Conceived and designed the experiments: PS AM CvzM KZ CB AZ. Performed the experiments: PS AM CW CC DW MB. Analyzed the data: PS AM CW NV CC DW MB AZ. Contributed reagents/materials/analysis tools: PS AM CvzM KZ CB AZ. Wrote the paper: PS AM CW DW MB AZ.
The authors have declared that no competing interests exist.
Tumor necrosis factor receptor-associated factors (TRAFs) are important signaling molecules for a variety of pro-atherogenic cytokines including CD40L, TNF α, and IL1β. Several lines of evidence identified TRAF6 as a pro-inflammatory signaling molecule
Lethally irradiated low density lipoprotein receptor-deficient mice (TRAF6+/+/LDLR−/−) were reconstituted with TRAF6-deficient fetal liver cells (FLC) and consumed high cholesterol diet for 18 weeks to assess the relevance of TRAF6 in hematopoietic cells for atherogenesis. Additionally, TRAF6+/−/LDLR−/− mice received TRAF6-deficient FLC to gain insight into the role of TRAF6 deficiency in resident cells. Surprisingly, atherosclerotic lesion size did not differ between the three groups in both aortic roots and abdominal aortas. Similarly, no significant differences in plaque composition could be observed as assessed by immunohistochemistry for macrophages, lipids, smooth muscle cells, T-cells, and collagen. In accord, in a small clinical study TRAF6/GAPDH total blood RNA ratios did not differ between groups of patients with stable coronary heart disease (0.034±0.0021, N = 178), acute coronary heart disease (0.029±0.0027, N = 70), and those without coronary heart disease (0.032±0.0016, N = 77) as assessed by angiography.
Our study demonstrates that TRAF6 is not required for atherogenesis in mice and does not associate with clinical disease in humans. These data suggest that pro- and anti-inflammatory features of TRAF6 signaling outweigh each other in the context of atherosclerosis.
Atherosclerosis, one of the leading causes of morbidity and mortality in Western countries
Tumor necrosis factor (TNF) receptor-associated factors (TRAFs) are intracellular adaptor proteins, which channel signaling for members of the TNF-/interleukin-1 (IL-1)-/toll-like-receptor (TLR)-superfamily such as TNFα, CD40L, and IL-1β, proteins known to promote inflammation and atherosclerosis
Data on the function of TRAF6 in vascular disease are scarce. Human monocytes transfected with TRAF6 binding protein inhibiting the association of TRAF6 with CD40, failed to activate ERK1/2, IKK, and cytokine production after stimulation with CD40L
Exploration of atherogenesis in TRAF6-deficient mice was hampered by the limited viability of homozygous TRAF6-deficient mice
LDLR−/− mice (CD45.2-positive/CD45.1-negative, A) were lethally irradiated (2×450 cGy) and reconstituted with fetal liver cells of 6–8 week old CD45.1-positive/CD45.2-negative mice (B). After an interval of 4 weeks, peripheral blood cells were immunostained with CD45.1-PE and CD45.2-FITC (exemplary donors are shown in C and D) or CD45.1-PE in combination with CD19-PECy (B-cell marker, E), CD3-APC (T-cell marker, F), and CD11b-FITC (monocytic marker, G) antibodies and analyzed by FACS.
Lethally irradiated 6 week old TRAF6+/+/LDLR−/− mice received TRAF6-deficient (N = 4) or competent fetal liver cells (N = 4), TRAF6+/−/LDLR−/− mice received TRAF6-deficient fetal liver cells (N = 2) only. RNA was isolated from spleens. Ratios of TRAF6/GAPDH mRNA as assessed by quantitative RT-PCR are shown as mean±SEM (A). RNA was isolated from spleens (B) and aortas (C) of TRAF6+/+/LDLR−/− (N = 5 and 3) TRAF6+/−/LDLR−/− mice (N = 3 and 2). Ratios of TRAF6/GAPDH mRNA as assessed by quantitative RT-PCR are shown as mean±SEM.
