SGLT2 Inhibition via Empagliflozin Improves Endothelial Function and Reduces Mitochondrial Oxidative Stress: Insights from Frail Hypertensive and Diabetic Patients

Abstract

Background:

Frailty is a multidimensional condition often diagnosed in older adults with hypertension and diabetes, and both these conditions are associated with endothelial dysfunction and oxidative stress. We investigated the role of the Sodium GLucose co-Transporter 2 (SGLT2) inhibitor empagliflozin on cognitive frailty in diabetic and hypertensive older adults.

Methods:

We studied the effects of empagliflozin in consecutive hypertensive and diabetic older patients with frailty presenting at the ASL (local health unit of the Italian Ministry of Health) of Avellino, Italy, from March 2021 to January 2022. Moreover, we performed in vitro experiments in human endothelial cells to measure cell viability, permeability, mitochondrial Ca²⁺, and oxidative stress.

Results:

We evaluated 407 patients; 325 frail elders with diabetes successfully completed the study. We propensity score matched 75 patients treated with empagliflozin and 75 with no empagliflozin. We observed a correlation between glycemia and Montreal Cognitive Assessment (MoCA) score and between glycemia and 5-meter gait speed (5mGS). At 3-month follow up, we detected a significant improvement in the MoCA score and in the 5-meter gait speed in patients receiving empagliflozin compared to non-treated subjects. Mechanistically, we demonstrate that empagliflozin significantly reduces mitochondrial Ca²⁺ overload and ROS production triggered by high-glucose in human endothelial cells, attenuates cellular permeability, and improves cell viability in response to oxidative stress.

Conclusions:

Taken together, our data indicate that empagliflozin reduces frailty in diabetic and hypertensive patients, most likely by decreasing mitochondrial ROS production in endothelial cells.

Graphical Abstract

Frailty is a multidimensional condition that leads to functional decline with physical and cognitive impairment^1–3. Frailty is often diagnosed in older adults with hypertension and diabetes, and both these conditions are associated with endothelial dysfunction and oxidative stress^4–6.

Our group has recently evidenced the detrimental effects of hyperglycemia in frail hypertensive elders⁷. Several investigators have shown that hyperglycemia drives inflammation and oxidative stress leading to endothelial dysfunction, with a well-established negative impact on frail patients^8–11.

In this scenario, Sodium GLucose co-Transporter 2 (SGLT2) inhibitors — which include the FDA approved drugs canagliflozin, dapagliflozin, empagliflozin, and ertugliflozin¹² — are considered among the most promising oral antidiabetic drugs to reach and maintain an optimal glycemic control in older subjects¹³. The main mechanism of action of these drugs is the blockage of SGLT2 proteins in the renal proximal convoluted tubules in order to reduce the reabsorption of filtered glucose and decrease the renal threshold for glucose, thereby promoting urinary glucose excretion^14–17. Empagliflozin is a SGLT2 inhibitor that has been shown to reduce mortality and re-hospitalization for heart failure in diabetic patients^18–23. Additional potential benefits of empagliflozin and other SGLT2 inhibitors include improved cardiovascular energetics, reduced vascular tone and blood pressure, decreased renal dysfunction, increased circulating levels of ketone bodies, protection against pulmonary ischemia/reperfusion injury, and overall reduced systemic inflammation^24–37.

On these grounds, we investigated the role of empagliflozin on cognitive frailty in diabetic and hypertensive older adults. To mechanistically confirm our results, we evaluated the effects of empagliflozin on mitochondrial oxidative stress and permeabilization in human endothelial cells.

METHODS

Data Availability

The data that support the findings of this study are available from the first author upon reasonable request.

