Leukocyte TLR5 deficiency inhibits atherosclerosis by reduced macrophage recruitment and defective T-cell responsiveness

Toll-like receptors (TLR) provide a critical link between innate and adaptive immunity, both important players in atherosclerosis. Since evidence for the role of TLR5 is lacking, we aimed to establish this in the immune axis of atherosclerosis. We assessed the effect of the TLR5-specific ligand Flagellin on macrophage maturation and T-cell polarisation. Next, we generated TLR5−/−LDLr−/− chimeras to study the effect of hematopoietic TLR5 deficiency on atherosclerosis formation. Flagellin stimulation did not influence wildtype or TLR5−/− macrophage maturation. Only in wildtype macrophages, Flagellin exposure increased MCP-1 and IL6 expression. Flagellin alone reduced T-helper 1 proliferation, which was completely overruled in the presence of T-cell receptor activation. In vivo, hematopoietic TLR5 deficiency attenuated atherosclerotic lesion formation by ≈25% (1030*103 ± 63*103 vs. 792*103 ± 61*103 μm2; p = 0.013) and decreased macrophage area (81.3 ± 12.0 vs. 44.2 ± 6.6 μm2; p = 0.011). In TLR5−/− chimeric mice, we observed lower IL6 plasma levels (36.4 ± 5.6 vs. 15.1 ± 2.2 pg/mL; p = 0.003), lower (activated) splenic CD4+ T-cell content (32.3 ± 2.1 vs. 21.0 ± 1.2%; p = 0.0018), accompanied by impaired T-cell proliferative responses. In conclusion, hematopoietic TLR5 deficiency inhibits atherosclerotic lesion formation by attenuated macrophage accumulation and defective T-cell responsiveness.

Atherosclerosis is a progressive multi-factorial disease affecting middle-sized and large arteries. It is the main culprit in the development of ischemic heart and cerebrovascular disease, together responsible for 29% of all deaths worldwide in 2013 1 . Although initially thought to be exclusively caused by lipids and risk factors such as diabetes, hypertension and smoking 2 , it is now generally acknowledged that inflammation plays a critical role in atherosclerosis 3,4 . Migration of leukocytes into the vessel wall is an essential step in atherosclerotic lesion initiation and progression 5,6 . Hence, inhibition or prevention of leukocyte recruitment towards the vessel wall may prevent or reduce plaque development. In addition to the recruitment of leukocytes, it has been shown that immune cell activation plays an important role in lesion development 7 . Best known leukocyte activators are the highly conserved toll-like receptors (TLR), which are expressed on both immune and non-immune cells 8 . Several TLR subfamilies have been linked to atherosclerotic lesion initiation and progression, of which TLR2 and TLR4 have been the most extensively studied 9 . Although TLR5 has been linked to inflammation in a variety of inflammatory diseases including atherosclerosis [10][11][12][13] , the role of hematopoietic TLR5 deficiency in atherosclerotic plaque formation remains unclear.
Scientific RepoRts | 7:42688 | DOI: 10.1038/srep42688 TLR5 is an extracellular receptor for bacterial Flagellin and ubiquitously expressed in almost all tissue types 14 . In addition to one or more exogenous stimuli, most TLRs also respond to specific endogenous ligands 15 . Whereas a broad variety of endogenous ligands have been described for most TLRs 9 , an equivalent for TLR5 is lacking. Since many exogenous TLR ligands are expressed in (human) atherosclerotic lesions 16 , Flagellin may also have a role in the development of atherosclerosis in the current study 17 . However, since TLR5 is an important contributor in diseases characterized by sterile inflammation (like cardiac ischemia-reperfusion injury 11 and rheumatoid artritis 18 ), it is to be expected that both endogenous as well as exogenous ligands can lead to TLR5-dependent inflammation.
