C57BL/6 N male mice aged 8–10 weeks (weight 22–25 g) were purchased from Gem Pharma Tech LLC (Nanjing, Jiangsu, China) and maintained in a 12/12-h light/dark cycle environment at a constant temperature of 22 °C with free access to standard laboratory chow and tap water. They were housed in groups and were divided into normal diet (ND) group and ketogenic diet (KD) group (n = 13 in each group), fed with standard chow and ketogenic chow respectively (Additional file 1: Table S1, purchased from Xietong Pharma Co., Jiangsu, China), with free access to food and water 24hs. After 2 weeks, mice underwent hindlimb ischemia by ligation of the unilateral femoral artery and were given continued access to the two different diets. After the procedure, mice were evaluated for ischemic hindlimb blood perfusion using a laser Doppler perfusion imager. All animal experimental procedures conformed 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) and were reviewed and approved by the Animal Ethics Committee at Zhongshan Hospital, Fudan University, China.

Blood glucose and ketone levels (represented by β-hydroxybutyrate levels) were measured using a glucose meter and ketone meter, respectively (Abbott Diabetes Care, Maidenhead, UK). Tissue β-HB content was measured by β-Hydroxybutyrate (β-HB) assay kits (MAK041, Sigma, Kawasaki, Kanagawa). Tissues were homogenized in cold β-hydroxybutyrate assay buffer and centrifuged at 13,000 g for 10 min at 4 °C to remove insoluble material.

To establish a hind limb ischemia model, we performed unilateral femoral artery ligation in mice. Following percutaneous injection of the anesthetic (4% chloral hydrate), a groin incision was made in the left adductor hind-limb region. The femoral artery was identified, ligated with 6–0 silk ties at the ends of the vascular trunk, and transected from the middle. The incision was then closed with interrupted non-absorbable sutures, and mice were closely monitored for 24 h post-procedure.

Mice were subjected to an inhaled anesthetic at body temperature maintained by a warming pad. A laser Doppler perfusion imager (Periscan PIM3, Perimed, Beijing, China), was used to evaluate the bilateral hind limbs. Perfusion was evaluated in the whole limb, gastrocnemius region, and hind-paw region on post-HLI days 0, 7, and 21.

Gastrocnemius muscle tissues were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin then dehydrated in graded ethanol solutions and toluene. Tissues were dissected into 5-μm-thick sections and stained with hematoxylin and eosin (H&E) and Masson. For immunofluorescence staining, the sections were blocked with 10% goat serum albumin (Invitrogen, Waltham, Massachusetts, USA) for 60 min before staining with CD31 monoclonal antibody (1:1500, CST).

Total protein was extracted from the gastrocnemius tissue. Equal amounts of protein extract were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% bovine serum albumin and probed with primary antibodies individually at 4 °C overnight. After subsequent washing, the blots were incubated with horseradish peroxidase-coupled anti-rabbit or anti-mouse secondary antibodies at room temperature for 2 h. The blots were visualized and detected by chemiluminescence reaction (LuminataTM Forte, Millipore, Burlington, Massachusetts, USA) and ChemiDoc™ Imaging System (Bio-Rad, Hercules, CA, US). The density of the protein blots was determined using Image J software (1.50i, Open Source, USA) and normalized to β-actin (1:1000, Kang Chen, Wuxi, China).

Total RNA from gastrocnemius tissues was extracted using TRIzol™ Reagent (#15,596,026, Invitrogen, Waltham, Massachusetts USA). The concentration and purity of RNA were determined by Nanodrop (Thermo Fischer, Waltham, Massachusetts, USA), and 1000 ng RNA was purified with an A260/A280 ratio of 1.8–2.0 and then reverse transcribed into cDNA using PrimeScript™ Reverse Transcription Master mix (# RR036A, TaKaRa, Kusatsu, Shiga, Japan). Reverse transcription polymerase chain reaction (RT-PCR) was performed using PrimeScript™ RT Master Mix (TaKaRa, Japan). A total of 20 μL reaction system was used, including DNA template 1.6 μL and SYBR 10 μL, and primers 0.4 μL and ddH2O 7.6 μL. Primers used in this study are listed in Additional file 1: Table S2. The PCR reaction cycles were set as follows: 30 s at 95 °C, then 5 s at 95 °C and 30 s at 60 °C for 40 cycles. Fluorescence signals were normalized to Actb using the 2 − ΔΔCT method.

All statistical analyses were performed using Prism 7.0 (GraphPad Software, Inc., La Jolla, CA, USA). Continuous variables are expressed as mean ± standard error of the mean (SEM). Normal distribution was determined using the Shapiro–Wilk test. Differences in normal variates were tested using the Student’s t-test (within two groups) or a one-way analysis of variance (ANOVA, among three groups or more), with post hoc comparisons using the Tukey’s multiple comparisons test. Non-normal data were analyzed using the Mann–Whitney U test or the Kruskal–Wallis H test. Statistical significance was defined as two-tailed P < 0.05(*), **P < 0.01(**), and ***P < 0.001(***).