If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Attenuated IFNγ response in macrophages with interferon gamma signaling deficiency.
Absence of myeloid interferon gamma signaling does not affect atherosclerosis.
Myeloid interferon gamma receptor deficiency is dispensable in liver inflammation.
Background and aims
Atherosclerosis is a chronic lipid-driven inflammatory disease of the arterial wall. Interferon gamma (IFNγ) is an important immunomodulatory cytokine and a known pro-atherosclerotic mediator. However, cell-specific targeting of IFNγ or its signaling in atherosclerosis development has not been studied yet. As macrophages are important IFNγ targets, we here addressed the involvement of myeloid IFNγ signaling in murine atherosclerosis.
Bone marrow was isolated from interferon gamma receptor 2 chain (IFNγR2) wildtype and myeloid IFNγR2 deficient mice and injected into lethally irradiated LDLR-/- mice. After recovery mice were put on a high fat diet for 10 weeks after which atherosclerotic lesion analysis was performed. In addition, the accompanying liver inflammation was assessed.
Even though absence of myeloid IFNγ signaling attenuated the myeloid IFNγ response, no significant differences in atherosclerotic lesion size or phenotype were found. Also, when examining the liver inflammatory state no effects of IFNγR2 deficiency could be observed.
Overall, our data argue against a role for myeloid IFNγR2 in atherosclerosis development. Since myeloid IFNγ signaling seems to be nonessential throughout atherogenesis, it is important to understand the mechanisms by which IFNγ acts in atherogenesis. In the future new studies should be performed considering other cell-specific targets.
]. Cell-specific targeting of IFNγ or its signaling in atherosclerosis development has not been performed yet. As macrophages are both numerous in an atherosclerotic lesion and are important IFNγ targets, we here addressed the involvement of myeloid IFNγ signaling in a mouse model of atherosclerosis by specifically deleting the interferon gamma receptor 2 chain (IFNγR2) on myeloid cells. Since the IFNγR2 chain of the IFNγR is the signal transducing component of this receptor complex and the limiting factor upon IFNγ stimulation [
], deleting this receptor chain would fully ablate IFNγ signaling. We could demonstrate for the first time that in contrast to systemic IFNγ or systemic IFNγ signaling, myeloid IFNγ signaling is dispensable for atherosclerosis development in mice.
2. Materials and methods
All mice were on a C57Bl/6 background. Female LDLR-/-, IFNγR2fl/fl (IFNγR2wt) and LysMCre-IFNγR2fl/fl (IFNγR2del) [
] littermates were house bred at the breeding facility of the Academic Medical Center (Amsterdam, The Netherlands) according to guidelines of the Animal Research Ethics Committee of the University of Amsterdam. IFNγR2−/− (IFNγR2KO) mice were generated at the Helmholtz Institute for Infection Research, Braunschweig, Germany. All animal experiments were approved by the Animal Welfare Committee of the University of Amsterdam.
2.2 Bone marrow transplantation
42 10 week old female LDLR-/- mice were placed in filter-top cages and were provided with antibiotics water (autoclaved tap water containing neomycin (100 mg/l, Sigma, Zwijndrecht, The Netherlands) and polymyxin B sulphate (60.000 U/l, Invitrogen, Bleiswijk, The Netherlands)) one week before the transplantation until 5 weeks after the bone marrow transplantation. The animals received 2 times 6 Gy total body irradiation on two consecutive days. Bone marrow was isolated from IFNγR2wt and IFNγR2del mice and 107 cells were injected intravenously per LDLR-/- mouse. Six weeks after the transplantation, mice were put on a high fat diet (0.15% cholesterol, 16% fat, Arie Blok Diets, The Netherlands) for a period of 10 weeks. Bone marrow transplantation efficiency was determined by q-PCR for the LDLR on DNA isolated from blood. Blood samples were taken before start of the diet and at 9 weeks of diet for lipid profiling and to determine blood leukocyte counts.
