Development of supramolecular anticoagulants with on-demand reversibility

General methods

Unless otherwise specified, all reagents and solvents for all organic synthesis procedures were purchased from commercial sources and were used without further purification. High-performance liquid chromatography (HPLC) purification was performed with an Agilent Technologies 1260 Infinity HPLC using a ZORBAX 300SB-C18 column (9.4 × 250 mm). LC–mass spectrometry (LC–MS) spectra were recorded on a DIONEX Ultimate 3000 UHPLC with a Thermo LCQ Fleet Mass Spectrometer System using a PINNACLE DB C18 column (1.9 µm, 50 × 2.1 mm) operated in positive mode. All the LC–MS spectra were measured by electrospray ionization. Matrix-assisted laser desorption/ionization–time of flight (MALDI–TOF) mass spectra were measured using a Bruker Daltonics Autoflex spectrometer operated in positive mode. High-resolution mass spectra were obtained on a Xevo G2 TOF spectrometer (ionization mode, electrospray ionization positive polarity; mobile phase, methanol at 100 µl min–1). Automated solid-phase synthesis was performed on an Intavis AG Multipep RS instrument.

Synthesis of PNA–peptide conjugates

Resin (5.0 mg) was swollen in dichloromethane (DCM) for 10 min and washed twice with dimethylformamide (DMF). Iterative cycles of amide coupling (procedure 1), capping of the resin (procedure 4) and deprotection of the protecting group (procedure 2 or procedure 3) were performed to synthesize the PNA probes. The compounds were deprotected and cleaved from the resin using procedure 5 and finally purified by HPLC. Characterization of the PNA–peptide conjugates was performed using MALDI (Bruker Daltonics Autoflex spectrometer with Flex control 3.4 software and analysis with FlexAnalysis 3.4) and/or LC–MS (DIONEX Ultimate 3000 UHPLC with a Thermo LCQ Fleet Mass Spectrometer System using a PINNACLE DB C18 column (1.9 µm, 50 × 2.1 mm) with Thermo Xcalibur 2.2.SP1.48 software and analysis with Thermo Xcalibur Qual Browser 2.2.Sp1.48). For MALDI analysis, 1.0 µl of the sample (in either water or water/acetonitrile (1:1)) was mixed with 1.0 µl of 2,5-dihydroxybenzoic acid (DHB) matrix solution (30 mg of DHB in 1.0 ml of 70:30:0.01 water/acetonitrile/trifluoroacetic acid (TFA)), and the mixture was spotted on a MALDI plate. The measurements were acquired in positive linear mode. For LC–MS analysis, 20 µl of sample in water or water/acetonitrile (1:1) was injected on the LC and further analyzed by MS in positive mode. Compounds containing the benzene disulfonic acid motif could only be analyzed by LC–MS due to fragmentation when analyzed by MALDI.

2-Chlorotrityl chloride resin loading

2-Chlorotrityl chloride resin (1.46 mmol g–1 loading) was swollen in dry DCM for 30 min, followed by washing with DCM + 1% N,N-diisopropylethylamine (DIPEA; 3 ml, one time) and DCM (3 ml, ten times). A solution of Fmoc-Xaa-OH (0.7 mmol g–1 resin) and DIPEA (4 equiv. relative to resin functionalization) in DCM (final concentration of 0.125 M amino acid) was added to the resin, which was shaken at room temperature for 16 h. The resin was then washed with DCM (3 ml, five times), DMF (3 ml, five times) and DCM (3 ml, five times). The resin was then capped via treatment with 17:2:1 (vol/vol/vol) DCM/methanol/DIPEA (5 ml) for 40 min at room temperature. The resin was then washed again with DCM (3 ml, five times), DMF (3 ml, five times) and DCM (3 ml, five times) before further use.

Rink amide resin loading

Nova PEG Rink amide resin (0.44 mmol g–1; Novabiochem) was swollen in DCM for 10 min and washed twice with DMF. Standard amide coupling (procedure 1) was performed, followed by capping of the resin (procedure 4). The resin was then washed again with DCM (3 ml, five times), DMF (3 ml, five times) and DCM (3 ml, five times) before further use.

