Blocking Interferon β Stimulates Vascular Smooth Muscle Cell Proliferation and Arteriogenesis

Abstract

Increased interferon (IFN)-β signaling in patients with insufficient coronary collateralization and an inhibitory effect of IFNβ on collateral artery growth in mice have been reported. The mechanisms of IFNβ-induced inhibition of arteriogenesis are unknown. In stimulated monocytes from patients with chronic total coronary artery occlusion and decreased arteriogenic response, whole genome expression analysis showed increased expression of IFNβ-regulated genes. Immunohistochemically, the IFNβ receptor was localized in the vascular media of murine collateral arteries. Treatment of vascular smooth muscle cells (VSMC) with IFNβ resulted in an attenuated proliferation, cell-cycle arrest, and increased expression of cyclin-dependent kinase inhibitor-1A (p21). The growth inhibitory effect of IFNβ was attenuated by inhibition of p21 by RNA interference. IFNβ-treated THP1 monocytes showed enhanced apoptosis. Subsequently, we tested if collateral artery growth can be stimulated by inhibition of IFNβ-signaling. RNA interference of the IFNβ receptor-1 (IFNAR1) increased VSMC proliferation, cell cycle progression, and reduced p21 gene expression. IFNβ signaling and FAS and TRAIL expression were attenuated in monocytes from IFNAR1^−/− mice, indicating reduced monocyte apoptosis. Hindlimb perfusion restoration 1 week after femoral artery ligation was improved in IFNAR1^−/− mice compared with wild-type mice as assessed by infusion of fluorescent microspheres. These results demonstrate that IFNβ inhibits collateral artery growth and VSMC proliferation through p21-dependent cell cycle arrest and induction of monocyte apoptosis. Inhibition of IFNβ stimulates VSMC proliferation and collateral artery growth.

Introduction

Collateral arteries alleviate symptoms of ischemia in patients with arterial obstructive disease. Circulating monocytes play an essential role in the growth of collateral arteries (1). We recently performed a monocyte mRNA expression study in patients with single vessel subtotal coronary artery stenosis. Genome-wide expression analysis demonstrated that increased expression of interferon (IFN)²-β as well as IFNβ downstream genes is associated with insufficient coronary collateral artery growth, suggesting a potential anti-arteriogenic effect of this glycoprotein. In a murine model, application of IFNβ indeed resulted in an attenuated arteriogenic response upon femoral artery occlusion (2).

In the current study, we investigated gene expression profiles of stimulated monocytes from patients with chronic total coronary occlusions. In contrast to patients with subtotal occlusions as previously studied, these patients can be considered to be in a stable phase of collateralization in which collateral artery growth has been maximally triggered and has reached its plateau phase.

The mechanisms via which IFNβ exerts its anti-arteriogenic effects are unresolved. It is hitherto unknown if inhibition of IFNβ signaling results in stimulation of collateral artery growth. Vascular smooth muscle cells (VSMC) display a high rate of proliferation during arteriogenesis (3). In growing arteriolar anastomoses, they change from a contractile toward a proliferative phenotype (4). We therefore analyzed the effects of IFNβ application and inhibition in in vitro models of VSMC cell cycling and proliferation and on monocyte apoptosis. In vivo, the effects of inhibition of IFNβ were tested in mice lacking the subunit 1 of the IFNα/β receptor (IFNAR1^−/− mice) (5).

MATERIALS AND METHODS

Patient Study

The investigation conforms to the principles outlined in the Declaration of Helsinki and was approved by the institutional medical ethics committee (Ref. no. MEC 06/186). After giving informed consent, 50 Caucasian patients were included who underwent percutaneous coronary intervention (PCI) of a total coronary occlusion. Patients were considered eligible if they had symptoms of angina pectoris for ≥4 weeks and total chronic occlusion (CTO) of a coronary artery. Exclusion criteria were previous myocardial infarction, cardiac surgery, depressed left ventricular function, diabetes mellitus, neoplastic, or inflammatory disease. Invasive coronary collateral flow index (CFI) measurements were performed as previously described (2). Briefly, a 0.014 inch pressure guide wire (BrightWire, Volcano, Rancho Cordova, CA) was used for intracoronary pressure measurements. During a 1-min balloon inflation, the pressure distal to the coronary occlusion (wedge pressure, Pw) as well as aortic (Pao) pressure was determined. CFI was calculated as (Pw-CVP)/(Pao-CVP), where CVP was estimated to be 5 mmHg.

