All animal experiments were approved by The Finnish National Animal Experiment Board (ELLA) (license number: ESAVI/5343/04.10.07/2014) and followed the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes or the NIH Guide for the Care and Use of Laboratory Animals. The animals in the present study were done at the same time using the same methods as a set of aged, healthy mice and aged, hyperlipidemic mice without gene transfer which were previously published (Wirth et al., 2023). Hyperlipidemic mice deficient of LDL receptor and expressing only apolipoprotein B100 (LDLR^−/−ApoB^100/100 female mice, C57Bl/6J background bred in-house, n = 58), with an average age of 9 months, were used in this study. In addition to having a human-like lipoprotein profile, these animals normally have a 3-fold increase in blood cholesterol levels, predisposing them to develop atherosclerosis even on a regular chow diet (Véniant et al., 1998; Heinonen et al., 2007). Mice underwent permanent ligation of both the common femoral artery and vein proximal to the origin of the profound femoral artery of the right hindlimb, resulting in acute distal ischemia. Post-operatively, mice were randomized and allocated into two groups for adenoviral vector transfers. Adenoviral vectors, produced in the National Virus Vector Laboratories at the University of Eastern Finland, either expressing human VEGF-A₁₆₅ (AdVEGF) or beta-galactosidase (AdLacZ) control gene (Korpisalo et al., 2011) were delivered as a single injection into the ischemic posterior calf muscles. A total dose of 2 x 10¹² vp/mL in 50 µL 0.9% NaCl was evenly distributed along the injection route, starting from the proximal end towards the distal end of the calf muscle, using 0.33 mL insulin syringe. Mice were anesthetized with isoflurane inhalation (induction: 4.5% isoflurane, 450 mL/min air, maintenance: 2.5% isoflurane, 250 mL/min air; Baxter International, Deerfield, IL, United States) during surgery and ultrasound imaging. After surgery, mice received subcutaneously analgesia (Rimadyl, 10 mg/kg, Pfizer, Dublin, Ireland). The animals were fed on a regular chow diet and kept in standard housing conditions in the National Animal Laboratory Center of the University of Eastern Finland, Kuopio, Finland.

Tissue edema was assessed with MRI using 7-T Bruker PharmaScan Magnet (Bruker BioSpin, Ettlingen, Germany) interfaced to Paravision 5.1, using a quadrature volume coil as a transmitter and a 20 mm single loop surface coil as a receiver pre-operatively and at 1, 4, 7 and 29 days after femoral artery ligation and gene transfer. Mice were anesthetized with isoflurane (Baxter Oy, Helsinki, Finland) in 70% N₂: 30% O₂ for imaging and placed on a holder carefully stretching the hindlimbs straight over the imaging coil. A pneumatic pillow was placed under the mouse to monitor respiration using a small animal monitoring unit (Model 1,025, SA Instruments, Stony Brook, NY, United States). A warm water circulating pad was set over the mouse to maintain a temperature close to the physiological level. The legs were imaged using multi-slice T2-weighted fast spin echo images [repetition time (TR) = 2 s, effective echo time (TE) = 22 m, echo train length (ETL) = 4, field of view (FOV) = 20 × 20 mm², matrix = 256 × 256, two averages, 20 slices, slice thickness 1 mm]. From these images, the whole leg cross section area (mm²), used to measure edema build-up, was determined by hand-drawn regions of interest (ROIs) at a medial position of the calf muscle and expressed as an average of ROIs from three consecutive slices, using Aedes software package (aedes.uef.fi) on Matlab platform (Math Works, Natick, MA).