2.3 Histological analyses of atherosclerosis and hepatic inflammation
Upon sacrifice, hearts were taken out and frozen in tissue-tek (Dako, Eindhoven, The Netherlands). Hearts were cut into sections of 7 μm at the aortic root, after which serial cross-sections of every 42 μm were stained with toluidin blue. Plaque size was measured in a blinded fashion using Adobe Photoshop software and was presented as the sum of three valves. On the toluidine blue stained slides necrosis was measured (as % of lesion size) and the lesions were typed according to severity by an experienced pathologist, as described before [
]. Lesion collagen in the aortic root was stained with 0.05% sirius red (direct red 80; Sigma, Zwijndrecht, The Netherlands) in saturated picric adic, for 30 min. Images were obtained using the Leica DM3000 microscope. Quantification of collagen content (as % of lesion size) was performed in a blinded fashion using Adobe Photoshop software.
For Oil red O staining at the aortic root the slides were shortly fixed with formalin and stained with a working solution of Oil red O for 45 min. Cells were then destained with 60% isopropanol, washed in PBS and counterstained with haematoxylin. Quantification of neutral lipid content (as % of lesion size) was performed using Adobe Photoshop software. For further immunohistochemistry at the aortic root slides were fixed in acetone, blocked with the Avidin/Biotin Blocking Kit (all immunohistological kits are from Vector Laboratories, Burlingame, USA) and incubated with an antibody against macrophages (MOMA-2, 1:4000, AbD Serotec, Uden, The Netherlands) or against the endothelium (VCAM-1, 1:200, BD Pharmingen, Breda, The Netherlands). Biotin-labeled rabbit anti-rat antibody (1:300, Dako, Hervelee, Belgium) was used as secondary antibody. Hereafter, the signal was amplified using the ABC kit and then visualized using an AEC kit. Quantification of macrophage content (as % of lesion size) was performed using the Adobe Photoshop software. Scoring of VCAM-1 was performed by an experienced pathologist, where score 1 indicated little staining and score 3 indicated severe staining.
Liver sections were H&E stained and stained for total leukocytes (CD45 1:200, BD Pharmingen, Breda, The Netherlands), macrophages/kupffer cells (CD68, 1:200, AbD Serotec, Uden, The Netherlands), infiltrating macrophages (CD11b, 1:50, EBioscience, Vienna, Austria), neutrophils (Ly6G, 1:200, BD Pharmingen, Breda, The Netherlands) and T cells (KT3, 1:250, AbD Serotec, Uden, The Netherlands), using the same protocol as was used for the MOMA-2 staining in the aortic root. Quantification of cells was performed in 3 liver sections per mouse and was noted as cells/mm2. Scoring of liver steatosis was performed on the H&E stained slides and scoring of Kupffer cell foam cell formation was performed on CD68 stained slides. Both scorings were performed by an experienced pathologist, where score 0 indicated no lipid deposition in the hepatocytes or Kupffer cells for steatosis or Kupffer cell foam cell formation respectively. Score 4 indicated severe lipid deposition in both the hepatocytes or in the Kupffer cells.
2.4 Mouse blood parameters
Blood was withdrawn via tail vein incision before the diet and right before sacrifice. Plasma cholesterol and triglyceride levels were enzymatically measured according to the manufacturer's protocol (Roche, Woerden, The Netherlands). Absolute leukocyte blood counts were determined using the Scil Vet abc Plus analyzer (Scil animal care company BV, Oostelbeers, The Netherlands) and relative counts were assessed by flow cytometry after antibody staining with CD11b and Ly6C for macrophages, Ly6G for neutrophils (all from EBioscience, Vienna, Austria), CD3 for T cells and CD19 for B cells (both from Biolegend, London, UK) and CD4 and CD8 for T cell subsets (both from BD Pharmingen, Breda, The Netherlands). Flow cytometry measurements were performed on a FACS Canto II (BD Biosciences, USA) and analyzed using FlowJo Software.
2.5 Peritoneal macrophage analysis
Four days before sacrifice 5 mice per group were intraperitoneally injected with 3% thioglycollate medium (Fisher, Bleiswijk, The Netherlands). Upon sacrifice the peritoneum was flushed twice with 10 ml ice-cold PBS to collect peritoneal macrophages. Flow cytometry was used to determine peritoneal macrophage status by staining for CD11b, F4/80 (both from EBioscience, Vienna, Austria), MHC2 and CD64 (both from Biolegend, London, UK). In addition, after extracellular staining with CD11b and F4/80, cells were washed and permeabilised using ice-cold methanol for 30 min at 4 °C. Then cells were intracellularly stained for phospho-STAT1 (Cell Signaling Technologies, Leiden, The Netherlands). Flow cytometry measurements were performed on a FACS Canto II (BD Biosciences, USA) and analyzed using FlowJo Software.