Mononuclear cells were collected from peripheral blood by density gradient centrifugation. Monocytes were negatively isolated using immunomagnetic beads and stimulated in culture with 10 ng/ml LPS for 3 h. mRNA from monocytes from 10 and 10 patients from the extreme ends of the spectrum (very high versus very low CFI and matched for age, sex, medication, and other factors that influence collateral artery growth), was amplified and biotinylated. Samples were randomly hybridized to HumanRef-8 Expression bead chip arrays (Illumina), followed by scanning and feature extraction, all performed at ServiceXS (Leiden, The Netherlands). Microarray data have been submitted to the Gene Expression Omnibus (GEO) under accession number GSE13290.

Validation of Gene Array Results

RNA from all 50 patients was reverse-transcribed into cDNA, and gene expression of CXCL9, CXCL10, CXCL11, CCL8, IL27, IFIT1, IL15RA, and GAPDH was assessed using real-time RT-PCR.

Animal Experiments

The investigation was approved by the Institutional Medical Ethics Committee (Ref. no. DKC 100847) and conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). 30 wild-type (129Sv/Ev) and 20 IFNα/β-receptor-1 knock-out (IFNAR1^−/−) mice underwent unilateral double femoral artery ligation. Ten wild-type mice received daily subcutaneous injections of 10⁵ IU/kg rmIFNβ.

Gene and Protein Expression Analysis of Murine Monocytes and Hindlimb Tissue

Three days after femoral artery ligation, blood was collected using cardiac puncture, and peripheral blood monocytes were isolated by density gradient centrifugation, taking the mononuclear cell fraction into culture for 2 h and washing away non-adherent cells. Adhering monocytes were subsequently stimulated with 10 ng/ml lipopolysaccharide (LPS) for 3 h. Monocyte gene expression was assessed by real-time RT-PCR of mm8S rRNA, mmIFNAR1, mmSTAT1, mmCXCL10, mmCXCL11, mmIL15, mmTNFSF10, mmFASL, mmFAS, and mmCASP7.

Hindlimb was dissected for RNA and protein isolation. Gene expression was analyzed by real-time RT-PCR of the following targets: mm18SrRNA, mmIFNAR1, mmIRF3, mmSTAT1, mmCXCL10, mmCXCL11 mIL15, mmTNFSF10, and p21. Protein content was measured spectroscopically. ELISA analysis for murine CXCL10 was performed from isolated protein.

Immunohistochemical Analysis of Hindlimb Tissue

Seven days after femoral artery ligation, hindlimb tissue was dissected, and frozen sections were prepared and stained with a monoclonal goat anti-mouse IFNAR1 antibody for its localization in growing collateral arteries. VSMC were visualized with an antibody against α-smooth muscle actin (Sigma), nuclei stained with Hoechst 33342 (Molecular Probes).

Hindlimb Perfusion Measurements

Seven days after femoral artery ligation, perfusion restoration was assessed using fluorescent microsphere infusion under conditions of maximal vasodilation by infusion of adenosine in an established mouse model of arteriogenesis as previously described (6). Hindlimb tissue was harvested, digested, and microspheres were counted in a flow cytometer. Perfusion restoration was expressed as percentage perfusion ligated versus non-ligated hindlimb.

In Vitro Analysis of Monocyte Apoptosis and Gene Expression upon IFNβ Treatment

THP-1 monocytes (ATCC) were treated with increasing concentrations of rhIFNβ, and apoptosis was measured after 24 and 48 h by staining with Annexin V and propidium iodine (PI) antibodies and detecting the percentage of Annexin V-positive PI-negative cells using flow cytometry. IFNβ-stimulated THP-1 monocytes also underwent gene expression analysis of P0, CXCL11, p15, p21, p27, FAS, FASL, and TNFSF10. Similarly, freshly isolated primary human monocytes from health donors were stimulated with IFNβ and underwent gene expression analysis of P0, CXCL11, p21, and FAS as well as measurement of apoptosis.