Amide coupling (procedure 1)

The corresponding Fmoc-protected PNA monomer40 or amino acid (4 equiv., 0.2 M in N-methylpyrrolidone (NMP)) was incubated for 5 min with HATU (3.5 equiv., 0.5 M in NMP) and base solution (1.2 M (4 equiv.) DIPEA and 1.8 M (6.0 equiv.) 2,6-lutidine in NMP). The mixture was then added to the corresponding resin. After 20 min, the mixture was filtered, the resin was washed with DMF, and a new premixed reaction solution was added to the resin and allowed to react for another 20 min. The resin was then washed sequentially with DMF, DCM and DMF two times each.

Fmoc deprotection (procedure 2)

A solution of 20% (vol/vol) piperidine in DMF was added to the resin and allowed to react for 5 min. The mixture was then filtered, the resin was washed with DMF, and the sequence was repeated for another 5 min. The resin was then washed sequentially with DMF, DCM and DMF two times each.

4-Methyltrityl deprotection (procedure 3)

A solution (made from 244 mg of hydroxybenzotriazole in 10 ml of hexafluoroisopropanol and 10 ml of 1,2-dichloroethane) was added to the prewashed resin to reach a volume of 10 ml g–1 of resin and allowed to react for 5 min. The solution was flushed, the resin was washed with DCM, and the sequence was repeated for another 5 min. Finally, the resin was washed sequentially with DCM and DMF two times each.

Capping (procedure 4)

The resin was treated with a capping mixture (0.92 ml of acetic anhydride and 1.3 ml of 2,6-lutidine in 18 ml of DMF; 10 ml of solution per g of resin) for 5 min. After flushing the solution, the resin was washed sequentially with DMF, DCM and DMF two times each.

Cleavage from the resin and final deprotection (procedure 5)

Resin (5.0 mg, 1.0 μmol) was treated with 125 μl of a mixture of TFA and scavengers (440 µl of TFA + 25 mg of phenol + 25 µl of water + 10 µl of triisopropylsilane) for 2 h. The resin was filtered and washed with TFA (50 μl), and the collected fractions of cleavage product were precipitated in cold ether (1.5 ml). After centrifugation, the pellet was vortexed again with cold diethyl ether (1.5 ml) and centrifuged (18,000g). The pellet was dissolved in water/acetonitrile (3:1; 1.5 ml) and lyophilized to obtain a white powder.

Microcleavage for quality control (procedure 6)

The minimum number of beads was picked with a pipette plastic tip and transferred to 50 µl of TFA. The solution was left for 1 h and transferred to 1.0 ml of ether. The ether solution was maintained at −20 °C for 5 min and centrifuged for 5 min at 18,000g. The ether supernatant was removed, and the pellet was dissolved in 20 µl of 1:1 acetonitrile/water, which was then used for analysis by MALDI–TOF and/or LC–MS.

On-resin copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC; procedure 7)

A solution of CuSO4 (15 µl, 64 mg ml–1 in water) was added to tris(benzyltriazolylmethyl)amine (2 mg) in 20 µl of DMF, followed by the addition of 50 µl of sodium ascorbate (396 mg ml–1 in water). Azide-containing peptide (2 equiv. in 60 µl of DMF) was added to the mixture, which was mixed before the addition of 5 mg of alkyne-derivatized Rink amide resin (0.0022 mmol). After 16 h of shaking, the mixture was filtered, and the resin was washed six times with 250 µl of sodium diethyl dithiocarbamate (0.02 M) in DMF, six times in 250 µl of DMF, six times in methanol and six times in DCM.

Coupling of Fmoc-l-F2Smp(nP)-OH (procedure 8)

Fmoc-l-F2Smp(nP)-OH was prepared as previously described22. A mixture of Fmoc-l-F2Smp(nP)-OH (0.003 mmol, 1.5 equiv.), hydroxybenzotriazole (0.003 mmol, 1.5 equiv.) and N,N′-diisopropylcarbodiimide (0.003 mmol, 1.5 equiv.) was added to the corresponding resin and shaken overnight. The mixture was filtered, and the resin was washed sequentially with DMF, DCM and DMF two times each.

Coupling of Arg(Pbf)-benzothiazole (procedure 9)

Arg(Pbf)-benzothiazole (0.0044 mmol, 2 equiv.) and HATU (0.0034 mmol, 1.5 equiv.) in NMP (100 μl) were added to 5 mg of resin (0.0022 mmol), followed by the addition of DIPEA (0.012 mmol, 6 equiv.). The reaction was shaken for 2 h, the mixture was filtered, and the resin was washed sequentially with DMF, DCM and DMF two times each.

