Different PET tau tracers bind to multiple binding sites on the tau fibril

Subscriber access provided by the Henry Madden Library | California State University, Fresno

Different PET tau tracers bind to multiple binding sites on the tau fibril : Insight from computational modeling Natarajan Arul Murugan, Agneta Nordberg, and Hans Ågren ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.8b00093 • Publication Date (Web): 09 Apr 2018 Downloaded from http://pubs.acs.org on April 10, 2018

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 34 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

ACS Chemical Neuroscience

Different PET tau tracers bind to multiple binding sites on the tau fibril : Insight from computational modeling N. Arul Murugan† , Agneta Nordberg‡,ν Hans ˚ Agren† ,q † Division of Theoretical Chemistry and Biology, School of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Center, S-106 91, Stockholm, Sweden ‡ Department of Neurobiology, Care Sciences and Society, Center of Alzheimer Research, Division of Clinical Geriatric, Karolinska Institutet, Huddinge, S-141 86, Stockholm, Sweden ν Theme Aging,Karolinska University Hospital, Huddinge, S-141 86, Stockholm q

Department of Physics and Astronomy, Uppsala University, SE-751 20 Uppsala, Sweden Abstract

Thanks to the recently reported cryo-EM structure for the tau fibril (Nature 547, 185-190 (2017)) which is a potential target concerning Alzheimer’s disease, we present the first molecular modeling studies on its interaction with various Positron Emission Tomography (PET) tracers. Experimentally, based on the binding assay studies, at least three different high affinity binding sites have been reported for tracers in the tau fibril. Herein, through integrated modeling using molecular docking, molecular dynamics and binding free energy calculations, we provide insight into the binding patterns of various tracers to the tau fibril. We suggest that there are four different high affinity binding sites available for many of the studied tracers showing varying binding affinity to different binding sites. Thus PBB3 binds most strongly to site-4 and interestingly this site is not a preferable site for any other tracers. For THK5351 our data show that it strongly binds to site-3 and site-1, the former one being more preferable. We also find that MK6240 and T807 bind to site-1 specifically. The modeling data also give some insight into whether a tracer bound to a specific site can be replaced by others or not. For example, the displacement of T807 by PBB3 as reported experimentally can also be explained and attributed to the larger binding affinity of the latter compound in all binding sites. The binding free energy results very well explain the small binding affinity of THK523 compared to all the aryl quinoline moiety containing THK tracers. The ability of certain tau tracers, like FDDNP and THK523, to bind to amyloid fibrils has also been investigated. Further such off-target interaction of tau tracers with amyloid beta fibrils has been validated using a quantum mechanical fragmentation approach.

Keywords: PET tracers, Tau fibril, Fibril imaging, Paired helical filaments, Molecular docking, Binding free energy calculations, Tauopathies. ACS Paragon Plus Environment

ACS Chemical Neuroscience 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

1

Page 2 of 34

Introduction

Alzheimer’s disease (AD) is contributing significantly to the healthcare burden in recent years making research for early diagnosis and therapy as of primary importance. While there are different views on selecting the most important potential targets for the early diagnosis of AD, currently the focus on tau imaging is steadily increasing based on that tau aggregation may occur before amyloid aggregation in the brain and that it shows a stronger correlation to cognitive impairment 1–3 . However, contemporary PET imaging of tau fibrils 4 imposes many challenges 5,6 as the aggregates are deposited in the intraneuronal compartments and are in particular deposited in the neurons, astrocytes, and oligodendroglia 7 . Although several tau PET tracers have been developed and also tested in different patient populations there are still many questions that remain unanswered regarding sensitivity, specificity and their applicability in clinical use. The design of tau specific tracers should target tau fibrils not only related to the occurrence in AD but also in other non-AD dementia disorders 8 such as frontotemporal lobe dementia, corticobasal degeneration, progressive supranuclear palsy and Down’s syndrome. Due to the intracellular deposition of tau aggregates, apart from binding affinity and specificity requirements, the tracers should also have the ability to pass through the blood brain barrier and possess cell permeability. Due to the similarity in the secondary structures and richness of beta sheet contents, the binding affinity of an effective tracer towards tau fibrils should be much larger, ideally an order of magnitude, than that of amyloid fibrils 6 and also because that concentrations of tau fibrils are 5-20 times lower when compared to amyloid fibrils when they coexist 6 . The specificity of the tracers is of major concern as many tracers have non-negligible binding to other off-targets 9 , for example, the amyloid tracer FDDNP has shown to be binding also to neurofibrillary tangles (NFTs) and was introduced as a tau tracer lately 6,10–12 . Three major chemical families of tau PET tracers 13 have been developed such as,

11

THK523,

C-PBB3,

18

18

F-T807,

F-THK5117,

18

18

F-T808, and different THK derivatives (18 F-THK5105,

18

F-

F-THK5351). More recently additional tau tracers referred to as

second generation tracers have been introduced. In particular, 18 F-MK6240, 18 F-RO6958948 and 18 F-PI26208 have been shown to bind to NFTs and paired helical filaments (PHFs) and have been clinically tested in tauopathies 14–18 . The T807 (also referred as AV-1451) and T808 tracers have been shown to have a 2527-fold larger binding affinity towards PHFs of the tau fibril than amyloid fibrils 14,19 . They also have lower molecular weight and suitable logP (1.5) indicating that they can effectively pass across the blood brain barrier. The pyridinyl-butadienyl-benzothiazole (PBB3) tracer has also shown promising affinity towards NFTs, in fact 40-50 times larger than that for senile plaques, possessing also an easy BBB penetrability. Most of these tracers have been ACS Paragon Plus Environment

