Aptamer generation Proteins used in binding experiments

Transcription

1 Aptamer generation A DNA template was synthesized with the sequence 5 -GGAGGGAAAAGTTATCAGGC-N4- GATTAGTTTTGGAGTACTCGCTCC-3, where N4 represents a 4-nucleotide sequence in which there is an equal probability of incorporating a da, dc, dg, or dt residue at each position and d denotes a 2 -H residue. The DNA template was amplified by polymerase chain reaction (PCR) with forward primer 5 -GACTGTAATACGACTCACTATAGGAGGGAAAAGTTATC- AGGC-3 and reverse primer 5 -GGAGCGAGTACTCCAAAACTAATC-3 and subsequently transcribed to generate a starting pool of approximately 1 14 different sequences comprised of dc, ma, mg, and mu residues, where m denotes a 2 -OCH 3 residue. Eleven rounds of selection were carried out by first incubating the pool of molecules with recombinant full-length TFPI (American Diagnostica, Inc.) at 37 C for 1 hr in Dulbecco s phosphate-buffered saline with Ca 2+ and Mg 2+ (DPBS; Mediatech) in a 1-μL reaction volume. The binding reaction was then passed over a nitrocellulose filter to capture protein and proteinaptamer complexes. The filter was washed two or four times with 1 ml of DPBS and bound aptamer was eluted two times with 2 μl of elution buffer (7M Urea, 1mM NaOAc, 3 mm EDTA, ph 5.7) at 95 C. The eluted aptamers were ethanol precipitated, reverse transcribed, PCR amplified, and subsequently forward transcribed as above to give rise to a pool for use in the next round of selection. Over the course of the selection, the selection stringency was increased by gradually decreasing the concentration of TFPI from nm to nm, through the addition of trna (Ambion) and salmon sperm DNA (Invitrogen), and increasing the number and length of the washes after the binding step. The concentration of the pool was 1.66 μm in round 1 and 1 μm for all subsequent rounds. The round 11 pool was cloned and sequenced. Individual clones were generated by chemical synthesis on a Mermade 192 RNA/DNA synthesizer (Bioautomation) using standard phosphoramidite chemistry. Clones were tested for binding to recombinant TFPI with a nitrocellulose dot blot binding assay and for inhibition of TFPI in the calibrated automated thrombogram (CAT) assay, as described in the Methods. From these experiments, the parent clone (5 -mgmgmamgmgmgmamamamamgmumumamudcmamgmgdcdcmumgmamamumumumgmgmamamumamumadcmumumgmgdcmudcmgmumumamgmgmumgdcmgmumamumamumamgmamumumamgmumumumumgmgmamgmumadcmudcmgdcmudcdc-3 ) was determined to bind to TFPI with nanomolar affinity and inhibit its activity in plasma at nanomolar concentrations. The core aptamer motif, ARC1748, was identified by design of molecules that contained a portion of the parent clone sequence and evaluation in the same assays. The core aptamer was synthesized with a hexylamine linker at the 5 -end which was conjugated postsynthetically to a branched 4 kda PEG moiety to give rise to ARC Proteins used in binding experiments Recombinant TFPI and human plasmin were purchased from American Diagnostica, Inc. Factor Va, factor XII, antithrombin III, heparin cofactor II, prothrombin, α-thrombin, factor VIIa, factor XIa, and factor Xa were from Haematologic Technologies, Inc. Kallikrein was purchased from Enzyme Research Laboratories. Serpin A1 and TFPI-2 were from R&D Systems.

