IRF3 and IRF7 mediate neovascularization via inflammatory cytokines

Abstract Objective To elucidate the role of interferon regulatory factor (IRF)3 and IRF7 in neovascularization. Methods Unilateral hind limb ischaemia was induced in Irf3−/−, Irf7−/− and C57BL/6 mice by ligation of the left common femoral artery. Post‐ischaemic blood flow recovery in the paw was measured with laser Doppler perfusion imaging. Soleus, adductor and gastrocnemius muscles were harvested to investigate angiogenesis and arteriogenesis and inflammation. Results Post‐ischaemic blood flow recovery was decreased in Irf3−/−and Irf7−/− mice compared to C57BL/6 mice at all time points up to and including sacrifice, 28 days after surgery (t28). This was supported by a decrease in angiogenesis and arteriogenesis in soleus and adductor muscles of Irf3−/− and Irf7−/− mice at t28. Furthermore, the number of macrophages around arterioles in adductor muscles was decreased in Irf3−/−and Irf7−/− mice at t28. In addition, mRNA expression levels of pro‐inflammatory cytokines (tnfα, il6, ccl2) and growth factor receptor (vegfr2), were decreased in gastrocnemius muscles of Irf3−/− and Irf7−/− mice compared to C57BL/6 mice. Conclusion Deficiency of IRF3 and IRF7 results in impaired post‐ischaemic blood flow recovery caused by attenuated angiogenesis and arteriogenesis linked to a lack of inflammatory components in ischaemic tissue. Therefore, IRF3 and IRF7 are essential regulators of neovascularization.

These growth factors induce endothelial cells from the pre-existing vasculature to migrate, grow and differentiate to shape new capillaries. 3,4 Arteriogenesis, the maturation of pre-existing collaterals into arterioles, [5][6][7][8] is induced by inflammation, shear stress and circumferential stretch on the vascular wall after occlusion, mainly proximal to the occlusion. Inflammatory cells such as macrophages, monocytes, but also CD4+ and CD8+ T cells adhere and invade into the vascular wall and produce inflammatory cytokines, chemokines and growth factors, which cause further maturation of arterioles and induce angiogenesis. Subsequently, smooth muscle cells proliferate and migrate towards the vessel wall, resulting in full functional arterioles. 7,9,10 Both angiogenesis and arteriogenesis are essential for accurate neovascularization, to restore blood flow after arterial occlusions.
For the production of inflammatory cytokines, chemokines and growth factors, inflammatory cells require activation.
However, it is observed that TLR3 can also activate pro-inflammatory transcription factors via IRF3 and IRF7 and may be potential regulators of cell proliferation and survival. [16][17][18][19] Since neovascularization is mainly induced by pro-inflammatory cytokines and growth factors, we hypothesize an important role for IRF3 and IRF7 in arteriogenesis and angiogenesis.
In the current study, we aimed to elucidate the role of IRF3 and IRF7 in vivo in post-ischaemic neovascularization in a hind limb ischia (HLI) mouse model, using wild-type C57BL/6, Irf3 −/− and Irf7 −/− mice.

| Mice
This study was performed in compliance with the Dutch govern-
Unilateral hind limb ischaemia (HLI) was performed. In brief, a single ligation model was performed by electrocoagulation of the left common femoral artery proximal to the bifurcation of the popliteal and saphenous artery. 21,22 After surgery, anaesthesia was antagonized with flumazenil (0.7 mg/kg, Fresenius Kabi). Buprenorphine (0.1 mg/ kg, MSD Animal Health) was given after surgery to relieve pain in a fixed regime and when circumstances required this was repeated.
Number of total operated mice was 11 per group, however, in the C57BL/6 group 5 mice deceased during or after surgery and 2 and 4 mice, respectively in the Irf3 −/− and Irf7 −/− group. Mice characteristics of Irf3 −/− , Irf7 −/− and C57BL/6 mice are shown in Table S1. Body weight at day of surgery (t = 0) and when killed (t = 28) was similar in the Irf3 −/− and Irf7 −/− mice compared to C57BL/6 mice.

