Oxime coupling of active site inhibited factor seven with a non-volatile

Subscriber access provided by OPEN UNIV OF HONG KONG

Article

Oxime coupling of active site inhibited factor seven with a non-volatile, water soluble fluorine-18 labelled aldehyde Troels Elmer Jeppesen, Lotte Kellemann Kristensen, Carsten Haagen Nielsen, Lars Christian Petersen, Jesper Bøggild Kristensen, Carsten Behrens, Jacob Madsen, and Andreas Kjær Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.8b00900 • Publication Date (Web): 24 Jan 2019 Downloaded from http://pubs.acs.org on January 27, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Title Oxime coupling of active site inhibited factor seven with a non-volatile, water-soluble fluorine-18 labelled aldehyde

Authors Troels E. Jeppesen1, Lotte K. Kristensen1,2, Carsten H. Nielsen1,2, Lars C. Petersen3, Jesper B. Kristensen3, Carsten Behrens3, Jacob Madsen1, Andreas Kjaer1,* 1

Dept. of Clinical Physiology, Nuclear Medicine & PET and Cluster for Molecular Imaging, Dept. of Biomedical Sciences, Rigshospitalet and University of Copenhagen, DK-2100; 2Minerva Imaging ApS, DK-2200; and 3Novo Nordisk A/S, DK-2760 *Correspondence: Prof. Andreas Kjaer: [email protected]; phone: +45 2725 8614

Abstract A non-volatile fluorine-18 aldehyde prosthetic group was developed from [18F]SFB, and used for site-specific labelling of active site inhibited factor VII (FVIIai). FVIIai has a high affinity for tissue factor (TF), a transmembrane protein involved in angiogenesis, proliferation, cell migration and survival of cancer cells. A hydroxylamine N-glycan modified FVIIai (FVIIai-ONH2) was used for oxime coupling with the aldehyde [18F]2 under mild and optimized conditions in an isolated RCY of 4.7  0.9%, and a synthesis time of 267  5 min (from EOB). Retained binding and specificity of the resulting [18F]FVIIai to TF was shown in vitro. TF-expression imaging capability was evaluated by in vivo PET/CT imaging in a pancreatic human xenograft cancer mouse model. The conjugate showed exceptional stability in plasma (>95% at 4 hours) and a binding fraction of 90%. In vivo PET/CT imaging showed a mean tumor uptake of 3.8 ± 0.2% ID/g at 4 hours postinjection, a comparable uptake in liver and kidneys and low uptake in normal tissues. In conclusion, FVIIai was labelled with fluorine-18 at the N-glycan chain without affecting TF binding. In vitro specificity and a good imaging contrast at 4 hours post-injection was demonstrated.

Introduction Bioactive proteins usually show high affinity and selectivity for their endogenous targets, which can be utilized in nuclear medicine.1 In particular, fluorine-18 (t1/2 109.8 min, 97% +, 635 keV) is interesting for positron emission tomography (PET) with small proteins and peptides due to the favorable decay properties and matching biological and isotope half-life. Direct 18F-fluorination of proteins is problematic, as harsh reaction conditions such as strong bases and elevated temperatures are needed. Alternatively, prosthetic groups containing 18F (for example [18F]SFB), have been introduced2,3. These prosthetic groups have the ability to react under conditions compatible with proteins, typically in water at room temperature and around neutral pH.4 Random labelling of proteins and antibodies has been performed for more than a decade, resulting in the first antibody-drug conjugate receiving FDA approval in 2001.5 However, random labelling of proteins and antibodies has several shortcomings due to the heterogeneity of the products. Heterogeneous labelling may especially affect pharmacodynamic and pharmacokinetic properties.5

