycle at G1 phase to inhibit cell proliferation; in contrast, tumor cells often escape the antiproliferative effects of TGFβ by acquiring loss-of-function mutations or deregulating the expression of various components in the TGFβ pathway (7). Aberrations in TGFβ pathway genes have been reported in NPC (8–10), and NPC cells often show a loss of TGFβ antiproliferative response (11, 12). However, it remains unclear what genetic mutations or the aberrant activities of regulatory molecules are the cause of resistance to TGFβ in NPC.

Using a genome-wide association study (GWAS), we have previously identified TNFRSF19 as a genetic susceptibility gene in NPC (13). Subsequently, TNFRSF19 has also been reported to be a lung cancer susceptibility gene in Han Chinese (14), indicating that germline mutations in TNFRSF19 confer a predisposition to certain cancers. TNFRSF19 (TNF Receptor Superfamily Member 19), also known as TROY, belongs to the TNF receptor superfamily, which commonly transduces cytokine signals via a specific adaptor protein bound to the intracellular domain (ICD). TNFRSF19 is unique because it does not bind to known TNF ligands and its ICD exhibits no sequence homology to any other characterized members of the TNF receptor superfamily (15, 16). High expression of TNFRSF19 is associated with poor prognosis in various types of cancer (17–21). However, the signal transduced by TNFRSF19 and molecular basis of TNFRSF19 in carcinogenesis have not been explored. In this study, we characterize TNFRSF19 as a potent negative regulator of the TGFβ receptor–induced signaling response and a key determinant of NPC pathogenesis.

Cell lines, transfection, and lentiviral infection

The immortalized nasopharyngeal epithelial (NPE) cells NPEC2-Bmi1 and NPEC5-TERT were provided by Dr. Mu-Sheng Zeng [Sun Yat-sen University Cancer Center (SYSUCC), Guangzhou, China] and were maintained in keratinocyte/serum-free medium (Invitrogen). CNE-1 and HNE-1 cells were provided by Dr. Chao-Nan Qian (SYSUCC) and maintained in DMEM (Invitrogen) with 10% FBS (Invitrogen) at 37°C and 5% CO2. All the NPE and NPC cell lines used in this study were authenticated using short tandem repeat profiling. All cell lines were tested Mycoplasma-free as determined by PCR-based method (16s rDNA-F: 5′-ACTCCTACGGGAGGCAGCAGTA-3′, 16s rDNA-R: 5′-TGCACCATCTGTCACTCTGTTAACCTC-3′). Mycoplasma testing was carried out every 2 or 3 weeks, and the cells were not cultured for more than 2 months.

Transfection using PEI as well as lentiviral packaging and infection were performed as described previously (22).

Constructs, reagents, and antibodies

cDNA fragments encoding TNFRSF19.1 (referred as TNFRSF19), TNFRSF19.2, mouse Tory, TNFRSF21, LMP1, TGFβRI, TGFβRII, Smad2, Smad3, and Smad4 were subcloned into pDONR201 (Invitrogen) entry clones and subsequently transferred to gateway-compatible destination vectors. Point mutants (T204D and K232R) and deletion mutants [ΔGS (delete 175-204 a.a.), Δkinase (delete 208-503 a.a.), and ΔICD (delete 151-503 a.a.)] of TGFβR1 were generated using site-directed mutagenesis PCR. All constructs were verified by sequencing.

Recombinant human TGFβ1 (240-B) was obtained from R&D Systems; SB-431542 was from Selleckchem.

Rabbit anti-TNFRSF19 antisera were raised by immunizing rabbits with GST-TNFRSF19 (residues 30-140) fusion proteins expressed in and purified from Escherichia coli (E. coli). The antisera were affinity-purified using the AminoLink Plus Immobilization and Purification Kit (Pierce). Antibodies against P21 (2947), PAI-1 (11907), p-Smad2 (3108), Smad2 (5339), p-Smad3 (9520), Smad3 (9523), Smad4 (9515), p-P38 (4511), P38 (8690), p-IκBα (2859), IκBα (4814), caspase-3 (9662), HA (3724), and GST (2624) were obtained from Cell Signaling Technology. Antibodies against TGFβRI (sc-398) and TGFβRII (sc-400) were from Santa Cruz Biotechnology. The antibody against GAPDH (60004-1-Ig) was from Proteintech. Mouse anti-FLAG (F3165) and rabbit anti-FLAG (F7425) antibodies were from Sigma-Aldrich.

Microarray assay

Total RNA was isolated from CNE-1 and HNE-1 cells and their TNFRSF19-knockout (KO) counterparts using TRIzol reagent (Invitrogen Corp.) according to the manufacturer's instructions. The concentration and purity of total RNA were determined by spectrophotometry. RNA integrity was confirmed by agarose gel electrophoresis. Control and TNFRSF19 KO cells were selected for microarray analysis. Human Genome U133 Plus 2.0 microarrays (Affymetrix Corp.) were used to monitor changes in gene expression. Total RNA was labeled and processed according to the manufacturer's instructions. The microarray analysis was performed by CapitalBio Corporation. A gene was considered to be differentially expressed if it was up- or downregulated by at least 2-fold. Online CapitalBio Molecule Annotation System (MAS) version 3.0 (http://bioinfo.capitalbio.com/mas3/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were used to perform pathway analyses of the differential genes. Microarray data are available publicly at http://www.ncbi.nlm.nih.gov/geo (GEO accession numbers: GSE113328).

GSEA assay

Microarray data were downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo/) using the accession numbers indicated in Fig. 5C. Gene set enrichment analysis (GSEA) was performed using GSEA 2.2.4 (http://www.broadinstitute.org/gsea/).

Luciferase reporter assay

(CAGA)12-Luc and the control vector pRL-TK (Promega) encoding Renilla luciferase were cotransfected into HEK293T cells or NPC cells using PEI. Luciferase activity was measured 24 hours later using the Dual-Luciferase Reporter Assay System (Promega). The firefly luciferase activity values were normalized to those of Renilla, and the ratios of firefly/Renilla activities were determined. The experiments were independently performed in triplicate.

Immunofluorescence analysis

Immunostaining was performed as described previously (23). Briefly, cells were incubated with primary antibodies against Smad2 and then with Alexa Fluor Plus 488–conjugated goat antibodies against rabbit (Invitrogen). The cells were counterstained with DAPI and imaged with a confocal laser-scanning microscope (Olympus FV1000). The data were processed with Adobe Photoshop 7.0 software.

Establishment of TNFRSF19 KO NPC cell lines

Gene knockout was performed in cells using the CRISPR/Cas9 as described previously (24). The sequences of guide RNAs (gRNA) targeting exon 3 of human TNFRSF19 gene were as follows: gRNA#1, CAAGAATTCAGGGATCGGTC and gRNA#2, GTGTTCCCTGCAACCAGTGT. Knockout clones were verified by Western blotting and Sanger sequencing (see the Supplementary Material for detail).

Tandem affinity purification and coimmunoprecipitation

Tandem affinity purification (TAP) and coimmunoprecipitation (co-IP) were carried out as described previously (23). Briefly, HEK293T cells were transfected with plasmids encoding C-terminal SFB-tagged (S-tag, flag epitope tag, and streptavidin-binding peptide tag) TNFRSF19 to establish stable cells via puromycin (2 μg/mL) selection. The cells were lysed in NETN buffer containing 50 mmol/L β-glycerophosphate, 10 mmol/L NaF, and 1 mg/mL each of pepstatin A and aprotinin. The lysates were centrifuged at 12,000 rpm to remove debris and then incubated with streptavidin-conjugated beads (Amersham) for 1 hour at 4°C. The beads were washed five times with NETN buffer and followed by elution with NETN buffer containing 2 mg/mL biotin (Sigma). The elutes were incubated with S-protein beads (Novagen) for 4 hours. After five washes, the bound proteins were analyzed by SDS-PAGE, and mass spectrometry (MS) was performed by PTM BioLabs.

For co-IP experiments, cells were washed with ice-cold PBS and then lysed in NETN buffer at 4°C for 30 minutes. The crude lysates were cleared by centrifugation at 12,000 rpm and 4°C for 30 minutes, and the supernatants were incubated with S-protein beads or anti-HA agarose (Sigma) at 4°C for 4 hours to precipitate SFB-tagged or HA-tagged proteins, respectively. For endogenous IP, the cell lysates were incubated with control IgG or a protein-specific antibody overnight at 4°C, followed by incubation with protein A/G PLUS-Agarose (Santa Cruz Biotechnology) at 4°C for 1 hour. The immunocomplexes were washed four times with NETN buffer and then subjected to SDS-PAGE and Western blotting.