2 Nonhuman primate model Anesthesia Animals were sedated using 1 mg/kg ketamine via intramuscular administration. For induction, animals were masked with isoflurane at 2 3% with an oxygen delivery rate of 2 L/min. Atropine at.4 mg/kg was administered intramuscularly. The animals were intubated with a mm Ruschelit Safety clear tracheal tube (Teleflex Medical) and maintained in surgical anesthesia with inhalant isoflurane, %, with an oxygen delivery rate of 2 L/min. Anesthetized animals were placed on circulating hot water blankets to aid in maintaining normal body temperature throughout the study. At the end of the study, buprenorphine was administered at.5 mg/kg for pain control and 3 35 ml of warm saline was injected via intravenous (IV) administration. The animals were removed from the anesthesia machine and allowed to partially recover. The endotracheal tube was removed and the animals were placed in their cages for monitoring until full recovery. Dosing and sampling schemes Sheep polyclonal anti-human FVIII antibody was administered to monkeys as a single slow IV bolus via a 22-gauge catheter in the saphenous vein. Each monkey received ~, Bethesda units of anti-fviii antibody, corresponding to 1 ml of 16 2 mg/ml antibody, depending on the lot used. Test substances (either saline or ARC19499) were administered as a single slow IV bolus via the same catheter 3.5 hours after antibody administration. Blood samples were drawn from the saphenous vein prior to antibody administration to determine baseline measurements. Subsequent samples were drawn from the femoral vein post antibody administration and post test substance administration. In initial studies, surgically-implanted ports were used occasionally for drug administration or blood sampling; however, most blood samples were drawn as follows. A 4. French introducer sheath was placed in the femoral vein, and ml of saline was flushed through the sheath. Approximately 3 4 ml of blood mixed with the saline was withdrawn and set aside. An undiluted blood sample was taken using a 3 or 5 ml syringe. The blood was injected into citrate-containing tubes, inverted repeatedly, and processed to obtain plasma. Plasma processing Citrated whole blood samples were kept at room temperature until plasma processing. Within 1 minutes after collection, the samples were centrifuged at 2, g for 15 minutes at room temperature. The plasma was removed and immediately stored at 8 C. Prior to analysis, plasma samples were thawed rapidly at 37 C. Bleeding model For the bleeding model, monkeys were anesthetized as described above. In addition, blood pressure was monitored during the study by placing a 5. French introducer line in the femoral artery of the monkey. Optimally, the femoral artery used was on the opposite side from the saphenous vein that was used for bleeding time assessment. Blood pressure was then monitored using a Datascope Passport instrument (Datascope). Blood pressure readings were taken 1 5 minutes before and after bleeding time assessment. Sheep polyclonal anti-human FVIII antibody was administered to the saphenous vein that was not being used for bleeding time assessment, as described above. Each monkey received 12,642 Bethesda units/kg of FVIII antibody. ARC19499

3 (1 mg/kg) was administered via IV as a single slow bolus dose by the same catheter. Prior to use,.5 1. ml of fluid was removed from the catheter. Next, the FVIII antibody or ARC19499 was administered and then the catheter was flushed with 2 3 ml warm saline. Determination of ARC19499 concentration in plasma samples from in vivo studies The concentration of ARC19499 in monkey plasma samples was determined via high performance liquid chromatography with ultraviolet absorbance detection (HPLC-UV). The instrumental conditions are found in the table below. µl of plasma was mixed with µl of digestion buffer (6 mm Tris-HCl, ph 8., 1 mm EDTA and.5% SDS) and µl of proteinase solution (1. mg/ml proteinase K in 1 mm Tris-HCl, ph 7.5, 2 mm CaCl 2, 1% glycerol v/v), and the samples were incubated overnight, shaking, at 55 C. Following incubation, samples were centrifuged (14, rpm; 4 C; 15 minutes), and 1 µl of each supernatant was transferred to an HPLC injection vial. The assay injection volume was approximately µl. The lower limit of quantification was.2 µg/ml with a linear concentration range of.2 to µg/ml. The HPLC method was calibrated relative to concentration reference standards of ARC19499 prepared in monkey plasma and extracted by the same proteinase method used to prepare the in vivo samples.

4 Table S1. Instrumental conditions for HPLC-UV Equipment: HPLC system equipped with an autosampler, columntemperature-controller and UV detector Column: Dionex DNAPAK PA-1 (4 2 mm) with Guard (4 mm) Column Temperature: 8 C Flow Rate:.9 ml/min Injection Volume: µl UV Detector: 26 nm Run Time: 3 minutes Retention Time: Approximately 15.8 min A: % mm sodium phosphate buffer (ph 7.) and % Acetonitrile. Mobile Phase: B: % mm sodium phosphate buffer (ph 7.) and % Gradient Table: Acetonitrile containing 4 mm NaClO 4 Time %A %B 1 1

5 A 2 Thrombin (nm) anti-tfpi antibody treated plasma Time (min) B Lag Time (min) ETP (nm) Peak Thrombin (nm) Figure S1. ARC19499 has no activity in human plasma treated with an anti-tfpi antibody (A) Representative thrombin generation curves in fresh-frozen PNP after treatment with 1 µm inhibitory polyclonal anti-tfpi antibody initiated with 1 pm TF. The dark line represents untreated PNP. The remaining lines represent antibody-treated plasma in the absence and presence of different concentrations of ARC (B) Effect of ARC19499 on thrombin generation with 1 pm ( ) or.1 pm ( ) TF. Left panel, Lag time; middle panel, ETP; right panel, Peak thrombin. The solid and dashed lines correspond to plasma in the absence of ARC19499 with 1. or.1 pm TF, respectively.