Cell pellets of 5 × 105 cells were made for gene expression analysis of individual mice and the remainder of the thioglycollate-elicited macrophages was pooled and cultured in RPMI-1640 containing 25 mM HEPES, 2 mM l-glutamine, 10% FCS, 100 U/ml penicillin and 100 μg/ml streptomycin (all Gibco, Breda, The Netherlands) at a density of 5 × 105 cells/well in a 24-wells plate (all plates from Greiner Bio-One, Alphen a/d Rijn, The Netherlands). The peritoneal macrophages adhered overnight after which the medium was replaced. Cells were then stimulated with IFNγ (100 U/ml), LPS (10 ng/ml) or left unstimulated for 6 h after which gene expression was analyzed.
To measure nitric oxide (NO) production, cells were seeded at a density of 1 × 105 cells/well in a 96-wells plate. Cells adhered for 2 h, then the medium was replaced after which the cells were stimulated with LPS (100 ng/ml), IFNγ (50 ng/ml), a combination of both or left unstimulated for 24 h. Hereafter NO production was measured by mixing 50 μl of culture medium with 50 μl of Griess reagent (1 part 0.1% N-(1-napthyl)ethyldiamine dihydrochloride/60% ethanoic acid plus 1 part 1% sulfanilamide/30% ethanoic acid (all Sigma–Aldrich, Gillingham, UK)). Then, the absorbance was measured at 570 nm on a plate reader.
To measure arginase-1 activity, cells were seeded at a density of 1 × 105 cells/well in a 96-wells plate. Cells adhered for 2 h, then the medium was replaced after which the cells were stimulated with LPS (100 ng/ml), IL-4 (50 ng/ml), a combination of both or left unstimulated for 24 h. Hereafter the cells were lysed with 50 μl 0.1% Triton X-100 containing 5 μg of the protease inhibitors pepstatin, aprotinin and antipain (all Sigma–Aldrich Gillingham, UK). Then, 50 μl of 10 mM MnCl2, 50 mM Tris–HCl, pH 7.5 (Sigma–Aldrich) was added to the cell lysate for 10 min at 55 °C. Hereafter arginine hydrolysis was started by adding 25 μl 0.5 M arginine, pH 9.7 (Sigma–Aldrich) to 25 μl of the activated lysate for 60 min at 37 °C. The reaction was stopped by adding 400 μl stop solution containing H2SO4, H3PO4 in H2O (1:3:7 v/v; Sigma–Aldrich). Arginase-1 activity was then assessed as the amount of conversion from arginine to urea, which was colorimetrically quantified at 540 nm on a plate reader.
2.6 Gene expression analysis
Total RNA from aortic arches and liver was isolated using the Qiagen RNeasy Mini Kit (Qiagen, Venlo, The Netherlands) according to the manufacturer's protocol. For the peritoneal macrophages, RNA was isolated from 5 × 105 cells per mouse or in vitro condition using the High Pure RNA Isolation kit (Roche, Woerden, The Netherlands). cDNA synthesis was performed using 500 ng total RNA with iScript (BioRad, Veenendaal, The Netherlands). Real-time PCR was performed with 4 ng cDNA, 300 nM of each primer and 300 nmol Sybr Green Fast Mix (Applied Biosystems, Bleiswijk, The Netherlands) on a ViiA7 PCR machine (Applied Biosystems, Bleiswijk, The Netherlands). Gene expression was corrected for ARBP and cyclophillin-A as housekeeping genes. Primer sequences are available on request.
2.7 Western blot
Equal amounts of protein samples were loaded on a SDS polyacrylamide gel and transferred to a PVDF membrane (Bio-Rad, Hemel Hempstead, UK). After blocking with 0.05% Tween-20 and 5% nonfat dry milk in PBS for 90 min, blots were incubated overnight with anti-STAT1 (1:1000; Santa Cruz, Heidelberg, Germany), anti-pSTAT1 (1:1000; Cell Signaling Technologies, Hitchin, UK) and anti-β-actin (1:1000; Sigma, Gillingham, UK). Blots were then washed and incubated with the appropriate HRP-conjugated secondary antibody (1:10000; Stratech Scientific, Newmarket, UK) for 90 min and visualized using the ECL substrate kit (Thermo Scientific, Hemel Hempstead, UK).