Pull-down assay

GST, GST-fused TNFRSF19 extracellular domain (ECD; 30-170 a.a.) or ICD (192-423 a.a.), GST-fused full-length TNFRSF19 without transmembrane domain (30-170 plus 192-423 a.a.), and MBP-fused TGFβRI ECD (34-134 a.a.) or ICD (148-504 a.a.) were expressed in E. Coli BL21 cells. The GST fusion proteins were purified with Glutathione Sepharose 4B (GE Healthcare), and the MBP fusion proteins were purified with amylose beads (New England Biolabs) according to the manufacturers' instructions. For the pull-down assay, bait fusion proteins were incubated with cell lysates or target proteins in NETN buffer for 2 hours at 4°C. The beads were washed five times with NETN buffer, and the bound proteins were separated by SDS-PAGE and analyzed by Western blotting or MS.

Patient enrollment and IHC

A cohort of 140 patients with NPC (median age, 44.8 years; range, 15–74 years) who had undergone definitive treatment with curative intent at our institute from 2003 to 2011 was evaluated. The cases were selected based on the following criteria: pathologically confirmed NPC with available biopsy specimens for tissue microarray construction; no previous malignant disease or second primary tumors; and no prior history of radiotherapy, chemotherapy, or surgery. All the selected NPC samples contained at least 70% carcinoma tissue as determined by the examination of frozen sections. All patients were treated with standard curative radiotherapy with or without chemotherapy. Protocols of the study were approved by Ethic Committees of SYSUCC (YB2013-04). This study was conducted under the provisions of the Declaration of Helsinki, and informed written consents were obtained from all patients before inclusion. The clinical NPC samples were fixed in 10% formalin and embedded in paraffin, and then sections of the embedded specimens were deparaffinized and rehydrated. The slides were subjected to appropriate antigen retrieval protocols, and endogenous peroxidase activity was blocked with 10% H2O2 for 10 minutes. The slides were then exposed to anti-TNFRSF19 antibodies at 4°C overnight. Immunostaining was performed using the Envision System (Dako). A semiquantitative scoring criterion was used for the IHC results, whereby both the staining intensity and positive areas were recorded. The staining index (values 0–12) was obtained by multiplying the intensity of TNFRSF19-positive stain (negative, 0; weak, 1; moderate, 2; or strong, 3) by the proportion of immunopositive cells of interest (<25%, 1; 25%–50%, 2; 50%–75%, 3; or ≥75%, 4). All scores were subdivided into two categories according to a cut-off value of the ROC curve in the study cohort: low expression (≤7) and high expression (>7).

Statistical analysis

SPSS software version 16.0 was used to perform all statistical analyses. Cumulative survival was calculated using Kaplan–Meier analysis, and comparison between groups was performed using the log-rank test. Bivariate correlations between study variables were determined using Pearson correlation coefficients. Each experiment was performed at least three times. The significance of variances between groups was determined by the t test. All statistical tests were two-sided, and P < 0.05 was considered statistically significant. The authenticity of this article has been validated by uploading the key raw data onto the Research Data Deposit public platform (www.researchdata.org.cn), with the approval RDD number as RDDB2018000310.

Soft agar colony formation, tumor spheroid formation, and xenograft studies

All procedures were performed as described previously (24, 25). All the animal experiments were performed with the approval of Institutional Animal Care and Use Committee of Sun Yat-sen University (reference no. GZR2016-105), and the animals were handled in accordance with institutional guidelines. For xenograft studies, female BALB/c nude mice (5–6 weeks old) were purchased from Shanghai Laboratory Animal Center (Shanghai, China).

High expression of TNFRSF19 in NPC

As our previous work suggested that TNFRSF19 is associated with NPC risk (13), we first evaluated TNFRSF19 expression in NPC patient samples. A specific polyclonal antibody recognizing TNFRSF19 was raised and used for IHC in NPC biopsies (Supplementary Fig. S1A) and Western blotting (Supplementary Fig. S1B). We found that TNFRSF19 was highly expressed in patient-derived NPC tissues, but it could be barely detected in normal NPE tissues (Fig. 1A). Similarly, TNFRSF19 was expressed in NPC cell lines but not in the normal NPE cell lines NPEC2-Bmi1, NPEC5-Tert, and NP69 (Fig. 1B). In addition, Oncomine expression analysis revealed high expression of TNFRSF19 in other human cancers (Fig. 1C), suggesting that the high expression of TNFRSF19 is characteristic of multiple human cancer types. To investigate whether TNFRSF19 expression serves as a novel prognostic marker, the correlation of TNFRSF19 expression with NPC prognosis was evaluated. Kaplan–Meier survival curves showed that patients with high TNFRSF19 expression had a significantly poorer overall survival and distant metastasis-free survival when compared with patients with low TNFRSF19 expression, as demonstrated by the log-rank test (P < 0.001, Fig. 1D; P = 0.032; Fig. 1E). There was no significant correlation between TNFRSF19 expression and recurrence-free survival (P = 0.191; Fig. 1F). In addition, the online database Kaplan–Meier Plotter also revealed a statistically significant inverse correlation between high TNFRSF19 expression and overall survival in lung and gastric cancers (Fig. 1G). These data indicate that TNFRSF19 may play an oncogenic role.

Figure 1.

TNFRSF19 is highly expressed in NPC. A, Left, IHC analysis of TNFRSF19 expression in 8 normal nasopharyngeal and 140 NPC tissues (scale bar, 50 μm), together with an enlarged view of each in the corresponding inset. Right, scatterplots representing the IHC scores are shown on the left. B, Western blot assay of TNFRSF19 expression in three normal NPE cells and 6 NPC cell lines. C, Oncomine box plots of TNFRSF19 expression levels in multiple advanced human cancers. D–F, Kaplan–Meier analysis of TNFRSF19 expression and overall survival (D), distant metastasis-free survival (E), and recurrence-free survival (F) in 140 patients with NPC. G, Kaplan–Meier analysis of overall survival in lung (1,145 patients) and gastric cancer (631 patients) stratified by TNFRSF19 expression. The P values were calculated using the log-rank test.

Figure 1.

TNFRSF19 is highly expressed in NPC. A, Left, IHC analysis of TNFRSF19 expression in 8 normal nasopharyngeal and 140 NPC tissues (scale bar, 50 μm), together with an enlarged view of each in the corresponding inset. Right, scatterplots representing the IHC scores are shown on the left. B, Western blot assay of TNFRSF19 expression in three normal NPE cells and 6 NPC cell lines. C, Oncomine box plots of TNFRSF19 expression levels in multiple advanced human cancers. D–F, Kaplan–Meier analysis of TNFRSF19 expression and overall survival (D), distant metastasis-free survival (E), and recurrence-free survival (F) in 140 patients with NPC. G, Kaplan–Meier analysis of overall survival in lung (1,145 patients) and gastric cancer (631 patients) stratified by TNFRSF19 expression. The P values were calculated using the log-rank test.

Close modal

Loss of TNFRSF19 decreases tumorigenicity of NPC

To gain insights into the role of TNFRSF19 in cancer development, the TNFRSF19 gene was knocked out in two NPC cell lines, CNE-1 and HNE-1, using the CRISPR-Cas9 genome editing system. Two independent sgRNAs with efficient cleavage activity were selected (Supplementary Fig. S2A), and the genetic ablation of TNFRSF19 was confirmed by Western blotting (Fig. 2A) and Sanger sequencing (Supplementary Fig. S2B). Knockout of TNFRSF19 resulted in a substantially reduced growth rate (Fig. 2B), lower plating efficiency (Fig. 2C) and colony-forming capacity on soft agar (Fig. 2D), and reduced primary and secondary tumor spheroid formation (Fig. 2E and F), indicating that the transformation ability was greatly impaired by TNFRSF19 loss in vitro. To determine the role of TNFRSF19 in tumorigenesis, we established xenograft tumors via subcutaneous inoculation of wild-type (WT) and TNFRSF19 KO NPC cells into the right and left flanks of nude mice, respectively. TNFRSF19 KO cells displayed a significant inhibition of tumor growth compared with control cells (Fig. 2G and H). Furthermore, knockout of TNFRSF19 resulted in a reduction of tumor-initiating ability, as determined by limiting dilution transplantation analysis of NPC xenografts (Supplementary Table S1). The above data indicate that TNFRSF19 is required for cell growth and tumorigenesis of NPC.

Figure 2.

TNFRSF19 is required for NPC tumorigenesis. A, Western blot analysis of TNFRSF19 in WT and knockout (KO) cells. TNFRSF19 KO cells were generated using the CRISPR-Cas9 system with two single-guide RNAs targeting exon 3, and clones 1# and 2# were from sgRNAs 1# and 2#, respectively. B, Proliferation curve of HNE-1 WT or TNFRSF19 KO cells as measured by CCK-8 assay. Error bars, SDs (n = 3). C and D, Knockout of TNFRSF19 reduced the ability to form colonies on conventional plates (C) and soft agar (D). E and F, Primary and secondary tumor sphere formation in ultralow attachment plates of CNE-1 and HNE-1 cells with or without TNFRSF19 KO. Secondary spheroids (F) were obtained from the dissociation of primary spheroids (E) into single cells and reseeded. Error bars, SD (n = 3). Scale bars, 50 μm. G, Wild-type or TNFRSF19 KO cells were subcutaneously injected into the right and left flanks of athymic BALB/c mice, respectively. Images were captured 7 weeks postimplantation. H, Growth curves of the tumors formed by wild-type or TNFRSF19 KO cells. Mean tumor volumes are plotted. Error bars, SD (n = 6).

Figure 2.

TNFRSF19 is required for NPC tumorigenesis. A, Western blot analysis of TNFRSF19 in WT and knockout (KO) cells. TNFRSF19 KO cells were generated using the CRISPR-Cas9 system with two single-guide RNAs targeting exon 3, and clones 1# and 2# were from sgRNAs 1# and 2#, respectively. B, Proliferation curve of HNE-1 WT or TNFRSF19 KO cells as measured by CCK-8 assay. Error bars, SDs (n = 3). C and D, Knockout of TNFRSF19 reduced the ability to form colonies on conventional plates (C) and soft agar (D). E and F, Primary and secondary tumor sphere formation in ultralow attachment plates of CNE-1 and HNE-1 cells with or without TNFRSF19 KO. Secondary spheroids (F) were obtained from the dissociation of primary spheroids (E) into single cells and reseeded. Error bars, SD (n = 3). Scale bars, 50 μm. G, Wild-type or TNFRSF19 KO cells were subcutaneously injected into the right and left flanks of athymic BALB/c mice, respectively. Images were captured 7 weeks postimplantation. H, Growth curves of the tumors formed by wild-type or TNFRSF19 KO cells. Mean tumor volumes are plotted. Error bars, SD (n = 6).

Close modal

TNFRSF19 interacts with TGFβ type-I receptor

The major signal transducers for the TNFR superfamily are TNF receptor–associated factors (TRAF), which are linked to NF-κB activation (16). As a member of TNFRs, some studies have shown that TNFRSF19 interacts with TRAF proteins and activates the NF-κB pathway (19, 21, 26). The human TNFRSF19 gene encodes two transcripts: TNFRSF19.1 and TNFRSF19.2, which share most of exons except the last one (Supplementary Fig. S3A) and are both expressed in human cells (Supplementary Fig. S3B and S3C). TNFRSF19.1 has 421 a.a. and lacks the TRAF-binding motif; TNFRSF19.2 is shorter (417 a.a.) and has a distinct C-terminus with a potential TRAF-binding consensus sequence (P/S/A/T)X(Q/E)E, 413SLQE416. Mouse Tnfrsf19, commonly named as Troy, only has one transcript and the encoded protein containing a different TRAF-binding motif 276TLQE279 (Supplementary Fig. S3A and S3D). However, transient transfection of these TNFRSF19 constructs into HEK293T cells did not lead to IκBα phosphorylation or caspase-3 cleavage, while a positive control LMP1, an EBV oncoprotein acting as a constitutively active mimic of TNFR CD40 and binding TRAFs (27), potently stimulated IκBα phosphorylation (Fig. 3A). Consistently, co-IP revealed that TNFRSF19 transcript 1 and 2 or the mouse ortholog did not associate with TRAF2 or TRAF6, whereas binding of TRAF2 to LMP1 was readily detectable (Fig. 3B). Therefore, it seems that TNFRSF19 does not share the same set of signal transducers with other TNFRs and may have different signaling outputs.

Figure 3.

TNFRSF19 interacts with type I TGFβ receptor. A, Lysates from HEK293T cells transfected with human TNFRSF19.1, 19.2, mouse Troy, and LMP1 constructs were immunoblotted with the indicated antibodies. LMP1 served as a positive control for NFκB activation. B, Top, SFB-tagged TNFRSF19.1, 19.2, and mouse Troy or LMP1 were transfected into HEK293T cells along with HA-TRAF2. The cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. LMP1 served as a positive control for TRAF2 binding. Bottom, the interaction of SFB-tagged TNFRSF19.1, 19.2, and mouse Troy with HA-TRAF6. C, HEK293T cells stably expressing SFB-tagged TNFRSF19 were used for TAP. Silver staining of TAP and the number of peptides identified by MS are shown. D, Co-IP of exogenous TGFβRI (TβRI) and TNFRSF19 (T19). HEK293T cells were transfected with plasmids encoding C-terminal SFB-tagged TGFβRI and C-terminal HA-tagged TNFRSF19. Cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. E, Interactions between endogenous TNFRSF19 and TβRI in HNE-1 and CNE-1 cells were assessed by immunoprecipitation (IP) with anti-TβRI antibody or control IgG and immunoblotting using anti-TNFRSF19 and TβRI antibodies. Arrow, IgG heavy chain. F, Lysates from HEK293T cells expressing HA-tagged Smad2, Smad4, and TβRI along with SFB-tagged TNFRSF19 were immunoprecipitated with anti-HA agarose and immunoblotted with antibodies as indicated. G, Co-IP of SFB-tagged TNFRSF19 or TNFRSF21 with HA-TβRI or TβRII in transfected HEK293T cells. H, The ICD of TNFRSF19 binds to TβRI. Bacterially expressed GST-ECD and ICD fragments of TNFRSF19 or GST alone were incubated with HNE-1 cell lysates. Proteins bound to glutathione sepharose beads were analyzed by immunoblotting. GST fusion proteins are shown by Coomassie Brilliant Blue (CBB) staining. I, TNFRSF19 directly binds to ICD of TβRI. MBP-tagged ECD or ICD of TβRI fragments were coincubated with GST-tagged TNFRSF19; proteins bound to amylose beads were immunoblotted with anti-GST antibody. J, Left, schematic representation of the TGFβRI domain structure. SP, signal peptide; TM, transmembrane domain. Right, lysates from HEK293T cells expressing SFB-tagged full-length (FL) or deletion mutants of TβRI with or without HA-tagged TNFRSF19 were subjected to immunoprecipitation with anti-HA agarose and immunoblotting with anti-FLAG and anti-HA antibodies.

Figure 3.

TNFRSF19 interacts with type I TGFβ receptor. A, Lysates from HEK293T cells transfected with human TNFRSF19.1, 19.2, mouse Troy, and LMP1 constructs were immunoblotted with the indicated antibodies. LMP1 served as a positive control for NFκB activation. B, Top, SFB-tagged TNFRSF19.1, 19.2, and mouse Troy or LMP1 were transfected into HEK293T cells along with HA-TRAF2. The cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. LMP1 served as a positive control for TRAF2 binding. Bottom, the interaction of SFB-tagged TNFRSF19.1, 19.2, and mouse Troy with HA-TRAF6. C, HEK293T cells stably expressing SFB-tagged TNFRSF19 were used for TAP. Silver staining of TAP and the number of peptides identified by MS are shown. D, Co-IP of exogenous TGFβRI (TβRI) and TNFRSF19 (T19). HEK293T cells were transfected with plasmids encoding C-terminal SFB-tagged TGFβRI and C-terminal HA-tagged TNFRSF19. Cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. E, Interactions between endogenous TNFRSF19 and TβRI in HNE-1 and CNE-1 cells were assessed by immunoprecipitation (IP) with anti-TβRI antibody or control IgG and immunoblotting using anti-TNFRSF19 and TβRI antibodies. Arrow, IgG heavy chain. F, Lysates from HEK293T cells expressing HA-tagged Smad2, Smad4, and TβRI along with SFB-tagged TNFRSF19 were immunoprecipitated with anti-HA agarose and immunoblotted with antibodies as indicated. G, Co-IP of SFB-tagged TNFRSF19 or TNFRSF21 with HA-TβRI or TβRII in transfected HEK293T cells. H, The ICD of TNFRSF19 binds to TβRI. Bacterially expressed GST-ECD and ICD fragments of TNFRSF19 or GST alone were incubated with HNE-1 cell lysates. Proteins bound to glutathione sepharose beads were analyzed by immunoblotting. GST fusion proteins are shown by Coomassie Brilliant Blue (CBB) staining. I, TNFRSF19 directly binds to ICD of TβRI. MBP-tagged ECD or ICD of TβRI fragments were coincubated with GST-tagged TNFRSF19; proteins bound to amylose beads were immunoblotted with anti-GST antibody. J, Left, schematic representation of the TGFβRI domain structure. SP, signal peptide; TM, transmembrane domain. Right, lysates from HEK293T cells expressing SFB-tagged full-length (FL) or deletion mutants of TβRI with or without HA-tagged TNFRSF19 were subjected to immunoprecipitation with anti-HA agarose and immunoblotting with anti-FLAG and anti-HA antibodies.