Neopentyl deprotection and characterization of PNA–peptide conjugates (procedure 10)

The precipitate collected after cleavage and ether precipitation was lyophilized. The remaining solid was dissolved in a solution of 1 M ammonium acetate and 6 M guanidinium chloride and shaken at 37 °C for 2 h. The solution was then diluted with water/acetonitrile (50:50) and purified by HPLC.

Thrombin inhibition assay

Inhibition of the activity of human α-thrombin (Haematologic Technologies, HCT0020) was followed spectrophotometrically using Phe-Pro-Arg-Coumarin (synthesis described in Supplementary Information) as the chromogenic substrate.

Inhibition assays were performed using 0.2 nM enzyme, 20 μM substrate and increasing concentrations of inhibitor. The concentration of each inhibitor variant was determined using the absorbance of the PNA at 260 nm, as measured by NanoDrop. All reactions were performed at 37 °C in 50 mM Tris-HCl (pH 8.0), 50 mM NaCl and 1 mg ml–1 bovine serum albumin in black 96-well microtiter plates (Thermo Fisher Scientific, 267342). Reaction progress was monitored by excitation at 339 nm and emission at 439 nm using a SpectraMax or Tecan Spark Plate Reader. Dose–response curves were used to determine the half-maximal inhibitory concentrations (IC50) using Prism 8.0 (GraphPad Software). For each inhibitor, the reactions were performed in triplicate, together with control reactions in the absence of enzyme. The initial velocity was calculated from the slope of the first 10 min of the assay. The curves were normalized to the well without inhibitor, where the initial velocity was set to 100% activity.

For the antidote assay, the plate was removed from the plate reader at the desired time of addition (usually 30 min). One microliter of antidote (100×) was added, and reading was resumed.

Fibrinogen assay

Human α-thrombin (Haematologic Technologies, HCT0020; final concentration of 2.5 nM) was incubated with compound (final concentration of 15 nM) at 37 °C for 30 min. Fibrinogen (final concentration of 1 mg ml–1) was added, and absorbance at 288 nm was measured using a SpectraMax Plate Reader. All reactions were performed at 37 °C in 50 mM Tris-HCl (pH 8.0), 50 mM NaCl and 1 mg ml–1 bovine serum albumin in clear 96-well microtiter plates (Greiner Bio-One, 650201).

For the antidote assay, the plate was removed from the plate reader at the desired time of addition (usually 30 min). One microliter of antidote (100×) was added, and reading was resumed.

Selectivity assays

Inhibitory activity of A1–E1 was tested against α-human thrombin, FXIa and FXa (Haematologic Technologies) and α-FXIIa and PK (Enzyme Research Laboratories). Chromogenic assays were followed spectrophotometrically using the following specific substrates: 100 μM Tos-Gly-Pro-Arg-PNA (Chromozym TH, Roche) for thrombin, 500 µM Pyr-Pro-Arg-PNA (L-2145, Bachem) for FXIa, 500 µM Moc-d-norleucine-Gly-Arg-PNA (L-1565, Bachem) for FXa and 200 µM or 400 µM d-Pro-Phe-Arg-PNA (Cayman Chemical) for α-FXIIa or PK, respectively. The assay buffers included 50 mM Tris-HCl (pH 8.0) and 50 mM NaCl for thrombin (0.2 nM); PBS (pH 7.4) for FXIa (0.5 nM); 25 mM Tris-HCl (pH 7.5), 100 mM NaCl and 5 mM CaCl2 for FXa (0.5 nM); 20 mM HEPES (pH 7.6), 150 mM NaCl, 0.1% (wt/vol) PEG 8000 and 0.01% (vol/vol) Triton X-100 for α-FXIIa (4 nM); and 50 mM Tris-HCl (pH 8.0) and 150 mM NaCl for PK (0.25 nM). Bovine serum albumin (Sigma) was added to all buffers at 1 g l–1. All reactions were initiated by the addition of the protease and were performed at 37 °C in 96-well, flat-bottom microtiter plates. Reaction progress was monitored at 405 nm for 30 min (60 min for FXa and α-FXIIa) on a multimode microplate reader (Synergy2, BioTek) with measurements taken every 5 min. All measurements were performed in duplicate. IC50 values were determined from the log dose–response curves with Prism 9 (GraphPad Software).

