AAV-mediated expression of PFKFB3 in myofibers, but not endothelial cells, improves ischemic muscle function in mice with critical limb ischemia

Abstract

6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 (PFKFB3) is a powerful driver of angiogenesis through its modulation of glycolytic metabolism within endothelial cells. Recent work has demonstrated that PFKFB3 modulates the response to muscle ischemia, however the cell specificity of these effects is not fully understood. In this study, we tested the impact of viral mediated expression of PFKFB3, driven by gene promoters specific for myofibers or endothelial cells, on ischemic hindlimb revascularization and muscle function. We hypothesized that both endothelium- and muscle-specific expression of PFKFB3 would attenuate limb pathology following femoral artery ligation. Male and female BALB/cJ mice were injected with adeno-associated virus encoding the either a green fluorescent protein (GFP) or PFKFB3 driven by either the human skeletal actin (ACTA1) or cadherin-5 (Cdh5) promoters. Four weeks after AAV treatment, mice were subjected to unilateral femoral artery ligation and limb perfusion and muscle function were assessed. Both endothelium- and muscle-specific PFKFB3 expression resulted in significantly more perfused capillaries within the ischemic limb muscle, but neither changed myofiber size/area. Muscle-, but not endothelium-specific, PFKFB3 expression significantly improved maximal force production in ischemic muscle (P=0.0005). Notably, there was a significant effect of sex on maximal force levels in both cohorts of mice (P=0.0075 and P=0.0481), indicating that female mice had higher ischemic muscle strength compared to male mice, regardless of treatment group. Taken together, these data demonstrate that while both muscle- and endothelium-specific expression of PFKFB3 enhanced ischemic revascularization, only muscle-specific PFKFB3 expression improved muscle function.

INTRODUCTION

Critical limb ischemia (CLI) is the most severe manifestation of peripheral artery disease (PAD) consequent to atherosclerotic occlusion of blood vessels in the lower limb. CLI patients often experience chronic limb pain at rest and commonly suffer from non-healing wounds and/or gangrene. Unfortunately, there are currently no approved pharmaceutical therapies that treat CLI pathophysiology directly and surgical interventions are burdened by unacceptably high failure rates (1-3). There is a growing need to consider the pathophysiological role of multiple tissue/cell types within the ischemic limb, including resident skeletal myocytes which are crucial for optimal physical function (4-8). The ischemic microenvironment within the limb places a tremendous bioenergetic burden on the resident cells/tissues to maintain a constant supply of ATP required to sustain cellular life. In most cells, the majority of ATP is produced via oxidative phosphorylation within the mitochondria. However, this process becomes compromised in conditions of low oxygen availability. In addition to low oxygen delivery, CLI patients suffer from a devastating mitochondrial myopathy which compromises the efficiency by which the myocyte can utilize the limited available oxygen (9, 10). In this context, uncovering mechanisms that facilitate metabolic flexibility and the ability to utilize alternative metabolic pathways (i.e. non-aerobic) could protect against necrotic cell death caused by ATP depletion.

To this end, we recently demonstrated that expression of 6-phosphofructo-2-kinase/fructose- 2,6-bisphosphatase 3 (PFKFB3), a known enhancer of glycolysis, was sufficient to reduce ischemic limb myopathy in mice and enhance muscle cell survival in hypoxia (11) . In this previous study, viral mediated expression of PFKFB3, under control of a ubiquitous CMV promoter, improved limb perfusion recovery and muscle function following femoral artery ligation in the CLI-susceptible BALB/c mouse. However, PFKFB3 has been shown to play crucial roles on angiogenesis via regulation of endothelial cell metabolism (12-16). Genetic knockdown of PFKFB3 significantly reduced the number and length of sprouts in endothelial cell spheroids (14) and PFKFB3 blockage has been shown to normalize pathological angiogenesis(15). Mechanistically these changes were attributed to reduced endothelial cell proliferation, impaired lamellipodia formation, and a reduction in tip cell activity – all processes involved in sprouting angiogenesis. Overexpression of PFKFB3 has been shown to improve sprouting angiogenesis (14, 16), a result that has been partially attributed to increased production and secretion of lactate (16, 17). Xu et al. provide evidence that exogenous lactate supplementation could abrogate the effects of PFKFB3 knockdown on endothelial cell tube formation (16). This observation is supported by a recent study demonstrating that endothelium cell specific deletion of PFKFB3 impaired limb perfusion recovery in mice subjected to hindlimb ischemia (17). The authors of this paper further linked endothelial cell PFKFB3 to lactate secretion which was found to serve a paracrine role promoting M2 macrophage polarization that enhanced ischemic muscle regeneration.

In the present study, we sought to determine whether the beneficial impact of PFKFB3 expression in the ischemic limb was derived from the PFKFB3-mediated changes in the endothelial cell or the muscle cell. To address this question, we employed muscle- and endothelium-specific viral mediated overexpression of PFKFB3 in mice subjected to a murine model of CLI. We hypothesized that both endothelial and muscle specific over-expression of PFKFB3 would improve limb pathology following femoral artery ligation.

