In vivo administration of urolithin A and B prevents the occurrence of cardiac dysfunction in streptozotocin-induced diabetic rats

The treatment started immediately after the documented increase in blood glucose levels (2 days after STZ injection; blood glucose cut-off: 250 mg/dl). To prepare UA and UB stock solutions, the compounds were dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich) at 50 mg/ml and stored in darkness at 4 °C. For each animal, immediately prior to i.p. injection, an appropriate aliquot was taken from the stock solution and diluted in PBS to reach the desired concentration in a final volume of 250 µl.

Hemodynamic measurements were performed under anaesthesia in all CTRL rats and in all D animals after 3 weeks of hyperglycaemia (D3, D3-UA and D3-UB). This early stage of diabetes is characterised by the occurrence of the first signs of cardiomyocyte contractile dysfunction associated with changes in myocardial tissue microenvironment, in the absence of marked cardiac structural damage, collagen accumulation or inflammatory cell infiltrate [2].

Plasma samples were defrosted, vortexed, and extracted as previously reported [14]. Briefly, 400 μl plasma aliquots were mixed with 1 ml of acidified acetonitrile (2% formic acid, Sigma-Aldrich). Samples were vortexed, ultrasonicated for 10 min, and centrifuged at 16,000g for 10 min. The supernatant was reduced to dryness using a Speedvac concentrator (Thermo Fisher Scientific Inc., San José, CA, USA), the pellet was then suspended in 100 μl of 50% (v:v) methanol (Sigma-Aldrich) acidified with 0.1% (v:v) formic acid, and centrifuged at 16,000g for 5 min prior to analysis by ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS/MS).

Urine was defrosted, centrifuged at 16,000g for 5 min, and diluted 1:4 with water containing 0.1% formic acid. Powdered tissues were extracted by using a protocol similar to that used for plasma samples but following repeated extractions. Briefly, from each tissue sample (~1 g), proteins were precipitated with 1500 μl of 2% formic acid in acetonitrile (the solvent volume for extraction was modified proportionately to the total weight of the samples). The samples were then vortexed vigorously for 5 min, placed in a sonicator bath for 10 min, and centrifuged (16,000g for 5 min). A second extraction was performed for each sample with 3 ml of acidified acetonitrile as described above. The two supernatants were pooled and dried under vacuum. Finally, the pellet was suspended in 100 μl of methanol:water:formic acid (50:50:0.1, v/v/v) and centrifuged at 16,000g for 5 min before analysis.

Plasma, urine, heart ventricles, liver, and pancreas samples were analysed according to Mele et al. [15] with minor modifications. Briefly, samples were run using an Accela UHPLC 1250 equipped with a linear ion trap-mass spectrometer (LTQ XL, Thermo Fisher Scientific) fitted with a heated-electrospray ionization probe (H-ESI-II). Separations were performed using a XSELECTED HSS T3 (50 × 2.1 mm), 2.5 μm particle size (Waters, Milford, MA, USA), with an injection volume of 5 μl, column oven temperature of 40 °C and elution flow rate of 0.3 ml/min. The initial gradient was 80 of 0.1% aqueous formic acid and 20% acetonitrile (in 0.1% formic acid), reaching 80% acetonitrile at 6 min. The MS conditions included: capillary temperature of 275 °C and source heater temperature of 250 °C, sheath gas flow of 60 units, auxiliary gas of 7 units, source voltage of 3.5 kV, and capillary voltage and tube lens of −39 and −88 V, respectively. Analyses were carried out using specific MS² full scans corresponding to the putative phase II metabolites of UA and UB (glucuronide, sulphate, methyl, or a combination thereof), with collision induced dissociation (CID) equal to 35 (arbitrary units) (Table 1). After this first step, further specific MS³ analyses were carried out to unambiguously identify the compounds revealed in the first step. Pure helium gas was used for CID. Data processing was performed using Xcalibur software from Thermo Scientific. Quantification was performed with calibration curves of pure standards in the case of UA, UB, and UB-glucuronide (provided by Prof. O. Dangles, INRA, Avignon, France). UA-sulphate was expressed as UA equivalents, while UB-sulphate as UB equivalents.

Invasive hemodynamic data were recorded in 39 rats (10 CTRL, 9 D3, 10 D3-UA, and 10 D3-UB). After anaesthesia with ketamine chloride (Imalgene, Merial, Milan, Italy; 40 mg/kg i.p.) plus medetomidine hydrochloride (Domitor, Pfizer Italia S.r.l., Latina, Italy; 0.15 mg/kg i.p.), the right carotid artery was cannulated with a microtip pressure transducer catheter (Millar SPC-320, Millar Instruments, Houston, TX, USA) connected to a recording system (Power Laboratory ML 845/4 channels, 2Biological Instruments, Besozzo, Italy) and systolic and diastolic blood pressures were determined. The catheter was then advanced into the left ventricle to measure the following parameters: left ventricular (LV) systolic pressure (LVSP), LV end-diastolic pressure (LVEDP), the maximum rate of ventricular pressure rise (+dP/dt_max) and reduction (−dP/dt_max), taken as indexes of myocardial mechanical efficiency, isovolumic contraction time (IVCT: duration of isovolumic contraction), and total cycle duration (Tcycle) (software package CHART B4.2).