In Vitro Analysis of VSMC Proliferation and Gene Expression

Following stimulation with IFNβ, gene expression of CXCL10 and IL15 was assessed in VSMC using real-time RT-PCR. Expression of STAT1 protein was assessed in VSMC using Western blot.

Proliferation of primary human VSMC was assessed in vitro by determining their BrdU-uptake after stimulation with rhIFNβ, rhCXCL10, or rhIL15. Cell cycle analysis was performed by flow cytometric analysis of cellular DNA content after staining with propidium iodine.

To investigate whether the effects of IFNβ on VSMC are depending on p21, VSMC were transfected with siRNA against p21. As a control, nonspecific siRNA was used. Expression of p21 and IFNβ target genes were measured at gene (real-time RT-PCR) and protein levels (Western blot), and proliferation and cell cycle were assessed as above.

Inhibiting IFNβ signaling in vitro, VSMCs were also transfected with siRNA against the IFNα/β receptor-1 (IFNAR1) or nonspecific siRNA. Down-regulation of IFNAR1 was assessed at mRNA (real-time RT-PCR) and protein level (Western blot). Proliferation and cell cycle status was assessed as described above, and gene expression of P0, IFNAR1, CXCL10, IL15, p15, p21, and p27 was analyzed.

Statistical Analysis

In vitro and animal data are presented as means ± S.E. of the mean. Intergroup comparisons were performed using Student's t test if normally distributed. Comparisons between three or more groups were performed using one-way analysis of variance (ANOVA) and Bonferroni post-hoc correction. Dose-dependent trends were calculated using probit regression analysis. A p value <0.05 was considered statistically significant. Normalization and statistical analysis of the gene array data were carried out using the limma package (7) and scripts in R/Bioconductor (8). Genes were considered significant if the p values, adjusted for multiple testing by using Benjamini and Hochberg's method (9), were <0.05. The false discovery rate was thereby controlled to be <5%. Gene set enrichment analysis (GSEA) (10) was used for analysis of differential expression of functional groups of genes and their promoters (c2, curated gene sets; c3, transcription factor targets, TRANSFAC version 7.4), employing MSigDB version 2.5. See supplemental data for a detailed description of the methods.

RESULTS

Expression of IFNβ-regulated Genes Is Increased in Patients with Hampered Collateral Artery Growth

Whole-genome expression of LPS-stimulated monocytes from patients with CTO was analyzed in relation to the capacity of the coronary collateral circulation. 10 patients with lowest CFI (0.27 ± 0.06) were compared with 10 patients with highest CFI (0.47 ± 0.08), who were well matched for factors potentially influencing collateralization (supplemental Table S1). After correcting for multiple testing, 120 genes were found to be differentially expressed between patients with low and high CFI, of which 65 genes were more strongly induced in the group with a low CFI. Pathway analysis (GSEA) showed a highly significant enrichment of several interferon-related pathways in patients with a low CFI (Table 1). Individual genes of these pathways were up-regulated (Fig. 1), and analysis of their promoter sequences confirmed the strong enrichment for IFNβ response elements (Table 1). GSEA core-enrichment of the individual genes within these gene sets is shown in supplemental Table S2. Results of the GSEA transcriptome comparison of patients with low and high CFI were projected onto the LPS-induced TLR4 pathway, visualizing the enhanced activation of specifically the IFNβ pathway in patients with a low CFI (supplemental Fig. S1). Real-time RT-PCR analysis of the whole patient group (n = 50) confirmed increased expression of IFNβ-regulated genes in stimulated monocytes from patients with low CFI. Negative correlations with CFI were found for gene expression of the IFNβ-regulated genes CXCL11, CCL8, IL27, and IL15RA (Table 2). Dichotomizing patients according to the median CFI detected in the whole group (0.37), fold-changes between low and high CFI were calculated. Individual gene fold-changes were low, but the consistent and coordinated up-regulation of the panel of IFNβ genes confirmed the GSEA pathway analysis results.