METHODS

Animals.

All animal experiments performed herein involved male and female mice and experiments began at 12 weeks of age. BALB/cJ mice (N=56 total mice), obtained from Jackson Laboratories (Stock # 000651), were used because of their susceptibility to CLI pathology (18-22). All mice were housed in a temperature (22°C) and light controlled (12-hour light/12-hour dark) room and maintained on standard chow with free access to food and water. All animal experiments adhered to the Guide for the Care and Use of Laboratory Animals from the Institute for Laboratory Animal Research, National Research Council, Washington, D.C., National Academy Press, and any updates. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Florida (Protocol 201810121).

Plasmid constructs and adeno-associated virus generation/delivery.

The PFKFB3 coding sequence was PCR amplified from cDNA generated from a gastrocnemius muscle obtained from a C57BL6J mouse. To accomplish muscle cell-specific expression of PFKFB3, the human ACTA1 promoter (1541 bp proximal to the transcription start site) was PCR amplified from human genomic DNA isolated from a donor muscle biopsy. To accomplish endothelium-specific expression of PFKFB3, a 2510 bp portion of the Cdh5 promoter (23-25) was PCR amplified from genomic DNA isolated from a C57BL6J mouse. The resulting promoters and PFKFB3 coding sequence were inserted into promoter-less AAV cloning vectors (CellBio Labs) using In-Fusion cloning (Takara Bio). AAV’s were packaged using AAV2/9 serotype and obtained from Vector Biolabs (Malvern, PA). AAV’s were injected intramuscularly to the tibialis anterior (TA, 50μl injection volume) and extensor digitorum longus (EDL, 12μl injection volume) muscles of the surgical limb at 5 x 10¹¹ vg/mouse (HSA vectors) or 4 x 10¹² vg/mouse (Cdh5 vectors) four weeks prior to surgery to induce CLI. Individuals involved in study procedures were blinded to the grouping until all data analyses were completed.

Animal model of critical limb ischemia.

Femoral artery ligation (FAL) (18) was performed by anesthetizing mice with intraperitoneal injection of ketamine (90 mg/kg) and xylazine (10 mg/kg) and surgically inducing unilateral hindlimb ischemia by ligation of the femoral artery just below the inguinal ligament. The inferior epigastric, lateral circumflex, and superficial epigastric artery branches of the femoral artery were left intact, thereby preserving collateral perfusion to the limb. Sham surgeries were performed accordingly and involved the complete dissection and separation of the neurovascular bundle, but no ligation was performed.

Laser Doppler Limb Perfusion Measurements.

Limb perfusion was measured using a laser Doppler flowmeter (moorVMS-LDF, Moor Instruments) prior to surgery, immediately post-surgery, 14 days post-surgery and just prior to euthanasia (day 28 post-surgery). The hindlimbs were shaved to remove hair and the laser Doppler probe was carefully placed ~1mm above the mid-belly of the TA muscle. Data were collected continuously for 60 seconds, and the average perfusion was calculated. Perfusion recovery in the ischemic limbs was calculated as a percentage of the non-ischemic control limb as previously described (6, 7).

RNA Isolation and qRT-PCR.

Total RNA was extracted from the TA and EDL muscles using Trizol-Phenol Reagent (Invitrogen; Cat. No. 15596026) as described by manufacturer’s instruction. UV-spectroscopy (ThermoFisher Scientific, Nanodrop 2000) was used to assess RNA quantity and quality. cDNA was generated by reverse transcribing total RNA using Superscript IV (ThermoFisher; Cat. No. 18091200). Real-time PCR (RT-PCR) was performed on Quantstudio 3 (ThermoFisher Scientific) using Taqman Fast Advanced Master Mix (ThermoFisher Scientific; Cat. No. 4444963) and Taqman FAM-labeled probe for PFKFB3 (ThermoFisher Scientific; Mm00504650_m1) multiplexed with VIC-labeled probe for 18s (ThermoFisher Scientific; Hs03003631_g1). mRNA values were normalized to 18s and relative gene expression was calculated using 2^−ΔΔCT from the GFP control group.

Skeletal muscle morphology and ischemic lesion area.

Skeletal muscle morphology was assessed by standard light microscopy. Transverse sections (10-μm thick) from the TA and EDL muscles were cut using a Leica 3050S cryotome and collected on slides for staining. For morphological analyses, standard methods for hematoxylin and eosin (H&E) histological staining were performed. Briefly, slides were dipped in Mayers’s Hematoxylin (Millipore-Sigma; Cat. No. MHS1) five times and left in solution for 15 minutes. A 15-minute wash was performed by placing slides in a circulating water bath. Slides were then dipped in an eosin (Millipore-Sigma; Cat. No. HT110132) solution five times followed by a one-minute incubation. Ethanol washes were performed at increasing ethanol concentration (60% in diH2O, 70%, 80%, 90%, and 100%). A 1:1 xylene/ethanol solution was used to clear the slides of excess stain and slides were dried and coversliped in Permount (ThermoFisher Scientific; Cat. No. SP15-500). Images were obtained at 20x magnification using automated image capture/tiling in order to image the entire muscle section using an Evos FL2 Auto microscope (ThermoFisher Scientific). All image analyses were conducted by a blinded investigator using Image J.

