Different Activation of TRAF4 and TRAF6 in Inflammatory Bowel Disease

In recent years, interests combining the exploration of tumor necrosis factor receptor-associated factor 4 (TRAF4) and TRAF6 in immune cells and transgenic mice are emerging. Although it has been found that TRAF4 and TRAF6 share the same TRAF binding sites, comprehensive study of TRAF4 and TRAF6 in inflammatory bowel disease (IBD) is still lacking. This paper shows similar and different expression patterns of TRAF4 and TRAF6 in patients with IBD. The results indicate that TRAF4 and TRAF6 are overexpressed in IBD. TRAF4 and TRAF6 play different roles in the pathogenesis of IBD. Moreover, TRAF4 may be an indicator of endoscopic disease activity of UC and TRAF6 preactivation can be detected in noninflamed colonic segments.


Introduction
Tumor necrosis factor receptor-associated factors (TRAFs) act as adapter molecules controlling signaling pathways, such as nuclear factor kappa B (NF-B), interleukin-1 receptor (IL-1R), toll-like receptor (TLR), and transforming growth factor-(TGF-) [1,2]. For a long time, TRAF6 has shown conserved function in activation of autoimmunity and in�ammation. It is important that TRAF6 contributes to the CD40-mediated activation of NF-B and c-Jun kinase (JNK). Association of TRAF6 with CD40 is essential for CD40-mediated IL-6 expression [3], which could explain the requirement for membrane-bound CD40 ligand to induce IL-6 production by immunocytes [4]. e survival, regulation, and activation of immunocyte and epithelial cell, signaling through cell surface receptors to activate NF-B and mitogenactivated protein kinases (MAPKs) through TRAF6, are critical regulations of immune response [5]. Unlike TRAF6, the molecular mechanism of TRAF4 in multiple signaling pathways triggered by TNFR-related proteins remains enigmatic. Moreover, the subcellular localization and functions of TRAF4 have been controversial for years. It has been indicated that TRAF4 augments NF-B activation through glucocorticoid-induced TNFR (GITR) expression on T cells, B cells, and macrophages [6]. As a unique TRAF family member mediating signal transduction by TNF, IL-1R, or TLR, it is found that TRAF4 acts as a positive effector of bone morphogenetic protein (BMP) and the TGF-signaling pathway [7].
e intestinal epithelium and immune cells in the gut establish active sites of immune reactivity. Breakdown of homeostasis between intestinal microbiota and the mucosal immune system, together with both environmental and genetic factors, leads to in�ammatory bowel disease (IBD). NF-B and TLR are considered as nodal points in the suppression and/or recruitment of immune responses in IBD [8]. Interestingly, although mechanisms of TRAFs in IBD are not yet fully studied, TRAF-related in�ammatory mediators as TGF-or CD40 play critical roles in a wide array of cellular functions in IBD [9,10].
Recent studies in immune cells and transgenic mice regarding the role of TRAF4 and TRAF6 have revealed that they share the same binding sites, yet comprehensive study of TRAF4 and TRAF6 in IBD is still lacking [11]. Based on the hypothesis that TRAF4 and TRAF6 may be activated prior to the clinical or endoscopic activation in patients with IBD, we sought to measure TRAF4 and TRAF6 expressions to explore their potential roles in IBD patients.