Close modal

To gain insight into the biological pathway transduced by TNFRSF19, we studied the protein interaction network of TNFRSF19. TAP and MS (TAP-MS) were carried out in HEK293T cells stably expressing SFB (S-FLAG-SBP) triple-tagged TNFRSF19 (transcript variant 1). Interestingly, MS data revealed peptides that corresponded to TβRI (Fig. 3C). TβRI was also identified via TAP in the NPC cell line HK1 (Supplementary Fig. S3E); besides, the pull-down assay using TNFRSF19 ICD as bait also captured TβRI peptides in HEK293T cell lysates (Supplementary Fig. S3F), but no TRAF proteins or death domain-containing proteins were found in these MS lists. To verify the interaction of TNFRSF19 with TβRI, reciprocal co-IPs were performed in HEK293T cells and showed that TβRI bound to TNFRSF19 in cooverexpression experiment (Fig. 3D). Furthermore, endogenous TNFRSF19 was coimmunoprecipitated with TβRI in HNE-1 and CNE-1 cells (Fig. 3E). In addition, TNFRSF19 did not bind Smad2 or Smad4 (Fig. 3F). To further ensure the binding specificity between TNFRSF19 and TβRI, another member of the TNF receptor superfamily, TNFRSF21, which contains the death domain (28), was used as a control, and the co-IP result suggested that only TNFRSF19 interacted with TβRI but not TβRII in transiently transfected HEK293T cells (Fig. 3G). Collectively, these results suggest that TNFRSF19 specifically binds TβRI in vivo.

Next, we sought to determine the regions responsible for the TNFRSF19–TβRI interaction. To achieve this goal, bacterially expressed ECDs and ICDs of TNFRSF19 and TβRI were used in pull-down assays. The GST-fused ICD, but not the ECD of TNFRSF19, was able to pull down TβRI (Fig. 3H). Conversely, GST-fused TNFRSF19 bound to MBP-fused ICD but not to the ECD of TβRI (Fig. 3I). These results suggest that TNFRSF19 binds directly to TβRI through the respective cytoplasmic domains. As TNFRSF19 ICD does not harbor any known functional motif, we further dissected the domains of TβRI ICD mediating the association with TNFRSF19. The cytoplasmic region of TβRI is composed of a glycine- and serine-rich sequence, termed the GS domain, followed by a kinase domain. A series of deletion mutants lacking the GS domain, kinase domain, or the entire ICD of TβRI was generated (Fig. 3J, right), and the co-IP assay demonstrated that the GS domain was dispensable for TβRI–TNFRSF19 interaction, but deletion of the kinase domain or the entire ICD of TβRI abolished the TNFRSF19 association (Fig. 3J, left). Taken together, these data suggest that TNFRSF19 binds to the kinase domain of TβRI in the cytoplasm.

TNFRSF19 blocks formation of the TGFβRI–Smad2/3 complex

Having established the interaction between TNFRSF19 and TβRI, we next asked how the TβRI complex is regulated by TNFRSF19. Upon ligand stimulation, TβRI forms a heteromeric complex with TβRII and is phosphorylated in the GS domain. Phosphorylation induces a conformational change of TβRI and its subsequent association and phosphorylation of the R-Smads, Smad2, and Smad3 via the kinase domain (29–32) (Fig. 4A). We first examined whether the TNFRSF19 and TβRI association is TGFβ-induced. As most TNFRSF19-positive NPC cells lost their response to TGFβ (see below), we chose HaCaT, an immortalized human keratinocyte cell line that is highly responsive to TGFβ and TNFRSF19-positive, to immunoprecipitate endogenous TβRI before and after TGFβ treatment. The TβRI–TβRII complex was induced after 1 hour of TGFβ treatment and declined after 6 hours of treatment, which correlated with the phosphorylation of Smad2; in contrast, a constitutive interaction between TNFRSF19 and TβRI was observed regardless of the presence of TGFβ (Fig. 4B). In addition, we compared the interaction of TNFRSF19 with either active or inactive forms of TβRI. The T204D mutation in the GS domain causes ligand- and TβRII-independent activation of TβRI, whereas the K232R mutation leads to kinase inactivation (29). Co-IP experiment suggested that TNFRSF19 bound to all forms of TβRI, and treatment with SB-431542, a small molecule that inhibits the catalytic activity of TβRI, did not affect their interactions (Fig. 4C). Thus, unlike TβRII and Smad2/3, the binding of TNFRSF19 to TβRI is TGFβ independent.

Figure 4.

TNFRSF19 blocks TGFβRI from binding to R-Smads. A, Schematic representation of TβRI activation and signal transduction via receptors. B, Cell lysates from HaCaT cell line treated with TGFβ (5 ng/mL) for the indicated times were immunoprecipitated (IP) using anti-TβRI antibody or control IgG, followed by immunoblotting with the indicated antibodies. C, HEK293T cells were cotransfected with plasmids encoding HA-tagged WT, T204D, and K232R mutants of TβRI, along with TNFRSF19-SFB. Following 1 hour of SB-431542 (10 μmol/L) treatment or no treatment, the cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. Phospho-Smad2 and PAI-1 represent early and late response to TGFβ, respectively. D, Co-IP of SFB-tagged TβRII and HA-tagged TβRI in the presence or absence of exogenous TNFRSF19 in transfected HEK293T cells. E, Co-IP of SFB-tagged Smad2 or Smad3 with HA-TβRI in the presence or absence of TNFRSF19 overexpression in transfected HEK293T cells. F, Co-IP of SFB-Smad2 with HA-tagged WT, T204D, or K232R mutant of TβRI with or without cotransfection of TNFRSF19 in transfected HEK293T cells. G, Interaction of endogenous TβRI and Smad2 in WT and TNFRSF19 KO HNE-1 cells with or without TGFβ (5 ng/mL) treatment was analyzed by immunoprecipitation with anti-TβRI antibody, followed by immunoblotting with the indicated antibodies. Arrow, IgG heavy chain. H, Co-IP of SFB-Smad2 and Myc-Smad4 with or without TNFRSF19 overexpression in transfected HEK293T cells. I, Endogenous Smad2–Smad4 interaction in WT and TNFRSF19 KO HNE-1 cells with or without TGFβ (5 ng/mL) treatment as analyzed by immunoprecipitation with anti-Smad2 antibody, followed by immunoblotting (IB) with anti-Smad4 antibody.

Figure 4.

TNFRSF19 blocks TGFβRI from binding to R-Smads. A, Schematic representation of TβRI activation and signal transduction via receptors. B, Cell lysates from HaCaT cell line treated with TGFβ (5 ng/mL) for the indicated times were immunoprecipitated (IP) using anti-TβRI antibody or control IgG, followed by immunoblotting with the indicated antibodies. C, HEK293T cells were cotransfected with plasmids encoding HA-tagged WT, T204D, and K232R mutants of TβRI, along with TNFRSF19-SFB. Following 1 hour of SB-431542 (10 μmol/L) treatment or no treatment, the cell lysates were precipitated with S-protein beads and immunoblotted with the indicated antibodies. Phospho-Smad2 and PAI-1 represent early and late response to TGFβ, respectively. D, Co-IP of SFB-tagged TβRII and HA-tagged TβRI in the presence or absence of exogenous TNFRSF19 in transfected HEK293T cells. E, Co-IP of SFB-tagged Smad2 or Smad3 with HA-TβRI in the presence or absence of TNFRSF19 overexpression in transfected HEK293T cells. F, Co-IP of SFB-Smad2 with HA-tagged WT, T204D, or K232R mutant of TβRI with or without cotransfection of TNFRSF19 in transfected HEK293T cells. G, Interaction of endogenous TβRI and Smad2 in WT and TNFRSF19 KO HNE-1 cells with or without TGFβ (5 ng/mL) treatment was analyzed by immunoprecipitation with anti-TβRI antibody, followed by immunoblotting with the indicated antibodies. Arrow, IgG heavy chain. H, Co-IP of SFB-Smad2 and Myc-Smad4 with or without TNFRSF19 overexpression in transfected HEK293T cells. I, Endogenous Smad2–Smad4 interaction in WT and TNFRSF19 KO HNE-1 cells with or without TGFβ (5 ng/mL) treatment as analyzed by immunoprecipitation with anti-Smad2 antibody, followed by immunoblotting (IB) with anti-Smad4 antibody.