Assessment of myofiber cross-sectional area.

Skeletal myofiber cross-sectional area (CSA) was assessed by immunofluorescence microscopy. 10-μm-thick transverse sections were cut from the mid-belly of the TA and EDL muscles frozen in liquid nitrogen cooled isopentane in optimum mounting medium (OCT) using a Leica 3050S cryotome. Muscle sections were fixed in ice cold acetone for 30 seconds for quantification of cross-sectional area and regenerating fibers. Sections were then blocked for one hour at room temperature with 1x PBS supplemented with 5% goat serum and 1% bovine serum albumin. Slides were then incubated overnight at 4°C with a primary antibody for laminin (Millipore-Sigma; Cat. No. L9393) at 1:100 in blocking solution. Following washes with PBS, muscle sections were incubated with AlexaFluor secondary antibodies (ThermoFisher Scientific; Alexa Fluor Goat Anti-Rabbit IgG 555 Cat. No. A32732, at 1:250 dilution). Coverslips were mounted with Vectashield hardmount (Vector Laboratories; Cat. No. H-1400). Images of the entire TA and EDL muscles were obtained at 20x magnification using automated imaging and tiling routines on an Evos FL2 Auto microscope (ThermoFisher Scientific). Skeletal myofiber CSA was calculated using the FIJI automated plugin, MuscleJ, as previously described (26).

Assessment of total and perfused capillary density.

To label perfused capillaries within the ischemic limb muscle, mice received an injection of 50μL of 1 mg/mL Griffonia simplicifolia lectin (GSL) isolectin B4, Dylight 594 (Vector Laboratories; Cat. No. DL-1207) via the retro-orbital sinus two hours prior to euthanasia. Following the injection, mice were returned to their cages and allowed normal cage activity during which the fluorescently-labeled isolectin was able to bind alpha-galactose residues on the surface of endothelial cells. Following euthanasia, muscles were frozen in OCT compound and cryosectioned as described above. Perfused and total capillaries were quantified on serial sections. Sections were fixed with 4% paraformaldehyde (ThermoFisher Scientific; Cat. No. J19943-K2) for five minutes and subsequently permeabilized with 0.25% triton X-100 (Millipore-Sigma; Cat. No. 93443) and washed three times with PBS and blocked for one hour at room temperature with 1x PBS supplemented with 5% goat serum and 1% bovine serum albumin. Slides for total capillary density quantification were incubated overnight at 4°C with a rabbit anti-CD31 antibody (Abcam; Cat. No. ab28364) at a 1:100 dilution. The next morning slides underwent four 3-minute washes with PBS and were then incubated in Alexa Fluor Goat Anti-Rabbit IgG 555 (ThermoFisher Scientific; Cat. No. A32732) for one hour at a 1:250 dilution. Slides were again washed four times with PBS and coverslipped with Vectashield hardmount (Vector Laboratories; Cat. No. H-1400). Images of the entire TA and EDL muscles were obtained at 20x magnification using automated imaging and tiling routines on an Evos FL2 Auto microscope (ThermoFisher Scientific). Quantification of the total and perfused capillary counts of the entire TA and EDL muscles were performed using manual counting in Fiji by a blinded investigator.

Immunofluorescence Detection of PFKFB3.

To confirm cell-specific PFKFB3 expression, muscle bundles were prepared by mechanically separating fibers in PFA-fixed specimens under a dissecting microscope. Following gentle separation, muscle bundles were permeabilized with 30μg/ml saponin (Millipore-Sigma; Cat. No. S7900) for 30 minutes in PBS. After washing permeabilized bundles in PBS with gentle rocking, muscle bundles were blocked in PBS supplemented with 5% goat serum and 1% BSA for one hour are room temperature with gentle rocking. Next, muscle bundles were immunolabeled with an anti-rabbit PFKFB3 (Proteintech; Cat. No. 13763-I-AP) antibody at a 1:100 dilution overnight at 4°C. The next morning after sequential washes with PBS, they were incubated with AlexaFluor Goat Anti-Rabbit IgG 555 secondard antibody (ThermoFisher Scientific; Cat. No. A32732; 1:250 dilution) and either 5μg/ml Wheat Germ agglutin-647 [muscle-specific AAV9 specimens] (ThermoFisher Scientific; Cat. No. W32466) or 5μg/ml Griffonia simplicifolia lectin (GSL) isolectin B4, Dylight 647 to counterstain endothelial cells [endothelium-specific AAV9 specimens] (Vector Laboratories; Cat. No. DL-1208) for one hour at room temperature. Bundles were then washed again with PBS and positioned on coverslips for confocal microscopy. Antibody specificity was confirmed by collection of images from specimens stained without the primary antibody incubation, but processed exactly through the other steps described above. Images were collected on a Stellaris 5 (Leica Microsystems) confocal microscope using a x40 oil immersion objective (NA = 1.3, Leica Plan Apochromat 11506359). All images were acquired at 8192 x 8192 pixel with a 187.5 ns/pixel dwell time, sequential scan mode, with 400Hz scan speed and 0.75x digital zoom. Z-stack max projections were generated in FIJI/ImageJ and images were pseudocolored using lookup tables. No adjustments to brightness or contrast were made.