Patients and Samples.
Patients were enrolled according to clinical and endoscopic diagnosis. Patients with pregnancy, colorectal resection for UC, disease involving only small bowel, poor bowel preparation as visible area of intestinal mucosa <90%, or use of steroids, immunosuppressants, or in�iximab before colonoscopic sampling were excluded. No diagnosis altered aer at least 3 months of followup. Endoscopic score was evaluated using simpli�ed endoscopic score in Crohn's disease (SES-CD) or Baron score for patients with CD or UC, respectively. Healthy controls were included without sign or symptoms of bowel disease.
Human intestinal biopsies and blood samples were collected at Division of Gastroenterology and Hepatology, Shanghai Jiao-Tong University School of Renji Hospital Medicine, in accordance, with guidelines of the Research Ethics Committee of Renji Hospital, Shanghai Jiao Tong University, School of Medicine. All patients and healthy controls agreed to provide written consents.
Human tissue specimens were taken from both macroscopically in�amed and non-in�amed regions of the colon. Biopsies from the colon of healthy donors were also analyzed. Tissue specimens were put into liquid nitrogen within 10 minutes aer biopsy for protein extraction or kept in RNAlater (Qiagen) for RNA isolation. Human peripheral blood was separated into plasma and peripheral blood mononuclear cells (PBMCs). Plasma was obtained using commercially EDTA-treated tubes (Gongdong Medical Technology Co., Ltd.) and PBMCs were isolated according to Lymphoprep (Axis-Shield PoC AS, Norway) protocol. Brie�y, diluted blood was overlayed over 3 mL Lymphoprep and centrifuged at 800 ×g for 20 mins. PBMCs were removed from a distinctive band at the sample interface aer centrifugation. en, PBMCs were kept in RNAlater (Qiagen) for RNA isolation according to manufacturer's protocol.

Enzyme-Linked Immunosorbent Assay (ELISA).
Plasma was obtained following centrifugation of whole blood for 15 minutes at 2,000 ×g. Samples were stored at −80 ∘ C prior to analysis via Elisa. Samples were analyzed using kits against TRAF4 and TRAF6, according to the manufacturer's speci�cations (Lanji Biochemical and Diagnostics, Shanghai, China) and a microtiter plate reader was used to read absorbance at 450 nm. Experiments were performed in triplicate.
cDNAs were produced with PrimeScriptTM RT reagent Kit (Takara Biotechnology Dalian Co., Ltd.). Brie�y, reverse transcripts were incubated at 37 ∘ C for 15 minutes and 85 ∘ C for 5 seconds. SYBR Premix Ex Taq kit was purchased from TakaRa and real-time PCR reactions were done using a StepOne Plus device (Applied Biosystems) at 95 ∘ C for 10 seconds followed by 40 cycles of 95 ∘ C for 5 seconds and 60 ∘ C for 20 seconds according to instruction of the SYBR Premix Ex Taq kit. e expression levels of the target genes were normalized to GAPDH with 2-ΔΔCt method [12].

Western Blot Analysis.
For Western blot analysis, PMBCs and tissue samples were lysed in RIPA buffer (Sigma) containing protease inhibitors (Roche) and agitated on ice for 30 minutes. Protein concentrations were determined using Pierce BCA Protein Assay Kit (Pierce, Wohlen, Switzerland). Protein electrophoresis was performed according to the protocol of Mini-PROTEA III (Bio-Rad). Brie�y, proteins were separated in 10% polyacrylamide gels (Tris/glycine) and transferred onto polyvinylidene �uoride membrane (Millipore). Membranes were sequentially labeled by primary and secondary antibodies. Western blots were probed with antibodies against TRAF4 (rabbit polyclonal anti-TRAF4; 1 : 1000, Santa Cruz), TRAF6 (mouse monoclonal anti-TRAF6, 1 : 1000, Santa Cruz), and -actin (mouse monoclonal anti--actin; 1 : 2500, Santa Cruz). Secondary antibody was purchased from GE Healthcare life Science. Detection was enhanced by SuperSignal West Pico Chemiluminescent Substrate (Pierce). Experiments were performed in triplicate.

Statistical Analysis.
Statistical signi�cance was determined using GraphPad Prism 5.0 for Windows (GraphPad Soware, San Diego, CA, USA). A < was considered signi�cant with either ANO�A analysis or Tukey's multicomparison. A two-sided Fisher's exact test or 2 test was performed to analyze discrete variables.

Characteristics of Included Subjects. From February 2007
to February 2010, 40 CD patients, 42 UC patients, and 40 healthy controls were included in our present study. Patients with IBD indicated signi�cantly lower body mass index (BMI) than healthy controls ( < ). e UC group contained signi�cantly more smokers than CD patients (22 : 8, 2 ). ree patients with ileitis, 15 patients with ileocolitis, and 22 patients with colitis were enrolled in CD group. irteen patients with proctosigmoiditis, 20 patients with le sided colitis, and 9 patients with pancolitis were enrolled in UC group. Characteristics of included subjects were described in Table 1. , signi�cance is the difference from healthy controls; * * , signi�cance is the difference from patients with Crohn's disease; † values are medians and 25%-75% percentile, simpli�ed endoscopic score in Crohn's disease (SES-CD) is used to validate endoscopic severity in Crohn's disease, and Baron score is used to validate endoscopic severity in ulcerative colitis, respectively.