ACS Paragon Plus Environment

Bioconjugate Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 19

To overcome these challenges, site-specific conjugation to proteins and antibodies has gained increased interest in recent years5. One option is to modify the protein or antibody of interest selectively in one single position by genetic code expansion techniques and use the non-natural amino acid introduced this way in a biorthogonal reaction.6 This is not always applicable for larger and more complex proteins or antibodies. A different approach is glycan chain modification.7 However, for many proteins, conserving the terminal sialic acid in the glycan chain is of paramount importance for the biological half-life.8 Therefore, techniques such as Glycoconnect,7 that removes this sialic acid are not applicable to these proteins. Instead, site-specific modification of the terminal sialic acid provides the opportunity to introduce an hydroxylamine group.9 A bioorthogonal reaction that can be used for site-specific conjugation to proteins10 with a hydroxylamine modification is the oxime coupling. Oxime coupling is a type of reaction that is highly chemo-selective, proceeds under mild reaction conditions and gives stable oxime products under physiological pH.11 However, it suffers from slow reaction rates at pH 7, and thus requires millimolar concentrations of the reactants or large excess of one reactant to proceed within an acceptable timeframe required for use of short-lived radioisotopes such as fluorine-18. The reaction rates can be increased by adding a nucleophilic catalyst in form of aniline or derivatives thereof.12– 16 This reaction rate acceleration allows convenient use of oxime coupling protocols for proteins in micromolar concentration (typically 200-300 µM)17,18 with radioactive prosthetic groups (typically in the nano- to picomolar range). Indeed, a reaction of a hydroxylamine-modified leptin was achieved with [18F]fluorobenzaldehyde using aniliniumacetate buffer pH 4.5.18 18

F labelled aromatic aldehydes produces stable oxime-products in general, but suffer from slower kinetics than their aliphatic counterparts due to the more electron-rich carbonyl carbon.19–22 Even 18 F labelled aliphatic aldehydes such as [18F]FDR17,23–25, [18F]SiFA26 and a PEG-aldehyde19 either needs elevated temperatures, anilinium acetate buffer, pH 4 or micromolar concentrations of hydroxylamine for oxime coupling to proceed. Tissue factor (TF) in complex with its natural ligand activated factor seven (FVIIa) initiates the extrinsic coagulation cascade. TF is normally expressed on cells outside of vasculature and is not exposed to FVIIa in the bloodstream. Vessel rupture exposes TF to FVIIa and initiates coagulation.27 In addition, signaling by the TF:FVIIa complex on several cancer cells affects angiogenesis, proliferation, cell migration as well as cell survival, and TF is, therefore, an interesting target within cancer.27,28 Furthermore, therapies directed towards TF has shown promising results29–32 for which PET-imaging of TF-expression could be attractive as a companion diagnostic. Blockade of the active site in the protease domain of FVIIa produces active site inhibited factor seven (FVIIai). FVIIai, compared to FVIIa forms a complex with TF with a fivefold higher affinity33,34, and this binding does not initiate the coagulation cascade. In this work, the terminal N-glycan sialic acid was modified to contain a hydroxylamine (FVIIaiONH2). The conditions previously mentioned for oxime coupling of peptides or proteins, i.e. low pH and high amounts of hydroxylamine, are not suitable for sensitive proteins, such as FVIIai. Optimized conditions, proposed in this work, include preparation of a non-volatile fluorine-18 labelled aldehyde, removal of organic reaction media, and reaction in aqueous buffer at neutral pH at room temperature with nanomole amounts of FVIIai-ONH2.


Page 3 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Results and Discussion FVIIai-ONH2

FVIIai-ONH2 was produced as described by Øie et al.9 In short, FVIIai was desialylated by neuraminidase treatment, and resialylated with a hydroxylamine-modified cytidine-5’monophospho-N-acetylneuraminic acid derivate by the N-glycan-specific sialyltransferase ST3GalIII.