| Laser Doppler perfusion imaging
Post-ischaemic blood flow recovery was measured in the left ischaemic and right non-ischaemic paw of Irf3 −/− , Irf7 −/− and C57BL/6 mice with the use of laser Doppler perfusion imaging (LDPI) (Moor Instruments). 23 Blood flow was measured before and directly after surgery and at day 3, 6, 9, 14, 21 and 28. Paw perfusion was expressed as a ratio of left ischaemic to right non-ischaemic paw. 24 Before LDPI, mice were anaesthetized as described before 25,26 with an intraperitoneal injection of midazolam (8 mg/kg) and medetomidine (0.4 mg/kg). After LDPI, anaesthesia was antagonized by subcutaneous injection of flumazenil (0.7 mg/kg). Mice were killed 28 days after surgery after the last LDPI via cervical dislocation after subcutaneous admission of analgesic fentanyl (0.08 mg/kg). The adductor, soleus and gastrocnemius muscles were harvested for either (immuno) histochemical analysis or RT-qPCR.

| Immunohistochemistry
Adductor muscles were fixed in 4% formaldehyde and embedded in paraffin, soleus and gastrocnemius muscles were snap frozen on dry ice and stored at −80°C.
Cross sections of 6 μm were made throughout the embedded adductor muscle. Adductor muscle sections were stained with alpha smooth muscle cell actin (aSMActin, DAKO) to visualize vascular smooth muscle cells (VSMC). Stained slides were photographed (20× magnification) with microscope photography software (Axiovision, Zeiss) and analysed with ImageJ (FIJI) by counting the number of arterioles and measuring the diameters of each arteriole with a visible lumen to determine arteriogenesis. 25 In addition, adductor muscle sections were stained with alpha smooth muscle cell actin (pink) and MAC3 (green) (BD Pharmingen) to visualize macrophages around arterioles. Images were acquired on a Philips Ultra Fast Slide Scanner (Philips Digital Pathology Solutions, Best) and analysed by counting the number of macrophages around the arterioles divided by the circumference of the arteriole (presented as macrophages/μm). However, 12 sections, in the same area in all mice were analysed per mice per leg and the average was used for analyses. Negative controls were performed by omitting either the secondary antibody of aSMActin or MAC3, or PBS.
Fresh-frozen soleus muscles were cross sectioned in 6 μm slices with a cryostat and sections were fixated in acetone. To visualize endothelial cells, soleus muscle sections were stained with CD31 (BD Pharmingen).
Stained slides were photographed (20× magnification) with microscope photography software. The number of CD31+ structures was analysed with ImageJ, which was used to determine angiogenesis and was presented in absolute numbers of CD31+ structures per area.

| RNA isolation, cDNA synthesis and RT-PCR from adductor muscles
RNA was isolated from snap frozen gastrocnemius muscles 28 days after surgery as described before. 27 In brief, muscle tissues were crushed with mortar and pestle, while using liquid nitrogen to preserve sample integrity. RNA was isolated according to manufacturer's protocol using TRIzol Reagent (Life

| Statistical analysis
All data are presented as mean ± SEM. In statistics software GraphPad Prism 7.0, statistical analyses on parametric data were performed by using a 2-tailed Student's t test to compare individual groups, Mann-Whitney test was used for nonparametric data. A 1way ANOVA was performed on parametric data comparing more than two groups and a Kruskal-Wallis test was performed on nonparametric data, followed by Tukey's post hoc test. P value of <0.05 was considered significant.

| Arteriogenesis in adductor muscles
In addition to angiogenesis, arteriogenesis also is an important process during neovascularization. To investigate arteriogenesis, adductor muscles were stained with aSMActin ( Figure 3A Figure 4H). In the right non-ischaemic adductor muscle, Irf3 −/− and Irf7 −/− mice showed no differences in macrophages per μm circumference compared to C57BL/6 mice ( Figure 4I).

| mRNA expression of essential genes for arteriogenesis and angiogenesis
RNA from the gastrocnemius muscles was isolated and the mRNA expression of tnfα, ccl2, il6, il1β, sdf1, vegf164, vegfr1 and vegfr2 was measured with qPCR as genes essential for angiogenesis and No differences were observed in the mRNA expression of tnfα, ccl2, il6, il1β, sdf1, vegf164, vegfr1 and vegfr2 in the right non-ischaemic gastrocnemius muscles of Irf3 −/− and Irf7 −/− mice compared to control mice ( Figure S3). These results demonstrate that IRF3 and IRF7 regulate the expression of several genes essential for angiogenesis and arteriogenesis.