Resting blood flow in calf muscles was monitored and quantitatively measured with Acuson Sequoia 512 system and a 15L8 transducer (Siemens, Malvern, PA, United States). Contrast enhanced Power Doppler (CED) and Cadence contrast pulse sequencing (CEU) applications were used to differentiate between blood flow in large arterial/venous vessels and the microvasculature, as previously described (Wirth et al., 2023). Measurements were made pre-operatively and at 0, 4, 7 and 29 days after surgery and gene transfer. Briefly, transverse plane perfusion video clips (20 s in length, 15 clip frames/1 s) were captured immediately following injection of 50 μL bolus of a second-generation contrast agent (a sulfur hexafluoride gaseous core in a phospholipid shell, approx. 2 x 10⁸ bubbles/mL, mean size 2.5 µm, SonoVue, Bracco, Milan, Italy) via the jugular vein. Arterial/venous and microvascular skeletal muscle blood flow were quantified with Datapro software (v2.13, Noesis, Courtaboeuf, France) using CED and CEU peak signal intensities (dB; relative to both blood flow and volume) of the video clips, respectively. The level of arterial driving pressure was determined based on CEU-contrast arrival time measurements, i.e., the time (in seconds) from bolus injection to the arrival of the contrast agent into the imaging plane (Wirth et al., 2023; Rissanen et al., 2008). Maximal arrival time was recorded in instances in which the contrast agent did not arrive to the imaging plane within the 20-seconds-time measurement window. Ultrasound data analysis was performed in a blind manner.

Microvascular oxygen saturation of hemoglobin in calf muscles was monitored and quantitatively measured with Vevo 2100 LAZR system and a linear-array transducer LZ250 (VisualSonics Inc., Toronto, ON, Canada) in a photoacoustic (PA) mode (frequency = 21 MHz, PA gain = 50 dB, 2D gain = 18 dB, depth = 20 mm, width = 23.04 mm and dual laser wavelength = 750/850 nm) as previously described (Wirth et al., 2023). Measurements were performed pre-operatively and at 0, 4, 7 and 29 days after surgery and gene transfer. Muscle level average microvascular oxygen saturation of hemoglobin (mHbO₂%) was quantified with Vevo LAB software (v1.7.2, VisualSonics Inc., Toronto, ON, Canada) from at least 30 transverse plane PA clip frames.

Mice were euthanized with carbon dioxide and perfusion-fixed with 1% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) through the left ventricle. After dissection, posterior calf muscles were immersion-fixed in 4% PFA in 7.5% sucrose (pH 7.4) for 4 h and kept in 15% sucrose in H₂O (pH 7.4) overnight. After fixation, calf muscle samples were embedded in paraffin and transversely cut to 4 µm thick sections used for immunohistochemical and hematoxylin-eosin (HE) staining.

To evaluate post-ischemic muscle capillary reactivity, endothelium was immunostained with a rat monoclonal primary antibody against CD31 [platelet endothelial cell adhesion molecule (PECAM-1), dilution 1:25, MEC 13.3, BD Biosciences Pharmingen, San Diego, CA, United States] and biotinylated rabbit anti-rat secondary antibody (dilution 1:200, Vector Laboratories, Burlingame, CA, United States). To visualize the immunoreactivity, avidin-biotin-horseradish peroxidase system (Vector Laboratories, Burlingame, CA, United States) and 3,3′-Diaminobenzidine (DAB) color substrate (Zymed, San Francisco, CA, United States) were used. Tyramide signal amplification system (TSA, dilution 1:50, Biotin System, PerkinElmer, Shelton, CT, United States) was used to enhance signal intensity. Images were taken with Nikon H550L light microscope (Tokyo, Japan) and NIS-Elements advanced research imaging software at ×20 magnification. The following parameters were measured from the immunostained images: capillary area (endomysium localized CD31⁺ mean vessel luminal area in µm²), capillary density (endomysium localized CD31⁺ vessels/muscle area mm²), capillary size distribution [endomysium localized CD31⁺ vessels with a luminal area more than 33 μm² reported as a percentage of the total vessel amount/time point as previously described (Wirth et al., 2023)] and total capillary area (capillary density/capillary area in mm²). These measurements were calculated from two to three fields using NIS-Elements software (v4.51.01, Nikon, Tokyo, Japan). Only areas with myofiber histopathological changes were selected for the evaluation of CD31-based capillary reactivity as previously described (Lee et al., 2020). CD31-stained calf muscle sections deriving from intact hindlimbs (n = 8) were used to determine capillary parameters at baseline.