From the heart of 6 CTRL, 6 D3, 5 D3-UA, and 4 D3-UB rats, individual LV myocytes were enzymatically isolated by collagenase perfusion. Briefly, the heart was rapidly excised through median sternotomy and mounted on a Langendorff isolated heart apparatus. The heart was then perfused at 37 °C by cannulating the aorta with the following sequence of solutions gassed with 100% oxygen: (1) calcium-free solution for 5 min to remove the blood, (2) low-calcium solution (0.1 mM) plus 1 mg/ml type 2 collagenase (Worthington Biochemical Corporation, Lakewood, NJ, USA) and 0.1 mg/ml type XIV protease (Sigma-Aldrich) for about 20 min, and (3) an enzyme-free, low-calcium solution for 5 min. Calcium-free solution contained the following (in mM, all chemicals from Sigma-Aldrich): 126 NaCl, 22 dextrose, 5.0 MgCl₂, 4.4 KCl, 20 taurine, 5 creatine, 5 Na pyruvate, 1 NaH₂PO₄, and 24 HEPES (pH  =  7.4, adjusted with NaOH). The left ventricle was minced and shaken for 10 min. The cells were filtered through a nylon mesh and re-suspended in low-calcium solutions for 30 min (0.1 and 0.5 mM, for 15 min each). Then, cells were used for measuring sarcomere shortening and calcium transients.

Mechanical properties of freshly isolated ventricular myocytes were assessed by using the IonOptix fluorescence and contractility systems (IonOptix, Milton, MA, USA). LV myocytes were placed in a chamber mounted on the stage of an inverted microscope (Nikon-Eclipse TE2000-U, Nikon Instruments, Florence, Italy) and superfused (1 ml/min at 37 °C) with a Tyrode solution containing (in mM; all chemicals from Sigma-Aldrich): 140 NaCl, 5.4 KCl, 1 MgCl₂, 5 HEPES, 5.5 glucose, and 1 CaCl₂ (pH 7.4, adjusted with NaOH). Only rod-shaped myocytes with clear edges and average sarcomere length ≥1.7 µm were selected for the analysis, using a 40× oil objective lens (NA: 1.3).

The cells were field stimulated at a frequency of 0.5 Hz by constant current pulses (2 ms in duration, and twice diastolic threshold in intensity; MyoPacer Field Stimulator, IonOptix). Load-free contraction of myocytes was measured with the IonOptix system, which captures sarcomere length dynamics via a Fast Fourier Transform algorithm. A total of 359 isolated ventricular myocytes were analysed (66 from CTRL hearts, 91 from D3, 100 from D3-UA, and 102 from D3-UB) to compute the following parameters: mean diastolic sarcomere length, fraction of shortening (FS), maximal rates of shortening (−dL/dt_max) and re-lengthening (+dL/dt_max), and time to 10, 50 and 90% of re-lengthening (time to basic length: TBL10%, TBL50%, and TBL90%). Steady-state contraction of myocytes was achieved before data recording by means of a 10 s conditioning stimulation. Sampling rate was set at 1 kHz.

In a subset of cells of each group (33 from CTRL hearts, 57 from D3, 63 from D3-UA, and 42 from D3-UB), Ca²⁺ transients were measured simultaneously with cell motion. Ca²⁺ transients were detected by epifluorescence after loading the myocytes with fluo-3-AM (10 µM; Invitrogen, Carlsbad, CA, USA) for 30 min. Excitation length was 480 nm, with emission collected at 535 nm. Fluo-3 signals were expressed as normalized fluorescence (f/f0: fold increase). The time course of the fluorescence signal decay was described by a single exponential equation, and the time constant (tau) was used as a measure of the rate of intracellular Ca²⁺ clearing [16].