To verify the assumption that TRAF6 heterozygous mice express lower levels of TRAF6, we also analyzed TRAF6 expression in spleens and aortas of TRAF6+/+/LDLR−/− and TRAF6+/−/LDLR−/− animals without transplantation. In both cases, TRAF6 heterozygous mice expressed lower TRAF6/GAPDH mRNA ratios (
After 4 weeks allowed for reconstitution, mice consumed a high cholesterol diet for 18 weeks. At the beginning of the study no significant difference in body weight, plasma cholesterol, triglyceride levels, and phenotype was observed (
Lethally irradiated 6 week old TRAF6+/+/LDLR−/− mice received TRAF6-deficient (hatched bars, N = 22) or competent fetal liver cells (white bars, N = 27), TRAF6+/−/LDLR−/− mice received TRAF6-deficient fetal liver cells (black bar, N = 23) only. Subsequently, all groups consumed high cholesterol diet (HCD) for 18 weeks. Weights (A), plasma cholesterol levels (B), triglycerides (C), and leukocyte counts (D) were taken before and after HCD.
TRAF6+/+/LDLR−/− | TRAF6+/−/LDLR−/− | TRAF6+/+/LDLR−/− in TRAF6+/+/LDLR−/− | TRAF6−/−/LDLR−/− in TRAF6+/+/LDLR−/− | TRAF6−/−/LDLR−/− in TRAF6+/−/LDLR−/− | |
CD4+% of T- cells | 51.2±6.7 | 55.9±3.5 | 65.9±1.1 | 63.6±2.5 | 65.0±1.3 |
CD8+% of T- cells | 37.0±2.6 | 38.1±3.5 | 28.2±1.0 | 28.3±5.0 | 28.5±3.4 |
T- reg% of T- cells | 3.7±2.3 | 3.4±2.3 | 3.4±2.4 | 4.5±2.0 | 6.6 |
B- cells% of leukocytes | 23.9±13.0 | 23.8±4.7 | 20.6±2.2 | 19.5±4.1 | 20.9 |
Inflammatory monocytes % of total monocytes | 65.6±19.7 | 67.4±13.8 | 65.0±12.2 | 63.4±23.4 | 70.6 |
Also, TRAF6+/+/LDLR−/− mice receiving TRAF6-deficient FLC (496±46 mg/dl, N = 21) had significantly lower plasma cholesterol levels than those receiving TRAF6-competent FLC (730 mg/dl±36 mg/dl, N = 23, p = 0.002) after 18 weeks of HCD while no significant difference in cholesterol levels could be detected when compared with TRAF6+/−/LDLR−/− mice receiving TRAF6-deficient FLC (620 mg/dl±41 mg/dl, N = 22, p = 0,0513
Intimal lesion size in the aortic roots of TRAF6+/+/LDLR−/− (0.199 mm2±0.04 mm2, N = 21, p = 0.2825) and TRAF6+/−/LDLR−/− mice (0.198 mm2±0.03 mm2 N = 22, p = 0.7178) reconstituted with TRAF6-deficient FLC did not differ compared with TRAF6+/+/LDLR−/− mice receiving TRAF6-competent FLC (0.213 mm2±0.03 mm2, N = 21,
Lethally irradiated 6 week old TRAF6+/+/LDLR−/− mice received TRAF6-deficient (hatched bars, N = 21) or competent fetal liver cells (white bars, N = 21), TRAF6+/−/LDLR−/− mice received TRAF6-deficient fetal liver cells (black bars, N = 22) only. Subsequently, all groups consumed high cholesterol diet (HCD) for 18 weeks. Intimal lesion area of the atherosclerotic plaques in aortic roots was quantified. Pooled mean intimal lesion area ± SEM are shown as graphs in the upper panel (A), representative sections stained with oil red O below (B).
Lethally irradiated 6 week old TRAF6+/+/LDLR−/− mice received TRAF6-deficient (hatched bars, N = 21) or competent fetal liver cells (white bars, N = 21), TRAF6+/−/LDLR−/− mice received TRAF6-deficient fetal liver cells (black bars, N = 22) only. Subsequently, all groups consumed high cholesterol diet (HCD) for 18 weeks. Sections of the aortic roots were analyzed for macrophage- (A), lipid- (B), smooth muscle cell- (C), collagen (D), and T cell-content (E). Mac-3-, oil-red-O-, α-actin-, picrosirius red, and CD4-positive staining in per cent of total wall area is displayed as mean±SEM.