2.8 Statistical analysis
All results are presented as mean ± S.E.M. Statistical analysis was performed using GraphPad Prism 5 software (GraphPad Software, San Diego, CA, USA) and SPSS Statistics 20 (IBM Software Group). All data were evaluated by the unpaired student t-test, except for scoring data, which were evaluated by the Pearson Chi–Square test. Data were considered significant if p < 0.05.
3. Experimental results
We studied the role of macrophage IFNγR2 in atherosclerosis development by use of a genetic approach. First a full IFNγR2−/− mouse (IFNγR2KO) was used to show that IFNγR2 deficiency ablates IFNγ responses in general. IFNγ is known to polarize macrophages towards a pro-inflammatory M1 phenotype, with NO production as one of its hallmarks [
]. As a tightly regulated balance exists between the different macrophage phenotypes, IFNγR2 deficiency was expected to decrease the M1 macrophage phenotype on the one hand and promote the M2 anti-inflammatory phenotype, with increased arginase-1 activity, on the other hand [
]. Indeed peritoneal macrophages lacking IFNγR2 were irresponsive to IFNγ and showed strongly reduced NO production after LPS + IFNγ (Fig. 1A). Moreover, we observed increased arginase-1 activity upon LPS and IL-4 stimulation of the cells lacking IFNγR2 (Fig. 1B), indicating that full IFNγR2 deficiency results in a shift towards the anti-inflammatory M2 macrophage phenotype. Next, a LysMCre-IFNγR2fl/fl (IFNγR2del) conditional knockout was generated to specifically ablate IFNγ signaling in myeloid cells only. We could then show that macrophages from both IFNγR2KO and IFNγR2del mice indeed had a decreased IFNγ signaling pathway as compared to IFNγR2wt mice. CD11b+F4/80+ peritoneal macrophages (PEMs) from both IFNγR2KO and IFNγR2del mice were unresponsive to IFNγ stimulation as compared to IFNγR2wt mice, as phosphorylation of STAT1 was absent, indicating an abrogated IFNγ signaling pathway (Fig. 1C). In addition, the non-myeloid CD11b−F4/80- cells from IFNγR2wt and IFNγR2del mice still had increased STAT1 phosphorylation upon IFNγ stimulation, while STAT1 phosphorylation was absent in PEMs from IFNγR2KO mice (Fig. 1C). This indicates that the conditional myeloid knockout has no effect on other inflammatory cell lineages. In non-T cells isolated from a mixed population of splenocytes and peripheral lymph nodes from IFNγR2del mice, a similar abrogation of IFNγ signaling was detected upon IFNγ stimulation, as decreased phosphorylation of STAT1 was detected (Supplemental Fig. 1). In contrast, their isolated T cell populations were still responsive to IFNγ, again indicating the conditional myeloid knockout has no effects on other inflammatory cell lineages (Supplemental Fig. 1).
Next, we transplanted atherosclerosis susceptible LDLR-/− mice with bone marrow from either IFNγR2fl/fl (IFNγR2wt) or IFNγR2fl/fl-LysMCre (IFNγR2del) mice and fed them a high fat diet for a period of 10 weeks. We could confirm that reconstitution of the transplanted bone marrow was efficient and was not different between IFNγR2wt and IFNγR2del mice (Fig. 2A). Throughout the experiment, IFNγR2wt and IFNγR2del mice showed a similar weight gain (Fig. 2B) and a similar plasma lipid profile (Fig. 2C and D). In addition, there were no differences in blood leukocyte subsets between both groups of mice (Fig. 2E–G). In spleen only very modest effects on some leukocyte subsets could be seen (Supplemental figure 2A–2C).
