The (R)-enantiomer of the 6-chromanol derivate SUL-121 improves renal graft perfusion via antagonism of the α1-adrenoceptor

SUL-compounds are protectants from cold-induced ischemia and mitochondrial dysfunction. We discovered that adding SUL-121 to renal grafts during warm machine reperfusion elicits a rapid improvement in perfusion parameters. Therefore, we investigate the molecular mechanisms of action in porcine intrarenal arteries (PIRA). Porcine kidneys were stored on ice overnight and perfusion parameters were recorded during treatment with SUL-compounds. Agonist-induced vasoconstriction was measured in isolated PIRA after pre-incubation with SUL-compounds. Receptor binding and calcium transients were assessed in α1-adrenoceptor (α1-AR) transgenic CHO cells. Molecular docking simulation was performed using Schrödinger software. Renal pressure during warm reperfusion was reduced by SUL-121 (−11.9 ± 2.50 mmHg) and its (R)-enantiomer SUL-150 (−13.2 ± 2.77 mmHg), but not by the (S)-enantiomer SUL-151 (−1.33 ± 1.26 mmHg). Additionally, SUL-150 improved renal flow (16.21 ± 1.71 mL/min to 21.94 ± 1.38 mL/min). SUL-121 and SUL-150 competitively inhibited PIRA contraction responses to phenylephrine, while other 6-chromanols were without effect. SUL-150 similarly inhibited phenylephrine-induced calcium influx and effectively displaced [7-Methoxy-3H]-prazosin in CHO cells. Docking simulation to the α1-AR revealed shared binding characteristics between prazosin and SUL-150. SUL-150 is a novel α1-AR antagonist with the potential to improve renal graft perfusion after hypothermic storage. In combination with previously reported protective effects, SUL-150 emerges as a novel protectant in organ transplantation.

SCientifiC RepoRTs | (2019) 9:13 | DOI: 10.1038/s41598-018-36788-0 preservation agent that protects from damage associated with hypothermia and reactive oxygen species (ROS) in adipose-derived stem cells 5 . Specifically, SUL-109 preserved the mitochondrial network structure and activated mitochondrial complexes I and IV, thereby preventing ROS formation while maintaining ATP production. Recently, we demonstrated the effectiveness of SUL compounds against hypothermic damage in the kidney as SUL-109 and the related SUL-121 compound ameliorated renal injury after deep hypothermia and rewarming in rats by preserving mitochondrial mass, function and ATP levels 6 . Moreover, the protective effects of SUL compounds have now been demonstrated in several disease models characterized by mitochondrial dysfunction. In type 2 diabetes, treatment with SUL-121 halted the progression of kidney damage and prevented endothelial dysfunction through a mechanism involving lowered ROS production 7 . Further, in chronic obstructive pulmonary disease, SUL-121 also reduced ROS production, resulting in reduced LPS-induced airway hyper reactivity and neutrophil influx and increased hydrogen sulfide (H 2 S) production 8 .
Although we thoroughly investigated the underlying protective mechanisms of SUL-109 and SUL-121 on the kidney during hypothermia, some observations were left unaddressed. Interestingly, SUL-121 caused a significant further reduction in blood pressure during the hypothermic phase compared to controls not receiving SUL-121 6 . In addition, we now report that SUL-121 induced a profound and immediate increase in renal flow during warm machine reperfusion of cold-stored porcine kidneys. To further explore these putative vascular effects of SUL-121, we here investigate the effects of SUL-121 and its enantiomers on renal machine reperfusion and on vasomotor function of porcine intra-renal arterial segments.

Results
Effects of SUL-121 and its enantiomers SUL-150 and SUL-151 on renal perfusion pressure and flow. Cold-stored porcine kidneys were rewarmed to 20 °C with warm oxygenated medium while flow-rate was kept constant and occasionally adjusted to maintain a pressure of 80 mmHg (data not shown). Addition of SUL-121 to the circulating medium caused a profound decrease in renal pressure ( To investigate how the observed decrease in renal resistance by SUL-150 would benefit renal perfusion flow, additional experiments were performed with fixed pressure (80 mmHg) and dynamic flow adjustment. SUL-150 caused an approximate 35% increase in renal flow from baseline (Fig. 1D,E; baseline: 16.2 ± 1.7 mL/min, SUL-150: 22.0 ± 1.4 mL/min, p < 0.05).