Assessment of Muscle Contractile Function.

To access skeletal muscle function, the left (ischemic) extensor digitorum longus (EDL) muscle was surgically removed with the proximal tendon mounted to a Dual-Mode Muscle Lever System (Aurora Scientific; 300C-LR). The distal tendon was looped around a metal hook in between two platinum electrodes submerged in an organ bath containing a bicarbonate-buffering solution (137 mM NaCl, 5 mM KCl, 1 mM MgSO₄, 1 mM NaH₂PO₄, 24 mM NaHCO₃, and 2 mM CaCl₂) which was continuously gassed with 95% O₂ - 5% CO₂ at 25°C and allowed 5-minute thermal equilibration period. Briefly, optimal muscle length (L_o) was assessed by stimulating the muscle every minute with a 1 Hz twitch. After each twitch, the length of the muscle was increased by 0.3mm using a Dual-Mode Lever System (Aurora Scientific) until peak twitch force was obtained. Isometric forces were measured at stimulation frequencies of 1, 20, 60, 100 and 150 Hz using a biphasic high-power stimulator (701C, Aurora Scientific) delivered with current of 600 mA, pulse duration 0.25 ms, train duration 500 ms. Each muscle stimulation was separated by a one-minute rest interval. The muscle stimulation and data collection were controlled through an automated software (DMC, Aurora Scientific). Isometric forces were normalized to muscle cross sectional area, which was estimated by measuring muscle weight and length at L_o. Muscle weight was divided by length multiplied by 1.06 g/cm³, the density of mammalian skeletal muscle(27).

Statistical Analysis.

Data are presented as mean ± SD. Normality of data was assessed using the Shapiro-Wilk test. Data were analyzed using two-way ANOVA with Tukey’s post-hoc multiple comparisons when significant interactions were detected. Comparisons between two groups were made using two-tailed unpaired Student’s t-tests. All statistical analysis was performed in GraphPad Prism (Version 8.0). In all cases, P < 0.05 was considered statistically significant.

Results

Source data for this study can be assessed at the following link: https://figshare.com/s/48cb84c4b243b5c3783c.

AAV-mediated expression PFKFB3 in myofibers improves perfused capillary density but not limb perfusion following FAL in both males and females.

To examine the role of PFKFB3 in skeletal muscle, mice received adeno-associated virus-containing a green fluorescent protein (AAV9-HSA-GFP) or PFKFB3 (AAV9-HSA-PFKFB3) driven by a human skeletal actin promoter (HSA, ACTA1). GFP expression was confirmed using microscopy (Fig. 1A) and qPCR analyses demonstrated a 110 ± 9.59 and 27 ± 6.13-fold increase in PFKFB3 mRNA expression in the TA and EDL muscles respectively (Fig. 1B). Immunolabeling of PFKFB3 also confirmed the increased expression of PFKFB3 within the myofibers of mice treated with AAV9-HSA-PFKFB3 (Fig. 1C). To investigate the impact of myofiber PFKFB3 expression on CLI limb pathology, mice were injected with AAV four weeks prior to FAL and were allowed to recovery from surgery for 28 days (Fig. 1D). Male mice treated with AAV9-HSA-PFKFB3 displayed a ~19% non-significant increase in TA muscle perfusion recovery at day 14 compared to control treated male animals, but no group or sex interactions were detected, and this improvement eroded by day 28 (Fig. 1E). There were no effects of biological sex on limb perfusion recovery following FAL surgery. Histological analysis of total capillary counts revealed that AAV9-HSA-PFKFB3 male and female mice had a 21% and 18% increase in total capillaries, although this was not statistically significant (P=0.074) (Fig. 1F). However, AAV9-HSA-PFKFB3 treated ischemic TA muscles displayed a significant increase in perfused capillaries (P=0.0485) when compared to GFP treated ischemic muscle (Fig. 1G). In contrast to the TA muscle, the EDL muscle of mice treated with AAV9-HSA-PFKFB3 do not display an increase in total or perfused capillaries (Fig. 1G, H). This discrepancy may be explained by the fact that qPCR analysis demonstrated that PFKFB3 expression was at least 3-fold greater in the TA muscle compared to the EDL.

Figure 1: Skeletal muscle-specific PFKFB3 expression increased perfused capillary density but not limb perfusion recovery following FAL.