TABLE 1.

Gene set enrichment analysis: interferon target sets in patients with low versus high CFI

The depicted gene sets are enriched in patients with low CFI and result from two separate GSEA comparing the transcriptomes of patients with either low (n = 10) or high (n = 10) CFI. We used gene sets consisting of (A) 1892 curated pathways from various sources and (B) 500 transcription factor binding sites defined in the TRANSFAC database, version 7.4. These sets are available online at the BROAD Institute (10). See supplemental Table S3 for a complete overview of publications on which these interferon target gene sets are based.

	Sets (n)	GSEA statistics
		NESa	FDRb	FWERc
			%	%
A. IFN target set
type II	2	2.5	0	0
type I	5	2.6	0	1
I and II	2	2.1	0	7
B. Transcription factor
IRF2	1	1.8	3	11
IRF3	2	1.9	2	15
IRF1	1	1.8	7	33
IRF	1	1.7	7	40
IRF8	1	1.7	6	41

^a NES, normalized enrichment score.

^b FDR, false discovery rate.

^c FWER, family wise error rate.

FIGURE 1.

Differential interferon response between patients with low and high CFI. Microarray mRNA expression levels of interferon related signature genes in stimulated monocytes of patients with low (gray bars) and high (black bars set to 1) CFI. The bar graphs compare the average relative mRNA levels of n = 55 interferon-related genes. Note that these 55 signature genes have an elevated expression level in patients with low CFI, as determined by GSEA. A, gene symbol marked with an asterisk indicates the presence of an inter-species conserved IRF-1 or -2 motif within the promoter region of that gene as determined by GSEA or rVISTA.

TABLE 2.

Real-time RT-PCR from all 50 CTO patients analysis confirms the consistent up-regulation of a panel of IFNβ target genes in patients with low CFI

Results are depicted as fold-changes of patients with low compared to high CFI. Gene expression of CCL8, CXCL11, IL27, and IL15RA significantly correlated inversely with CFI.

Symbol	-Fold change	p value for -fold change	Correlation coefficient	p value for correlation
CCL8	1.53	0.093	−0.371	0.010
CXCL9	1.44	0.086	−0.218	0.137
CXCL10	1.25	0.298	−0.169	0.252
CXCL11	1.45	0.020	−0.353	0.014
IL27	1.34	0.046	−0.355	0.013
IFIT1	1.37	0.055	−0.259	0.075
IL15RA	1.31	0.043	−0.353	0.014

IFNβ Treatment Induces TRAIL and p21 in Vivo

To investigate the mechanisms of IFNβ-induced inhibition of collateral artery growth, a murine hind limb model of femoral artery ligation was employed. In collateral-containing hindlimb tissue from IFNβ-treated mice, gene expression of pro-apoptotic TNF-related apoptosis-inducing ligand (TRAIL), previously shown to mediate IFNβ-induced apoptosis in lymphoma cells (11), and anti-proliferative IL15 (12) were increased compared with control group (TRAIL: 17.22 ± 2.62 versus 7.33 ± 0.93, p < 0.01; IL15: 38.97 ± 6.13 versus 10.29 ± 1.04, p < 0.001). IFNβ-treated mice also displayed enhanced expression of the cell cycle regulator cyclin-dependent kinase inhibitor-1A (CDKN1, p21), known to negatively regulate G1- to S-phase transition (59.49 ± 12.50 versus 7.37 ± 1.65, p < 0.01).

IFNβ Induces Apoptosis of Monocytes in Vitro

To analyze its effects on mononuclear cells, THP1 cells were treated with increasing concentrations of IFNβ. FACS analysis showed enhanced apoptosis after 24 h and after 48 h exposure (4.6 versus 38.1% apoptosis after 48 h stimulation with 0 versus 100 ng/ml IFNβ, Fig. 2A). Real-time RT-PCR for CXCL11 confirmed increased IFNβ signaling (Fig. 2B). TRAIL, FAS, and FASL were found up-regulated upon stimulation with IFNβ, indicating cell death signaling (Fig. 2, C–E). IFNβ stimulated expression of p21 but not p15 or p27 (Fig. 2, F–H).