To study the effect of myeloid IFNγR2 deficiency in atherosclerosis, we first tested characteristics and responses of isolated macrophages. We could demonstrate that gene expression of the IFNγR2 was almost completely absent on PEMs under unstimulated conditions confirming excellent gene deletion (relative gene expression: IFNγR2wt 1 ± 0.05 vs. IFNγR2del 0.02 ± 0.004, p < 0.0001) (Fig. 3A). When stimulated ex vivo with IFNγ or LPS the expression of IFNγR2 was upregulated but remained strongly reduced in the IFNγR2del macrophages (relative gene expression after IFNγ: IFNγR2wt 5.48 ± 0.44 vs. IFNγR2del 3.11 ± 0.28, p = 0.017. LPS: IFNγR2wt 14.87 ± 0.10 vs. IFNγR2del 5.94 ± 0.27, p = 0.0001) (Fig. 3A). We then analyzed whether gene expression of IFNγ target genes was also reduced due to the absence of the IFNγR2. Indeed, gene expression of Cxcl10, Nos2 and Cd86 was significantly downregulated in PEMs from IFNγR2del animals as compared to IFNγR2wt animals (relative gene expression Cxcl10: IFNγR2wt 1 ± 0.06 vs. IFNγR2del 0.04 ± 0.004, p < 0.0001. Nos2: IFNγR2wt 1 ± 0.08 vs. IFNγR2del 0.006 ± 0.003, p = 0.0002. Cd86 IFNγR2wt 1 ± 0.28 vs. IFNγR2del 0.40 ± 0.08, p = 0.112) (Fig. 3B). Also after ex vivo stimulation with IFNγ or LPS this downregulation was present (Fig. 3C and D). When PEMs were analyzed by flow cytometry we observed that expression of the Fc-gamma receptor CD64, which is induced upon IFNγ stimulation [
], was reduced in IFNγR2del mice compared to IFNγR2wt (median: IFNγR2wt 11920 ± 857.5 vs. IFNγR2del 8967 ± 873.8, p = 0.050) (Fig. 3E). In addition, we observed a trend towards decreased expression of IFNγR2 in the aortic arch (relative gene expression: IFNγR2wt 1 ± 0.24 vs. IFNγR2del 0.57 ± 0.14, p = 0.128) (Fig. 3F). As many different non-myeloid cell types are present in the aortic arch, we did not see such a strong deletion as the deletion observed in the PEMs. Similar for the PEMs, we could show a downregulation of the IFNγ target genes Stat1, Cxcl10 and Cd86 in the aortic arch (relative gene expression Stat1: IFNγR2wt 1 ± 0.16 vs. IFNγR2del 0.82 ± 0.14, p = 0.385. Cxcl10: IFNγR2wt 1 ± 0.27 vs. IFNγR2del 0.45 ± 0.08, p = 0.048. Cd86: IFNγR2wt 1 ± 0.15 vs. IFNγR2del 0.65 ± 0.08, p = 0.048) (Fig. 3G). No downregulation of IFNγR2 or of IFNγ target genes was observed in the spleen of IFNγR2del compared to IFNγR2wt animals (Supplemental figure 2D and 2E). As most spleen leukocytes are lymphocytes, this indicates that the bone marrow transplantation had no effect on non-myeloid cells. Altogether these data demonstrate that the myeloid deletion of IFNγR2 was successful, also in a complex in vivo setting.
We then set out to determine whether myeloid deletion of IFNγR2 affected atherosclerosis development. Upon sacrifice, we assessed gene expression of inflammatory cytokines in the aortic arch as a first indication of atherogenesis. Gene expression of several inflammatory cytokines was found to be similar in both groups of mice (Fig. 3H). In addition, gene expression of the macrophage marker CD68, which was used as a surrogate marker for atherosclerotic lesion development, did also not differ between both groups (relative gene expression Cd68: IFNγR2wt 1 ± 0.22 vs. IFNγR2del 0.65 ± 0.17, p = 0.216). Hereafter, we further analyzed atherosclerosis development in a second atherosclerotic location, the aortic root. A comparable lesion size in the aortic root was observed between IFNγR2 wildtype and deleted animals (lesion size in μm × 104: IFNγR2wt 32.20 ± 1.60 vs. IFNγR2del 27.22 ± 2.38, p = 0.101) (Fig. 4A and B). In addition, plaque severity scoring showed no difference between both groups (p = 0.535) (Fig. 4C). Plaque phenotype analysis showed no effects on collagen or macrophage content either, as well as on the amount of necrosis or the lipid content (Data represent % of lesion area; Collagen: IFNγR2wt 33.08 ± 2.87 vs. IFNγR2del 36.60 ± 2.40, p = 0.350. Macrophage content: IFNγR2wt 58.68 ± 2.50 vs. IFNγR2del 59.67 ± 2.63, p = 0.786. Necrosis: IFNγR2wt 17.70 ± 2.38 vs. IFNγR2del 17.40 ± 2.45, p = 0.931. Lipid content: IFNγR2wt 42.95 ± 3.60 vs. IFNγR2del 40.79 ± 4.39 p = 0.713) (Fig. 4D–H). Furthermore, endothelial activation as measured by VCAM1 staining was also similar between both groups (IFNγR2wt 2.22 ± 0.11 vs. IFNγR2del 2.30 ± 0.11 p = 0.645) (Fig. 4I). These data thus exclude a role for myeloid IFNγR2 in atherosclerosis development.