(A) Representative images of the TA muscle injected with either AAV9-HSA-Empty or AAV9-HSA-GFP labeled with DAPI (nuclei) and wheat germ agglutin (myofibers). (B) Quantification of PFKFB3 mRNA levels within the TA and EDL muscles (n=3/group). (C) Representative immunolabeling of PFKFB3 demonstrates myofiber expression following AAV treatment. (D) Graphical depiction of study design created using Biorender.com. (E) Quantification of limb perfusion before and after FAL (n=5-6/group/sex). Measurements are reported as a percentage of the non-ischemic contralateral limb. *P<0.05 and ***P<0.001 vs. GFP using an unpaired two-tailed t-test. (F) Quantification of total capillaries within the TA muscle assessed via immunofluorescence staining of CD31⁺ vessels (n=5-6/group/sex). (G) Quantification of perfused capillaries within the TA muscles assessed via immune-labeling of endothelial cells with systemic delivery of a fluorescent isolectin (n=5-6/group/sex). (H) Quantification of total capillaries within the EDL muscle assessed via immunofluorescence staining among male and female mice treated with AAV9- HSA-GFP or AAV9-HSA-PFKFB3 post HLI (n=5-6/group/sex). (I) Quantification of perfused capillaries within the EDL muscle assessed via immunofluorescence staining among male and female mice treated with AAV9-HSA-GFP or AAV9-HSA-PFKFB3 post HLI (n=5-6/group/sex). The solid line represents the average values of sham treated animals; the dotted lines represent the standard error of the mean within the sham groups. Error bars represent SD. Panels D-H were analyzed using two-way ANOVA.

AAV-mediated skeletal muscle PFKFB3 expression improves ischemic force production in CLI mice.

To examine the impact of increased PFKFB3 expression on muscle contractile function, submaximal and maximal muscle contraction of the EDL was performed ex vivo (Fig. 2A). Maximal specific force in mice treated with AAV9-HSA-PFKFB3 was significantly higher compared to GFP-treated animals (P=0.0005) (Fig. 2B). Interestingly, there was a significant effect of sex on peak specific force demonstrating that female mice exhibited higher force production in ischemic muscles compared to male mice (P=0.0075) (Fig. 2B). Representative images of ischemic muscles stained with hematoxylin and eosin showed classical signs of ischemic injury including fibers with centralized nuclei and clusters of small irregular shaped myofibers (Fig. 2C). Although AAV9-HSA-PFKFB3 treated animals displayed a modest improvement in force production, this was not due to differences in myofiber cross sectional area in the EDL or TA muscle (Fig. 2D-G). Consistent with force production, there was a significant effect of biological sex on myofiber cross sectional area within the EDL (Fig. 2G) demonstrating female mice exhibited higher mean cross-sectional area compared to male mice (P= 0.0386).

Figure 2: Skeletal muscle-specific PFKFB3 expression increased ischemic muscle force levels but did not change myofibers area.

(A) Force-frequency curves of the EDL muscle in male and female mice treated with either AAV9-HSA-GFP or AAV9-HSA-PFKFB3 (n=5-6/group/sex). *P<0.05 vs. AAV9-HSA-GFP using two-way ANOVA. (B) Peak specific force in the EDL muscle (n=5-6/group/sex). (C) Representative H&E images of the TA muscle from sham (non-ischemic) and AAV-treated ischemic muscles. (D) Histograms of the TA myofiber CSA presented as a percentage of frequency distribution in both male and female mice (E). Mean TA myofiber area in male and female mice (n=5-6/group/sex). (F) Histogram of EDL myofiber CSA presented as a percentage of frequency distribution in both male and female mice (G) Mean EDL myofiber area in male and female mice (n=5-6/group/sex). Error bars represent SD. Panels B, E and G were analyzed using two-way ANOVA.

Endothelial cell-specific expression of PFKFB3 increased capillary density and enhanced limb perfusion recovery following femoral artery ligation.

Because PFKFB3 is a well-established regulator of endothelial cell-driven angiogenesis (12-16, 28), we delivered AAV encoding PFKFB3 or GFP driven by the endothelial cell-specific cadherin 5 promoter (AAV9-Cdh5-PFKFB3) to mice prior to FAL. Immunofluorescence microscopy demonstrated endothelium-specific GFP expression evidenced by co-localization with Dylight-594 conjugated Griffonia Simplicifolia Lectin I isolectin B4 which labels alpha-galactose residues on endothelial cells (Fig. 3A). Further to this, immunolabeling of PFKFB3 confirmed the effectiveness of the Cdh5 promoter to drive enhanced expression of PFKFB3 in the limb muscle vasculature (Fig. 3B). Using an identical experimental timeline to muscle-specific PFKFB3, male and female BALB/cJ mice were treated with AAV four weeks prior to FAL surgery and allowed to recover for 28 days following surgery (Fig. 3C). After FAL, both male and female AAV9-Cdh5-PFKFB3 treated animals displayed similar perfusion recovery as GFP-treated animals, indicating that endothelium-specific PFKFB3 did not enhance TA muscle perfusion recovery in CLI mice (Fig. 3D). Notably, female mice had significantly higher TA perfusion rates compared to male mice at day 28 post-surgery (P=0.0271) (Fig. 3D). AAV9-Cdh5-PFKFB3 treated mice demonstrated no statistical differences in total capillary counts (group effect P= 0.1403), but interestingly, displayed significantly higher perfused capillary counts (group effect P= 0.0310) within the TA muscle when compared to AAV9-Cdh5-GFP treated mice (Fig. 3E,F). In the EDL muscle, AAV9-Cdh5-PFKFB3 treated mice displayed no differences in total capillaries (group effect P=0.5265) but displayed a non-significant increase in perfused capillary density (group effect P= 0.0768) (Fig. 3G,H).