FIGURE 2.

IFNβ induces apoptosis in THP1 monocytes in vitro. Percentage of apoptotic THP1 monocytes dose-dependently increased (A), with a greater effect after 48 h than after 24 h of stimulation with IFNβ. Increased gene expression of CXCL11 confirmed enhanced IFNβ signaling (B). TRAIL, FAS, and FASL (C–E) were up-regulated, supporting the anti-proliferative and pro-apoptotic effect of IFNβ on monocytes. Expression of cyclin-dependent kinase inhibitor 1A (p21) was found up-regulated upon IFNβ stimulation (F), indicating a cell cycle inhibiting effect of the cytokine. p15 and p27 were not significantly differentially expressed (G, H). In primary human monocytes, IFNβ similarly induced apoptosis (I) and increased expression of CXCL11 (J), FAS (K), and p21 (L). *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with 0 ng/ml IFNβ.

To confirm the effects of IFNβ on THP1 cells in primary cells, the experiments were repeated in freshly isolated human monocytes. Comparable to THP1 cells, IFNβ dose-dependently induced apoptosis after 48 h incubation (27.7% ± 5.2 versus 10.4 ± 1.2 apoptotic cells, Fig. 2I) and increased gene expression of CXCL11, FAS, and p21 (Fig. 2, J–L).

IFNβ Attenuates Proliferation of VSMC via Cell Cycle Arrest

Immunohistochemistry showed expression of the type I interferon-receptor IFNAR1 in the vascular media of collateral arteries (Fig. 3A). IFNβ has been shown to inhibit VSMC proliferation in vitro (2). As demonstrated by real-time RT-PCR, IFNβ treatment up-regulated gene expression of the anti-angiogenic factors CXCL10 and IL15 as downstream targets of IFNβ in VSMC (Fig. 3, B and C). Also, STAT1 was increased at protein level (Fig. 3D). However, application of increasing doses of recombinant human CXCL10 and IL15 did not affect VSMC proliferation as measured by BrdU incorporation (Fig. 3, E and F). Instead, cell cycle analysis showed increased percentages of cells arrested in the G1 phase of the cell cycle after IFNβ treatment (97.8 versus 66.7% for 100 versus 0 ng/ml IFNβ, Fig. 3G), and p21 gene expression was found to be up-regulated, already at low concentrations of IFNβ (Fig. 3H). Expression of two other cell cycle regulators p15 and p27 were not affected by IFNβ (data not shown).

FIGURE 3.

IFNβ attenuates proliferation of VSMC in vitro. Immunohistochemistry showed expression of IFNAR1 (red) in the collateral arterial media (VSMC, green, nuclei, blue) (A). IFNβ treatment up-regulated CXCL10 and IL15 (B, C) at gene level and STAT1 at protein level (D). Application of either of either CXCL10 or IL15 to VSMC in vitro remained without an effect on proliferation, as measured by BrdU uptake (E, F). Cell cycle analysis showed increased cell numbers arrested in G-phase and reduced cell in S-phase after treatment with IFNβ (*, p < 0.05 for dose-dependent trends) (G), and p21 expression was enhanced by IFNβ treatment (H). *, p < 0.05; **, p < 0.01 compared with 0 ng/ml IFNβ.

VSMC Cell Cycle Inhibition of IFNβ Is Mediated by p21

Subsequently, we tested if IFNβ-induced VSMC cell cycle inhibition was dependent on p21. Using RNA-interference, p21 protein expression was attenuated in VSMC in vitro. Whereas expression of p21 protein was induced at low concentrations of IFNβ (1 ng/ml), this effect was abolished by previous siRNA treatment against p21. BrdU incorporation experiments confirmed the anti-proliferative effect of IFNβ, which was abrogated by anti-p21 siRNA. Accordingly, anti-p21 treatment reversed the cell cycle inhibiting effect of IFNβ. See also Fig. 4.