SPR experiments

SPR experiments were performed on a Biacore T200 instrument (GE Healthcare) at 25 °C in PBS-P+ buffer (10× stock; Cytiva Life Sciences, 28995084). Biotin-PNA (8-mer) was immobilized on a Streptavidin Series S sensor chip (Cytiva Life Sciences, 29104992). Before immobilization, the two channels were conditioned with 1 M NaCl in 50 mM NaOH. After stabilization, the compound (solution in PBS-P+) was flowed over one of the flow cells of the sensor chip at a concentration of 50 nM at a flow rate of 10 μl min−1 with a response unit target of 500. Biotin-PNA (8-mer) reached a response unit value of 513.7. The system (not including the flow cells) was washed with 50% isopropanol in 1 M NaCl and 50 mM NaOH after each ligand injection. Kinetic measurements consisted of injections (association, 400 s; dissociation, 450 s; flow rate, 30 μl min−1) of decreasing concentrations of PNA (4-, 6- and 8-mer; twofold cascade dilutions from the starting concentration). The chip was regenerated between cycles by one injection of regeneration solution (50 mM NaOH) for 10 s at a flow rate of 20 μl min−1, followed by a 10-s stabilization period. Binding was measured as resonance units over time after blank subtraction, and the data were interpreted using Biacore T200 software (version 3.2). All measurements were performed in duplicate. The KD values were calculated based on steady-state affinity (1:1 binding).

aPTT in vitro

aPTT measurements were performed on a BFT II benchtop analyzer as per the manufacturer’s instructions. Dade Actin FSL Activated PTT Reagent (23-044-647) and calcium chloride solution (10446232 ORHO37) were both sourced from Siemens Healthcare Diagnostics Products, and lyophilized pooled human reference plasma (Pooled Norm., 00539) was purchased from Diagnostica Stago. Pooled human plasma was reconstituted as per the manufacturer’s instructions (Milli-Q water, 30 min, room temperature). Pooled mouse plasma was prepared by collection of whole blood from three to four C57BL/6 mice (Australian BioResources) into sodium citrate (3.8%), with plasma isolated by centrifugation at 5,000g for 15 min and stored on ice until required.

Human or mouse plasma was incubated with inhibitors at the indicated concentrations and prewarmed to 37 °C. Fifty microliters of each plasma/inhibitor mixture was incubated with Actin FSL (50 ml) in a stirred reaction vessel for 3 min before addition of 50 ml of calcium chloride solution to initiate coagulation. The time taken for fibrin clot formation was recorded in a semiautomated fashion using a BFT II Analyzer, which uses a turbodensitometric detection technique.

Ex vivo aPTT

All procedures involving the use of animals were performed as approved by the University of Sydney Animal Ethics Committee (protocol 2021/1912). C57BL/6 mice (25–30 g) were anesthetized using a mixture of ketamine (125 mg per kg (body weight)) and xylazine (12.5 mg per kg (body weight); intraperitoneal delivery) and administered A1–E1 as a single bolus delivered intravenously via the femoral vein at either 2.5 or 5.0 mg per kg (body weight). Blood was drawn from the inferior vena cava at the indicated times into citrate anticoagulant (3.8%), plasma was isolated as described above for in vitro aPTT studies, and aPTT was assessed via changes in plasma opacity at 405 nm using a CLARIOstar plate reader fitted with dual injectors heated to 37 °C using a modified version of the aPTT protocol described above. Briefly, injectors were primed for Dade Actin FSL Activated PTT Reagent (line A) and calcium chloride solution (line B), and mouse plasma was aliquoted in duplicate (25 μl) into wells of a Nunc 368-well polystyrene plate (Z723010, Sigma-Aldrich). Following injection of 25 ml of Dade Actin FSL, the plate was mixed using the orbital shaking function for 2 s (500 rpm) and incubated for 182 s at 37 °C. At this time (designated t = 0 s), 25 ml of calcium chloride solution was injected, the plate was mixed as described above, and absorbance measurements were taken at 405 nm for 360 intervals (22 flashes per well, interval time of 0.5 s). Clotting time was denoted by the timing of initial inflection point, denoting transition of plasma from transparent to opaque.