Study Participants

We designed a prospective study to enroll consecutive hypertensive and diabetic older adults presenting at the ASL (local health unit of the Ministry of Health) of Avellino, Italy, from March 2021 to January 2022 with a diagnosis of frailty. Inclusion criteria: Age >65 years, a diagnosis of frailty, presence of diabetes and primary hypertension, Montreal Cognitive Assessment (MoCA) Score <26, Glomerular Filtration Rate (GFR) <30. Exclusion criteria: age <65 years, history of previous stroke. The study was approved by the local Ethical Committee (Campania Nord). All subjects or their legal representatives signed an informed consent. The study had been registered in clinicaltrials.gov (#NCT04962841).

Clinical evaluation and assessment of cognitive and physical impairment

Blood glucose, HbA1c, and creatinine were measured in all patients. Clinical assessment was completed at baseline and after three months. Diabetes was defined according to the American Diabetes Association (ADA) guidelines³⁸. Hypertension was defined as systolic blood pressure (SBP) ≥140 mmHg, and/or diastolic blood pressure (DBP) ≥90 mmHg on repeated measurements, or use of antihypertensive medications³⁹.

Frailty Assessment

A diagnosis of physical frailty was made with at least three out of the five Fried Criteria⁴⁰, as we previously described^{41, 42}:

• Weight loss (unintentional loss ≥4.5 kg in the past year);

• Weakness (handgrip strength in the lowest 20% quintile at baseline, adjusted for sex and body mass index);

• Exhaustion (poor endurance and energy, self-reported);

• Slowness (walking speed under the lowest quintile adjusted for sex and height);

• Low physical activity level (lowest quintile of kilocalories of physical activity during the past week).

Additionally, we performed a 5-meter gait speed (5mGS) test⁴³ in all patients. Cognitive function was evaluated by applying the MoCA test (cognitive impairment was defined by values <26⁴¹).

Ca²⁺ imaging

Ca²⁺ imaging experiments to measure mitochondrial Ca²⁺ were performed as we previously described described^{44, 45}. Human umbilical vascular endothelial cells (HUVECs, passages 3–7) were plated in glass-bottom dishes using the “Ca²⁺ imaging solution”, which consists of: 138 mM NaCl, 5.3 mM KCl, 1.2 mM NaH₂PO₄, 1.2 mM MgCl₂, and 20 mM HEPES (pH 7.38, adjusted with NaOH); to this solution we added 5 mM EGTA, 1.8 mM CaCl₂, 5 or 30 mM glucose, according to the experimental settings. Cells were loaded with Rhod-2 AM (3 μM, Thermo Fisher Scientific, Waltham, MA; catolog number: R1244) at 37 °C for 30 min, followed by washout and 1 h rest at room temperature for de-esterification: indeed, since Rhod-2 AM has a delocalized positive charge, it preferentially accumulates within the mitochondrial matrix, where it is hydrolyzed and trapped. Fluorescence was detected using a pass-band filter of 545–625 nm in response to excitation at 542 nm.

Mitochondrial oxidative stress

HUVECs were plated on glass bottom culture dishes and treated with different concentrations of glucose and empagliflozin for 24h; mitochondrial reactive oxygen species (ROS) were quantified by MitoSOX™ Red (Thermo Fisher Scientific; catalog number: #M36008) as we described^{7, 46}, using MitoTracker™ Green FM (Thermo Fisher Scientific, catalog number: #M7514) to identify mitochondria.

Real time PCR (RT-qPCR)

RT-qPCR was performed using SYBR Green mix as previously described^{45, 47, 48}; GAPDH was used as an internal standard; primer sequences are listed in Supplementary Table S1.

Endothelial permeability assay

Human brain microvascular endothelial cells (hBMECs, passages 3–6) were cultured as we previously described^{7, 49}. In some experiments, cells were treated with empagliflozin (1 μM), adding glucose at 5 or 30 mM, for 48 h. The endothelial permeability assay was performed using fibronectin-coated trans-well filters (Corning Inc., Corning, NY, USA), as we previously reported^{7, 49, 50}.

Cell viability assay

Cell viability was evaluated in human brain microvascular endothelial cells using the [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay, as we previously described^{44, 51}. Briefly, cells were pretreated for 24 h with 1 μM empagliflozin and then incubated with increased doses of H₂O₂ for 5 h.