Since effects on atherogenesis may be site- and stage-specific we also analyzed lesion formation
Lethally irradiated 6 week old TRAF6+/+/LDLR−/− mice received TRAF6-deficient (hatched bars, N = 10) or competent fetal liver cells (white bars, N = 10), TRAF6+/−/LDLR−/− mice received TRAF6-deficient fetal liver cells (black bars, N = 10) only. Subsequently, all groups consumed high cholesterol diet (HCD) for 18 weeks. Abdominal aortas were fixed in formalin, pinned, and stained with oil red O to detect lipid deposition. Oil red O-positive staining in per cent of total area is shown as mean±SEM in the upper panel (A), representative images are shown below (B).
Since TRAF6+/+/LDLR−/− and TRAF6+−/LDLR−/− mice showed decreased levels of cholesterol at the end of the study (
Bone marrow-derived macrophages were isolated from 6 week old TRAF6+/+/LDLR−/− mice receiving TRAF6-deficient (hatched bars, N = 4) or competent fetal liver cells (white bars, N = 4), were stimulated with 4 mg/ml cholesterol or 0.75 µM palmitic acid, and assayed for expression of IL-6 (A), MCP-1 (B), TNFα (C), and IL12-p70 (D) by cytometric bead array.
Since we previously demonstrated overexpression of TRAF6 protein in human plaques, we tested the hypothesis that TRAF6 expression associates with acute or chronic coronary heart disease in humans
325 patients undergoing coronary angiography were divided into three groups: no coronary heart disease (No CHD, N = 77), stable coronary heart disease (CHD, N = 178), and acute coronary syndromes (ACS, N = 70). TRAF6 and GAPDH mRNA was analyzed by quantitative real-time PCR in total blood RNA. Results are presented as mean±SD computed from the average measurements obtained from each group.
No CHD (n = 77) | CHD (n = 178) | ACS (n = 70) | |
Age (years) | 62±1 | 65±0,6 |
64±1 |
BMI (kg/m2) | 28.2±0.5 | 27.5±0.3 | 27.8±0.5 |
% men | 71 | 83 | 79 |
% diabetes | 7,6 | 24.6 |
25.7 |
% hypertension | 12,8 | 40.2 |
15.5 ∫ |
% smokers | 9,1 | 27.4 |
14 |
SBP | 131±1 | 131±1 | 133±2 |
DBP | 77±1 | 77±1 | 79±1 |
Glucose | 110±4 | 114±3 | 120±6 |
Cholesterol | 202±7 | 182±21 |
196±9 |
Triglycerides | 150±19 | 151±9 | 171±18 |
LDL | 116±5 | 96±4 |
99±7 |
VLDL | 35±4 | 33±2 | 40±3 ∫ |
HDL | 52±3 | 48±1 | 53±5 |
Creatine kinase | 109±9 | 117±12 | 709±155 |
Pro-BNP | 389±145 | 534±98 | 1330±509 |
*, p<0.05 vs controls.
∫, p<0.05 vs CHD.
The current study presents the novel and unexpected finding that TRAF6 deficiency on FLC-derived cells does not alter atherogenesis in either TRAF6+/+/LDLR−/− or TRAF6+/−/LDLR−/− mice. Our data challenge the common view of TRAF6 as pro-inflammatory signaling molecule in the context of atherosclerosis.
Multiple reports identified TRAF6 as positive regulator of CD40L- and IL-1 but not TNFα-induced NFκB signaling
Recently, Lutgens
Several studies implicate TRAF6 in the recruitment and function of mononuclear cells
Since we previously observed increased expression of TRAF6 in human carotid plaques we tested whether TRAF6 mRNA levels in blood associate with chronic or acute coronary heart disease. We observed no significant difference in TRAF6 expression between the tested groups: no coronary heart disease (no CHD), stable coronary heart disease (CHD) and acute coronary syndrome (ACS), corroborating our findings obtained in mice.