In addition to specific tissue cells, TLR5 is also present on different immune cells [19][20][21] , of which macrophages and T-cells are of main importance in the context of atherosclerosis. Upon activation of the TLR5 receptor, MyD88 recruitment and activation of different intracellular kinases eventually leads to nuclear localization of NF-κB, resulting in a pro-inflammatory respons 22 . Since imbalanced inflammation is key in the development and progression of atherosclerosis, we investigated whether monocytes and T-cells lacking TLR5 differ in migratory and inflammatory behavior and whether hematopoietic TLR5 deficiency influences atherosclerotic plaque formation in vivo.
Stimulation of bone marrow derived macrophages. Bone marrow cells were isolated from the tibias and femurs of male TLR5 −/− and C57Bl/6 mice. To obtain bone marrow-derived macrophages, cells were cultured for 7 days in RPMI supplemented with 10% FCS, 100 U/mL penicillin/streptomycin, 0.1 mM nonessential amino acids, 1% pyruvate and 2 mM L-glutamine (Thermo Fisher Scientific) in the presence of 10 ng/mL macrophage colony-stimulating factor.
For in vitro stimulation, 0.5 × 10 6 macrophages were added per well (24-well cell culture plates (Greiner Bio-One)). Cells were stimulated with 1 ng/mL Flagellin Ultrapure (Invivogen, Toulouse, France) or 1 ng/mL LPS (Sigma Aldrich, Zwijndrecht, the Netherlands) as a positive control. After 24 hours, cells were harvested for RNA isolation.
Luminex. Circulating levels of IL6, IL17A and interferon-γ (IFN-γ) were determined by luminex according to the manufacturers protocol. A multiplex panel (eBioscience Mouse Th1/Th2 essential 6-plex and Mouse IL17A Simplex) was used in combination with the "Bio-Plex Multiplex system (Bio-Rad)" to perform luminex analyses.
Ex vivo proliferation and T-cell polarisation. Spleens were isolated from WT and TLR5 −/− chimeras to assess ex vivo proliferation or C57Bl/6 WT mice to assess T-cell polarization. Spleens were gently squeezed through a 70 μm mesh cell strainer (Becton Dickinson, San Diego, CA, USA) to obtain a single cell suspension. Cells were washed and resuspended in RPMI1640 (supplemented with 10% FCS, 20 mM L-glutamine, 100 U/ ml penicillin and 100 μg/mL streptomycin) and seeded at a density of 3 × 10 5 cells/well in a 96 well u-bottom cell culture plate (Greiner Bio One, Alphen aan den Rijn, the Netherlands). Cells were stimulated for 72 hours with medium alone, 1 ng/mL Flagellin Ultrapure, 1 ng/mL Flagellin Ultrapure in combination with 2 μg/mL anti-CD3 (eBioscience) or 2 μg/mL anti-CD3 in combination with 2 μg/mL anti-CD28 as a positive control. To assess proliferation, cells were incubated with 0.5 μCi [3 H]Thymidine during the last 16 hours of 3 days in culture. To quantify thymidine incorporation the cells were washed with PBS and lysed with 0.1 M NaOH and cell-associated Scientific RepoRts | 7:42688 | DOI: 10.1038/srep42688 radioactivity was determined by liquid scintillation counting. To assess T-cell polarisation, after 72 hours of stimulation, 10 μg/mL Brefeldin-A was added overnight to retain the proteins inside the cell. The next morning cells were harvested for flow cytometric analysis.

Intracellular flow cytometry.
Cultured T-cells were harvested, centrifuged and resuspended for intracellular staining according to the manufacturers protocol (eBioscience). In short, cells were resuspended in PBS supplemented with 1% normal mouse serum and stained for the cell surface marker CD4 (BD Biosciences) for 30 minutes in the dark at 4 °C. Cells were washed to remove unbound antibody and resuspended in IC fixation buffer. After 30 minutes, permeabilisation buffer was added, cells were centrifuged (1500 rpm, 5 min) and resuspended in permeabilisation buffer. Antibodies for intracellular antigens Tbet, IFN-γ, RORγt, IL17A and Foxp3 (eBioscience) were added and incubated for 45 minutes at room temperature in the dark. Cells were washed twice in permeabilisation buffer and resuspended in flow cytometry staining buffer for cell acquisition (FACSCanto, BD Biosciences).