To assess post-ischemic myofiber responses, calf muscle sections were stained with HE as previously described (Wirth et al., 2023). The sections were then first carefully examined with Olympus AX-70 light microscope (Olympus Optical, Tokyo, Japan) at ×20 magnification, after which analysis images were obtained at ×1.25 magnification using analySIS imaging software (v3.00, Soft Imaging System GmbH, Muenster, Germany). Six irregularly distributed areas representing 1) normal area, 2) rounded myofibers, 3) necrosis, 4) early regenerative changes, 5) advanced regeneration and 6) late regeneration were measured as previously characterized in detail (Wirth et al., 2023). The damaged muscle area was calculated as the sum of areas 2–3. The regenerative muscle area was calculated as the sum of areas 4–6. In addition, area with mixed myofiber atrophy was identified by the presence of very small angulated myofibers often devoid of nuclei in addition to scattered very small rounded myofibers with/without peripherally oriented condensed, dark nuclei. Fibrotic area was defined by the presence of connective tissue together with scattered severely atrophied myofibers, of which some were almost reduced to clusters of nuclei, occasional fibroblasts, and mononuclear cells. Mononuclear pockets were considered large clusters of dark basophilic mononuclear cells located within the skeletal muscle tissue. All defined areas were reported as a percentage of the whole cross-sectional muscle area. All measurements were performed in a blind manner.

Statistical analyses were performed using SPSS software (v27.0 for Windows, SPSS Inc., Chicago, IL, United States). Results were expressed as mean ± standard error of the mean (SEM). The Kruskal–Wallis test followed by the Mann-Whitney U test were applied for the assessment of significant changes between the control and treated group as well as within each group. The difference in the prevalence of macroscopical ischemic symptoms between the two groups was assessed using the Chi-squared test. Differences were considered significant at ^*, ^#P < 0.05, ^**, ^##P < 0.01, ^***, ^###P < 0.001.

As quantified from CD31-immunostainings of ischemic muscle samples, AdVEGF gene transfer significantly increased mean capillary area of hyperlipidemic LDLR^−/−ApoB^100/100 mice as compared to the AdLacZ control (Figure 1A). A statistically significant increase was observed 7 days after the gene transfer although there was a visible enlargement present already on day 4 in some capillaries. In contrast to previous reports (Rissanen et al., 2005; Korpisalo et al., 2011; Kholová et al., 2007), the enlargement in AdVEGF stimulated capillaries persisted at least until the end of the follow-up on day 29. At that time, close to 30% of the capillaries in the AdVEGF group were still enlarged (Figure 1B). Total capillary area in the AdVEGF group was significantly increased only on day 29 as compared to AdLacZ control (Figure 1C). Capillary density was not affected by the AdVEGF gene transfer as it remained unchanged throughout the 29-day follow-up when compared to control values (Figure 1D). On days 4–7, many of the enlarged vessels in the AdVEGF group seemed aberrant in structure (Figure 1E). In contrast, on day 29 the enlarged vessels in AdVEGF muscles were structurally round and displayed thick walls. The angiogenic effect of the AdVEGF gene transfer in the ischemic hyperlipidemic muscle, hence, mainly seemed to result in long-term enlargement of pre-existing capillaries without increasing capillary number.