TRAF4 and TRAF6 Expressions in Plasma.
To investigate the diagnostic value of TRAF4 and TRAF6 in IBD, we detected levels of soluble TRAF4 and TRAF6 in plasma of IBD patients. We found that TRAF4 and TRAF6 were signi�cantly higher both in patients with CD and UC than in healthy controls (Figures 1(a) and 1(b)). However, only overexpression of soluble TRAF4 showed a signi�cantly positive correlation with endoscopic disease activity index (Baron score) in UC patients (spearman's , ). Furthermore, we observed that TRAF4 showed a signi�cantly diagnostic value in differentiating active IBD patients from healthy controls ( , Figures 2(a) and 2(b)). Although TRAF6 also showed a signi�cantly diagnostic value in differentiating active CD, UC from healthy controls, the lower area under the curve (AUC) predicted a less diagnostic value than TRAF4 (Figures 2(c) and 2(d)).

TRAF4 and TRAF6 Gene Expressions in Peripheral Blood
Mononuclear Cells. To identify gene expressions of TRAF4 and TRAF6 in PBMCs in patients with CD and UC, we isolated RNA from PBMCs. Similar to their expression in plasma, TRAF4 and TRAF6 showed signi�cantly higher levels both in patients with CD and UC than in healthy controls (Figures 1(c) and 1(d)) (all ).

Different Upregulation and Preactivation of TRAF4 and TRAF6 Expressions in Colonic Tissues.
Given that segmental changes can exhibit in�ammation in the colon in IBD patients and that intestinal segments in endoscopic remission can appear as histologic colitis, we determined the expressions of TRAF4 and TRAF6 both in in�amed and non-in�amed intestinal mucosae. Unfortunately, 3 patients with only ileitis and nine patients with pancolitis were excluded based on exclusion criteria. 5 CD patients and three UC patients were also excluded for refusal to biopsy. Quantitative real-time PCR (qRT-PCR) was used to determine the gene expressions of TRAF4 and TRAF6 in in�amed and non-in�amed intestinal mucosae of IBD. It was found that TRAF4 and TRAF6 expressions were signi�cantly higher in in�amed intestinal mucosa of patients compared to normal mucosa of healthy controls (all ) (Figures  3(a) and 3(b)). Interestingly, TRAF6 expressions were also signi�cantly higher in non-in�amed tissue of IBD patients than in healthy controls, which may indicate potential preactivation of TRAF6 in IBD (Figure 3(b)).
Western blotting was used to measure protein expressions of TRAF4 and TRAF6 in in�amed and non-in�amed intestinal mucosae of IBD ( Figure 4). Our data indicate that only TRAF6 expressions were signi�cantly higher in non-in�amed tissue of IBD patients than in healthy controls, although TRAF4 and TRAF6 protein expressions were sig-ni�cantly higher in in�amed intestinal mucosa of patients than in normal mucosa of healthy controls (all ) (Figures 3(c) and 3(d)). Similar to their gene expressions, TRAF4 and TRAF6 protein expressions were signi�cantly increased in in�amed intestinal mucosa compare to the non-in�amed mucosa or healthy controls. Also, TRAF6 protein expression was signi�cantly higher in in�amed intestinal mucosa of IBD patients compared to healthy controls.