Statistical Analysis

Data are presented as n (%) for categorical data (compared using the χ2 test), mean ± SD for normally distributed continuous data (compared via Student’s t test or ANOVA followed by Bonferroni post hoc correction. Normality was verified using the Anderson-Darling test. A propensity score analysis was performed as described⁵² in order to minimize any selection bias due to the differences in clinical characteristics between the empagliflozin and the no-empagliflozin groups. Briefly, for each patient a propensity score was calculated by the use of a non-parsimonious multivariable logistic regression, applying the 1:1 nearest neighbor propensity score matching method. Variables that could potentially affect treatment assignment or outcomes were selected, including sociodemographic characteristics, comorbidities, and concomitant drugs. We developed a dispersion model using Pearson analysis to assess the correlation between glycemia and MoCA score and between glycemia and 5mGS. In order to adjust for potential confounders (selected a priori based on their clinical significance and possible confounding effect), we performed linear regression analyses with MoCA score or 5mGS test as dependent variables.

Statistical significance was considered based on 2-tailed P<0.05. All calculations were computed using SPSS version 26 (IBM Corporation, Armonk, NY), and GraphPad Prism version 9 (GraphPad by Dotmatics, Boston, MA).

RESULTS

Patient Characteristics

We evaluated 407 patients. Since 22 patients were unwilling to provide clinical information and 60 subjects did not fulfill eligibility criteria, 325 patients were enrolled. To minimize potential selection bias and confounding variables, we propensity-score-matched 150 patients, 75 treated with 10 mg empagliflozin in addition to standard therapy and 75 not receiving empagliflozin (Figure 1). At baseline, there were no significant differences in age, BMI, sex distribution, comorbidities, and laboratory parameters between the two groups (Table 1). Notably, MoCA score and 5mGS were not different among the two study arms at baseline (Table 1).

Figure 1.

Flow chart of the study.

Table 1.

Baseline characteristics of our population.

	Empagliflozin	No Empagliflozin
N	75	75
Sex (M/F)	36/39	35/40
Age (years)	78.05±7.5	78.62±6.2
BMI (kg/m2)	27.7±1.6	27.9±1.4
SBP (mmHg)	121.7±7.3	121.0±7.6
DBP (mmHg)	76.6±6.5	76.4±6.7
Heart rate (bpm)	82.3±9.4	82.6±9.0
5mGS (m/s)	0.60±0.12	0.59±0.13
MoCA score	20.32±4.2	20.35±4.3
Comorbidities, n (%)
Dyslipidemia	39 (52.0)	40 (53.0)
COPD	32 (42.7)	31 (41.0)
CKD	38 (50.7)	39 (52.0)
Previous Stroke	12 (16.0)	11 (15.0)
Laboratory analyses
Plasma glucose (mg/dl)	153.87±71.6	153.36±73.1
HbA1c (mmol/mol)	61±7.5	61±6.2
Creatinine (mg/dl)	1.0±0.2	1.0±0.2

5mGS: 5-meter gait speed; BMI: body mass index; CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; DBP: diastolic blood pressure; HbA1c: glycated hemoglobin; HR: heart rate; MoCA: Montreal Cognitive Assessment; SBP: systolic blood pressure.

Empagliflozin significantly attenuates cognitive and physical impairment

In our population, we observed a strong correlation between glycemia and MoCA score (r²: 0.91; p<0.001) (Figure 2A) and between glycemia and 5mGS (r²: 0.79; p<0.001) (Figure 2B).

Figure 2.

Correlation between MoCA score and glycemia (A, r²: 0.91; p<0.001) and between 5mGS and glycemia (B, r²: 0.79; p<0.001). Differences in Montreal Cognitive Assessment (MoCA) score (C), 5-meter gait speed (5mGS, D), and frailty (E) at baseline and at 3-month follow-up in empagliflozin treated and not-treated groups (*:p<0.05; **:p<0.01; ***:p<0.001).