Flow cytometry. Blood, spleen and bone marrow cells were isolated from WT and TLR5 −/− mice that had been fed a Western-type diet for 1 week. Bone marrow was isolated by flushing the femurs and tibias of WT and TLR5 −/− mice with PBS. Subsequently, the cell suspension was gently filtered through a 70 μm cell strainer to obtain a single cell suspension (70 μm pores, BD Bioscience). Spleens were harvested and single-cell suspensions of splenocytes were prepared by gently mincing the spleen through a cell strainer (70 μm pores, BD Bioscience). Bone marrow cells and splenocytes were incubated at 4 °C with erythrocyte lysis buffer (155 mM NH4CL in 10 mM Tris/HCL, pH 7.2) for 5 minutes. Cells were centrifuged for 5 minutes at 1500 rpm and then resuspended in lysis buffer to remove residual erythrocytes. Cells were washed twice with PBS. 50 μL whole blood, bone marrow (300.000 cells), or spleen cell suspensions (300.000 cells) were stained for the surface markers CD11B (eBioscience), Ly6G (1A8) (Biolegend), Ly6C (Biolegend) and CCR2 (R&D Systems) and incubated for 30 minutes in the dark. Subsequently cells were either washed in PBS and subjected to flow cytometric analyses (Gallios, Beckman Coulter) or prepared for intracellular staining.
For intracellular assessment of cytokine content, cells were washed to remove unbound antibody and resuspended in IC fixation buffer (4° Celsius). After 60 minutes, cells were centrifuged (1500 rpm, 5 min) and resuspended in permeabilisation buffer, washed once more and resuspended in permeabilisation buffer. Antibodies for intracellular antigens IL10 and IL12 (eBioscience, San Diego, CA, USA) were added and incubated for 30 minutes at room temperature in the dark. Cells were washed twice in permeabilisation buffer and resuspended in flow cytometry staining buffer for cell acquisition (Gallios, Beckman Coulter). FACS data were analyzed with Kaluza software (Beckman Coulter) and gated according to the strategy depicted in supplemental Fig. 1.

Histological analyses.
Cryosections of the aortic root (10 μm) were collected and stained with Oil-red-O for morphometric analyses. Corresponding sections on separate slides were used for immunohistochemical stainings. Macrophages were stained with an antibody directed against MOMA-2 (monoclonal rat IgG2b, dilution 1:50; Serotec, Oxford, United Kingdom). Goat anti-rat IgG-AP (dilution 1:100; Sigma, St. Louis, MO) was used as a secondary antibody and NBT-BCIP (Dako, Glostrup, Denmark) as an enzyme substrate. CD4 + T-cells were stained using CD4 (monoclonal rabbit, dilution 1:50; Cell Marque, Darmstadt, Germany). BrightVision poly-AP anti-rabbit (Immunologic, Duiven, The Netherlands) was used as a secondary antibody and liquid permanent red (DAKO) as an enzyme substrate. CD8 + T-cells were stained using CD8a (monoclonal rat IgG2a,λ, dilution 1:50, eBioscience). The ImmPRESS HRP anti-Rat IgG, Mouse adsorbed (Peroxidase) Polymer Detection Kit (Vector, Burlingame, CA, USA) was used as a secondary antibody and DAB (Sigma, St. Louis, MO, USA) as an enzyme substrate. Collagen was visualized by a Masson's trichrome staining (Sigma). The section with the largest lesion and at least four flanking sections were used to determine lesion size. At least two flanking sections were used to analyse macrophage, T-cell and collagen content. The percentage of macrophages and collagen in the lesions was determined by dividing the area positive for MOMA-2 or collagen by the total lesion surface area. Necrotic core size was determined in the Oil-red-O stained sections and defined as an a-cellular, lipid-poor region. CD4 + and CD8 + T-cells were manually counted and expressed as counts/mm 2 .