Discussion
In the current study, we demonstrated that two members of the TRAF family, TRAF4 and TRAF6, were activated in patients with IBD. Although both TRAF4 and TRAF6 showed potentially diagnostic value in differentiating active CD and UC from healthy controls, only TRAF6 could be preactivated in non-in�amed tissue of IBD patients. Although TRAFs have similar overall structural features including a C-terminal receptor-binding domain and a leucine-zipper domain, their own structural difference leads to distinctive interaction with receptors [13,14]. Growing evidence indicates that TRAFs are regulated not only by their own structural features but also by the nature of their interactions, recruitment, or localization. e understanding of TRAF functions increased much more rapidly for TRAF6 than for TRAF4. Early overexpression studies clearly indicated that TRAF6 contributes to the CD40-mediated activation of NF-B and other signaling molecules [15]. NF-B activity has been upregulated in lamina propria immune cells and in epithelia of the in�amed gut in IBD [16]. e NF-B pathway regulates in�ammation, regulatory T-cell production, and DC function. However, activation of important transcriptional regulators including NF-B and the stress-activated protein kinases (SAPKs) mediated by TRAF6 requires binding to CD40. CD40 mutants and transgenic mice have been established to address the roles of TRAF6. e cytoplasmic CD40 binding domain for TRAF6 is necessary for the CD40-mediated activation of IL-6 production in monocytes and macrophages [17], and the proin�ammatory IL6 biologic networ� is upregulated in active IBD [18]. Moreover, experiments with immune cells or �broblasts isolated from TRAF6-de�cient mice indicate that TRAF6 is required for the CD40-mediated activation of not only NF-B but also JNK and p38 MAPK signals [19,20]. Although CD40-mediated JNK activation in B cells seems to require cytoplasmic TRAF6, TRAF6 mutants defective in CD40 binding were able to activate the JNK pathway and upregulate CD80, indicating that TRAF6 may be able to contribute to certain JNK signals without the binding of CD40 [21]. JNK and MAPK signaling pathways are involved in governing lymphocyte in�ux into the gut in IBD patients by regulating lymphocyte adhesion and transmigration [22]. Furthermore, only one study screened potential alterations of the TRAF6 gene in a large number of CD and UC patients but failed to identify apparent disease-causing mutations [23]. eir data suggested that TRAF6 may have essential roles in human biology that it might not tolerate any signi�cant structural alterations. TRAF4 possesses several unique characteristics that are different from other members of TRAFs. e subcellular localization and molecular mechanisms of action of TRAF4 are controversial. Few studies have implicated TRAF4 in IB� or colitis. �endritic cells from TRAF4-de�cient mice exhibited reduced migration both in vitro and in vivo experiments [24]. is result indicates that TRAF4 could participate in immune functions by facilitating immune cell migration. Interestingly, TRAF4 increases NF-B activation through the GITR via a TRAF-binding site located in the cytoplasmic domain of GITR. is domain is responsible for the inhibition of Treg cells and the promotion of T-cell activation. Although no NF-B activity has ever been detected via TRAF4 overexpression, TRAF4 has been implicated as an upstream molecule that regulates the JNK pathway via interaction and activation of Misshapen (Msn), a member of the SPS1 protein kinase family [25]. Moreover, TRAF4 positively regulates transforming growth factor (TGF)-via potentiating bone morphogenetic protein (BMP) and Nodal signaling [26]. TGF-plays a role in the pathogenesis of IBD by activating its speci�c receptors [27]. All TRAFs except TRAF4 have been identi�ed to directly or indirectly act with CD40. Accordingly, TRAF4 is upregulated in B cells following CD40 signaling, which suggests that TRAF4 affected downstream of CD40 pathway [28]. is may explain our result that TRAF4 expressions were not signi�cantly different in non-in�amed tissue of IBD patients compared to healthy controls. Our understanding of TRAF4 and TRAF6 functions in the present study indicates that overlapping and unique functions will ultimately be attributed to each of the TRAFs. To our knowledge, this paper may be the �rst one to combine analyzing TRAF4 and TRAF6 in IBD patients. However, there are still some limits to the present study. First, the present study did not include histological criteria to distinguish non-in�amed tissue from in�amed tissue, which may confuse the microscopically in�amed tissues with non-in�amed tissues. Second, the present study could not include laboratory data as ESR or CRP to assess activity of disease, which may lead to the potential bias of overreliance on clinical manifestations and endoscopic assessment. Collectively, our data showed similar and different expression patterns of TRAF4 and TRAF6 in patients with IBD. To date, correlations between TRAF4 and TRAF6 have not been explored to elucidate signaling cascades. Future studies to determine multifaceted roles may offer targets in the treatment of IBD.