As increased dietary cholesterol intake is also associated with liver inflammation [
], we assessed the inflammatory state of the liver in both groups of mice. High fat diet induced liver steatosis as an accumulation of lipid droplets was observed in H&E stained sections. However, the amount of steatosis was comparable between wildtype and IFNγR2 deleted mice (p = 0.682) (Fig. 5A and B). Also, when Kupffer cell foam cell formation was scored, no significant differences between the groups were found (p = 0.564) (Fig. 5C and D). Further quantification of hepatic inflammatory cells showed no effect of IFNγR2 deletion (Data represent cells/mm2; Leukocytes: IFNγR2wt 1121 ± 61.49 vs. IFNγR2del 1171 ± 70.78, p = 0.595. Infiltrated macrophages: IFNγR2wt 200.6 ± 15.07 vs. IFNγR2del 242.9 ± 22.86, p = 0.142. T cells: IFNγR2wt 250.1 ± 18.33 vs. IFNγR2del 223.9 ± 15.23, p = 0.279. Neutrophils: IFNγR2wt 63.15 ± 5.17 vs. IFNγR2del 58.42 ± 4.77, p = 0.505) (Fig. 5E–H). Further analysis of inflammation by assessing gene expression of inflammatory cytokines and chemokines showed no difference either (relative gene expression Il-1β: IFNγR2wt 1.03 ± 0.12 vs. IFNγR2del 1.18 ± 0.19, p = 0.508. Il-6: IFNγR2wt 1 ± 0.14 vs. IFNγR2del 1.03 ± 0.15, p = 0.886. Tnf: IFNγR2wt 0.95 ± 0.14 vs. IFNγR2del 0.87 ± 0.13, p = 0.676. Mcp1: IFNγR2wt 0.97 ± 0.14 vs. IFNγR2del 0.96 ± 0.15, p = 0.975) (Fig. 5I). This indicates that myeloid IFNγR2 is also dispensable for the hepatic inflammation that accompanies atherosclerosis development.
This study clearly demonstrates that myeloid IFNγR2 is dispensable for both atherosclerosis development and its accompanying hepatic inflammation. So far IFNγ was thought to be an essential pro-atherosclerotic cytokine [
]. But unlike several reports showing that targeting systemic IFNγ or IFNγ signaling strongly inhibits atherosclerosis development, we now demonstrate that focusing solely on myeloid IFNγ signaling does not affect atherosclerosis [
]. So, even though myeloid cells itself are critically involved in atherosclerosis, IFNγ signaling is dispensable in those cells. As systemic IFNγ has been shown to be critical in atherosclerosis development, IFNγ signaling is thus essential in non-myeloid cells in atherogenesis. T-cells are obvious candidates as the IFNγ-producing Th1-cells for instance are present in the human atherosclerotic plaque and are also known IFNγ targets [
]. As in our experiment IFNγ signaling was still intact in all these cells, the numerous immunomodulatory actions of IFNγ here could have counterbalanced the expected anti-atherogenic outcome of a myeloid IFNγR2 deletion.