We then evaluated the differences in MoCA score (Figure 2C) and 5mGS (Figure 2D) at baseline and at 3-month follow-up in empagliflozin treated and non-treated patients, observing significantly different favorable effects of the empagliflozin treatment.

Lastly, we assessed how many patients in the two study arms still had frailty at 3-month follow-up; the empagliflozin-treated group included only 25.3% of patients with frailty (n: 19), while in the non-empagliflozin group 73.3% of patients (n: 55) had frailty (p<0.001, Figure 2E).

Importantly, the significant correlation linking glycemia with MoCA score (Figure 2C) and 5mGS (Figure 2D) was confirmed after adjusting for potential confounders in multivariable regression analyses with MoCA Score or 5mGS as dependent variable (Table 2).

Table 2.

Linear regression analysis in the empagliflozin group using basal MoCA score (top) or 5mGs (bottom) as dependent variable.

	OR	S.E.	95% C.I.		p
			Lower Bound	Upper Bound
Age	−.024	.019	−.062	.014	.211
BMI	.017	.062	−.105	.139	.781
SBP	−.009	.014	−.036	.018	.530
DBP	−.002	.017	−.035	.031	.903
HR	.022	.013	−.004	.047	.093
Dyslipidemia	−.099	.212	−.518	.319	.640
CKD	−.140	.221	−.577	.297	.527
COPD	.051	.227	−.397	.500	.821
HbA1c	−.766	.422	−1.600	.069	.072
Glycemia	−.046	.005	−.056	−.037	<0.001
	OR	S.E.	95% C.I.		p
			Lower Bound	Upper Bound
Age	−.001	.001	−.003	.001	.418
BMI	.002	.003	−.004	.008	.451
SBP	.000	.001	−.001	.001	.739
DBP	.000	.001	−.001	.002	.620
HR	.001	.001	.000	.002	.055
Dyslipidemia	−.010	.010	−.029	.010	.339
CKD	−.001	.010	−.021	.020	.946
COPD	.004	.011	−.017	.025	.711
HbA1c	−.002	.020	−.041	.038	.936
Glycemia	−.002	.000	−.002	−.001	<0.001

BMI: body mass index; CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; DBP: diastolic blood pressure; HbA1c: glycated hemoglobin; HR: heart rate; SBP: systolic blood pressure.

Empagliflozin reduces mitochondrial Ca²⁺ overload and mitochondrial oxidative stress

To test whether empagliflozin could attenuate mitochondrial Ca²⁺ overload and ROS production, we cultured human endothelial cells in normal glucose (5 mM), high-glucose (30 mM), and high-glucose plus empagliflozin for 24h. High-glucose condition produced mitochondrial Ca²⁺ overload (Figure 3AE) and significantly augmented ROS production in these organelles (Figure 3F), while empagliflozin significantly attenuated these responses (Figure 3).

Figure 3.

Evaluation of mitochondrial Ca²⁺ (A-E) and mitochondrial reactive oxygen species (ROS) generation (F-G) in human umbilical vascular endothelial cells (HUVECs, passages 3–7) treated for 24 hours with the indicated concentrations of glucose (Glu) and empagliflozin (Empa). Mitochondrial Ca²⁺ (A-E) was measured in at least 30 cells per group, in triplicate independent biological replicates, in response to thapsigargin (Tg, 1 μM) or after the addition of 1.8 mM Ca²⁺. To assess the production of mitochondrial ROS, cells were stained with MitoTracker Green and mitoSOX Red (F, representative images from quadruplicate experiments; dimensional bar: 20 μm), and these assays have been quantified in the panel on the right; in the violin plot, median (solid line) and quartiles (dotted lines) are indicated; *: p<0.05 vs. 5 mM Glu + vehicle; #: p<0.05 vs. 30 mM Glu + Vehicle.