Close modal

Next, we questioned whether TNFRSF19 affects the interaction of TβRI with the upstream TβRII and the downstream R-Smads. Overexpression of TNFRSF19 did not affect TβRI/II heterotetrameric receptor complex formation (Fig. 4D); however, the interactions of Smad2/3 with TβRI were severely impaired by TNFRSF19 overexpression (Fig. 4E), and both wild-type and the constitutively active mutant T204D of TβRI lost their abilities to associate with Smad2 in the presence of exogenous TNFRSF19 (Fig. 4F). In agreement with the overexpression data, knockout of TNFRSF19 in HNE-1 cells greatly induced endogenous TβRI–Smad2 complex formation (Fig. 4G). And as expected, the downstream event of TβRI/R–Smads interaction, oligomerization of Smad2 with the Co-Smad, Smad4 was inhibited by TNFRSF19 overexpression (Fig. 4H), and depletion of TNFRSF19 in HNE-1 cells substantially increased Smad2–Smad4 complex (Fig. 4I). In summary, TNFRSF19 competes with R-Smads for binding to the kinase domain of TβRI and thereby blocks the recognition and activation of R-Smads and the subsequent formation of an active R-Smad/Co-Smad complex.

TNFRSF19 inhibits TGFβ signaling

To investigate the biological significance of TNFRSF19 in the TGFβ pathway, we compared the global gene expression profiles of control and TNFRSF19 KO cells using microarrays. A total of 143 differential genes were found between WT and TNFRSF19 KO HNE-1 cells (Fig. 5A). Strikingly, functional profiling of these differentially expressed genes suggested that the TGFβ signaling pathway was one of the most significantly affected pathways by the loss of TNFRSF19 (Fig. 5B). Furthermore, after analysis of TNFRSF19 expression and TGFβ-regulated gene signatures via GSEA in GEO public NPC patient expression datasets (33), we found that TNFRSF19 levels were inversely correlated with the gene signatures activated by TGFβ (Fig. 5C).

Figure 5.

TNFRSF19 suppresses TGFβ signaling in NPC. A, Double-log scatter plot comparing the differential expression of mRNAs in control and TNFRSF19 KO HNE-1 cells. Green, downregulated genes; red, upregulated genes. B, KEGG pathway enrichment analysis of differentially expressed genes between WT and TNFRSF19 KO cells. C, GSEA plot showing that TNFRSF19 expression is inversely correlated with TGFβ-activated gene signatures in published NPC patient gene expression datasets (GSE12452, n = 31). ES, enrichment score; NES, normalized enrichment score. D, Luciferase assay of the TGFβ reporter (CAGA)12-Luc in WT or TNFRSF19 KO cells treated or not with TGFβ (5 ng/mL) for 24 hours. Error bars, SD (n = 3). E, Effects of TNFRSF19 KO on TGFβ-induced P21 and PAI-1 expression. WT, TNFRSF19 KO CNE-1, or HNE-1 cells were starved overnight and then treated with TGFβ for 6 hours; the induction of target genes was examined by Western blot analysis. F, Effects of TNFRSF19 KO on TGFβ-induced phosphorylation of Smad2/3 and P38. Cells with or without TNFRSF19 deletion were starved overnight and then stimulated with TGFβ for 1 hour, and the phosphorylated and total proteins were immunoblotted with the indicated antibodies. G, IHC analysis of TNFRSF19 and phosphorylated Smad2 and P21 in xenografts generated from WT and TNFRSF19-KO cells as shown in Fig. 2H. Scale bars, 100 μm. H, Expression levels of p-Smad2 and P21 were inversely associated with TNFRSF19 levels in 40 primary human NPC specimens. Two representative cases are shown. Scale bar, 100 μm. The inset shows a magnified view.

Figure 5.

TNFRSF19 suppresses TGFβ signaling in NPC. A, Double-log scatter plot comparing the differential expression of mRNAs in control and TNFRSF19 KO HNE-1 cells. Green, downregulated genes; red, upregulated genes. B, KEGG pathway enrichment analysis of differentially expressed genes between WT and TNFRSF19 KO cells. C, GSEA plot showing that TNFRSF19 expression is inversely correlated with TGFβ-activated gene signatures in published NPC patient gene expression datasets (GSE12452, n = 31). ES, enrichment score; NES, normalized enrichment score. D, Luciferase assay of the TGFβ reporter (CAGA)12-Luc in WT or TNFRSF19 KO cells treated or not with TGFβ (5 ng/mL) for 24 hours. Error bars, SD (n = 3). E, Effects of TNFRSF19 KO on TGFβ-induced P21 and PAI-1 expression. WT, TNFRSF19 KO CNE-1, or HNE-1 cells were starved overnight and then treated with TGFβ for 6 hours; the induction of target genes was examined by Western blot analysis. F, Effects of TNFRSF19 KO on TGFβ-induced phosphorylation of Smad2/3 and P38. Cells with or without TNFRSF19 deletion were starved overnight and then stimulated with TGFβ for 1 hour, and the phosphorylated and total proteins were immunoblotted with the indicated antibodies. G, IHC analysis of TNFRSF19 and phosphorylated Smad2 and P21 in xenografts generated from WT and TNFRSF19-KO cells as shown in Fig. 2H. Scale bars, 100 μm. H, Expression levels of p-Smad2 and P21 were inversely associated with TNFRSF19 levels in 40 primary human NPC specimens. Two representative cases are shown. Scale bar, 100 μm. The inset shows a magnified view.

Close modal

To validate the participation of TNFRSF19 in the TGFβ pathway, we used (CAGA)12-Luc, a TGFβ-responsive luciferase reporter containing a Smad3/4–binding box on the PAI-1 gene promoter (34), to determine whether TNFRSF19 affects TGFβ-mediated transcriptional responses. Luciferase assays showed that loss of TNFRSF19 dramatically increased the activity of the TGFβ/Smad–responsive reporter in TGFβ-treated HNE-1 cells (Fig. 5D). Consistent with the reporter data, the protein levels of P21 and PAI-1, two well-characterized direct transcriptional targets of TGFβ pathway, were robustly upregulated in TNFRSF19 KO cells than in control cells, the latter responded poorly to TGFβ (Fig. 5E). Next, we examined the early mediators of TGFβ signaling, for example, TGFβ-induced phosphorylation of R-Smads. Smad2/3 phosphorylation was almost undetectable in HNE-1 WT cells even after TGFβ stimulation; in contrast, much higher levels of basal and TGFβ-induced Smad2/3 phosphorylation were observed in TNFRSF19 KO cells, while p38 phosphorylation was slightly increased (Fig. 5F). These data were in line with the finding that TNFRSF19 prevented R-Smads from being recruited and activated by TβRI (Fig. 4).

To further examine the role of TNFRSF19 in the TGFβ pathway in vivo, we analyzed the levels of P21 and phospho-Smad2 by IHC in NPC tumor xenografts, as shown in Fig. 2H. The staining results revealed that tumors derived from TNFRSF19 KO cells had higher levels of p21 and phosphorylated Smad2 compared with tumors derived from control cells (Fig. 5G). Consistently, IHC in clinical specimens demonstrated a negative correlation between TNFRSF19 and P21 and phospho-Smad2 levels (Fig. 5H). Taken together, these results indicate that TNFRSF19 inhibits TGFβ-induced R-Smads phosphorylation and transcriptional responses in NPC.

Overexpression of TNFRSF19 confers resistance to growth-inhibitory effect of TGFβ

TGFβ acts as a tumor suppressor by inhibiting the growth of normal and premalignant cells. We next determined the effects of TNFRSF19 gain of function on the antiproliferative role of TGFβ in normal cells. Overexpression of TNFRSF19 abolished TGFβ-induced (CAGA)12-Luc reporter activation in HEK293T cells (Fig. 6A). When TNFRSF19 was transduced into a normal NPE cell line, NPEC2-Bmi1, which does not express TNFRSF19 (Fig. 1B), cell proliferation was accelerated (Fig. 6B), and inhibition of TGFβ-induced growth was significantly alleviated compared with the control cells without TNFRSF19 transduction (Fig. 6C). In epithelial cells, TGFβ stimulates the transcription of p15INK4b and/or p21Cip1, two cyclin-dependent kinase inhibitors, to arrest cell-cycle progression. Consequently, overexpression of TNFRSF19 substantially reduced the induction of P21 and PAI-1 by TGFβ (Fig. 6D), while P15 was undetectable due to homozygous deletion of the p15INK4b locus. The data suggest that TNFRSF19 reduces transcriptional activation of TGFβ to bypass TGFβ-induced growth inhibition. Consistently, TGFβ-induced phosphorylation of Smad2 was substantially inhibited by ectopic expression of TNFRSF19 (Fig. 6E). Furthermore, the TGFβ-induced cytoplasmic to nuclear translocation of Smad2 was suppressed by TNFRSF19 in NPE cells (Fig. 6F), Thus, the overexpression data were consistent with the results obtained from NPC knockout cells.

Figure 6.