T2-weighted MRI was used to visualize edema build-up after the gene transfers (Figure 2A). On day 7, the operated hindlimb in the AdLacZ group showed mostly intramuscular edema formation as compared to the intact hindlimb. At the same time, the operated hindlimb in the AdVEGF group displayed not only intramuscular edema but also massive subcutaneous fluid deposits. When quantified as total hindlimb area, the MR-images in both groups showed significant swelling as compared to their respective contralateral intact legs (Figure 2B). However, significantly more hindlimb swelling was observed in the AdVEGF limbs as compared to the AdLacZ control limbs. By day 29, there were no signs of edema in either group. In the ischemic limbs of the hyperlipidemic mice AdVEGF gene transfer thus caused transient edema formation that on average doubled the cross-sectional hindlimb area on days 4–7.

Unlike previously in young and healthy animals, gene transfer of AdVEGF in aged and diseased mice did not improve muscle blood flow in this study. High-resolution CEU and CED ultrasound imaging were performed to differentiate between blood flow in the microvasculature (CEU) and large arterial/venous vessels (CED). As expected, femoral artery ligation significantly reduced microvascular flow similarly in both groups (Figure 3A, P < 0.001). 4 days after the operation, CEU contrast signal intensity was restored to baseline level in the AdLacZ group, but not in the AdVEGF group, indicating that AdVEGF transduced muscles had a slower microvascular flow recovery than the control muscles (Figure 3A, P < 0.05). The average CEU signal intensity in the AdVEGF group reached baseline values only by day 29. Arrival time of CEU contrast agent inversely reflects the level of arterial driving pressure in the operated hindlimbs. Femoral artery ligation significantly increased CEU arrival time in both groups (Figure 3B, P < 0.05). In the AdLacZ control, arrival time remained significantly increased throughout the follow-up, indicating that arterial driving pressure did not recover in the AdLacZ control. In contrast, CEU arrival time in the AdVEGF group gradually decreased, recovering to baseline levels by day 29. Large arterial/venous CED flow significantly decreased upon the ischemic insult and remained significantly reduced equally in both groups during the 29-day follow-up (Figure 3C, P < 0.01). PAI-based quantification of mHbO₂% displayed that femoral artery ligation significantly reduced ischemic muscle microvascular hemoglobin oxygenation (Figure 3D, P < 0.001). During the later follow-up, PAI-based quantification of mHbO₂% showed no significant differences between the treatment groups and almost no recovery to baseline levels during the follow-up, except with AdVEGF on day 29. Based on the blood flow parameters, AdVEGF gene transfer therefore delayed post-ischemic recovery of microvascular blood flow. Nevertheless, on day 29 the AdVEGF group displayed significantly higher microvascular blood flow than the AdLacZ control group and also displayed recovery of both arterial driving pressure and microvascular haemoglobin oxygenation in contrary to the AdLacZ control.

Based on visual examination, AdVEGF transduced hindlimbs were more often stiff and red than limbs in the AdLacZ group on days 4–7 (Table 1, P < 0.01). Also, macroscopic signs of severe ischemia, such as gangrene or paw autoamputation, were only seen in the AdVEGF group. Quantitation of HE-based muscle pathology (Figures 4A, B) displayed that on days 4–7 AdVEGF transduced ischemic calf muscles were significantly more damaged than AdLacZ muscles (Figure 4A, P < 0.01). The damage was predominantly presented as necrosis and scattered rounding of myofibers (Figure 4C). Regardless of differences in damage, areas of regeneration were similar in the groups during the 29-day follow-up (Figure 4B). In both groups, % of the muscle area under regeneration significantly increased over time, leaving around half of the calf muscle under regeneration on day 29. Beyond the common features of satellite cell-mediated muscle regeneration (Figure 4C), areas of mixed myofiber atrophy (Figure 4D) and fibrosis (Figure 4E) were found in muscle samples of both groups on day 29, indicative of chronic muscle damage. Additionally, pockets of mononuclear cell infiltrates (Figure 4F) were commonly found in muscle samples of both groups, possibly suggesting adaptive immune responses towards cells transduced by the adenoviral vector. Beyond possible damage caused by immune response towards the adenovirus on day 29, the angiogenic effects of the AdVEGF seemed to cause significant aggravation of ischemic damage during days 4–7.