FIGURE 4.

Cell cycle-inhibiting effect of IFNβ is dependent on p21. siRNA against cell-cycle inhibitor p21 dose-dependently reduced p21 protein expression in VSMC as compared with nonspecific control-siRNA (A). p21 protein expression is stimulated by IFNβ (compare gene expression Fig. 3H). This up-regulation is attenuated after preincubation with p21 siRNA (B). BrdU incorporation analysis confirmed the anti-proliferative effect of IFNβ, which was abrogated by pretreatment with p21 siRNA (C). Similarly, the cell cycle-arresting effect of IFNβ is attenuated by inhibiting p21 (D). Interestingly, p21 siRNA even increased cell cycling in the presence of IFNβ (enhanced numbers of cells in S phase, less in G1 phase). *, p < 0.05 compared with 0 ng/ml IFNβ; #, p < 0.05 compared with same concentration IFNβ, ctrl siRNA.

In Vitro Blockade of IFNβ Signaling Leads to Increased VSMC Proliferation

Having found that IFNβ attenuates VSMC proliferation via its effects on cell cycle regulation, we next hypothesized that inhibition of IFNβ signaling could stimulate VSMC proliferation. In VSMC in vitro, IFNAR1 gene (RT-PCR) and protein expression (Western blot) were found down-regulated after transfection with siRNA (Fig. 5, A and B). After 48 h of transfection, BrdU incorporation was increased in VSMC, indicating enhanced proliferation in these cells (Fig. 5C). Cell cycle analysis indicated increased cell numbers in S-phase in IFNAR1 siRNA compared with control siRNA-transfected cells (35.8 versus 54.2%, Fig. 5D). Concomitantly, p21 expression was reduced in VSMC treated with IFNAR1 siRNA compared with nonspecific siRNA (Fig. 5E).

FIGURE 5.

Blockade of IFNβ signaling leads to increased VSMC proliferation in vitro. Real-time RT-PCR (A) and Western blot (B) confirmed strongly reduced expression of IFNAR1 in VSMC transfected with siRNA against IFNAR1 as compared with cells transfected with nonspecific control siRNA. Proliferation was enhanced in VSMC in which IFNAR1 was knocked down (C). Cell cycle analysis showed increased progression to S-phase in siIFNAR1-treated cells (D). IFNAR1 knockdown resulted in decreased mRNA expression of the cell cycle inhibitor p21 (E). *, p < 0.05; **, p < 0.01 compared with control siRNA.

Arteriogenesis Is Increased in IFNAR1^−/− Mice

Finally, we tested whether blocking IFNβ signaling stimulates arteriogenesis in vivo. Using microsphere infusion at different pressure levels under conditions of maximum vasodilation 1 week after femoral artery ligation we calculated hindlimb perfusion per gram tissue. Perfusion restoration (ligated versus non-ligated hindlimb) was found improved in IFNAR1^−/− mice compared with control group (54.3 ± 2.5% versus 38.3 ± 2.7%, p < 0.001, Fig. 6).

FIGURE 6.

Arteriogenesis is enhanced in IFNAR1^−/− mice. Hindlimb perfusion was assessed 1 week after femoral artery ligation using infusion of fluorescent microspheres under conditions of maximal vasodilation (n = 10). Perfusion restoration, expressed as percentage ligated versus non-ligated hindlimb, was increased in IFNAR1^−/− compared with wild-type mice (***, p < 0.001).

Gene Expression of Stimulated Circulating Monocytes of IFNAR1^−/− Mice Is Comparable to Patients with Well-developed Collateral Arteries

To mimic the activation status of monocytes in the clinical situation, murine circulating mononuclear cells were stimulated with LPS ex vivo. Real-time RT-PCR showed strongly reduced expression of IFNAR1, STAT1, CXCL10, CXCL11, TRAIL, and IL15 as signs of abrogated IFNβ signaling in LPS-stimulated monocytes of IFNAR1^−/− compared with wild-type mice (Fig. 7, A–F), resulting in an expression profile comparable to that of stimulated monocytes from patients with a good collateralization. Apoptosis related molecules FASL, FAS, and CASP7 were also found reduced (Fig. 7, G–I).