As IFNγ is known to polarize macrophages to a pro-inflammatory M1 phenotype [
], deleting myeloid IFNγ signaling would result in a less pro-inflammatory state. At the same time, M1 polarization can also be induced by Toll-like receptor (TLR) stimuli. Oxidized LDL is abundantly present in the atherosclerotic lesion and can act as a TLR4 ligand [
]. Its presence could therefore sustain the macrophage M1 phenotype and the accompanying inflammatory response, which might be an explanation for the discrepancy between our initial hypothesis and the obtained results. However, it must be noted that we still observed downregulation of specific IFNγ-target genes upon LPS stimulation of PEMs in vitro. Data on gene expression of actual LPS-dependent genes is missing, which could turn out to be comparable in both groups of mice, supporting the above mentioned assumption. Not only in these PEMs, but also in the aortic arch we observed downregulation of IFNγ-target genes, which is important as several IFNγ-target genes, like for instance CXCL10 and MCP1, have been proven to be potent stimulators of atherosclerosis development [
], the observed downregulation of those genes would thus suggest a reduction in atherosclerosis development. But, despite the observed downregulation and the excellent gene deletion, we did not observe any phenotypic effects on atherosclerosis or hepatic inflammation of myeloid IFNγR2 deficiency. Compensation for the loss of myeloid IFNγR2 might be an explanation for these unexpected results. However, compensation by IFNγR1 itself is impossible as IFNγR2 is required for signal transduction [
]. Perhaps alternative signaling mechanisms might have come into play in our in vivo setting, with intercellular vesicle trafficking being one of them. Recently it has been shown that the IFNγ/IFNγR1 complex can be transferred to neighboring cells in order to activate their STAT1 signaling [
]. It could thus be that surrounding cells in the atherosclerotic lesion produce IFNγ and activate macrophages via this alternative mechanism of IFNγ signaling. However, this intercellular communication mechanism might only partially explain the lack of phenotype we observed, as we still observed downregulation of IFNγ-target genes. In addition, cross-talk exists between the type I and type II interferons, which explains some overlapping functions of the two interferon families. It has for instance been shown that the type I interferons can also induce STAT1 homodimerization, resulting in gene expression of downstream IFNγ-target genes [
]. However, in our study no increased TNF gene expression was observed, and again, as we observed a strong downregulation of IFNγ-target gene expression upon deletion of myeloid IFNγR2, it seems unlikely that the lack of myeloid IFNγ signaling is fully compensated by the action of other cytokines.
Recently the paradigm that macrophages accumulate in the atherosclerotic lesion solely by monocyte recruitment has been brought into question, as it has been shown that maintenance of lesional macrophages also depends on local proliferation of previously recruited macrophages [
]. However, in our study atherosclerosis development was initiated 6 weeks after the bone marrow transplantation was performed. As it has been shown that early atherosclerosis is predominantly dependent on monocyte influx [
], it is thus likely that in our study all lesional macrophages were monocyte-derived and thereby of donor origin. So, although the contribution of proliferating resident macrophages is of importance in advanced atherosclerotic lesions, it did likely not affect our bone marrow transplantation strategy.
The contribution of gender in atherosclerosis development is also an issue to address. The influence of gender in murine atherosclerosis is still under debate, but it is known that gender specific effects on atherosclerosis exists [
]. However, in a mixed group of both male and female IFNγ−/−LDLR-/− mice a reduction in atherosclerosis development could be observed compared to controls, suggesting that effects on atherosclerosis were present in both genders [
]. Gender differences are thus important to consider, which we cannot exclude in our experiments. Taken all above mentioned factors in consideration, we now assume that the lack of myeloid IFNγ signaling is somehow compensated by IFNγ-induced immunomodulatory actions on other cell types in the atherosclerotic lesion.
Next to atherosclerotic lesion analysis we analyzed the amount of hepatic inflammation in both groups of mice, as it has been shown that dietary cholesterol intake can induce liver inflammation [
]. Deleting myeloid IFNγ signaling might therefore reduce the degree of hepatic inflammation present under the course of a high fat diet. But in contrast to our expectations, myeloid IFNγR2 deletion did not influence this. Again other cell type-specific approaches, perhaps lymphocyte or vascular smooth muscle cell-specific, might be more effective in reducing the liver inflammation that is present during atherosclerosis.
In conclusion, we show that absence of myeloid IFNγ signaling has no effect on atherosclerosis. As myeloid IFNγ signaling is not essential throughout atherosclerosis development, we now suggest that the role of IFNγ in this disease should be investigated in other cell types.
Source of funding
Menno P. J. de Winther is an established investigator of the Netherlands Heart Foundation (2007T067). He is supported by the Netherlands Heart Foundation (#2010B022), NWO (#TOP91208001), the Netherlands CardioVascular Research Initiative (CVON2011-19) and holds an AMC fellowship.
Conflict of interest
Appendix A. Supplementary data
The following is the supplementary data related to this article: