The SGLT2 inhibitor Empagliflozin promotes post-stroke functional recovery in diabetic mice

Eighty C57BL/6JRj mice (Janvier Labs, France) were used in this study. Mice were housed in environmentally controlled conditions (22 ± 0.5 °C, 12/12 h light/dark cycle with ad libitum access to food and water). The mice were kept under pathogen free conditions in type III size individually ventilated cages with wood chip bedding and nest material.

Group sizes were determined based on ≈ 20% effect size between groups in functional recovery with α = 0.05 and a statistical power of 90%. Standard deviation used in sample size calculation was obtained from pilot experiments. The analyses suggested the sample size of minimum n = 5 per group. However, after taking into consideration the success rate of stroke surgery, mortality and likelihood of statistical outliers, the experimental groups were set at n = 10–15 each.

Stroke was induced by tMCAO using the intraluminal filament technique as described previously [42]. Briefly, mice were anesthetized by inhalation of 3% isoflurane and throughout surgery, anesthesia was maintained by 1.5% isoflurane. Using a heated pad with feedback from a thermometer, body temperature of animals was kept at 37–38 °C. Left external (ECA) and internal (ICA) carotid arteries were exposed and a 7–0 silicone-coated monofilament (total diameter 0.17–0.18 mm) was inserted into the ICA until the origin of the MCA was blocked. The occluding filament was removed after 30 min. Cerebral blood flow in the vicinity of MCA was monitored by Laser Doppler Blood Flow Monitor (Moor Instruments Ltd, UK), and no differences between the groups were observed (data not shown). Stroke induction was considered unsuccessful when the occluding filament could not be advanced within the internal carotid artery beyond 7–8 mm from the carotid bifurcation, or if mice lacked symptoms of neurological impairment based on the neurological severity score [43]. After surgery, all mice were given analgesic (Carprofen, 5 mg/kg) and soft food.

Fasting glycemia was measured after an overnight (ON) fasting via blood from a tail tip puncture and a glucometer. For insulin tolerance tests (ITT), mice were fasted for 2 h. Hereafter, baseline glucose levels were measured. Then, mice were injected intraperitoneally (i.p.) with 0.5 U/kg human insulin and blood glucose levels were measured at 15, 30, 45, 60, 75 and 90 min after injection. Area under the curve was computed for statistical analysis.

To assess sensorimotor function, forelimb grip strength was tested as previously described [44]. Briefly, mice were held firmly by the body and allowed to grasp the grid with the affected forepaw. Hereafter, they were dragged backwards until their grip was broken. Grip strength was measured using a grip strength meter (Harvard apparatus, MA, USA) at 3 days and 1–5 weeks after stroke induction. Ten trials were performed, and the highest value was recorded.

Mice were anesthetized using an i.p. injection with an overdose of sodium pentobarbital. Hereafter, blood was collected via cardiac puncture and mice were perfused transcardially using phosphate-buffered saline (PBS) followed by a 4% ice-cold paraformaldehyde (PFA) solution. Brains were harvested and stored ON in 4% PFA at 4 °C. After 24 h of fixation, brains were transferred to PBS containing 25% sucrose and stored at 4 °C until they sank. Then, 30 μm thick coronal sections were cut using a sliding microtome, and sections were stored at − 20 °C in anti-freeze solution.

Immunohistochemical stainings were performed using the free-floating method. Briefly, sections were washed in PBS three times to remove anti-freeze solution. For visualization with 3–3′-Diaminobenzidine (DAB), quenching of endogenous peroxidases was performed by a 20 min incubation at RT in a PBS-solution containing 3% H₂O₂ and 10% methanol. For immunofluorescent DCX/Ki67 staining, sections were subjected to antigen retrieval by a 15 min incubation in citrate solution (pH = 6.0) at 95 °C. For immunofluorescent CD13, podocalyxin (PDXL), and NG2 stainings, sections were blocked with 5% serum with 0.25% triton-X100-PBS for 1 h at RT. For both DAB and immunofluorescent stainings, sections were incubated ON at 4 °C in PBS containing primary antibody, 3–5% normal serum and 0.25% Triton-X-100. For NeuN staining, sections were incubated in primary antibody solution for 48 h. The following primary antibodies were used: mouse anti-NeuN, a neuronal marker (1:500, #MAB377, Millipore, RRID:AB_2298772); goat anti-Iba1, a marker for microglia (1:1000, Ab5076, Abcam, RRID:AB_2220422); anti-DCX (doublecortin), a marker for migrating neuroblasts (1:200, sc-271390, Santa Cruz Biotechnology, RRID:AB_10610966); anti-Ki67, a marker for cell proliferation (1:300, ab15580, Abcam, RRID:AB_443209); goat anti-PDXL, a marker for endothelial cells (1:200, #AF1658, R&D Systems, RRID:JLD0117051); rat anti-CD13, a marker for pericytes (1:200, #MCA2183, Biorad, RRID:152021); rabbit anti-NG2, a marker for activated pericytes (1:200, #AB5320, Millipore, RRID:3517406); rabbit anti-albumin (1:200, ab19196, Abcam, RRID:GR3271594-14); rabbit anti-fibrinogen (1:400, ab27913, Abcam, RRID:GR28863-14). After incubation in primary antibody solution, sections were washed and then incubated 2 h at RT in PBS containing secondary antibody, 3–5% normal serum and 0.25% Triton-X-100. The following secondary antibodies were used: biotinylated horse anti-mouse (1:200, #BA-2000, Vector Laboratories, RRID:AB_2313581); biotinylated horse anti-goat (1:200, #BA-9500, Vector Laboratories, RRID:AB_2336123); Alexa-488 conjugated horse anti-rabbit (1:200, DI-1088, Vector Laboratories, RRID:AB_2336403); Alexa-584 conjugated horse anti-mouse (1:200, DI-2594, Vector Laboratories, RRID:AB_2336412); Alexa-488 conjugated donkey anti-rabbit (1:500, #711-545-152, JacksonImmuno Research, RRID:156009); Alexa-546 conjugated donkey anti-goat (1:500, #A11056, Invitrogen, RRID:997810); Cy5 conjugated donkey anti-rat (1:500, #712–175-150, JacksonImmuno Research, RRID:148159). For DAB-visualization, incubation with biotinylated secondary antibody was followed by incubation with avidin–biotin complex according to manufacturer’s instructions (Vectastain Elite ABC kit, Vector Laboratories), followed by visualization by DAB. For DCX/Ki67 staining, incubation with secondary antibody solution was followed by a 30 s incubation with PBS containing 1:10 000 DAPI.

Serum concentrations of insulin, FGF-21 and β-hydroxybutyrate (BHB) were quantified according to manufacturer’s instructions in serum samples obtained before stroke and at 2 and 5 weeks after tMCAO (90080, CrystalChem for insulin; R&D Systems MF2100, R&D Systems for FGF-21, ab93380, Abcam for β-hydroxybutyrate).

Data were checked for statistical outliers by using the ROUT method, and for normality by using the Shapiro–Wilk normality test.

Parametric tests: For pre- and post-stroke metabolic parameters and NeuN analysis, Brown-Forsythe and Welch ANOVA test, followed by two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli was used. For behavioral tests, two-way repeated measures ANOVA with Geisser-Greenhouse's correction followed by Dunnett T3 was used. For neuroinflammation, neurogenesis and vascular analysis, two-way repeated measures ANOVA followed by two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli was used. All data were analyzed by GraphPad Prism Version 9.0. Data are expressed as mean ± SD. p-values less than 0.05 were considered statistically significant.

As recently published [15, 16, 39], 8 months of HFD-feeding induced obesity (Fig. 2a), decreased insulin sensitivity (Fig. 2b, c), hyperinsulinemia (Fig. 2d) and hyperglycemia (Fig. 2e).

To assess the potential efficacy of Empagliflozin to improve stroke recovery, forepaw grip strength recovery was followed up for 5 weeks after tMCAO. After tMCAO, non-T2D mice recovered fully within 5 weeks while T2D-VH mice remained significantly impaired (Fig. 2f, g). Importantly, Empagliflozin treatment completely normalized the T2D-induced worsening of stroke recovery (Fig. 2f, g). No differences in stroke volume were observed between groups (Fig. 2h), demonstrating that the improved recovery was not due to differences in infarct size mediated by Empagliflozin-induced neuroprotection.

To investigate the potential association between Empagliflozin-induced improved recovery and metabolic changes, we analyzed several metabolic parameters after stroke. In accordance with previous studies [49‐51], tMCAO and subsequent switch from HFD to SD induced significant weight loss in the first two weeks after tMCAO in all T2D mice, without significant differences between T2D-VH (− 34 ± 3%) and T2D-E (− 28 ± 7%) groups (Fig. 2i). At 2 weeks after stroke, all T2D mice were still IR (Fig. 2j, k) and hyperinsulinemic (Fig. 2l), irrespective of treatment. However, T2D-E mice became normoglycemic while hyperglycemia was still present in T2D-VH mice, showing that Empagliflozin efficiently reduced hyperglycemia, but not IR after stroke in our model (Fig. 2m).

Unlike T2D mice subjected to tMCAO, T2D mice subjected to sham surgery only lost 12% of their initial body weight during the first two weeks post-diet change, resulting in differences in the metabolic state between stroke and sham mice (Additional file 1: Fig. S1a). In sham-operated animals treated with Empagliflozin, a trend (p = 0.112) towards attenuated hyperglycemia was observed, whereas no effect on insulin sensitivity was observed (Additional file 1: Fig. S1b–d).

In summary, our results demonstrate that a post-stroke treatment with Empagliflozin significantly improves post-stroke recovery, in association with the normalization of hyperglycemia.

Increased FGF-21 levels have been associated with post-stroke recovery [34]. Moreover, Jiang and colleagues recently demonstrated that a therapeutic administration of FGF-21 improves post-stroke recovery in diabetic mice [35]. Since recent literature has shown that SGLT2i can increase FGF-21 levels [52], we investigated whether the improved recovery in the T2D-E group was associated with increased serum levels of FGF-21. Before stroke, no difference in FGF-21 levels between non-T2D and T2D mice was recorded (Fig. 3a). At two weeks after stroke, FGF-21 levels were significantly decreased in both non-T2D and T2D-VH mice (Fig. 3a) and remained significantly lower than pre-stroke levels in T2D-VH (p = 0.01) at five weeks post-stroke (Fig. 3a). Interestingly the post-stroke treatment with Empagliflozin resulted in a significant increase of serum FGF-21 levels, both at two and at five weeks after stroke (Fig. 3b). Taken together, these results indicate that stroke decreases serum FGF-21 levels independently of the metabolic state of the animals, and that Empagliflozin prevents this stroke-induced reduction, in association with improved recovery.

In accordance with existing literature [53], HFD-induced T2D significantly increased serum BHB-levels (Fig. 3c). After stroke, there was a trend towards a decrease in serum BHB (p = 0.198 at 2 weeks and p = 0.056 at 5 weeks after stroke) in T2D-VH mice compared to pre-stroke levels (Fig. 3c). We found no difference between T2D-VH and T2D-E mice at either 2 or 5 weeks after stroke (Fig. 3d), indicating that after stroke, SGLT2i-treatment does not upregulate ketone production in T2D mice.

Stroke-induced neurogenesis has been associated with improved stroke recovery (reviewed in [39]. Therefore, we next assessed whether improved functional recovery by Empagliflozin after stroke was associated with the regulation of this process. Neural stem cell proliferation and neuroblast formation were analyzed by quantifying Ki67⁺ cells in the SVZ and DCX⁺ neuroblasts in the striatum, respectively. No differences in Ki67⁺ cells were recorded between groups in the SVZ (Fig. 4a). In accordance with existing literature [51], stroke induced a significant increase in DCX⁺ cells in the ipsilateral, stroke-damaged striatum in all three groups (Fig. 4b). However, there was no difference in the number of DCX⁺ cells between the groups (Fig. 4b), suggesting that improved stroke recovery in the T2D-E group was not due to increased neurogenesis.