[18F]Fluorobenzaldehyde [18F]Fluorobenzaldehyde is a well-known and easy to synthesize radioactive prosthetic group, that is well suited for oxime coupling.35,36 As a model compound, O-benzylhydroxylamine, was used as a reaction partner. [18F]Fluorobenzaldehyde reacts with O-benzylhydroxylamine forming a 10:1 E/Z ratio of the resulting 18F-labelled oxime (Supporting Information). However, purification of [18F]Fluorobenzaldehyde and reaction with sensitive hydroxylamine modified proteins is complicated as [18F]Fluorobenzaldehyde is in an organic solvent after purification. As [18F]Fluorobenzaldehyde is volatile, the organic solvent cannot be evaporated and sensitive proteins tolerate only low amounts of organic solvent. A distillation approach was used to circumvent this problem, yielding high-quality [18F]Fluorobenzaldehyde, see Supporting Information. The distillation approach is hampered by [18F]Fluorobenzaldehyde having poor solubility in water and was therefore abandoned.

Synthesis of [18F]2 (4-[18F]fluoro-N-(4-oxobutyl)benzamide) [18F]SFB was already set up in our lab2, and a non-volatile, water-soluble aldehyde [18F]2 (4[18F]fluoro-N-(4-oxobutyl)benzamide) was developed from [18F]SFB (Figure 1). [18F]SFB was reacted with 4-aminobutyraldehyde diethyl acetal, giving the acetal [18F]1 (4-[18F]fluoro-N-(4,4diethoxybutyl)benzamide) which was deprotected using TFA/H2O. The isolated radiochemical yield was 2109  264 MBq (20.3  4.2%, n=3) after HPLC purification of aldehyde [18F]2 (from EOB, starting from 28.9  5.4 GBq of fluorine-18). The radiochemical purity was 98.3  0.4% (HPLC), and the synthesis time was 150  14 min.



Page 4 of 19

Figure 1: Synthesis of [18F]FVIIai via aldehyde [18F]2. a) 4-aminobutyraldehyde diethyl acetal (4.15 µL, 24 µmol), MeCN (1.5 mL), 5 min, r.t. b) TFA/H2O (10% v/v, 1 mL), 5 min, r.t. c) [18F]2 (10 mM HEPES, 10 mM CaCl2, 50 mM NaCl, pH 7.5), Aniline (1% v/v), 30 min. r.t.

Cyclization of acetals 1 and 2 As both the acetal 1 and the aldehyde 2 has the potential to cyclize,37 (Figure 2) it was investigated whether cyclization affects further reaction with hydroxylamines. NMR indicates that acetal 1 exists as a mixture of 1 and a cyclized structure in the ratio of approximately 3:1 (Supporting Information). Deprotection by either 10% TFA/H2O or 4M HCl, yielded 2, which was analyzed by LC-MS. Full deprotection to 2 was shown by LC-MS as only one UV-signal and one mass peak was observed. But as the cyclized product has the exact same mass as the aldehyde 2 and since coelution with the aldehyde is possible, the formation of a cyclization product could not be excluded by this method. Deprotection by TFA/H2O gave fewer side-products than 4M HCl. The reaction of 2 with O-benzylhydroxylamine and aniline as a catalyst, gave one distinct product by LC-MS. Two substances, presumably the E and Z isomers, were observed by NMR, see Supporting Information for a detailed description. This indicates that even if aldehyde, 2 cyclizes to some extent, it does not interfere with the subsequent oxime coupling reaction.

Figure 2: Possible cyclization of acetal 1 and aldehyde 2


Page 5 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To further examine the deprotection step an experiment, with either 10% TFA/H2O, 10% HCl/H2O, 10% AcOH/H2O or H2O (as control) to deprotect [18F]1 to [18F]2 was conducted. TFA and HCl worked equally good to form aldehyde [18F]2 within 5 minutes, while AcOH showed no formation of [18F]2 (compared to H2O) within 5 minutes (Supporting Information). TFA/H2O was preferred due to fewer side-products by LC-MS.