Overexpression of TNFRSF19 confers resistance to TGFβ-mediated growth inhibition in normal NPE cells. A, Effect of TNFRSF19 on the (CAGA)12-Luc transcriptional response induced by TGFβ (5 ng/mL) for 24 hours in 293T cells. B, NPEC2-Bmi1 cells were infected with control lentivirus or lentivirus expressing TNFRSF19. The cell proliferation was determined by CCK-8 assay. C, NPEC2-Bmi1 cells with or without TNFRSF19 overexpression were cultured in the absence or presence of TGFβ. After 48-hour incubation, the cell viability was determined by CCK-8 assay. D, Control or TNFRSF19-overexpressing NPEC2-Bmi1 cells were treated or not with various concentrations of TGFβ for 6 hours; the inductions of P21 and PAI-1 are shown. E, TNFRSF19 overexpression inhibits the phosphorylation of Smad2 induced by 30 minutes of TGFβ treatment in NPE cells. F, TNFRSF19 inhibits the nuclear translocation of Smad2 induced by TGFβ. Left, immunofluorescence images representing the subcellular localization of Smad2 before and after 30 minutes of stimulation with TGFβ in control or TNFRSF19-overexpressing NPE cells. Scale bars, 50 μm. Right, histogram representing the percentage of cells displaying Smad2 distributed in the nuclear, cytoplasmic, or both compartments.

Figure 6.

Overexpression of TNFRSF19 confers resistance to TGFβ-mediated growth inhibition in normal NPE cells. A, Effect of TNFRSF19 on the (CAGA)12-Luc transcriptional response induced by TGFβ (5 ng/mL) for 24 hours in 293T cells. B, NPEC2-Bmi1 cells were infected with control lentivirus or lentivirus expressing TNFRSF19. The cell proliferation was determined by CCK-8 assay. C, NPEC2-Bmi1 cells with or without TNFRSF19 overexpression were cultured in the absence or presence of TGFβ. After 48-hour incubation, the cell viability was determined by CCK-8 assay. D, Control or TNFRSF19-overexpressing NPEC2-Bmi1 cells were treated or not with various concentrations of TGFβ for 6 hours; the inductions of P21 and PAI-1 are shown. E, TNFRSF19 overexpression inhibits the phosphorylation of Smad2 induced by 30 minutes of TGFβ treatment in NPE cells. F, TNFRSF19 inhibits the nuclear translocation of Smad2 induced by TGFβ. Left, immunofluorescence images representing the subcellular localization of Smad2 before and after 30 minutes of stimulation with TGFβ in control or TNFRSF19-overexpressing NPE cells. Scale bars, 50 μm. Right, histogram representing the percentage of cells displaying Smad2 distributed in the nuclear, cytoplasmic, or both compartments.

Close modal

In addition to its tumor-suppressing activity, TGFβ also exhibits tumor-promoting effects by assisting cell migration and cancer metastasis via epithelial-to-mesenchymal transition (7). However, loss of TNFRSF19 did not change the level of E-cadherin, although N-cadherin levels increased. Besides, the expression of E- and N-cadherin was insensitive to TGFβ in HNE-1 cells (Supplementary Fig. S4A). Furthermore, TNFRSF19-deficient cells exhibited similar migration ability compared with control cells (Supplementary Fig. S4B), suggesting that at least in the context of NPC, TNFRSF19 mainly regulates the growth-inhibitory function of TGFβ.

In summary, our results support a model in which TNFRSF19 functions as a key repressor of TGFβ receptor–induced signaling responses and of TGFβ-dependent antiproliferative effects in NPC (Fig. 7).

Figure 7.

Working model of TNFRSF19-mediated inhibition of TGFβ signaling in cancer.

Figure 7.

Working model of TNFRSF19-mediated inhibition of TGFβ signaling in cancer.

Close modal

NPC is a special type of head and neck cancer. Strong ethnic clustering and familiar aggregation of NPC indicates that genetic susceptibility plays a significant role in this disease (1, 35). However, even though numerous mutations/variations have been discovered in NPC by multiple whole-genome or exome sequencing studies over the years, the exact genetic perturbations resulting in NPC are far from being fully understood (3).

We have previously identified TNFRSF19 as a susceptibility gene for NPC (13), and another GWAS has reported that TNFRSF19 is a susceptibility factor in lung cancer (14). TNFRSF19 belongs to the TNFR superfamily. The family members are characterized by four conserved cysteine-rich domains in their ECD and a distinct ICD that is responsible for TNFR signaling. In human, a total of 29 TNFRs have been described to date. In general, members of the TNFR superfamily can be divided into two groups: survival receptors and death receptors. Survival receptors activate the NF-κB pathway by binding to TRAF proteins. Death receptors mediate signal-induced cell death through their death domains (16). TNFRSF19 is far less characterized in this family. TNFRSF19 does not bind to TNF-related ligands and remains an orphan receptor to date (15). Its cytoplasmic domain lacks death domain; in addition, despite the presence of a putative TRAF-binding motif in the cytoplasmic region of human TNFRSF19.2 and mouse Troy (Supplementary Fig. S3), they were unable to bind TRAF2 and TRAF6; accordingly, overexpression or knockout of TNFRSF19 did not affect NF-κB activity (Figs. 3A and B and 5E). Our results are in accordance with previous studies conducted in other cell lines and mouse models (15, 36). Moreover, unlike most TNFRs, which are expressed in the immune system and play roles in innate and adaptive immunity, TNFRSF19 is not present in lymphoid tissues but is highly expressed in skin and hair follicles (15, 37). Collectively, these findings indicate that TNFRSF19 is functionally distinct from canonical TNFRs.

TNFRs utilize adaptor proteins to transduce and amplify receptor information to different cell fates. To identify the adaptors for TNFRSF19, we performed affinity purification of TNFRSF19. Unexpectedly, TGFβ type I receptor was captured as a specific binding partner for the ICD of TNFRSF19. Domain mapping revealed that TNFRSF19 bound to the kinase domain of TβRI, which overlapped with the binding region of R-Smads on TβRI (Fig. 3J; refs. 30–32). However, unlike TGFβRI/R-Smad complex formation, which is induced by TGFβ, TNFRSF19 constitutively associates with TβRI, thereby blocking the formation of the active TβRI/R-Smad complex and the downstream R-Smad/Co-Smad complex (Fig. 4). The constitutive interaction of TNFRSF19 with TβRI and the competition of R-Smads for binding to the kinase domain inhibit leaky activation of the receptor in the absence of TGFβ, thereby eliminating spurious signaling caused by receptor oligomerization-induced R-Smads recruitment in the absence of ligand. Indeed, depletion of TNFRSF19 results in active TβRI/R-Smad complex formation and TGFβ pathway activation, even in the absence of ligand (Figs. 4G and I and 5). To our knowledge, this is one of the very few examples of a repressor of the TGFβ pathway via direct binding to TGFβ receptors, and this finding may advance our understanding of the functional divergence of the TNFR superfamily.

Genetic alterations in the TGFβ pathway, such as the biallelic inactivation of TGFBRII, Smad2 and Smad3 mutations, and Smad4 deletion, are often found in human cancers (7, 38). Notably, targeted deletion of Smad4 in the head and neck epithelium of mice is sufficient to drive spontaneous head and neck squamous cell carcinomas (39). Loss of sensitivity to TGFβ-induced growth suppression has been found in NPC (11, 12). We identified TNFRSF19 as a key repressor of the TGFβ pathway that is often overexpressed in NPC, and high levels of TNFRSF19 are inversely correlated with TGFβ pathway activity in vivo; these evidences indicate that the inactivation of TGFβ signaling in NPC could result from the gain of function of TNFRSF19 rather than mutations in canonical TGFβ components. In addition to TNFRSF19, other NPC susceptibility loci we have identified in 2010 include MDS1-EVI1 and the CDKN2A-CDKN2B gene cluster (13). Interestingly, CDKN2B (p15INK4b) is a TGFβ target gene that participates in the mediation of TGFβ-induced cell-cycle arrest (40), and MDS1-EVI1 is an oncoprotein that suppresses TGFβ signaling by binding to Smad3 (41). The three NPC susceptibility genes seem to be involved in the regulation of TGFβ signal transduction at different levels. In addition, TGFβRII has been shown to be downregulated in more than 50% of NPC (10). Various genetic perturbations leading to the dysfunction of the same biological pathway are common in cancers. Further study of the TGFβ pathway status in a large cohort of patients with NPC and evaluation of its association with prognosis are warranted.

The discovery that TNFRSF19 antagonizes TGFβ signaling by interacting with the type I receptor and preventing its activation in a ligand-independent manner explains how TNFRSF19 controls the sensitivity of cells to TGFβ signals. It is conceivable that reactivation of TNFRSF19 in normal or premalignant epithelial cells can protect them against small amounts of autocrine and paracrine TGFβ and eventually cause uncontrolled cell growth. Although expression of TNFRSF19 in human normal NPE cells is not associated with transformation, such as soft agar growth and the formation of tumors in xenograft mouse models, TNFRSF19 does exert growth-promoting effects in NPE cells (Fig. 6B). Given that the etiology of NPC is complex and involves predisposed genetic and environmental factors, it is conceivable that multiple factors orchestrate the initiation of NPC. For example, EBV has been considered as another key player in NPC pathogenesis, and its latent to lytic switch is induced by TGFβ (42, 43). Thus, TNFRSF19 may also play roles in the establishment of persistent EBV infection in the early stage of NPC development. Future studies are needed to acquire a better understanding of gene–gene and gene–environment interactions in the development of NPC.