FIGURE 7.

Inhibition of IFNβ- and apoptosis signaling in stimulated monocytes of IFNAR1^−/− mice. In LPS-stimulated monocytes from IFNAR1^−/− mice, expression of IFNAR1, signaling molecule STAT1 and target genes CXCXL10 and CXCL11 was strongly suppressed compared with wild-type mice (A–D). TRAIL and IL15, both part of the IFNβ signaling cascade, were strongly down-regulated in stimulated monocytes of IFNAR1^−/− mice (E, F), as were apoptosis-related molecules FAS, FASL, and CASP7 (G–I). *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with wild-type mice.

Reduced Expression of IFNβ Targets in Collateral-containing Hindlimb Tissue of IFNAR1^−/− Mice

Gene expression analysis of collateral-containing tissue showed reduced mRNA expression of the IFNβ pathway in the IFNAR1^−/− mice. IRF3, CXCL10, STAT1, CXCL11, and IL15 were found decreased compared with control animals (Fig. 8, A–F). At the protein level, ELISA analysis showed reduced expression of IFNβ downstream cytokine CXCL10 in IFNAR1^−/− mice (5.2 ± 1.8 versus 54.5 ± 8.9 relative to total protein (Fig. 8G).

FIGURE 8.

Gene expression data of hindlimb tissue from IFNAR1^−/− mice. In collateral-containing hindlimb tissue from IFNAR1^−/− mice, gene expression of IFNβ pathway components IFNAR1, IRF3, CXCL10, STAT1, and CXCL11 and IL15 (A–F) was down-regulated. ELISA showed decreased protein expression of IFNβ downstream target CXCL10 in IFNAR1^−/− hindlimb tissue (G). *, p < 0.05; **, p < 0.01; ***, p < 0.001 compared with wild-type mice.

DISCUSSION

IFNβ signaling was enhanced in stimulated monocytes from patients with chronic total coronary artery occlusion and hampered coronary collateralization. In vitro, IFNβ exerted a direct pro-apoptotic effect on monocytes. Proliferation analysis showed a p21-dependent inhibiting effect of IFNβ on VSMC, which were arrested in the G1-phase. Knockdown of IFNAR1 had the opposite effect, resulting in an increased number of VSMC in the S-phase and enhanced proliferation. Finally, we report decreased IFNβ- and apoptosis-related gene expression as well as increased collateral artery growth in IFNAR1^−/− mice, showing that inhibition of IFNβ signaling augments arteriogenesis.

Following our earlier study on IFNβ signaling and collateral artery growth (2), we here confirmed increased IFNβ signaling in an independent patient population (Fig. 1). In patients with CTO, the driving forces of arteriogenesis are less variable because the stenosis grade of the original coronary artery is constant (i.e. 100%). CFI measurements are therefore less susceptible to interindividual variations of coronary baseline hemodynamics. Besides, the present data demonstrate that increased IFNβ signaling is also found in patients with a low arteriogenic response despite the presence of a chronic total coronary occlusion, where collateral artery growth has reached a plateau phase. Despite this maximal response, a large number of patients remain symptomatic indicating a need for further growth of coronary collateral arteries.

Subsequently, in vitro and animal experiments were employed to disclose the mechanisms of the anti-arteriogenic effect of IFNβ. Because the IFNβ receptor was detected in the vascular media, effects of its stimulation and blockade were investigated in VSMC in vitro (Fig. 3). We report that IFNβ exerts its negative effects on VSMC proliferation by inducing cell cycle arrest in a p21-dependent fashion (Fig. 4). The cell cycle regulator p21 is known as the major mediator of cell cycle arrest (13). Its regulation by IFNβ has previously been observed in tumor cells (14), and inhibition of p21 has been shown to regulate endothelial cell proliferation (15). Antiproliferative, pro-apoptotic effects of IFNβ are known in oncology but have not yet been examined in cardiovascular pathophysiology. p21-mediated cell cycle inhibition by IFNβ has not been described previously in VSMC (16).