SGLT2 inhibition mitigates glucose-induced endothelial permeabilization

Hyperglycemia has been linked to an augmented vascular permeability^{53, 54} and endothelial leakage has been associated to cognitive decline and frailty^55–60. Therefore, we sought to test whether inhibiting SGLT2 in human brain microvascular endothelial cells could attenuate endothelial leakage elicited by high-glucose concentrations. We observed that empagliflozin significantly attenuated the endothelial permeability induced by high-glucose (Figure 4A). To further confirm the effects of empagliflozin on endothelial leakage, we measured the expression levels of mRNA encoding for established junctional proteins (i.e. Claudin-5, Occludin) and we found that high-glucose concentrations significantly reduced the mRNA levels of both Claudin-5 (Figure 4B) and Occludin (Figure 4C), and these reductions were mitigated by empagliflozin (Figure 4BC).

Figure 4.

Effects of empagliflozin on cell permeability assessed in human brain microvascular endothelial cells (A) and mRNA levels of Occludin (B) and Claudin-5 (C); in the violin plot, median (solid line) and quartiles (dotted lines) are indicated; *: p<0.05 vs. 5 mM Glu + vehicle; #: p<0.05 vs. 30 mM Glu + vehicle. Effects of empagliflozin (1 μM) on cell viability (D) assessed in human brain microvascular endothelial cells in response to increasing doses of H₂O₂; the solid line indicates the median, whereas the dotted lines indicate the quartiles; *: p<0.05 vs. vehicle.

SGLT2 inhibition protects human endothelial cells from oxidative stress

Since we demonstrated in HUVECs that empagliflozin attenuates oxidative stress induced by high-glucose concentrations, we sought to confirm in human brain microvascular endothelial cells that inhibiting SGLT2 could improve cell viability in response to oxidative stress. Therefore, we pretreated human brain microvascular endothelial cells with empagliflozin and then we treated them with increasing concentrations of H₂O₂; after 5h, we observed that cell viability was significantly improved by empagliflozin (Figure 4D).

DISCUSSION

The compelling combination of clinical data and in vitro assays provided in the present study indicates that SGLT2 inhibition significantly improves cognitive and physical impairment in diabetic and hypertensive patients, most likely reducing oxidative stress in endothelial cells (Graphical Abstract).

Oxidative stress is known to be a common pathogenic substrate in hypertension⁶¹, diabetes⁶², and frailty^{63, 64}. Specifically, a dysregulated mitochondrial fitness has been proposed as a potential root of age-related frailty⁶⁵, and mitochondrial free radicals are crucial players in endothelial dysfunction^{66, 67}. Henceforth, we sought to determine whether the favorable effects detected in the clinical setting could be attributable to an action on mitochondrial ROS generation at the endothelial level.

Our findings are in agreement with previous reports showing that hyperglycemia causes an augmented generation of mitochondrial ROS^{68, 69}; in fact, mitochondrial ROS are recognized as a major cause of clinical complications associated with diabetes^{70, 71}.

Hyperglycemia, alongside its actions on endothelial function, has been also associated with memory disturbances and a potential role in the development of cognitive impairment has been proposed^{72, 73}. Thus, a stable glycemic control is seen as imperative to reduce the incidence of functional decline and to avoid complications^74–78. Herein, we demonstrate that empagliflozin significantly improves cognitive impairment, assessed via MoCA, and physical decline, evaluated via 5mGS test, and we prove that these parameters, which we had previously shown to correlate with each other⁷⁹, independently correlate with blood glucose levels. Remarkably, our data also indicate a direct effect of empagliflozin on endothelial cells, which goes beyond its actions merely attributable to a reduction of blood glucose^80–82 or to the offsetting of insulin resistance^83–86. Specifically, we show that SGLT2 inhibition significantly attenuates mitochondrial Ca²⁺ overload and the subsequent increase in ROS production in human endothelial cells.