Effects of SUL-150 on constriction of isolated porcine intrarenal arteries.
To further investigate the putative vasodilatory effects of SUL-150, we explored the effects of pre-incubation with SUL-150 on vasoconstriction of isolated intrarenal arteries to various agonists (Fig. 2). For this, the effects of SUL-150 (30 µM and 100 µM) on cumulative concentration response curves to α 1 -adrenoceptor (α 1 -AR) agonists PE and methoxamine ( Fig. 2A,B, respectively), histamine (Fig. 2C) and U46619 (a synthetic thromboxane A2 (TP) receptor agonist, Fig. 2D) were determined. SUL-150 caused a concentration dependent competitive inhibition of the concentration response curves to PE and methoxamine ( Fig. 2A,B). SUL-150 did not affect responses to histamine and induced a small, albeit statistically significant shift in EC 50 and E max of U46619-induced vasoconstriction (Supplementary Tables 1 and 2). Further, these effects are not mediated through the endothelium as removal of the endothelium did not abrogate the effects of SUL-150 on methoxamine induced vasoconstriction ( Supplementary Fig. 1A). Rather, our findings suggest a competitive antagonistic effect of SUL-150 that is specific for α 1 -AR. We previously reported that vascular constriction via α 1 -ARs is partly dependent on transactivation of EGF receptors 9,10 . To exclude that inhibition of vasoconstriction by SUL-150 is mediated through altered epidermal growth factor receptor (EGFR) transactivation, additional experiments were performed in isolated blood vessels and cell culture (Supplemental Fig. 2). These experiments demonstrated that SUL-150 antagonism towards α 1 -ARs is indeed not mediated through EGFR transactivation.
Schild analysis of SUL-121, SUL-150 and SUL-151 antagonism in α 1 -AR mediated vasoconstriction. To further characterize the effects of SUL-121 and its enantiomers on α 1 -AR mediated vasoconstriction, Schild analysis was performed. Concentration response curves to PE in the presence of 10 µM, 30 µM and 100 µM SUL-121, SUL-150 or SUL-151 are shown in Fig. 3. SUL-121 and SUL-150 caused a dose-dependent right shift of the curves with significant changes in EC 50 (Table 1). Schild comparison of SUL-121 and SUL-150 constriction response curves revealed a lower potency of SUL-121 (pA 2 = 4.92 ± 0.02) over SUL-150 (pA 2 = 5.37 ± 0.03), with Schild slopes not deviating significantly from unity (Fig. 3B,D, respectively). Pre-treatment with SUL-151 was without effect and therefore our findings are in line with the results obtained from isolated perfused kidneys.  Table 2). Schild analysis for the three α 1 -AR subtypes is shown in Fig. 4B,D and F. For α 1 -AR subtypes A and B, pA 2 values were similar (5.31 ± 0.03 and 5.30 ± 0.10 respectively) to the pA 2 obtained from the inhibitory effects of SUL-150 on vascular contraction (5.37 ± 0.03). A significantly lower pA 2 was found for α 1 -AR subtype D (5.06 ± 0.10, p < 0.05). SUL-151 did not affect calcium transients in α 1 -AR transfected CHO cells ( Supplementary Fig. 1B). In line with our data from vascular measurements, SUL-150 did not have profound effects on U46619 and histamine induced calcium transients in Hela cells ( Supplementary Fig. 1C,D).