Figure 3: Endothelial cell-specific PFKFB3 expression increased perfused capillary density but not limb perfusion recovery following FAL.

(A) Representative TA muscle images from mice that received AAV9-Cdh5-GFP or saline injections were counterstained with DAPI (nuclei) and Dylight-594 GS-Isolectin B4 to label endothelial cells. (B) Representative images of immunolabeling of PFKFB3 shows endothelium-specific expression. Scale bar represents 20μm. (C) Graphical depiction of study design created using Biorender.com. (D) Quantification of limb perfusion recovery following FAL in male and female mice (n=5-6/group/sex). Measurements are presented as a percentage of the non-ischemic contralateral limb. (E) Quantification of total capillaries within the TA muscle assessed via immunofluorescence staining of CD31⁺ vessels (n=5-6/group/sex). (F) Quantification of perfused capillaries within the TA muscle assessed via immunolabeling of endothelial cells with systemic delivery of a fluorescent isolectin (n=5-6/group/sex). (G) Quantification of total capillaries within the EDL muscles assessed via immunofluorescence staining in male and female mice (n=5-6/group/sex). (H) Quantification of perfused capillaries within the EDL muscle of male and female mice (n=5-6/group/sex). The solid line represents the average values of sham treated animals; the dotted lines represent the standard error of the mean within the sham groups. Error bars represent SD. Panels D-H were analyzed using two-way ANOVA.

Endothelial-specific expression of PFKFB3 does not improve muscle force production following femoral artery ligation.

Because muscle function is a strong predictor of outcomes in PAD/CLI, we also assessed submaximal and maximal muscle contractile function in the EDL muscle ex vivo. Endothelial cell-specific PFKFB3 expression did not improve maximal and submaximal muscle force production in AAV9-Cdh5-PFKFB3 treated mice (Fig. 4A, B). Consistent with results in the muscle-specific AAV cohort (Fig. 2A, B), a significant effect of biological sex on peak specific force was observed, with female mice displaying significantly greater peak force compared to male mice (P=0.0481) (Fig. 4B). Representative images of ischemic muscles stained with hematoxylin and eosin showed classical signs of ischemic injury including fibers with centralized nuclei and clusters of small irregular shaped myofibers (Fig. 4C). There were no group effects of AAV9-Cdh5-PFKFB3 treatment on the myofiber cross-sectional area in the TA or EDL muscles (Fig. 4D-G). There was a significant interaction detected in the EDL mean myofiber cross-sectional area (P=0.0047). Post-hoc testing revealed a significant increase in myofiber cross-sectional area in female mice treated with AAV9-Cdh5- PFKFB3 compared with female mice that received AAV9-Cdh5-GFP (P=0.0239). As described above, the increased myofiber cross-sectional area of the EDL muscle in female AAV9-Cdh5- PFKFB3 treated mice, this did not result in improved peak specific force (Fig. 4B).

Figure 4: Endothelial-specific PFKFB3 expression does not improve ischemic muscle force levels or myofibers area.

(A) Force-frequency curves of the EDL muscle in male and female mice treated with either AAV9-Cdh5-GFP or AAV9-Cdh5-PFKFB3 (n=5-6/group/sex). (B) Peak specific force in the EDL muscle (n=5-6/group/sex). (C) Representative H&E images of the TA muscle from sham (non-ischemic) and AAV-treated ischemic muscles. (D) Histograms of myofiber CSA presented as a percentage of frequency distribution in both male and female mice (E). Mean TA myofiber CSA in male and female mice (n=5-6/group/sex). (F) Histogram of EDL myofiber CSA presented as a percentage of frequency distribution in both male and female mice. (G) Mean EDL myofiber CSA in male and female mice (n=5-6/group/sex). Error bars represent SD. Panels B, E and G were analyzed using two-way ANOVA. *P<0.05 using Tukey’s post-hoc testing.

Discussion

Peripheral artery disease impacts nearly 240 million people worldwide²⁶ and is the third leading cause of death among vascular diseases²⁷. The pathological burden is further amplified by the lack of approved therapies and current surgical interventions having with high failure rates^1-3, establishing the necessity for efficacious treatment options. The current study was designed to determine if the beneficial effect of PFKFB3 in limb ischemia was mediated through its endothelial cell-derived angiogenic effects or through its impact on skeletal muscle metabolism¹⁰. To accomplish this, adeno-associated viruses were generated to express PFKFB3 in a muscle- and endothelial-cell specific manners. These gene therapies were subsequently injected into adult BALB/c mice which model CLI and respond poorly to femoral artery ligation^{7, 18, 19}. Consistent with the wealth of literature establishing the role of PFKFB3 in endothelial cell biology(14-16), mice that received endothelium-specific PFKFB3 gene therapies displayed significantly higher perfused capillary densities compared to control-treated mice (Fig. 3F). However, this angiogenic response was not sufficient to improve EDL muscle contraction measured 28 days after FAL (Fig. 4A,B). In contrast, muscle-specific expression of PFKFB3 significantly improved muscle force production (Fig. 2A,B) and was also found to increase perfused capillary density (Fig. 1G) following FAL when compared to GFP-treated mice. Notably, improved force production occurred in the absence of changes in myofibers size suggesting that muscle-specific PFKFB3 expression attenuated the intrinsic contractile dysfunction caused by ischemic injury.