The hearts of 2 CTRL, 3 D3, 4 D3-UA, and 4 D3-UB were excised, and the left and right ventricles were weighed and immediately frozen at −80 °C, in order to determine the expression levels of the pro-inflammatory cytokine fractalkine (rabbit polyclonal ANTI-CX3CL1 antibody, 1:1000, Abcam, Cambridge, UK), and the following proteins involved in the intracellular calcium dynamics: the sarcoplasmic reticulum calcium ATPase 2 (SERCA2; rabbit polyclonal ANTI-SERCA2 ATPase antibody, 1:8000, Abcam), the regulatory protein phospholamban (PLB; mouse monoclonal ANTI-PHOSPHOLAMBAN antibody, 1:5000, Abcam) and the phosphorylated form of phospholamban (PLB-P; rabbit polyclonal ANTI- PHOSPHOLAMBAN (phospho S16) antibody, 1:1000, Abcam). The left and right ventricular tissue was mechanically fragmented in liquid nitrogen, and lysed with 300 µl of lysis buffer containing the following (all from Sigma-Aldrich): protease (1:100) and phosphatase (1:100) inhibitors, NaCl (150 mM), Tris–HCl (50 mM), EDTA (5 mM), Nonidet P-40 (1%), sodium fluoride solution (10 mM), sodium diphosphate dibasic (10 mM), SDS (0.1%) and sodium deoxycholate (0.5%). For each animal, equivalents of 50 µg of protein were separated on a 10% gradient SDS-PAGE gel for fractalkine and 4–20% gradient SDS-PAGE gel (Bio-Rad, Hercules, CA, USA) for SERCA2, PLB, and PLB-P. Membranes were blocked with 5% milk in Tris-Buffered Saline Tween-20 (TBS-T, Sigma-Aldrich) and incubated overnight at 4 °C with the primary antibody. After washing the membranes, a second incubation was performed for 1 h at RT with peroxidase-conjugated affinity purified goat anti-rabbit (goat Anti-Rabbit IgG Horseradish Peroxidase Conjugate, 1:8000; Bio-Rad, Hercules, CA, USA) or anti-mouse (goat Anti-Mouse IgG Horseradish Peroxidase Conjugate, 1:5000; Bio-Rad) secondary antibodies. Peroxidase activity was developed using the ECL Western blotting system (Amersham, Rahn AG, Zürich, Switzerland), according to the instructions of the manufacturer. Blots were scanned and the intensity of bands was quantified by means of the ImageJ software (NIH, Bethesda, MD, USA). All measurements were performed in technical triplicate. Actin (anti-actin rabbit polyclonal antibody, 1:5000, Sigma-Aldrich) was used as the loading control.

The IBM SPSS statistical package (International Business Machines Corporation, Armonk, NY, USA, version 23) was used. Normal distribution of variables was checked by means of the Kolmogorov–Smirnov test. Data are reported as mean values ± SEM, except the urolithin concentrations in fluids and tissues which were not normally distributed and were thus expressed as median and interquartile range (IQR). Comparisons among groups involved two-way ANOVA followed by post hoc individual comparisons with the Bonferroni or Games-Howell test, when appropriate (hemodynamics), 2-factor Nested ANOVA for repeated measurements (cell mechanics, and Ca²⁺ transients), and non-parametric statistical test Kruskal–Wallis and U Mann–Whitney test (western blot data and urolithin concentrations). Differences were considered statistically significant at p < 0.05.

The administration of urolithin aglycones by i.p. injection guaranteed the presence of urolithin metabolites in circulation, as assessed by evaluating the plasma concentration of UA and UB and their putative phase II conjugates (Tables 1, 2). In the case of UB treatment (D3-UB), only UB metabolites were found: UB-sulphate was the predominant metabolite and was found in all plasma samples from rats treated with UB, while UB-glucuronide was present in some but not all samples (4/9). The profile of urolithin metabolites in circulation in the case of the animals treated with UA (D3-UA) was more complex: besides the aglycone form (UA) and UA-sulphate, some metabolites originating from the dehydroxylation of UA forming UB and further phase II conjugation were identified. In particular, both UB–glucuronide and UB-sulphate were detected in some D3-UA samples (Table 2). Similarly to what was reported for UB, UA-sulphate was the predominant urolithin metabolite in the circulation for D3-UA rats. A high inter-individual variability was observed in the plasma concentrations of urolithins for both D3-UB and D3-UA animals (Table 2). Control plasma samples from the CTRL and D3 groups did not contain detectable amounts of any urolithins. The urinary excretion of urolithin derivatives was very limited and only UB-sulphate was detected in D3-UB rats (Table 2).

