**3. Discussion**

In this study we have shown that modulation of IL10, in particular elevated levels of circulating IL10, alters collateral reperfusion after femoral artery ligation (FAL) as well as the distribution of macrophage activation types in the perivascular bed of growing collaterals. Although, the inhibitory effect of an IL10 antibody on circulating IL10 blood levels could not be directly detected due to technical limitations, contrary to elevated levels of circulating IL10, the opposite pertaining to both collateral reperfusion and the distribution of macrophage activation types was observed. The remodeling processes in arteriogenesis involve a controlled destruction of vessel components, activation of endothelial and smooth muscle cell de-differentiation, proliferation and migration, and adventitial restructuring in which monocytes/macrophages were shown to be key orchestrators [1,14]. These seemingly disparate tasks mediated largely by a singular cell type are explained by their heterogeneity and plasticity in response to environmental and situational needs. Previous findings have demonstrated that M1 and M2 macrophages form around growing collateral vessels in a distinct temporal and spatial pattern suggesting a collaborative mechanism of action in collateral artery growth [11]. The M1 phenotype, found proximate to the vessel lumen, is associated with the production of the pro-inflammatory cytokines IL1 and IL6, both described to have autocrine growth e ffects on vascular smooth muscle cells (VSMC) [15]. Other arteriogenic stimulants expressed by the M1 phenotype include NOS, TNF and MCP-1 [1,2,15,16]. M2 macrophages, on the other hand, are found distal to the vessel lumen [11] and have been ascribed a more prominent role in mediating the growth processes. Takeda et al. [7] showed that improved collateral perfusion after FAL was due to an expansion of the M2 phenotype among tissue-resident macrophages in Phd2 haplodeficient (Phd2+/-) mice and that soluble factors secreted by Phd2+/- macrophages in vitro led to increased smooth muscle cell (SMC) proliferation and migration. IL10 is a known activator of the M2c phenotype and we have shown that increased levels of IL10 also led to an improved collateral reperfusion after FAL. Furthermore, an immunohistochemical analysis using the established macrophage polarization marker CD163 [17,18], revealed that perivascular macrophages were proportionally skewed towards the M2 phenotype. The supposed proarteriogenic e ffects of M2 macrophages induced by IL10 are further supported by the observation that the application of an IL10 antibody led to a transient impairment of collateral reperfusion and was accompanied by an increased onset of ischemic symptoms on ligated hind limbs. Also, the distribution of perivascular macrophages was conversely skewed towards the M1 phenotype. These findings do not undermine the role of M1 macrophages in arteriogenesis. The delayed and transient dip in reperfusion recovery we observed after the application of an IL10 antibody may support the hypothesis, that adluminally located M1 macrophages contribute to collateral vessel growth particularly in early phases through the recruitment of circulating monocytes and expression of arteriogenic relevant cytokines and proteins (IL1, IL6, TNF, NOS, MCP-1). They do, however, highlight the potential role played by M2 macrophages with regard to VSMC di fferentiation and adventitial restructuring to accommodate the growing collateral vessel. M2 macrophages induced by IL10 secrete high levels of TGFb1, known to stimulate VSMC di fferentiation and extracellular matrix (ECM) deposition [15,19,20]. They also produce high levels of MMP9 [21], which along with MMP2 was significantly increased in the adventitia of growing coronary collateral vessels [22]. This supports the hypothesis that IL10 induced M2 macrophages may play an active role in the augmentation of adventitial ECM proteolysis and remodeling, thus, facilitating arteriogenic growth. Our findings sugges<sup>t</sup> that varying levels of IL10 in vivo influence collateral reperfusion, which may be explained by a shift in the distribution of macrophages towards the M2 phenotype at the site of collateral growth. From a clinical standpoint, this presents itself as a new therapeutic approach to promote collateral vessel growth in patients su ffering from peripheral artery disease. The expression of CD163 on M2 macrophages, however, also poses concerns regarding their use in revascularization therapy. CD163+ macrophages in human atherosclerotic lesions were found to increase plaque instability by promoting angiogenesis within areas of intraplaque hemorrhage, in itself thought to be a result of plaque neovascularization and increased microvessel permeability, resulting in a vicious cycle [23]. Seen as a whole, these observations underline the diversity of macrophage functions and activation states with regard to tissue specific cues and needs. A mere distinction between classically activated M1 and alternatively activated M2 macrophages limited by CD163 as an M2 marker alone will not su ffice to fully describe the roles played by the macrophage activation states we observed around growing collateral vessels. Further studies utilizing other macrophage markers will be required to elucidate the underlying mechanism involved in our findings. While these remain hypothetical, the results presented in our study shed light on new therapeutic strategies in promoting collateral vessels growth.