Catalyst for oxime coupling The radiochemical conversion (RCC) of aldehyde [18F]2 was tested with varying concentrations of O-benzylhydroxylamine using two different catalysts, aniline and 5-methoxyanthranilic acid (Figure 3A).13,15 These experiments show similar results for increasing amounts of both catalysts. Having established both catalysts as equal in reactivity for this reaction, aniline was selected as it is the easiest to handle, and does not influence the pH of a weakly buffered mixture as much as 5methoxyanthranilic acid (for 5-methoxyanthranilic acid, pH had to be adjusted subsequently to addition). The RCC of [18F]2 as a function of the reaction time using 50 µM Obenzylhydroxylamine was tested (Figure 3B). The increased RCC from 30 min to 1 hour 45 min reaction time does not make up for the loss due to decayed fluorine-18.

Figure 3: Catalyst screening for oxime coupling with [18F]2 and O-benzylhydroxylamine. A, oxime formation with varying amounts of aniline or 5-methoxyanthranilic acid at 30 min reaction time at r.t. B, time dependence of oxime coupling with varying amounts of aniline or 5-methoxyanthranilic acid O-benzylhydroxylamine (50 µM).

Stability of the TF-binding domain of FVIIai Oxime formation is slower at pH 7 than at more acidic pH. The preferred pH is 4.5.22 However, acidic media is not always compatible with proteins. Intact TF-binding of FVIIai at different pH values along with the thermal stability was tested. As FVIIai binds with 5-fold higher affinity to TF than FVIIa, FVIIai will displace FVIIa in a competition assay (details in the experimental section and Supporting Information). The assay indicates that the TF-binding domain of FVIIai is intact at pH 4-9 for 30 min, and at temperatures lower than 50C, Figure 4.



Page 6 of 19

Figure 4: Stability of TF-binding domain of FVIIai. A, pH stability. Competition assay with FVIIai, FVIIa, and sTF. 30 min. reaction time. Readout of the proteolytic activity from the S-2288 substrate by an ELISA reader. Performed at room temperature. B, thermal stability. Competition assay with FVIIai, FVIIa and sTF. 30 min. reaction time. Readout of the proteolytic activity from the S-2288 substrate by an ELISA reader. Performed at pH 7.4. FVIIai retained full binding to TF at all pH values tested and at temperatures below 50C. Since FVIIai is stable from pH 4-9 led us to test if the oxime reaction in our case would be faster at lower pH. Thus aldehyde [18F]2 was coupled to O-benzylhydroxylamine with aniline as a catalyst at pH 4.9 and 7.5. These two reactions were followed over time, and in this model setup, the reaction at pH 4.9 is faster than the reaction at pH 7.5 (Figure S4).

Oxime reaction conditions with FVIIai-ONH2

All the aforementioned tests were made with aldehyde [18F]2 in MeCN added to the reaction in HEPES buffer (excluding the FVIIai stability studies). This will not be possible for coupling to FVIIai-ONH2, as FVIIai-ONH2 will not tolerate MeCN present in the mixture. Evaporation of MeCN from aldehyde [18F]2 resulted in degradation (Figure S5). To solve this problem, a small amount of DMSO was added to aldehyde [18F]2 before evaporation of MeCN. An amount of 20 µL DMSO was found to be ideal, as the aldehyde does not degrade and DMSO does not interfere with the subsequent oxime coupling (Figure S6). Small molecules, like O-benzylhydroxylamine, do not always behave like proteins even though they have the same reactive groups. This is due, in part, to steric effects, predicted by collision theory.38 Therefore, testing with small molecules can provide general information about reaction conditions, but cannot necessarily be predictive for the reactivity of proteins. As oxime coupling reactions are favored at lower pH, the effect of pH in the reaction mixture for oxime coupling of FVIIai-ONH2 and aldehyde [18F]2, was tested at pH 5.2 and pH 7.5. MeCN was evaporated from aldehyde [18F]2 with 20 µL DMSO present. FVIIai-ONH2 was added and reacted at room temperature for 30 minutes. The resulting mixtures were analyzed by HPLC, purified by PD-10 and analyzed by HPLC (details in the experimental section). At pH 5.2 several side-products appears that could not be separated from [18F]FVIIai with PD-10 purification (RCP 91%), while at pH 7.5 a single product appeared (>99%, Figure 5A and B). The peak in the chromatograms at retention time 6.2 min. is remaining [18F]2, and the peak at retention time 9.7 is [18F]FVIIai. A Lindmo assay (performed as described in the experimental section) gave a binding fraction towards TF of 69% for synthesis at pH 5.2 versus 90% for synthesis at pH 7.5 (Figure S7 and 8). The reason for poor conversion of [18F]2 to [18F]FVIIai, side-product formation and reduced binding was not