The analysis of tissue samples from the four experimental groups revealed the presence of urolithins in heart, pancreas, and liver of both D3-UB and D3-UA rats (Fig. 1), while these metabolites were absent in control samples. Aglycone and sulphate forms of urolithins, but not glucuronides, were detected in these tissues, although they were not detected in all the samples (Fig. 1), in line with the inter-individual variability highlighted for plasma and urine. Bioaccumulation of urolithin metabolites in the heart of D3-UB rats occurred in the form of UB and UB-sulphate (Fig. 1), without statistically significant differences between the concentrations of these two compounds. When D3 rats were treated with UA (D3-UA), UA, UA-sulphate and UB were found in the heart tissue. Again, no statistically significant differences were shown among metabolites for D3-UA heart samples. Urolithin A and B were also present in rat pancreatic and hepatic tissue, as both aglycone and sulphate. UB and UB-sulphate were detected in D3-UB pancreas and liver samples, while up to four metabolites (UA, UA-sulphate, UB, and UB-sulphate) were detected in D3-UA pancreas and livers. No statistical differences among metabolite concentrations for each treated group (D3-UB or D3-UA) and tissue (pancreas or liver) were observed, accounting for the absence of a prevailing compound. In general, UB, or its sulphated form, were found in tissues at higher concentrations with respect to UA or its sulphated form (Fig. 1).

The heart rate measured under anaesthesia was similar in the four groups (193 ± 7 beats per minute, 205 ± 11, 204 ± 7, 209 ± 9 in CTRL, D3, D3-UA, and D3-UB, respectively). Compared with CTRL animals, D3 animals exhibited a global deterioration of hemodynamic performance, as indicated by the significant decrease in left ventricular systolic pressure (LVSP; Fig. 2a), the maximal rate of LV pressure rise (+dP/dt_max) and decline (−dP/dt_max) (Fig. 2c, d), and the marked prolongation of isovolumic contraction time (IVCT; Fig. 2e) and total cycle duration (Tcycle; Fig. 2f). The treatment with UB induced a significant recovery of most parameters, during both contraction and relaxation. Specifically, the values of LVSP, ± dP/dt_max, IVCT and Tcycle approached those measured in CTRL animals (Fig. 2a, c–f). A lower effect was observed for UA, limited to the contraction phase (+dP/dt_max and IVCT; Fig. 2c, e). No significant differences among groups were found for LVEDP (Fig. 2b).

The average diastolic sarcomere length was comparable in all groups (1.732 ± 0.002 µm), as well as the fraction of shortening (Fig. 3c). Conversely, in accordance with the hemodynamic data recorded in vivo, contraction/relaxation properties and intracellular calcium dynamics were worsened in unloaded ventricular myocytes isolated from D3 hearts in comparison with CTRL (Fig. 3d–f). Specifically, D3 cardiomyocytes exhibited a significant decrease in the maximal rate of shortening (−dL/dt_max; Fig. 3d) and re-lengthening (+dL/dt_max; Fig. 3e), and a prolongation of re-lengthening times (TBL10%, TBL50%, and TBL90%; Fig. 3f). The impaired contractility in D3 cells was accompanied by a significant increase in calcium transient amplitude (+33%, Fig. 3g) and a prolonged tau (+37%, Fig. 3h).

Both urolithins led to an almost complete recovery of cellular mechanical properties and calcium dynamics, with only a slightly higher effect of UB (Fig. 3d–h).

SERCA2, PLB, and PLB-P expression. Sarcoplasmic reticulum (SR) calcium reuptake rate is mainly regulated by the SERCA2 activity which, in turn, is modulated by a small molecular weight protein, the phospholamban (PLB). The phosphorylation status of PLB (PLB-P) is able to remove the inhibitory effect of PLB on SERCA2, resulting in increased SR-Ca²⁺ uptake and enhanced myocyte contractility and relaxation [17]. On the other hand, alterations in SERCA2 expression and/or activity are recognized among the major factors contributing to the ventricular dysfunction observed in diabetic cardiomyopathy [18‐21].

A slight decrease (−19%; not significant) was observed in the expression level of SERCA2 associated with only negligible changes in PLB level, in D3 group as compared with CTRL (Fig. 4a, b), leading to a significant reduction in the SERCA2/PLB ratio (−35% vs CTRL, Fig. 4c). On the other hand, a reduced SERCA2/PLB ratio was shown to be associated with a parallel lessening of SERCA2 affinity for Ca²⁺, resulting in altered cardiomyocyte mechanics and depressed left ventricular function [22, 23]. Furthermore, diabetic hearts showed a pronounced decrease in PLB-P/PLB ratio (−21% vs CTRL, Fig. 4d) that, in the absence of substantial changes in the total PLB expression (Fig. 4b), suggested a reduction in the phosphorylated form of the protein. All these changes can account for the altered cellular contractile properties observed in diabetic cardiomyocytes. Although the effect of UB was slightly higher, both urolithins induced a marked amelioration of the two ratios whose values approached those measured in control hearts (Fig. 4c, d).

Fractalkine expression. Fractalkine, beside its action as a chemoattractant, is known to possess a detrimental effect on cardiac cells independently of inflammatory cell recruitment [24, 25]. The expression level of fractalkine (CX3CL1) was increased by hyperglycaemia (+48%; Fig. 4e) and significantly decreased by both urolithins (−33% for UA and −28% for UB; Fig. 4e).