In contrast to human TNFRSF19, its mouse ortholog Troy has been well studied in mouse models. Using lineage tracing technology, Troy has been proposed as a stem cell marker for stomach (44), kidney (45), intestine (20), and brain (46) in mice. Troy also interacts with TGFβRI as its human orthologs (Supplementary Fig. S5A), and therefore, we generated tnfrsf19/troy KO mice using CRISPR-Cas9 technology (Supplementary Fig. S5B and S5C). The isolated mouse embryonic fibroblasts (MEF) with a homozygous deletion of the tnfrsf19 gene were more sensitive to the antiproliferative effects of TGFβ than the WT and heterozygous MEFs (Supplementary Fig. S5D and S5E) and exhibited enhanced TGFβ signaling activity (Supplementary Fig. S5F and S5G). However, the tnfrsf19−/− mice were alive at birth and grew into adulthood without gross abnormalities compared with their littermates, which is consistent with previous reports of tnfrsf19 KO mouse models generated by different gene-editing strategies and in different mice strains (36, 47). It could be a result of the redundancy of other TNFR family members such as Edar (36). Nonetheless, we showed that human TNFRSF19 is essential for the proliferation and tumorigenicity of NPC cells. Future assessment of chemically induced tumor formation, such as 4NQO (4-nitroquinoline 1-oxide)-induced head and neck squamous cell carcinoma is needed in the Troy KO mouse model. One the other side, given the gain-of-function property of TNFRSF19 in NPC, a transgenic mouse model would be more valuable to define the role of TNFRSF19 in NPC development.

The cause of aberrant expression of TNFRSF19 in cancers is still elusive. Tumor-specific expression of TNFRSF19 has been observed in NPC and other cancers (18, 19, 21). However, whether the genetic variations cause reactivation of TNFRSF19 remains uncertain. GWAS identified three SNPs within the TNFRSF19 region, rs1572072, rs9510787, and rs753955, but none of these SNPs are located in the coding region or near the promoter of TNFRSF19 (13, 14). Moreover, using quantitative RT-PCR in various cell lines with or without TNFRSF19 expression, we did not find concordance between the mRNA and protein expression levels (Supplementary Fig. S3C). The reasons for the poor correlation of mRNA and protein levels include posttranscriptional and posttranslational modifications, necessitating further investigations.

In conclusion, our results suggest that gain of function of TNFRSF19 in NPC inhibits the tumor-suppressive role of the TGFβ pathway and promotes tumorigenesis. These findings help pave the way toward a better understanding of the molecular basis and therapeutic potential of NPC.

No potential conflicts of interest were disclosed.

Conception and design: J.-X. Bei, Y.-X. Zeng, L. Feng

Development of methodology: C. Deng, H.-J. Zhang, L. Feng

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Deng, G.-P. He, H.-J. Zhang, Q.-S. Feng, J.-X. Bei

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Deng, Y.-X. Lin, X.-K. Qi, L. Feng

Writing, review, and/or revision of the manuscript: C. Deng, Y.-X. Lin, X.-K. Qi, J.-X. Bei, L. Feng

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G.-P. He, Y. Zhang, M. Xu, Q.-S. Feng