CAT

Normal lyophilized human pooled plasma (Pool Norm., 00539, Diagnostica Stago) was reconstituted and incubated for 30 min at 37 °C. Vehicle and various inhibitors at different concentrations were then incubated in plasma for 30 min. Thrombin assays were performed using a Hemker Calibrated Automated Thrombinoscope (Diagnostica Stago) and a Fluoroskan Ascent plate reader (Thermo Fisher Scientific). All experiments were conducted in triplicate in 96-well microplates for fluorescence-based assays (M33089, Thermo Fisher Scientific) and calibrated using untreated plasma and a thrombin calibrator (86192, Diagnostica Stago). Thrombinoscope experiments were conducted following patented commercial protocols. In brief, each sample well was filled with 20 μl of PPP reagent containing a mixture of phospholipids and tissue factor (86193, Diagnostica Stago). Eighty microliters of plasma (untreated/ treated) was then added to each of these wells and mixed using reverse pipetting, and the well plate was incubated in the plate reader at 37 °C for 10 min. Meanwhile, a FluCa kit (86197, Diagnostica Stago) containing Fluo-Buffer and Fluo-Substrate was warmed to 37 °C. Following incubation, the thrombinoscope dispenser was flushed, emptied and filled with a FluCa mixture consisting of the Fluo-Buffer and Fluo-Substrate. Twenty microliters of the FluCa mixture was dispensed into each well containing plasma samples, initiating the coagulation reaction. Thrombin activity (nM) was measured over 1 h, with thrombogram parameters including lag time (min), velocity index (nM min–1), time to peak (min), peak height (nM), endogenous thrombin potential (nM × min) and time to tail (min).

Needle injury thrombosis model

C57BL/6J mice were purchased from Australian BioResources and housed at the Laboratory Animal Services facility (University of Sydney). All animals were maintained on a 12-h light/12-h dark cycle with access to food and water ad libitum. For intravital mouse studies, male mice aged between 5 and 8 weeks old (15–20 g) were used. All studies were approved by the University of Sydney Animal Ethics Committee (protocol 2021/1912) in accordance with the requirements of the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes41.

A clinical preparation of argatroban (Argatra/Exembol) was purchased from Mitsubishi Tanabe Pharma (Germany) and prepared in sterile saline with 25% (vol/vol) propylene glycol. Synthesized PNA inhibitors and PNA inhibitors + antidote solutions were prepared in sterile saline at a concentration of 2 mg ml–1. C57BL/6J male mice (15–25 g) were anesthetized with ketamine (150 mg per kg (body weight)) and xylazine (15 mg per kg (body weight)), supplemented with oxygen and subjected to intravital needle injury, as previously described42. Systemic injection of DyLight 649-conjugated anti-GP1bβ (X649, Emfret; 100 µg kg–1) and Alexa 546-conjugated anti-fibrin (0.31 mg per kg (body weight)) was performed before vessel injury to monitor thrombus formation and fibrin generation, respectively. Argatroban (80 µg kg–1 bolus, 40 µg kg–1 min–1, 60-min infusion) was delivered via a jugular catheter using a Harvard apparatus pump (704504, Pump 11 Elite I/W Single Syringe Pump). Injections of PNA inhibitors or PNA inhibitors + antidote (5 mg per kg (body weight) bolus every 30 min) were delivered intravenously. Two to four successive injuries were created in multiple vessels in each mouse from each treatment group. Following each injury, platelet thrombus formation and fibrin generation were monitored over a 15-min period using a confocal intravital microscopy platform (Nikon A1R-Si with an Apo LWD, ×40/1.15-NA water immersion objective; sequential excitation: 488-, 561- and 638-nm lasers; emission: 525/50-, 595/50- and 700/75-nm filters) and NIS Elements Advanced Research acquisition software. The microscope stage and objective were maintained at 37 °C throughout the experiment via a Peltier heater (OkoLab). Surface renders of confocal stacks representing thrombi from separate groups were generated using Imaris (version 9.8, Bitplane).

Quantitative analysis of thrombus volume over time

NIS Elements software (version. 5.02; Nikon) was used to apply a threshold to DyLight 649-conjugated anti-GP1bβ signal for each xyz stack in a time series and was used to calculate the volume for each time point.

Quantitation of change in fibrin amount over time

The signal obtained from DyLight 649-conjugated anti-GP1bβ for each xyz stack in a time series was thresholded to create a mask. The total signal (arbitrary units) from Alexa Fluor 546-conjugated anti-fibrin within this mask (that is, the fibrin signal within the thrombus) for each time point was then quantified using NIS Elements software (Nikon).

Statistical analysis

Statistical significance between multiple treatment groups was analyzed using a one-way ANOVA with Tukey’s post hoc testing with a single pooled variance (Prism software version 10.2; GraphPad Software for Science). Data are presented as mean ± s.e.m., where ‘n’ equals the number of independent experiments performed.