Effects of structural analogues of SUL
Radioligand binding assay in α 1A -AR overexpressing CHO cells. To explore whether SUL-150 directly interacts with the α 1 -AR as a receptor antagonist, a competitive binding assay was performed on α 1A -AR transgenic CHO cells using radiolabelled prazosin, an established α 1A -AR antagonist. Incubation with SUL-150 induced a significant displacement of the radioligand starting at 25 µM concentrations indicating higher potency than SUL-151, which demonstrated significant displacement only at the highest concentration (100 µM, Fig. 5A). Induced fit molecular docking simulation. To explain experimentally determined differences between SUL-150 and SUL-151 in competitive binding to the antagonist site on the α 1A -AR, Induced Fit molecular docking simulation was performed using prazosin as a reference. As no crystal structure is yet available for α 1 -AR a homology model was built using SWISS-MODEL. Manual evaluation of the resulting models identified an optimal homology between α 1 -AR and D3 dopamine receptor, with 36% sequence identity and 68% coverage. Validation using Ramachandran plot suggested 0.96% of residues as outliers, specifically PRO242, ASP228 and PRO299. In addition, 3.21% of residues were in allowed regions and 95.83% in the preferred region. Prazosin coexists in two protonation forms at pH 7.4. In the protonated form, the N1 assumes a positive charge, subsequently forming a salt bridge with the negatively charged side chain of ASP106, ultimately causing this form to assume an inverted orientation relative to its non-protonated form (Fig. 5B). The quinazoline scaffold of non-protonated prazosin was docked close to TM5 to form a hydrogen bond between the 6,7-methoxy groups and SER188 11,12 , a hydrogen bond between the furan oxygen and SER83, and between the prazosin carboxamide and GLN177 side chain; van der Waals interactions with side chains of PHE86, VAL107, ILE178, SER192 11 , PHE289 13 , MET292 and PHE312 12,14 ; π-π interactions with PHE288 and PHE312 12 ; and a π-hydrogen bond interaction with the side chain carboxyl group and backbone peptide carboxamide of ASP106 12,15 . Thus, the proposed binding mode of non-protonated prazosin utilized interactions that were in accord with those described in previous literature.
Following the generation of a reference prazosin binding site, an induced fit of SUL-compounds was performed ( Fig. 5C,D). Both SUL-150 and SUL-151 demonstrated alignment of the chromane scaffold with prazosin quinazoline, 6-hydroxy groups (SUL) and 6-metoxy (prazosin) as well as over their common piperazine moiety. Residues which were involved in forming contacts with all three compounds were VAL107, ILE178, SER188, PHE288, PHE289. Glide scores computed using the Schrödinger Small-Drug Discovery Suite were −10.8 kcal × mol −1 , −10.2 kcal × mol −1 and −9.4 kcal × mol −1 for prazosin, SUL-150 and SUL-151, respectively, which corresponded to a difference in potency (prazosin pA 2 for PE 8.93 ± 0.04 16 , SUL-150 5.37 ± 0.03). Interestingly, prazosin and SUL-150 ( Fig. 5C) formed contacts with PHE312 and ASP106 confirmed in prazosin binding by mutagenesis 11,15 , whereas SUL-151 ( Fig. 5D) did not show interactions with these residues. Additionally, the chirality of SUL-150 enables the orientation of its carboxamide towards ASN179, effectively forming a hydrogen bond. Furthermore, the protonated N-terminal of SUL-150 formed an additional hydrogen bond with TYR316 and a π-cation interaction with TRP102, known to frequently appear among aromatic amino acids 17 . Taken together, our model suggests that SUL-150 has a similar α 1 -AR binding profile to prazosin, considering the orientation of core ligand scaffolds in the antagonist binding pocket and formation of contacts with key receptor residues.

Discussion
Previously, we demonstrated that 6-chromanol derived SUL-compounds are protective in models of type 2 diabetes 7 , long-term hypothermic storage 5 and chronic obstructive pulmonary disease 8 . Further, the molecular mechanisms of SUL-compounds were studied in detail in adipose-tissue-derived stem cells demonstrating that compound SUL-109 activated mitochondrial membrane complexes I and IV thereby alleviating mitochondrial ROS production and preserving ATP production 5 . In addition to the effects of SUL-compounds on the mitochondrial respiratory chain, we now demonstrate that SUL-121, particularly through its (R)-enantiomer SUL-150, has antagonistic effects on α 1 -AR mediated vasoconstriction in isolated porcine kidneys. The present finding that SUL-150 improved renal machine perfusion parameters through antagonism of α 1 -AR, indicates active α 1 -AR signalling during rewarming of cold-stored kidneys. At present, only limited data  is available on the role of α 1 -AR in renal machine perfusion. However, it was reported that treatment of living donors with the α 1 -AR antagonist phentolamine improved flow and decreased resistance of machine-perfused graft kidneys, and promoted early allograft function after transplantation 18 . In a separate study, inhibition of the Rho-kinase pathway in porcine kidneys during cold storage resulted in higher flow upon reperfusion, as well as improved glomerular function and tubular cell integrity 19 . Considering the involvement of α 1 -ARs in Rho-kinase signalling, these results may further highlight the importance of vascular α 1 -AR in the regulation of kidney perfusion. In addition, in an experimental model of donation after brain death, massive sympathetic discharge as a result of brain death was shown to induce hemodynamic injury, resulting in increased ischemia reperfusion  to the transplanted lung 20 . Interestingly, these effects could be blocked by treatment with the α 1 -AR antagonist prazosin. Taken together, these findings support the importance of α 1 -AR as a target to treat vascular injury in transplantation.