Page 7 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


investigated further. In subsequent experiments, the oxime reaction with FVIIai-ONH2 was performed at pH 7.5.

Figure 5: pH dependence on oxime coupling. Retention time 6.2 min: Aldehyde [18F]2. Retention time 9.7 min: [18F]FVIIai. A, comparison of reaction between FVIIai-ONH2 and aldehyde [18F]2 at pH 7.5 and 5.2, crude reaction mixture. B, comparison of reaction between FVIIai-ONH2 and aldehyde [18F]2 at pH 7.5 and 5.2 after PD-10 purification. A typical coupling procedure gave [18F]FVIIai in an isolated RCY of 21  9% (269  93 MBq) from aldehyde [18F]2, a protein bound fraction of 98  0.5% and an RCP of 99  0.5% (HPLC). The molar activity was determined as 38.2  18.4 GBq/µmol, and the synthesis time was 116  11 min from [18F]2. The RCY from EOB was 4.7  0.9%, starting from 28.9  5.4 GBq of fluorine-18, and the total synthesis time was 267  5 min. The RCY is slightly lower but comparable and the synthesis time is longer compared to random labelling of FVIIai with [18F]SFB (RCY: 7.5  2.3% and synthesis time:  205 min.)2.

Plasma and buffer stability The stability of [18F]FVIIai was tested in gly-gly buffer and in mouse plasma. For stability in buffer, samples were analyzed with HPLC at different time-points. For stability in plasma, [18F]FVIIai was mixed with mouse plasma and left at 37C. Samples were diluted and frozen at 1-4 hours, and analyzed with SDS-PAGE by coomassie staining and autoradiography (Figure 6). Stability in buffer stability showed no degradation over 5 hours, and stability in plasma was >95% up to 4 hours.



Page 8 of 19

Figure 6: Stability of [18F]FVIIai in buffer and plasma. A, stability of [18F]FVIIai over 4 hours in buffer and plasma. B, SDS-PAGE with coomasie staining and autoradiography as an overlay. L: Seeblue plus 2 prestained standard. 1: FVIIai. 2: [18F]FVIIai. Left gel, lane 4-6: Plasma stability 1 hour. Left gel, lane 8-10: Plasma stability 2 hours. Right gel, lane 4,5 and 7: Plasma stability 3 hours. Right gel, lane 9-11: Plasma stability 4 hours.

Site-selectivity of oxime coupling Oxime coupling with aldehyde [18F]2 should give a homogeneous product, with the radioactive label at either one of the two hydroxylamine functionalized N-glycan chains of FVIIai-ONH2. The N-glycan chains are present in both the heavy and the light chain of FVIIai. The heavy and light chains become separated upon using reducing conditions for SDS-PAGE as described before.2 Glycan cleavage was applied to determine the distribution of aldehyde [18F]2 in [18F]FVIIai. PNGase F was used, where PNGase F cleaves the N-glycans from FVIIai. The proteins were analyzed using SDS-PAGE, and is depicted in Figure 7. PNGase F cleaves the glycan chains leaving only a small amount of radioactivity left (

Oxime coupling of active site inhibited factor seven with a non-volatile

Recommend Documents