The inhibitory effect of IFNβ on endothelial cell sprouting (angiogenesis) is mediated by CXCL10 (17). Its influence on collateral artery growth has not been studied. Our data show that the anti-arteriogenic effect of IFNβ is not mediated by CXCL10, because IFNβ but not CXCL10 inhibited VSMC proliferation (Fig. 3).

We reasoned that inhibition of IFNβ signaling might stimulate VSMC proliferation and arteriogenesis. Reduction of p21 expression and VSMC cell cycling and proliferation upon inhibition of IFNβ signaling showed that inhibition of cytokine signaling can stimulate cell growth (Fig. 5). Vascular regeneration has hitherto been linked to increased inflammatory signaling. For the first time, we show that reduced cytokine signaling can also stimulate arterial growth.

As anti-IFNβ treatment stimulated cellular proliferation in vitro, we hypothesized that a blockade of IFNβ signaling may boost adaptive collateral artery growth in vivo. In a first approach, treatment with IFNβ neutralizing antibodies showed no effect on perfusion restoration (data not shown). Subsequent in vitro measurements indicated that concentrations of ≥5,000 IU/ml of IFNβ neutralizing or anti-IFNAR1 antibody were necessary to accomplish a 50% blockade of IFNβ signaling. Achieving these concentrations in vivo would have resulted in unfeasibly high costs. We therefore decided to use a murine knock-out model of the interferon α/β receptor. In IFNAR1^−/− mice, a neo-marker is inserted in Exon III, resulting in dysfunctional mRNA and a loss of functional receptor protein (5). Because the promoter of the IFNAR gene is intact, mRNA is still detectable.³ This is reflected by our real-time RT-PCR data, showing decreased but still existent IFNAR mRNA in monocytes from IFNAR1^−/− mice. Interestingly, concerning the IFNβ pathway, gene expression profile of the stimulated murine monocytes and collateral-containing hindlimb tissue of IFNAR1^−/− mice was comparable to that of stimulated monocytes from patients with high CFI (Fig. 7).

Both gene expression data from murine monocytes and functional assays in vitro indicate increased monocyte apoptosis upon simulation with IFNβ. Increased apoptosis signaling pathways were also found in patients with low CFI. Our data (Fig. 2) corroborate earlier reports on IFNβ-mediated induction of apoptosis in monocytes from patients with multiple sclerosis (18). IFNβ has previously been demonstrated to induce up-regulation of surface-bound TRAIL and release of soluble TRAIL in human monocytes (19), and IFNβ-induced apoptosis has been reported to be mediated by TRAIL in lymphoma cells (11). Regulation of apoptosis by p21 through its caspase-mediated cleavage from cyclin-dependent kinase 2 has been described in endothelial cells (20). Decreased apoptosis-related gene expression in monocytes from IFNAR1^−/− mice suggests reduced susceptibility for programmed cell death in these cells.

We here show that inhibition of IFNβ signaling can stimulate cell proliferation in vitro and arteriogenesis in vivo. Pro-arteriogenic substances are mostly cytokines (growth factors, chemoattractants or colony-stimulating factors) which have pro-inflammatory or bone-marrow cell releasing effects, both of which potentially aggravate atherosclerosis or destabilize plaques (21). Recent data suggest that IFNβ attenuates angiotensin II-induced atherosclerosis in ApoE^−/− mice, while IFNβ alone did not have any effect on atherosclerosis in that study (22). Another study showed aggravation of atherosclerosis by myeloid type I IFN (23). These conflicting data require further study when considering anti-IFNβ therapy for the promotion of arteriogenesis. Therapeutic inhibition of interferons to promote blood vessel growth could also activate autoimmune processes, which are normally inhibited by IFNβ, as has been discussed earlier (24).

We conclude that inhibition of IFNβ signaling might serve as an approach to stimulate collateral artery growth by stimulating VSMC proliferation and cell cycling. Further investigations of the modulation of IFNβ signaling using pharmacological inhibitors of IFNβ or its receptor are required in atherosclerotic models before clinical approaches can be envisaged.