Study supervision: J.-X. Bei, L. Feng

We thank Drs. Junjie Chen and Kwok Wai Lo for helpful discussions. This work was supported by the National Key R&D Program of China (nos. 2016YFC0902000 to Y.-X. Zeng and 2017YFA0505600 to L. Feng), Major Project of Chinese National Programs for Fundamental Research and Development (no. 2013CB910301 to Y.-X. Zeng), the National Natural Science Foundation of China (nos. 81672980 to L. Feng and 81372882 to J.-X. Bei), the Key Program of the National Natural Science Foundation of China (no. 81430059 to Y.-X. Zeng), the Health & Medical Collaborative Innovation Project of Guangzhou City, China (no. 201803040003 to Y.-X. Zeng), and the Foundation of the Ministry of Science and Technology of Guangdong Province (no. 2015B050501005 to Y.-X. Zeng).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1.
Wei
WI
,
Sham
JS
.
Nasopharyngeal carcinoma
.
Lancet
2005
;
365
:
2041
54
.
2.
Razak
AR
,
Siu
LL
,
Liu
FF
,
Ito
E
,
O'Sullivan
B
,
Chan
K
.
Nasopharyngeal carcinoma: the next challenges
.
Eur J Cancer
2010
;
46
:
1967
78
.
3.
Dai
W
,
Zheng
H
,
Cheung
AK
,
Lung
ML
.
Genetic and epigenetic landscape of nasopharyngeal carcinoma
.
Chin Clin Oncol
2016
;
5
:
16
.
4.
Bei
JX
,
Zuo
XY
,
Liu
WS
,
Guo
YM
,
Zeng
YX
.
Genetic susceptibility to the endemic form of NPC
.
Chin Clin Oncol
2016
;
5
:
15
.
5.
Hildesheim
A
,
Wang
CP
.
Genetic predisposition factors and nasopharyngeal carcinoma risk: a review of epidemiological association studies, 2000–2011: rosetta stone for NPC: genetics, viral infection, and other environmental factors
.
Semin Cancer Biol
2012
;
22
:
107
16
.
6.
Derynck
R
,
Zhang
YE
.
Smad-dependent and Smad-independent pathways in TGF-beta family signalling
.
Nature
2003
;
425
:
577
84
.
7.
Massague
J
.
TGFbeta in Cancer
.
Cell
2008
;
134
:
215
30
.
8.
Wei
YS
,
Zhu
YH
,
Du
B
,
Yang
ZH
,
Liang
WB
,
Lv
ML
, et al
.
Association of transforming growth factor-beta1 gene polymorphisms with genetic susceptibility to nasopharyngeal carcinoma
.
Clin Chim Acta
2007
;
380
:
165
9
.
9.
Zhang
W
,
Zeng
Z
,
Fan
S
,
Wang
J
,
Yang
J
,
Zhou
Y
, et al
.
Evaluation of the prognostic value of TGF-beta superfamily type I receptor and TGF-beta type II receptor expression in nasopharyngeal carcinoma using high-throughput tissue microarrays
.
J Mol Histol
2012
;
43
:
297
306
.
10.
Lyu
X
,
Fang
W
,
Cai
L
,
Zheng
H
,
Ye
Y
,
Zhang
L
, et al
.
TGFbetaR2 is a major target of miR-93 in nasopharyngeal carcinoma aggressiveness
.
Mol Cancer
2014
;
13
:
51
.
11.
Xiao
J
,
Xiang
Q
,
Xiao
YC
,
Su
ZJ
,
Huang
ZF
,
Zhang
QH
, et al
.
The effect of transforming growth factor-beta1 on nasopharyngeal carcinoma cells: insensitive to cell growth but functional to TGF-beta/Smad pathway
.
J Exp Clin Cancer Res
2010
;
29
:
35
.
12.
Lo
AK
,
Dawson
CW
,
Lo
KW
,
Yu
Y
,
Young
LS
.
Upregulation of Id1 by Epstein-Barr virus-encoded LMP1 confers resistance to TGFbeta-mediated growth inhibition
.
Mol Cancer
2010
;
9
:
155
.
13.
Bei
JX
,
Li
Y
,
Jia
WH
,
Feng
BJ
,
Zhou
G
,
Chen
LZ
, et al
.
A genome-wide association study of nasopharyngeal carcinoma identifies three new susceptibility loci
.
Nat Genet
2010
;
42
:
599
603
.
14.
Hu
Z
,
Wu
C
,
Shi
Y
,
Guo
H
,
Zhao
X
,
Yin
Z
, et al
.
A genome-wide association study identifies two new lung cancer susceptibility loci at 13q12.12 and 22q12.2 in Han Chinese
.
Nat Genet
2011
;
43
:
792
6
.
15.
Hu
S
,
Tamada
K
,
Ni
J
,
Vincenz
C
,
Chen
L
.
Characterization of TNFRSF19, a novel member of the tumor necrosis factor receptor superfamily
.
Genomics
1999
;
62
:
103
7
.
16.
Locksley
RM
,
Killeen
N
,
Lenardo
MJ
.
The TNF and TNF receptor superfamilies: integrating mammalian biology
.
Cell
2001
;
104
:
487
501
.
17.
Spanjaard
RA
,
Whren
KM
,
Graves
C
,
Bhawan
J
.
Tumor necrosis factor receptor superfamily member TROY is a novel melanoma biomarker and potential therapeutic target
.
Int J Cancer
2007
;
120
:
1304
10
.
18.
Paulino
VM
,
Yang
Z
,
Kloss
J
,
Ennis
MJ
,
Armstrong
BA
,
Loftus
JC
, et al
.
TROY (TNFRSF19) is overexpressed in advanced glial tumors and promotes glioblastoma cell invasion via Pyk2-Rac1 signaling
.
Mol Cancer Res
2010
;
8
:
1558
67
.
19.
Loftus
JC
,
Dhruv
H
,
Tuncali
S
,
Kloss
J
,
Yang
Z
,
Schumacher
CA
, et al
.
TROY (TNFRSF19) promotes glioblastoma survival signaling and therapeutic resistance
.
Mol Cancer Res
2013
;
11
:
865
74
.
20.
Fafilek
B
,
Krausova
M
,
Vojtechova
M
,
Pospichalova
V
,
Tumova
L
,
Sloncova
E
, et al
.
Troy, a tumor necrosis factor receptor family member, interacts with lgr5 to inhibit wnt signaling in intestinal stem cells
.
Gastroenterology
2013
;
144
:
381
91
.
21.
Schon
S
,
Flierman
I
,
Ofner
A
,
Stahringer
A
,
Holdt
LM
,
Kolligs
FT
, et al
.
beta-catenin regulates NF-kappaB activity via TNFRSF19 in colorectal cancer cells
.
Int J Cancer
2014
;
135
:
1800
11
.
22.
Feng
L
,
Chen
J
.
The E3 ligase RNF8 regulates KU80 removal and NHEJ repair
.
Nat Struct Mol Biol
2012
;
19
:
201
6
.
23.
Feng
L
,
Huang
J
,
Chen
J
.
MERIT40 facilitates BRCA1 localization and DNA damage repair
.
Genes Dev
2009
;
23
:
719
28
.
24.
Lian
YF
,
Yuan
J
,
Cui
Q
,
Feng
QS
,
Xu
M
,
Bei
JX
, et al
.
Upregulation of KLHDC4 predicts a poor prognosis in human nasopharyngeal carcinoma
.
PLoS One
2016
;
11
:
e0152820
.
25.
Liang
Y
,
Zhong
Z
,
Huang
Y
,
Deng
W
,
Cao
J
,
Tsao
G
, et al
.
Stem-like cancer cells are inducible by increasing genomic instability in cancer cells
.
J Biol Chem
2010
;
285
:
4931
40
.
26.
Ding
Z
,
Roos
A
,
Kloss
J
,
Dhruv
H
,
Peng
S
,
Pirrotte
P
, et al
.
A novel signaling complex between TROY and EGFR mediates glioblastoma cell invasion
.
Mol Cancer Res
2018
;
16
:
322
32
.
27.
Kaye
KM
,
Devergne
O
,
Harada
JN
,
Izumi
KM
,
Yalamanchili
R
,
Kieff
E
, et al
.
Tumor necrosis factor receptor associated factor 2 is a mediator of NF-kappa B activation by latent infection membrane protein 1, the Epstein-Barr virus transforming protein
.
Proc Nat Acad Sci U S A
1996
;
93
:
11085
90
.
28.
Pan
G
,
Bauer
JH
,
Haridas
V
,
Wang
S
,
Liu
D
,
Yu
G
, et al
.
Identification and functional characterization of DR6, a novel death domain-containing TNF receptor
.
FEBS Lett
1998
;
431
:
351
6
.
29.
Wieser
R
,
Wrana
JL
,
Massague
J
.
GS domain mutations that constitutively activate T beta R-I, the downstream signaling component in the TGF-beta receptor complex
.
EMBO J
1995
;
14
:
2199
208
.
30.
Feng
XH
,
Derynck
R
.
A kinase subdomain of transforming growth factor-beta (TGF-beta) type I receptor determines the TGF-beta intracellular signaling specificity
.
EMBO J
1997
;
16
:
3912
23
.
31.
Chen
YG
,
Hata
A
,
Lo
RS
,
Wotton
D
,
Shi
Y
,
Pavletich
N
, et al
.
Determinants of specificity in TGF-beta signal transduction
.
Genes Dev
1998
;
12
:
2144
52
.
32.
Huse
M
,
Chen
YG
,
Massague
J
,
Kuriyan
J
.
Crystal structure of the cytoplasmic domain of the type I TGF beta receptor in complex with FKBP12
.
Cell
1999
;
96
:
425
36
.
33.
Sengupta
S
,
den Boon
JA
,
Chen
IH
,
Newton
MA
,
Dahl
DB
,
Chen
M
, et al
.
Genome-wide expression profiling reveals EBV-associated inhibition of MHC class I expression in nasopharyngeal carcinoma
.
Cancer Res
2006
;
66
:
7999
8006
.
34.
Dennler
S
,
Itoh
S
,
Vivien
D
,
ten Dijke
P
,
Huet
S
,
Gauthier
JM
.
Direct binding of Smad3 and Smad4 to critical TGF beta-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene
.
EMBO J
1998
;
17
:
3091
100
.
35.
Lo
KW
,
Huang
DP
.
Genetic and epigenetic changes in nasopharyngeal carcinoma
.
Semin Cancer Biol
2002
;
12
:
451
62
.
36.
Pispa
J
,
Pummila
M
,
Barker
PA
,
Thesleff
I
,
Mikkola
ML
.
Edar and Troy signalling pathways act redundantly to regulate initiation of hair follicle development
.
Hum Mol Genet
2008
;
17
:
3380
91
.
37.
Kojima
T
,
Morikawa
Y
,
Copeland
NG
,
Gilbert
DJ
,
Jenkins
NA
,
Senba
E
, et al
.
TROY, a newly identified member of the tumor necrosis factor receptor superfamily, exhibits a homology with Edar and is expressed in embryonic skin and hair follicles
.
J Biol Chem
2000
;
275
:
20742
7
.
38.
Siegel
PM
,
Massague
J
.
Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer
.
Nat Rev Cancer
2003
;
3
:
807
21
.
39.
Lu
SL
,
Herrington
H
,
Wang
XJ
.
Mouse models for human head and neck squamous cell carcinomas
.
Head Neck
2006
;
28
:
945
54
.
40.
Hannon
GJ
,
Beach
D
.
p15INK4B is a potential effector of TGF-beta-induced cell cycle arrest
.
Nature
1994
;
371
:
257
61
.
41.
Kurokawa
M
,
Mitani
K
,
Irie
K
,
Matsuyama
T
,
Takahashi
T
,
Chiba
S
, et al
.
The oncoprotein Evi-1 represses TGF-beta signalling by inhibiting Smad3
.
Nature
1998
;
394
:
92
6
.
42.
Liang
CL
,
Chen
JL
,
Hsu
YP
,
Ou
JT
,
Chang
YS
.
Epstein-Barr virus BZLF1 gene is activated by transforming growth factor-beta through cooperativity of Smads and c-Jun/c-Fos proteins
.
J Biol Chem
2002
;
277
:
23345
57
.
43.
Iempridee
T
,
Das
S
,
Xu
I
,
Mertz
JE
.
Transforming growth factor beta-induced reactivation of Epstein-Barr virus involves multiple Smad-binding elements cooperatively activating expression of the latent-lytic switch BZLF1 gene
.
J Virol
2011
;
85
:
7836
48
.
44.
Stange
DE
,
Koo
BK
,
Huch
M
,
Sibbel
G
,
Basak
O
,
Lyubimova
A
, et al
.
Differentiated Troy(+) chief cells act as reserve stem cells to generate all lineages of the stomach epithelium
.
Cell
2013
;
155
:
357
68
.
45.
Schutgens
F
,
Rookmaaker
MB
,
Blokzijl
F
,
van Boxtel
R
,
Vries
R
,
Cuppen
E
, et al
.
Troy/TNFRSF19 marks epithelial progenitor cells during mouse kidney development that continue to contribute to turnover in adult kidney
.
Proc Nat Acad Sci U S A
2017
;
114
:
E11190
E8
.
46.
Basak
O
,
Krieger
TG
,
Muraro
MJ
,
Wiebrands
K
,
Stange
DE
,
Frias-Aldeguer
J
, et al
.
Troy+ brain stem cells cycle through quiescence and regulate their number by sensing niche occupancy
.
Proc Natl Acad Sci U S A
2018
;
115
:
E610
E9
.
47.
Shao
Z
,
Browning
JL
,
Lee
X
,
Scott
ML
,
Shulga-Morskaya
S
,
Allaire
N
, et al
.
TAJ/TROY, an orphan TNF receptor family member, binds Nogo-66 receptor 1 and regulates axonal regeneration
.
Neuron
2005
;
45
:
353
9
.

Supplementary data