To evaluate stroke-induced neuroinflammation, we quantified Iba-1 immunoreactivity in ipsilateral, stroke-damaged striatum versus the intact contralateral hemisphere. Stroke induced a significant upregulation of Iba-1 in ipsilateral striatum compared to contralateral in all three groups. Notably, this increase was significantly higher in T2D-VH mice than in non-T2D mice (Fig. 5). However, Empagliflozin treatment did not significantly decrease Iba-1 immunoreactivity compared to T2D-VH animals, although a trend was observed (p = 0.103) (Fig. 5). Moreover, we observed no apparent morphological differences of striatal microglia between groups. In sham-operated animals, T2D upregulated striatal Iba-1 compared to non-T2D controls, but no differences between groups were detected between sham-T2D-VH and sham-T2D-E animals (Additional file 1: Fig. S2). Taken together, these data indicate that, at least at 5 weeks after stroke, improved stroke recovery in Empagliflozin treated animals is likely not associated with attenuated post-stroke Iba-1 immunoreactivity.

To investigate whether Empagliflozin treatment has an impact on the vascular system after stroke, we evaluated cerebral vascular changes in terms of vessel (PDXL⁺), total pericyte density (CD13⁺) and coverage (CD13⁺/PDXL⁺ ratio), vessel length and branching, markers of pericyte activation (CD13⁺/NG2⁺), and BBB leakage by assessing extravascular albumin and fibrinogen.

In non-T2D mice subjected to stroke, the injury significantly increased vascular density, total pericyte density, parenchymal pericyte density, pericyte coverage, and density of activated pericytes in the ipsilateral striatum compared to the contralateral, indicating post-stroke angiogenesis (Fig. 6a–e and Additional file 1: Fig. S3). In T2D-VH stroke mice, total and parenchymal pericyte density, coverage and activation were increased in the ipsilateral striatum (Fig. 6a–e, Additional file 1: Fig. S3). In T2D-E animals, stroke upregulated parenchymal pericyte density and coverage in the ipsilateral striatum compared to the contralateral hemisphere (Fig. 6a–e and Additional file 1: Fig. S3).

When comparing ipsilateral hemispheres between groups, two-way ANOVA revealed a significant increase in the total pericyte density in the T2D-VH group compared to the non-T2D group, which was normalized by Empagliflozin treatment (Fig. 6b). Interestingly, groups did not differ in vascular density (Fig. 6c) and had similar pericyte coverage of the vessels (Fig. 6d), implicating that the increased overall pericyte density observed in the T2D-VH group is due to pericytes located in the parenchyma. Indeed, we observed a significant increase in parenchymal pericyte density in the T2D-VH group compared to the non-T2D group (Fig. 6e). Moreover, there was a strong trend towards a decreased parenchymal pericyte density in the T2D-E group compared to the T2D-VH group (p = 0.061), and no difference was detected between non-T2D and T2D-E animals (Fig. 6e). No differences were seen between the three experimental groups within the contralateral striatum, with the exception of increased parenchymal pericyte density in the T2D-VH vs. non-T2D groups, indicating that T2D impacts the balance between perivascular and parenchymal pericytes (Fig. 6a–e). Pericyte activation was similarly activated in all groups after stroke (Additional file 1: Fig. S3). Taken together, these data indicate that T2D alters parenchymal pericyte density, and that Empagliflozin-treatment can normalize this effect.

To assess BBB integrity, we examined the presence of plasma proteins in the brain parenchyma analyzing two different molecular sizes, albumin (~ 65 KDa) and fibrinogen (~ 340 KDa). We observed no significant differences in albumin or fibrinogen extravasation between the groups, (Additional file 1: Fig. S4). In sham-operated animals, treatment with Empagliflozin led to a significant increase in total and parenchymal pericyte density and coverage, whereas no differences were found in BBB integrity between groups (Additional file 1: Fig. S5).

We next determined whether a post-stroke intervention with Empagliflozin could improve recovery independently of its glycemia-regulating properties. Non-T2D mice were treated daily with either VH or Empagliflozin starting from 3 days after stroke until Empagliflozin-treated mice fully recovered (Exp. design Fig. 1b). There was no difference in forepaw grip strength between the groups (Fig. 7), indicating that the improved stroke recovery induced by Empagliflozin in the diabetic study was likely mediated by the anti-T2D properties of Empagliflozin.