The findings that SUL-121 and SUL-150 significantly improved renal perfusion were further corroborated in isolated porcine intrarenal arteries, where we assessed the effects of SUL-121 and its enantiomers on vasoconstrictions induced by various agonists. SUL-121 and SUL-150, but not SUL-151 inhibited PE mediated contractions, confirming that the vascular effects of SUL-121 are predominately mediated through its (R)-enantiomer SUL-150. SUL-121 and SUL-150 enantiomers only had profound effects on α 1 -AR and not on histamine, and U46619 mediated constrictions. Further, none of the metabolites of SUL-150, including Trolox, showed similar effects as SUL-150. As Trolox is a strong antioxidant 21 , these results demonstrate that the effects of SUL-150 are not mediated through scavenging of ROS. In addition, none of the other SUL-compounds investigated (including the enantiomers of SUL-109) inhibited PE-induced contractions. Since SUL-109 has similar effects on mitochondrial function as SUL-121 6 , these findings indicate that the vascular effects of SUL-150 are also not mediated through altered mitochondrial function. The effects of SUL-150 also persisted after removal of the endothelium. Taken together, these results indicate a specific competitive inhibition of α 1 -AR on the vascular smooth muscle by SUL-121 and SUL-150.
To further demonstrate the specificity of SUL-150 for the α 1 -AR we measured calcium signalling in CHO cells overexpressing specific α 1 -AR subtypes. SUL-150 inhibited calcium signalling through α 1 -AR subtypes A and B, whereas the effect on α 1 -AR subtype D was less pronounced. Although the relative expression of α 1 -AR subtypes is unknown in porcine intrarenal arteries, data obtained from rat indicate high mRNA expression levels of α 1A and α 1B adrenergic receptors in renal blood vessels, whereas α 1D -AR mRNA was only found in minimal quantities 22 . If these findings may be translated to pig, SUL-121 and SUL-150 therefore antagonize the most predominant α 1 -AR subtypes in porcine intrarenal arteries. Finally, calcium signalling mediated by histamine and U46619 in HeLa cells was unaffected by SUL-150, confirming our findings in renal arteries.
Previously, our group demonstrated that upon stimulation with PE, α 1 -AR transactivate the EGF receptors leading to a stronger contraction of rat aorta 9,10 . Interestingly, the EGFR inhibitor AG1478 caused a competitive inhibition of PE mediated contraction in aorta, similar to the competitive inhibition observed for SUL-150 in porcine intra-renal artery. We therefore investigated the potential involvement of SUL-150 in EGFR transactivation. In porcine intra-renal arteries, AG1478 also caused an inhibition of PE mediated contraction. However, the inhibitory effects of AG1478 were not strictly competitive as no further inhibition could be observed at concentrations higher than 20 µM. Nevertheless, in the presence of a maximal inhibitory concentration of AG1478, SUL150 still caused a further inhibition of PE mediated contraction, indicating that the effects of SUL-150 are independent of EGFRs. The lack of effects of SUL-150 on EGFR was confirmed in HEK293 cells where EGF signalling was found to be unaffected by SUL-150. We therefore conclude that the inhibitory effects of SUL-150 on PE mediated contraction of porcine intra-renal arteries are not mediated through EGFR. The direct interaction of SUL-150 with the α 1 -AR was demonstrated in a radio ligand binding assay where SUL-150 displaced the α 1 -AR antagonist prazosin at effective concentrations similar as in the vascular and calcium signalling experiments. These findings suggest that SUL-150 and prazosin compete for the same binding site on the α 1 -AR. These findings were further corroborated by our in silico modelling of ligand-receptor interactions. Although our simulation was based on a homology model of the α 1 -AR rather than crystal structure, the estimated free binding energy and involvement of protein residues corresponded with experimentally determined displacement capabilities of SUL-150.