Histological analyses were performed in a blinded fashion using a Leica DMRE Microscope equipped with a Leica DC500 camera and Qwin quantification software (Leica, Rijswijk, the Netherlands). Histological analyses of CD4 + and CD8 + T-cell quantification were performed on an Olympus BX53 microscope equipped with a DP71 camera and CellSens quantification software (Olympus Corporation, Tokyo, Japan).

Statistical analysis.
Data are expressed as mean ± standard error of the mean (SEM). Data distribution was evaluated using a Shapiro-Wilk test. A two-tailed student's t-test was used to compare individual groups if data were normally distributed, whereas non-parametric data were analyzed using a Mann-Whitney U test. Multiple groups were compared by one-way ANOVA with a post-hoc two-tailed student's t-test between significantly different groups. A level of p < 0.05 was considered significant. Statistical analyses were performed using GraphPad Prism 6.
Ex vivo T-cell proliferation was performed to assess complete replacement of T-cells upon bone marrow transplantation (Fig. 4E). Proliferation under control, non-stimulatory conditions did not differ between WT and TLR5 −/− cells (p = 1.00). Proliferation upon stimulation with Flagellin alone or Flagellin in combination with αCD3 is displayed as fold induction to control proliferation. Stimulation with Flagellin led to an increased proliferation index in WT mice    . Atherosclerotic plaque size and composition. After 12 weeks of high-fat diet, the atherosclerotic plaque area was significantly larger in WT mice compared to TLR5 −/− mice (A). No difference between both groups was observed with respect to collagen content (B), yet the plaque of TLR5 −/− mice showed a decrease in macrophage influx (C,D). Although not significant, TLR5 −/− mice showed higher numbers of CD4 + (E) and CD8 + T-cells (F). The necrotic core expressed as a percentage of the plaque was significantly higher in WT mice (G); n = 12-13 per group (5-8 for CD4-8 staining). WT: wildtype, TLR5 −/− : toll-like receptor 5 knockout, *p < 0.05.

Discussion
Over the past decades there has been a widespread interest in the role of TLRs in the onset and progression of cardiovascular diseases. In atherosclerosis, TLR2 and TLR4 have been the most prominently studied 9 and evidence for the role of TLR5 has been previously shown 12,13 . In this regard, the current paper is the first to show that hematopoietic TLR5 deficiency attenuates atherosclerosis formation in LDLr −/− mice. In addition, the plaques of these mice contain less macrophages and a smaller necrotic core compared to mice that received WT bone marrow. These findings are in line with previous studies that show the effect of TLR5 in atherosclerotic plaque formation and inflammatory cell accumulation 12,13 . It also confirms the observation that inflammation -and TLRs in particular -are key drivers of atherosclerosis 3,9,12 .
In general, lack of TLRs is believed to decrease atherosclerotic burden through the impairment of pro-inflammatory signaling. In this respect, macrophages from WT mice showed an increased expression of MCP-1 and IL6 upon Flagellin stimulation, whereas no difference was observed upon stimulation of TLR5 deficient macrophages. Also, TLR5 −/− mice showed a decrease in circulating and bone-marrow derived Ly6C high monocytes and a decrease in pro-inflammatory cytokine production (IL10, IL12). Correspondingly, plasma levels of several pro-inflammatory cytokines were higher in WT mice than in TLR5 −/− bone marrow chimeras.