Endothelial cell metabolism has emerged as a critical regulator of the processes required for angiogenesis(14-16, 29-33). With this knowledge, manipulating glycolytic flux in endothelial cells could provide a therapeutic strategy to alter angiogenesis(12-15). PFKFB3 belongs to a family of enzymes that phosphorylate fructose-6-phosphate to fructose-2,6-bisphosphate(15), which is a potent allosteric activator of 6-phosphofructo-1-kinase a rate limiting enzyme in glycolysis. A series of elegant studies using cell/tissue culture and mice demonstrated that PFKFB3 expression promotes angiogenesis via enhanced glycolytic flux in endothelial cells(14-16). Further to this, endothelium-specific deletion of PFKFB3 was recently shown to impair limb perfusion recovery and ischemic muscle regeneration in mice following FAL, a finding that was mechanistically linked to the ability of endothelial-derived lactate to drive an M2-like phenotype in macrophage(17). Findings from the current study are consistent with the proposed role of PFKFB3 in endothelial cells. For example, mice that received endothelium-specific PFKFB3 AAVs displayed significantly more perfused capillaries in the ischemic limb compared to GFP-control treated animals (Fig. 3F). Interestingly, laser Doppler flowmetry did not show improvements in perfusion in mice treated with AAV9-Cdh5-PFKFB3. There are several plausible explanations for this apparent discrepancy in these results. First, laser Doppler perfusion was assessed using a flowmetry probe that measures only a small (~1mm) portion of the muscle and the laser does not penetrate deep into tissues, and thus measures only superficial tissues. In contrast, perfused capillaries were counted across the entire TA muscle and thus likely results in a more robust detection of whole muscle perfusion levels.

While PFKFB3 has been established to promote sprouting angiogenesis, a recent study has provided evidence suggesting that vessel splitting, known as intussusception, may be the primary mechanism of angiogenesis under low blood flow conditions(34). Intussusception is a process in which transluminal pillars begin to develop within a capillary and develop to span the diameter of the lumen creating two distinct vessel structures(35). In this recent study, the authors using intravital microscopy in the ischemic muscle of mice to document the appearance of large caliber neo-conduits that received low amounts of blood flow. These neo conduits over time branched into a vessel network that was primarily driven by intussusception. Endothelial cells with reduced VEGFR2 receptor signaling initiated this response by entering the lumen to form the pillar and generation of a new vessel. While much has been discovered about the role of metabolism in sprouting angiogenesis, far less is known about the role of endothelial cell metabolism in intussusceptive angiogenesis. Thus, future work is needed to evaluate if PFKFB3 expression facilitates intussusceptive angiogenesis.

Skeletal muscle demonstrates remarkable plasticity in which it can recover from a variety of insults. This plasticity requires coordination between the vasculature, the extracellular matrix, and the muscle fiber(36, 37). To facilitate intercellular communication, the muscle fiber can secrete growth factors including vascular endothelial growth factor (VEGF), angiopeoietin-1/2, and fibroblast growth factor among others that were well-known to regulate muscle capillarization(38). Intriguingly, muscle-specific PFKFB3 expression significantly increased perfused capillary density in the ischemic TA muscle, but not the EDL muscle (Fig. 1G and 1I). The discrepancy between muscles may be a result of the lower levels of PFKFB3 mRNA (~30-fold increase compared with GFP vs. ~110-fold increase in the TA muscle). This difference in AAV infection is most likely a product of the non-surgical approach to the intramuscular AAV injection where direct visualization of the EDL injection was not possible. Nonetheless, the finding of increase capillarity is consistent with the notion that alterations in muscle cell metabolism can evoke changes in the vasculature(39). Importantly, while both muscle- and endothelium-specific PFKFB3 expression were found to increase perfused capillary counts, muscle function (i.e. force production) was only improved in mice with muscle-specific PFKFB3 expression. Considering that muscle strength/function is the strongest predictor of outcomes in PAD/CLI(40-46), the results herein support the need to further investigate therapies that alter muscle cell metabolism for the treatment of PAD/CLI(5, 6).