In conclusion, our study demonstrates that antagonism of vascular α 1 -AR can be added to the molecular mechanisms of action of SUL-121 and specifically to its (R)-enantiomer SUL-150. This makes SUL-150 unique given that, as of yet, no other SUL compounds demonstrated enantiomer specificity and binding to any vascular receptor. Such new feature is an addition to previously reported effects shared with the rest of the SUL family, including the prevention of ROS formation, maintenance of ATP production and direct ROS scavenging. Interestingly, the criteria proposed for the prevention of ischaemia-reperfusion injury in graft kidneys include the use of vasodilators, antioxidants and anti-inflammatory agents 3 . SUL-150 suits all these criteria and this multi-potency is a distinct advantage over standard antihypertensive agents used to improve graft perfusion and to prevent delayed graft function.
Limitations of the current study include the relatively high concentrations of SUL-150 required to achieve α 1 -AR antagonism. Furthermore, these concentrations are higher than reported for effectively protecting cells against hypothermia 6 . Although we are currently investigating the toxicological parameters of SUL compounds, present data indicate no negative effects of SUL compounds on cell viability at concentrations up to 10 mM 5,6 . Strengths of this study include the use of porcine kidneys which are comparable in size and physiology to human kidneys 23 . As a consequence, our experimental setup is similar in scale to human kidneys, indicating that sufficiently high concentrations of SUL-150 can be reached to improve graft perfusion in clinical machine perfusion setups.

Methods
Isolated kidney perfusion. Porcine kidneys from female Dutch Landrace pigs were collected at a local slaughterhouse (Kroon Vlees, Groningen, The Netherlands). Both kidneys were excised together with a section of aorta, ureter and bladder. The aorta was cut open longitudinally and a catheter with a thickened section of silicone was inserted into the renal artery. Using a 50 ml syringe, the kidney was subsequently slowly perfused with 150 ml of cold physiological saline solution supplemented with heparin (5 U/mL). After flushing, 50 ml of cold UW with heparin (5 U/mL) was slowly infused. Subsequently, the same procedure was immediately performed on the contralateral kidney. Next, kidneys were dissected free from aorta and bladder and individually stored in plastic bags on ice for 24 h.
The following day, a single kidney was attached to an in-house developed isolated kidney perfusion system and first flushed with approximately 300 ml of warm saline solution with heparin (5 U/mL). After flushing, kidneys were perfused with warm (37 °C) and oxygenated medium (RPMI 1640) enriched with heparin (5 U/mL) and bovine serum albumin (50 g/L). When medium was detected in the venous effluent of the kidney, perfusion was switched to recirculation with a circulatory volume of approximately 250 ml. The kidneys were allowed to warm by perfusion only, while measuring flow and temperature at a fixed flow pressure of 80 mmHg. Once the renal temperature reached 20 °C, flow was set to constant and the effects of 50 µM SUL-121, SUL-150 and SUL-151 on kidney perfusion pressure were recorded. In separate experiments, pressure was set to a constant (80 mmHg) and flow and temperature were measured before and after addition of 50 µM SUL-150. To minimize the amount of compound needed, the circulatory volume was reduced to 100 ml just before addition of vasoactive compounds to the perfusion medium.