| D ISCUSS I ON
In the present study, we observed a decreased post-ischaemic blood flow recovery in Irf3 −/− and Irf7 −/− mice compared to C57BL/6 mice, using a HLI model. This was supported by a decrease in angiogenesis and arteriogenesis in, respectively, soleus and adductor muscles of  Since angiogenesis is induced via VEGF, it is interesting that we found a down-regulation of VEGR2 in the ischaemic muscles of IRF3 and IRF7-deficient mice. VEGF can bind to VEGR1, VEGFR2 and VEGFR3, but VEGFR2 is the main receptor transmitting VEGF signals. 30 The lack of VEGFR2 in the ischaemic tissue can (partly) explain the decrease in angiogenesis in IRF3 and IRF7-deficient mice compared to C57BL/6 mice. Interestingly, previously a relation between interferons, regulated by IRF3 and IRF7, and VEGF was observed.
IFNs were able to mediate VEFG production from human mast cells 31 and also in chronic myeloid leukaemia patients, imatinib in combination with pegylated-IFN-α2a was able to regulate VEGF levels. 32 Arteriogenesis is induced by inflammation, which results in the maturation of pre-existing collaterals into arterioles. 6 For optimal neovascularization and thus post-ischaemic blood flow recovery, both angiogenesis and arteriogenesis are essential. 21 We observed reduced arteriogenesis in the adductor muscles of Irf3 −/− and Irf7 −/− mice compared to C57BL/6 mice. The number of arterioles was only increased in Irf3 −/− mice compared to C57BL/6 mice. However, the diameter of the arterioles was significantly decreased in Irf3 −/− and Irf7 −/− mice compared to C57BL/6 mice, which is characteristic for arteriogenesis since this is defined as the maturation of pre-existing collaterals. 21 Interestingly, we showed a decrease in inflammatory cells around the arterioles and decreased mRNA expression of inflammatory cytokines tnfα, ccl2 and il6 in the gastrocnemius muscles.
It is shown that several inflammatory cytokines and components are essential in the regulation of arteriogenesis such as TNFα, 33 as well as the CCR7-CCL19/CCL21 Axis, 7

NK cells and CD4+ T cells 34 and
CD27-CD70 T cell co-stimulation. 25 We previously demonstrated that IRF3 regulates pro-inflammatory cytokines downstream TLR4 activation in vitro, in VSMCs and macrophages, prominent inflammatory cell types involved in arteriogenesis. 15 Together with the diminished inflammatory response observed here, this partly explains how IRF3 and IRF7 affect neovascularization.
Although from the current data on cytokine expression and macrophage presence at 28 days, it cannot be excluded that one might also conclude that due to the reduced neovascularization response, inflammatory cytokine expression is reduced and not vice versa.
Previous data from our and other groups have clearly shown that in early stage after induction of hind limb ischaemia hampered inflammatory responses lead to reduced neovascularization responses. 9,10,25,34 The regulation of pro-inflammatory cytokines via IRF3 and IRF7, in general goes via the TRIF-NFĸB pathway. 12,35 TRIF is also involved in TLR3-mediated type-I IFN production via IRF3 and IRF7 activation. It is suggested that activation of TLR3-TRIF can also suppress angiogenesis via initiation of apoptotic cell death. 36  showed, TRIF-NFĸB-mediated pro-inflammatory cytokines are important regulators of neovascularization. 9 In this respect, our results can only be partly explained by the effects of IFNβ and their IFNAR receptor. In addition, post-ischaemic blood flow recovery was measured in the paw of the mice, which was used to determine arteriogenesis. Both angiogenesis and arteriogenesis are essential for accurate neovascularization and post-ischaemic blood flow recovery. 21 In addition to arteriogenesis, which we determined by measuring the diameter of the arterioles in the adductor muscles of the Irf3 −/− , Irf7 −/− and C57BL/6 mice, we also measured angiogenesis in the soleus muscles.
The findings of the present study indicate that deficiency of IRF3 and IRF7 results in impaired post-ischaemic blood flow recovery caused by an attenuated angiogenesis and arteriogenesis linked to lack of inflammatory components in ischaemic tissue. In conclusion, IRF3 and IRF7 are essential regulators of neovascularization.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors confirm that there are no conflicts of interest.