The new 2022 guidelines on the management of heart failure⁸⁷ of the American-Heart-Association, the American-College-of-Cardiology, and the Heart-Failure-Society-of-America, recommend SGLT2-inhibitors as Class 1A in HFrEF and Class 2A in HFmrEF and HFpEF (conditions in which ARNI and MRA have weaker recommendations: Class 2B). These recommendations are mainly based on the protective effects of SGLT2 inhibitors on cardiovascular outcomes, independent of the glucose-lowering effects, revealed by recent clinical trials: the DAPA-HF (Dapagliflozin and Prevention of Adverse Outcomes in Heart Failure)⁸⁸, the EMPEROR-HF (Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Preserved Ejection Fraction)⁸⁹, and the EMPEROR-PRESERVED (Empagliflozin Outcome Trial in Patients with Chronic Heart Failure with Preserved Ejection Fraction)²¹ trials. However, the exact mechanism underlying these beneficial effects on cardiovascular outcomes have remained hitherto elusive.

Our results are consistent with a pilot study testing the acute effects of dapagliflozin in 16 diabetic patients, showing a marked reduction of urinary isoprostanes, a surrogate of systemic oxidative stress 48 hours after administration compared to baseline values⁹⁰. In our study, we also proved that empagliflozin counteracts the increased endothelial permeability induced by high-glucose, by regulating the expression of Occludin and Claudin-5 at the transcriptional level, tight junction proteins that have been shown to partake in cognitive impairment^91–93. Claudin-5 levels were also shown to be reduced in the nuclus accumbens of patients with depression^{94, 95}. Additionally, empagliflozin significantly prevented the death of cerebral endothelial cells induced by oxidative stress, and these data are particularly relevant when considering that oxidative stress is fundamental in process like senescence and aging^96–100. Therefore, our findings are highly suggestive for a potential role of SGLT2 inhibitors in preventing neurodegenerative disorders.

Strengths of our study include its prospective nature, the analysis of a real-world homogeneous population of elderly patients with diabetes and hypertension, the integration of clinical and in vitro data, and the consistency of the beneficial effects of empagliflozin in two different types of human endothelial cells, namely cells obtained from the umbilical vein and brain microvascular endothelial cells. Nevertheless, our study is not exempt from limitations, including the relatively brief follow-up, having tested the effects of only one SGLT2 inhibitor, and having enrolled only Caucasian patients.

Novelty and Relevance.

What Is New?

• The Sodium GLucose co-Transporter 2 (SGLT2) inhibitor Empagliflozin improves both cognitive and physical impairment in frail patients with hypertension and diabetes.

• Empagliflozin ameliorates endothelial dysfunction induced by high-glucose.

• Empagliflozin reduces mitochondrial Ca²⁺ overload and oxidative stress in endothelial cells.

What Is Relevant?

• SGLT2 inhibitors have been shown to have beneficial effects on the cardiovascular system but the underlying mechanisms were not fully understood.

• This work indicates that SGLT2 inhibitors have favorable effects on different types of human endothelial cells, i.e. umbilical and brain microvascular endothelial cells.

Clinical/Pathophysiological Implications?

The management of frailty in older adults should consider the expected benefits given by the treatment with SGLT2 inhibitors especially taking into account the pathophysiological rationale provided by this study, in terms of improvement of endothelial function and reduced oxidative stress.

Perspectives.

Based on our results and on the emerging pleiotropic effects of SGLT2 inhibitors²⁶, we speculate that SGLT2 inhibitors may be considered as an anti-frailty drug. This view is strongly supported by the beneficial effects of empagliflozin on endothelial cells, in terms of mitigated mitochondrial Ca²⁺ overload, reduced oxidative stress, and improved cell viability.

Nonstandard abbreviations

5mGS

5-meter Gait Speed

BMI

Body Mass Index

Ca²⁺

Calcium

CKD

Chronic Kidney Disease

COPD

Chronic Obstructive Pulmonary Disease

HbA1c

Glycated Hemoglobin

hBMECs

Human Brain Microvascular Endothelial Cells

HR

Heart Rate

HUVECs

Human Umbilical Vascular Endothelial Cells

MoCA

Montreal Cognitive Assessment

ROS

Reactive Oxygen Species

SGLT2

Sodium GLucose co-Transporter 2