The fact that we observe a decrease in the Ly6C high monocyte pool in both the bone marrow and circulation suggests that monocyte recruitment to the plaque may be impaired. In addition, CCR2 expression on the bone marrow monocyte pool is significantly decreased in TLR5 deficient mice, which partly reflects previous research showing that CCR2 deficiency attenuates macrophage influx into the plaque 23 . Decreased macrophage CCR2 expression could have had a similar effect in the current manuscript. In addition to the impairment in pro-inflammatory signaling, this may very well be one of the primary mechanisms in the reduction of atherosclerotic plaque formation. These findings are in accordance with previous studies showing that TLR5 stimulation of mouse endothelial cells enhanced the expression of pro-inflammatory molecules and increased the adhesive and migratory capacity for monocytes 12 . In addition, TLR5 inhibition showed decreased serum levels of several pro-inflammatory cytokines 13 . Similarly, ApoE −/− TLR4 −/− mice on high-cholesterol feeding showed a reduction in serum MCP-1 levels, plaque size and macrophage infiltration when compared to ApoE −/− mice 24 . Again, this emphasizes the importance of migratory proteins (MCP-1, IL6, etc.) in atherogenesis and inflammatory cell deposition 5,6,23,25 .
In addition to TLR5 being abundantly expressed on cells of myeloid origin, we observed expression of TLR5 on CD4 + T-cells as well. When exposing T-cells to Flagellin in a non-inflammatory condition, an increase in the number of Tregs and a decrease in Th1-cells were observed. These data are similar to a previous study showing that TLR5 stimulation enhances the suppressive capacity of regulatory T-cells 26 . However, when performing these experiments under more inflammatory conditions by adding αCD3 to simulate antigen-T-cell receptor interaction, the suppressive effect of Tregs was lost and actually proliferation of pro-atherogenic Th1-and Th17-cells was stimulated. This implicates a pro-atherogenic effect of stimulation of TLR5 on T-cells in a pro-inflammatory environment, like observed in atherosclerosis. Indeed, this was reflected in vivo by lower numbers of Tregs, activated CD4 + CD25 + cells and Th1-cells and decreased IL6, IL17A and IFN-γ plasma levels in TLR5 −/− bone marrow chimeras. Despite this general effect on CD4 + T-cells, no significant reduction in the number of CD4 + or CD8 + T-cells in the atherosclerotic plaque was observed.
Apart from their importance as pattern recognition receptors (PRR) of exogenous (mostly bacterial) ligands, TLRs can also potently respond to specific disease-associated endogenous ligands 15 . Whereas a broad variety of endogenous ligands have been described for several TLRs 9 , an equivalent for TLR5 is lacking. Since many exogenous TLR ligands are expressed in (human) atherosclerotic lesions 16 , Flagellin may indeed be (partially) responsible for the results observed in the current study, implicating a role of (subclinical) infection in the development of atherosclerosis 17 . However, this phenomenon may be debated on by accumulating evidence showing that TLR5 also plays an important role in diseases characterized by sterile inflammation, such as cardiac ischemia-reperfusion injury 11 and rheumatoid artritis 18 .
Although it remains unclear which ligand is responsible for the observed effect in the current study, we do show the importance of hematopoietic TLR5 deficiency in the early stage of atherosclerosis formation. Whether this also holds true for a model of atherosclerosis with larger and more complex lesions remains to be elucidated. In this regard, hematopoietic TLR2 deficiency has also been shown to be important in the early stage of atherosclerosis formation, but was not evident when feeding mice a 1.5% instead of a 0.15% cholesterol-enriched diet 27 . Moreover, although baseline TLR5 deficiency was able to attenuate atherosclerotic plaque formation, from a clinical perspective it remains of interest to see whether TLR5 inhibition at a later stage would also be able to reduce or stabilize already manifest atherosclerosis.

Conclusion
In conclusion, the present manuscript is one of the first to show that hematopoietic TLR5 deficiency in mice attenuates atherosclerotic plaque development through decreased macrophage accumulation and defective T-cell responsiveness. Still, more knowledge has to be gathered on the possibility to antagonize TLR5 signaling to reduce or stabilize previously established atherosclerotic lesions.