There is a wealth of literature demonstrating the efficacy of angiogenic-based therapies in preclinical models of PAD (47-53). New microvasculature can develop through endothelial-mediated sprouting of pre-existing capillaries, a process that involves angiogenic growth factors such as VEGF and fibroblast growth factor among others. Gene based therapies (plasmid or viral) of angiogenic growth factors improves recovery from hindlimb ischemia in multiple species (49-51, 54-57). Recruitment of certain types of circulating cells with angiogenic potential to the ischemic limb has also been explored as a potential therapeutic avenue in CLI. Cell-based therapies including the exogenous delivery of bone marrow-derived mononuclear cells (58, 59), mesenchymal stem cells (60), and endothelial progenitor cells (61-63) have proven effective in animal models on PAD. Beyond traditional VEGF-VEGFR2 angiogenic signaling (64-66), emerging evidence suggests that splice variants of VEGF (67, 68), non-coding RNAs (69, 70), and cellular metabolism (14, 16) also play a role in regulating vascular remodeling in ischemic tissues. Unfortunately, angiogenic gene and cell-based therapies have failed to achieve primary outcomes in clinical trials (52, 71) which reinforces that viewpoint that we have an incomplete understanding of the pathobiology of PAD/CLI. Consistent with failure in clinical trials, endothelium-specific PFKFB3 expression did not improve ischemic muscle function despite increasing the number of perfused capillaries within the ischemic muscle.

To date, women have been inadequately represented both in pre-clinical (22, 72, 73) and clinical PAD/CLI research(74). For this reason, the National Institutes of Health (NIH) mandated the inclusion of women and minorities via the NIH Revitalization Act of 1993, PL 103-43. Even with this mandate, a study conducted revealed that among clinical research studies conducted after 1994, there were not differences in frequency of sex reporting before or after this date(74). Adequate representation of women is critical for establishing efficacious therapies that aid in the treatment of PAD. The incidence of PAD in women below the age of 60 years is as low as 3% but drastically jumps to nearly 20% in women over the age of 70 (75) and up to nearly 40% in females over the age of 85 (76). This exponential increase after age 60 could be attributed to the completion of menopause and severe reductions in estrogen levels. This coincides with the fact that women are 3 years older on average and experience higher rates of CLI than men(77, 78). These age disparities lead to women presenting with more asymptomatic PAD(75) and leads to significantly reduced pain free walking distance, absolute claudication distance and overall less physical activity(73). There is great necessity for inclusion of women to help address these differences. The current study utilized both male and female animals and further demonstrated that sex differences do exist. Specifically, female mice displayed greater levels of contractile force following FAL compared to males (Fig. 2B, 4B), despite no differences in perfusion recovery or capillary density. The differences demonstrated between male and female mice could potentially be explained by the effects of estrogen. Estrogen has been shown to promote vasodilation of blood vessels by stimulating nitric oxide production in the endothelium to help increase blood flow (79).This effect has been shown pre-clinical where female mice who underwent ovariectomy (OVX) had significantly reduced blood flow recovery following FAL (80). The lack of blood flow recovery was also accompanied by a decrease in capillary/fiber ratio and decreased endothelial nitric oxide synthase (eNOS) in OVX female mice. The current study utilized normally cycling, young female mice which is a current limitation as these animals do not adequately represent the patient population in PAD.

Study Limitations

The current study is not without its limitations. First, AAV-mediated gene therapy was targeted only to dorsiflexor muscles using intramuscular delivery. Because of this, only a small proportion of the ischemic limb tissue was subjected to treatment. This approach was performed due to the prohibitive cost of systemic delivery of the required viral dosages to achieve adequate viral infection. Moreover, infection of the EDL with AAV, without a surgical exposure of the muscle, is less efficient due to its anatomical location beneath the TA muscle. The was evident in our mRNA analysis of PFKFB3 levels. Related to this, ex vivo analysis of muscle function was not possible in the TA muscle because of the large size and inability to adequately oxygenate the muscle in an organ bath. Limb perfusion was assessed using laser Doppler, an established method that is unfortunately limited in its sampling depth. As a result, perfusion measurements include only the superficial tissues, and a proportion of the signal is derived from the skin and subcutaneous tissue which was not infected with AAV. In addition to laser Doppler, we also systemically deliver fluorescent lectins to label perfused capillaries in the muscle. Although injection timing was consistent across mice, we did not control for cage activity during the two hours between injection and euthanasia. As discussed in detail above, this study was performed on male and female mice, albeit at a young age and with females that were normally cycling. Both characteristics are not the most relevant to the PAD population. Finally, therapeutic expression of PFKFB3 was performed in a preventative setting (i.e., before induction of PAD) which may not model the ideal clinical scenario for intervention.

Conclusions

Whereas AAV-based gene therapies targeting expression of PFKFB3 in muscle- or endothelial cells both resulted in increased perfused capillary counts in mice with CLI, only muscle-specific PFKFB3 expression improved muscle function – a strong predictor of outcomes in human patients (45). These findings support the notion that enhancing metabolic flexibility in the ischemic limb muscle can improve recovery and alleviate myopathic symptoms that devastate the CLI population. Notably, this study also uncovered sex differences in the ischemic muscle function with female mice displaying greater force levels than male mice following FAL regardless of treatment group; an observation that highlights the need for inclusion of both sexes in pre-clinical PAD/CLI studies.