### **4. Materials and Methods**

### *4.1. Animal Models*

Animal handling and all experimental procedures carried out were in full compliance with the Directive 2010/63/EU of the European Parliament on protection of animals used for scientific purposes. Approval was given by the responsible local authority, the hessian governmental council for animal protection and handling (permit reference numbers V54-19c20/15-B2/1152, permit date: 23.05.2017). Throughout this study all mice had access to water and food ad libitum.

### *4.2. Mouse Model of Hind-Limb Ischemia*

To evaluate collateral vessel growth perfusion recovery was measured after femoral artery ligation (FAL) using the model described in [24]. For all experiments 10-14 weeks old male C57BL/6 mice from our own breeding program were used with an approximate bodyweight of 30 g. Prior to each experiment all mice were inspected to ensure a healthy state. Anesthesia was performed by intraperitoneal injection using ketamin hydrochloride (120mg/kg bodyweight) and xylazine hydrochloride (16 mg/kg bodyweight). Pre- and post-operative analgesia was performed by subcutaneous injection with buprenorphine (0.1 mg/kg bodyweight). FAL was carried out on the left hind-limb by ligating the femoral artery immediately distal to the origin of the deep femoral branch to redirect blood flow to the collateral arteries. After termination of experiments the mice were euthanized by an anesthetic overdose using ketamin hydrochloride (180mg/kg bodyweight) and xylazine hydrochloride (16 mg/kg bodyweight) followed by exsanguination.

### *4.3. Pharmacological Stimulation*

FAL was performed on C57BL/6 mice. Mice were randomly allocated to each group receiving either recombinant murine interleukin 10 (IL10) (20 μg/kg bodyweight) or purified anti-mouse IL10 antibody (anti-IL10) (0.5 μg/kg bodyweight) diluted in sodium chloride solution (NaCl 0.9%) immediately after FAL, on day 3 and 7 after FAL. The control group received NaCl 0.9%. The application was carried out via intravenous injection into the tail vein. IL10 was purchased from PeproTech (Hamburg, Germany). Anti-IL10 was purchased from BioLegend (Koblenz, Germany).

### *4.4. Measurement of Blood Concentration Levels of IL10*

Retro orbital blood samples were obtained from mice before and 24 h after pharmacological stimulation with IL10 and anti-IL10 as described above. The control group received NaCl 0.9%. Anesthesia war provided as described above. A Mouse Magnetic Luminex Assay (Thermo Fisher Scientific, Waltham, MA, USA) was used to measure the concentration of endogenous IL10 in the blood samples at baseline and 24 h after pharmacological stimulation.

### *4.5. Hind-limb Perfusion Measurement after Pharmacological Stimulation and FAL*

Hind-limb perfusion was assessed and quantified via erythrocyte motion detection through Laser-Doppler-Imaging using a PeriScan PIM3 System (Perimed Instruments, Järfällä, Sweden, Software: LDPIwin for PIM3 3.1.3) before and immediately after FAL, on day 3, 7 and 14 after FAL. For each measurement mice were positioned on a heating plate at 37 ◦C for 3min prior to and during each measurement to ensure standardized conditions at a distance of 10 cm and a pixel resolution of 256 × 256. A 2 cm × 3 cm area including both feet was scanned and a region of interest (ROI) of approximately 80 mm<sup>2</sup> containing each foot was defined. Mean perfusion (arbitrary units) was used to calculate hind-limb perfusion and expressed as the ligated limb to non-ligated limb ratio as described in [25]. Follow-up measurements were performed under anesthesia as described above.