Our study has several limitations: First, we cannot rule out that the lower cholesterol levels observed in animals receiving TRAF6-deficient bone marrow mask a putative effect of TRAF6 deficiency. This is, however, unlikely since previous reports mainly suggest a pro-inflammatory function of TRAF6 and therefore one would expect reduced levels of atherosclerosis in mice deficient in TRAF6. Lower cholesterol levels also predispose for smaller lesions. Therefore, these should not impair the results of our study. Furthermore, we found no evidence for a TRAF6-dependent difference in inflammatory reactivity of macrophages, which could also mask a putative effect of TRAF6 in our model. Secondly, since γ-irradiation itself profoundly influences the development of atherosclerotic lesions we cannot rule out that this affects our results
In summary, we present the novel and surprising finding that TRAF6 deficiency does not influence atherogenesis in mice and does not associate with atherosclerosis in humans. Therefore, overall targeting of TRAF6 may not be a promising treatment strategy for atherosclerosis and probably also other chronic inflammatory diseases.
All animal procedures were approved by the Animal Board of Freiburg (Regierungspräsidium Freiburg, permit number G05/41). TRAF6+/− mice were kindly provided by Dr. T. W. Mak and fully backcrossed to C57/BL6 background as verified by background strain characterization at Jackson laboratories. Mice were crossbred with LDLR−/− mice (Jackson) to generate TRAF6+/−/LDLR−/− and TRAF6+/+/LDLR−/− mice. Genotyping of each mouse used polymerase chain reaction employing the following primers: LDLR, 5′-CCA TAT GCA TCC CCA GTC TT-3′ (common primer), 5′-GCG ATG GAT ACA CTC ACT GC-3′ (wild-type primer), 5′-AAT CCA TCT TGT TCA ATG GCC GAT C-3′ (mutant primer); TRAF6, 5′-CTG CAG TGA AAG ATG ACA GCG TGA GT-3′ (wild-type) ; 5′-
Fetal livers were obtained 17 days after conception from fetuses of TRAF6+/−/LDLR−/− mice. One arm was used for genotyping. Four week-old male TRAF6+/+/LDLR−/− and TRAF6+/−/LDLR−/− recipient mice were lethally irradiated with two doses of 450 cGy at a 6 h interval (Gammacell Exactor 40). Fetal livers were suspended with a pipette, filtered through a 100 µm cell strainer (BD bioscience), centrifuged, resuspended, and injected at 106 cells/300 µl into the tail vein. Transplanted mice received chow diet for four weeks allowing for reconstitution. To verify reconstitution, fetal liver cells from CD45.1 mice were transplanted into CD45.2 mice and reconstitution rates were assessed by FACS after 4 weeks.
FACS analysis was performed as described previously
After four weeks of reconstitution recipient mice consumed a high-cholesterol diet (HCD) for 18 weeks (Ssniff based on Research Diets D12108). Subsequently, mice were euthanized, hearts and aortas were removed, and histologically prepared as described previously
Blood samples were collected by retro-orbital puncture before and at the end of HCD after an overnight starvation. Serum total cholesterol and triglyceride concentrations were assayed by commercially available enzymatic assays according to the manufacturer's protocols (CHOL-H L and Triglyceride L-Type from WAKO).
Harvested organs were stored in RNAlater (Qiagen) at −80°C. RNA was extracted from murine aortas and spleens using TRIzol Reagent (invitrogen) utilizing a modified protocol. Homogenization was performed using a rotor-stator dispergator (IKA). 1 µg of total RNA was transcribed into cDNA using the Transcriptor 1st Strand cDNA Synthesis Kit (Roche). The cDNA obtained was subjected to quantitative real-time PCR with a Roche LightCycler 480 using the LightCycler 480 SYBR Green I Master (Roche). mGapdh served as endogenous control. Amplification of potential genomic DNA contamination was ruled out by using intron-spanning primer pairs and subsequent reassurance through melting curve analysis. The following primers were employed: mGapdh: 5′-TGC ACC ACC AAC TGC TTA G-3′ (foward) and
Frozen sections were air dried, fixed in 10% formalin for 10 min, washed, submerged in 100% propyleneglycol (Fisher scientific), incubated in oil red O (Sigma-Aldrich) for 25 min at 60°C, dipped into 0.25% ammonia H2O (EM Science), and coverslipped with glycerol gelatine (Sigma-Aldrich). Abdominal aortas were fixed with 10% formalin, opened longitudinally, pinned, stained with oil red O solution (2.5 h, RT), and washed with 85% propylene glycol.