Tissue preparation and contraction studies in isolated porcine renal artery rings. In an additional series of experiments, the effects of SUL-121, SUL-150 and SUL-151 were assessed in isolated porcine renal artery rings. For this, porcine kidneys were obtained from the local slaughterhouse and transported in normal physiological Krebs buffer (pH 7.4) on ice. The renal artery tree was dissected from the kidney, cleaned of surrounding tissue and cut into equally-sized ring segments (2 mm in length). In some rings, endothelium denudation was performed by gentle rubbing of the intimal surface with a paper clip. Rings were mounted in organ baths for measurement of isotonic displacement as described previously 24 . Arterial rings were washed thoroughly by replacing Krebs buffer and allowed to equilibrate for a period of 60 min under 1.4 g of resting tension before they were assessed for viability by inducing 2 subsequent constrictions with potassium chloride (KCl 60 mM). Rings that failed to produce a threshold displacement of 100 µm were excluded. After washout and stabilization, rings were treated for 30 minutes by incubation with either vehicle (0.1% DMSO), SUL-121, SUL-150 or SUL-151, followed by subsequent stimulation with cumulative doses of phenylephrine (PE, 10 nM -100 µM). Similarly, following pre-incubation with SUL-150, isotonic constriction to methoxamine (10 nM -100 µM), the thromboxane A 2 analogue U46619 (0.3 nM -10 µM) and histamine (30 nM -1 mM) was recorded in additional Radioligand binding assay. Intact CHO cells stably overexpressing the α 1A -AR were cultured to confluence in 96-well plates. Prior to the experiment, cells were washed and the culture medium was replaced by HBSS. Competition curves were built by simultaneous incubation with [7-Methoxy-3 H]-prazosin (100 nM) and concentrations of SUL-121 enantiomers in the range of 0.5µM-100µM. To calculate specific binding, control reactions were performed using a high dose of unlabelled prazosin (10 µM). Incubations were allowed to proceed for 1 hour at 37 °C. Subsequently, cells were placed on ice, washed three times with ice-cold PBS and lysed by adding 50 µl of 1 N sodium hydroxide (NaOH). Finally, the lysate was transferred to a scintillation vial and counted for 3 minutes.
Induced fit molecular docking simulation. The primary sequence of human α 1A -AR was obtained from UniProt database 25 using reference code P35348 and uploaded to SWISS-MODEL in order to build a homology model 26,27 , resulting in 373 templates. Subsequently, template ligand codes were used to query the PDB database 28 to obtain structural data in SMILES format, which were processed by Chemmine 29 to detect similarities with our compound of interest (SUL-121). The SWISS-MODEL template that contained a ligand with the highest similarity score (a D3 dopamine receptor in complex with, Eticlopride, ETQ) was used to align the α 1A -AR sequence onto a modelled backbone. The resulting homology model was validated by Ramachandran plot 30 and prepared with Protein Preparation Wizard 31 by the addition of hydrogens, bond order assignment, generation of partial charges to heteroatoms 32 and disulphide bonds. Final refinement was performed by hydrogen bond assignment at pH 7.4 and restrained minimization at 0.3 root-mean-square deviation of atomic positions (RMSD). Prazosin, SUL-150 and SUL-151 structural files were converted from SMILES to 3D structures using LigPrep 31,33 . Protonation states were generated with Epik at pH 7.4 and small molecule energy parameters were computed using the OPLS3 Force Field 34 .
SCientifiC RepoRTs | (2019) 9:13 | DOI:10.1038/s41598-018-36788-0 Flexible molecular docking simulation was performed using Induced Fit, a part of the Schrödinger Small-Molecule Drug Discovery suite 35 . A binding centroid was defined between residues involved in antagonist binding confirmed by mutagenesis (PHE312, PHE308 and ASP106) and ligands were docked within 15 Å, using an extended sampling protocol without constrains 12,15 . Residues within 5 Å of resulting ligand poses were refined using Prime to improve ligand conformational sampling 36 . Finally, the Scorpion server was used for the assessment and classification of small molecule-protein interactions 37 and the final results were rendered using PyMOL 2.0 38 .
Data and statistical analysis. The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology 39 . Values in figures were displayed as mean ± SEM of n independent experiments. Vascular constriction responses were expressed as percentage of final response to KCl to account for differences in contractile capability of individual vascular rings. The pEC 50 values were calculated using a four-parameter logistic regression in Prism 7 (GraphPad Software, Inc., La Jolla, CA) and compared using the Extra sum-of-squares F test with p < 0.05 considered statistically significant. Schild analysis was performed using Gaddum/Schild EC 50 shift in Prism 7 and pA 2 values compared using an unpaired parametric two-tailed t-test with Welch's correction. Fluorescent measurement data was processed in SoftMax Pro 7 (Molecular Devices LLC, Sunnyvale, CA) and expressed as % of baseline AUC with a 3-fold multiplier using a mean of first 10 measurement points as baseline. Western blot relative protein expression was quantified using Fiji, a distribution of ImageJ 40 .