Cryostat sections (6 µm) of mouse aortic roots were air-dried, fixed in acetone at −20°C, incubated with 0.3% H2O2, blocked with 4% rabbit serum (Vector Laboratories), incubated with primary antibodies (anti-mac-3, anti-α-actin, and anti-CD4 from Pharmingen), incubated with corresponding secondary antibodies (Vector Laboratories and Sigma-Aldrich), washed, incubated with avidin-biotin complex (Vector Laboratories), developed with 3-amino-9-ethylcarbazole (DAKO), counterstained with hematoxylin (Sigma-Aldrich), and coverslipped with glycerol gelatine (Sigma-Aldrich) as described previously
Air dried and formalin-fixed frozen sections were incubated for 3 h in 0.1% solution of picrosirius red (Polysciences) in saturated aqueous picric acid (Sigma-Aldrich). Slides were rinsed twice in 0.01 N HCl and distilled water, dehydrated in 70%, 95%, 100% ethanol, incubated in xylene, and mounted in Permount (Vector Laboratories). Picrosirius red staining was analyzed by polarization microscopy (Edmund Industrial Optics).
6 weeks after transplantation mice were euthanized and bones were removed. Bone marrow was flushed out, cells were cleaned up using ficoll (Biochrom AG, Biocoll Separating Solution), and differentiated to macrophages with 50 ng/ml and subsequently 25 ng/ml M-CSF for 3 days each. Finally, macrophages were stimulated after 24 h starvation with 0.75 µM palmitic acid (Sigma Aldrich) respectively 4 mg/ml cholesterol diluted in ethanol and BSA (Bovine Serum Albumin). The appropriate amount of ethanol and BSA was added to the control. Supernatant was collected and analyzed with cytometric bead array as previously described
Morphometric calculations of the tissue sections were analyzed by a blinded observer using image pro plus 5.1 (MediaCybernetics). Data were presented as mean±SEM. Comparison of the respective study groups used the Student's two-tailed t-test. The p-value refers to the control group and P<0.05 was considered statistically significant.
325 patients undergoing coronary angiography were included in the Tumor Necrosis Factor Receptor associated factors in Cardiovascular Risk Study (TRAFICS) approved by the local Institutional Review Board (ethic committees: Ethikkommission der Albert- Ludwigs- Universität Freiburg, permit numbers EK 57/06 and EK 379/09). After written informed consent, blood was drawn from all patients and total blood RNA was isolated by Qiagen PAXgene blood RNA kit according to the manufacturer's instructions. Demographic and clinical characteristics were documented. Patients were divided into three groups: no coronary heart disease (No CHD), stable coronary heart disease (CHD), and acute coronary syndrome (ACS). 1 µg RNA was transcribed into cDNA with use of the transcriptor 1st strand cDNA synthesis kit (Roche). The cDNA obtained was subjected to quantitative real time-PCR with a Roche Light Cycler using the Light Cycler 480 SYBR Green I Master (Roche). As endogenous control, GAPDH was employed. Conditions for quantification of TRAF6 mRNA were
We thank Dr. Tak Wah Mak from the University of Alberta, Toronto for providing us with TRAF6+/− mice, Dr. Peter Libby from Brigham and Women's Hospital in Boston for financial support, expert advice on experimental design, and critical review of the manuscript, Dr. Michael Reth from the Max-Planck-Institute in Freiburg for co-mentoring the PhD thesis of Dr. Anna Missiou, Dr. Uwe Schönbeck from Pfizer, USA for important and fruitful discussions on the topic, Dr. Marie Follow for her expert advice on qPCR, and Dr. Gabriele Niedermann from the Department of radiation therapy of the University of Freiburg for providing the irradiator.
We thank Sandra Ernst, Benjamin Sommer, and Christian Münkel for their technical support.