Page 3 of 34 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


reported to have little non-specific binding to brain white matter. Recent investigations with a few of these tracers, like THK5317, THK5351, T807, and PBB3, with brain homogenates of AD patients or/and autoradiographies have demonstrated their preferential binding to tau deposits 20 . The binding of certain tracers like T807 showed that the binding can be dependent on the morphology of tau deposits. The PET tracer THK523 has also been reported to have 10 fold times larger binding affinity to recombinant tau fibrils than amyloid beta(1-42) fibrils 21 . The binding affinity of the THK523 in its high affinity site in tau fibril (1.7 nM) while in amyloid fibril was around 20.7 nM 21 . The recently developed tau tracer MK6240 has shown higher specificity towards NFTs 18 . Moreover, it binds poorly to amyloid plagues (in the µM range) and other 118 commonly screened proteins in the central nervous system 18 and so has limited off-target binding. The off-target binding of tau tracers to amyloid fibrils and other targets like monoamine oxidase-B and monoamine oxidase-A 22 is a major problem that limits their clinical applications. Another important feature, lately captured, and being at the focus of ongoing research, is the binding of these tracers to specific sites on the tau fibrils (NFTs and PHFs). Cai et al. reported such preferential binding sites in NFTs for T-807 and THK-523 23 . A further interesting observation is that certain tracers are competitively binding to specific binding sites while for other tracers there was no such observation. It has been suggested that MK6240 and T-807 are binding to the same site 18 . In the temporal cortex of AD patients, PBB3 binding was better fitted to a two site model suggesting that it had at least two binding sites (one with the binding affinity 1.3 nM and another with 25.3 nM) 24 . However, T-807 was only binding to a specific site with larger binding affinity which has been shown to be blocked by PBB3 24 . Overall, from in vivo competitive binding assay studies using PBB3 and T807 tracers it has been deduced that there are at least three different high affinity binding sites available in tau fibrils. Further, based on competition binding assay studies using the tau tracer, 3 H-THK5117 in post-mortem AD brain tissue, Lemoine et al. showed the presence of at least three different binding sites 25 . Two sites with nM binding affinity and a third site with pM range binding affinity were reported for this tracer. In a recent study Lemoine et al. further investigated the binding profile of other tracers, such as THK5351, T807 and PBB3, head-to-head in brain tissue. Multiple binding sites were reported for all the tracers along with the identification of super-high affinity binding sites for THK5351, T807 20 . Binding of tau tracers to multiple binding sites has remained untested so far computationally due to lack of tau fibril structure. In order to explain the competitive and non-competitive binding of various tau traces we need to understand the number of binding sites and relative binding affinities in these sites and the microscopic nature of the interaction between the tracers and the tau fibrils. So far, the PET tracer designACS forParagon tau imaging has mostly been based on trial and error Plus Environment


approaches or compound screening 14 in laboratory conditions. Computational modeling has here the potential to provide the knowledge necessary for a rational design of such tracers. However, data on computational modeling have remained scarce because the 3D structure for the tau fibril has not been available until recently. This is in rather sharp contrast to the case of amyloid fibrils for which a variety of 3D structures for various polymorphic forms have been reported showing richness in the beta sheets contents. Thanks to the availability of the amyloid fibril structures, many detailed computational modeling studies have been carried out for the purpose of understanding binding mechanisms of potential amyloid beta tracers. Recent advances in structural elucidation of insoluble biomacromolecules using the cryogenic-electron microscopy (cryo-EM) technique has helped to solve the structures of paired helical filaments, straight filaments and protofibril structures of tau which have been reported very recently 26 . In particular, the structures for the dodecamer and pentamer units of tau fibrils are available which can be used for structure based tracer design and for getting insight into the binding mechanism of various tau tracers. Similarly, the cryo-EM technique has also been used recently in the structural elucidation of amyloid fibrils made of whole peptide fragments (1-42) 27 . Most of the earlier structures reported for amyloid fibrils were for partial peptide aggregates and so the recent one appears to be a more reliable target for computational studies on tracers binding to amyloids. Unlike enzymes with well defined binding sites, fibrils are known to have multiple binding sites owing to their polymorphic forms and conformational flexibility. The structural variance of the fibrils is contributing to the non-specific binding of tracers making it desirable to improve the binding affinity and specificity for various fibril tracers. We have earlier reported that for the amyloid beta (1742) and (9-40) fibrils there are many core sites (the sites buried within fibril and surrounded by protein residues and exposed only to negligible amount of solvent) and surface sites 28 (the sites that are exposed to larger amount of solvents). The best way to improve the binding affinity and specificity towards a specific fibril is to optimize their interaction with the key residues around the core binding sites (as they have fibril specific microenvironment) for which computational modeling can be indispensable. In this work we have undertaken such a study and investigated the binding of various tracers with fibrils of tau protein. The aim of the study is thus to investigate the binding mechanisms of various PET tracers with tau fibrils by utilizing the very recent publication of the structure for tau fibrils, with in paired helical filament and straight filaments, obtained from the cryo-electron ˚. The protofilaments microscopy method 26 reported with a resolution in the range 3.4-3.5 A structures reported are derived from Alzheimer’s patients making them a reliable target for designing new tracers for tau imaging. We have studied several tau tracers belonging to different chemical families namely napthylethylidene derivative (FDDNP), benzimidazolepyrimidine derivative (T807, T808), pyridinyl-butadienyl-benzothiazole derivative (PBB3), ACS Paragon Plus Environment

Page 4 of 34

Page 5 of 34 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


arylquinoline derivatives (THK5105, THK523, THK5351, THK5117) and pyrrolo-pyridineisoquinolineamine (MK6240) and other compounds such as JNJ64349311 (will be referred as JNJ311), RO69558948 (will be referred as R06955). Due to the limited accuracy in the calculation of binding free energies, we have not attempted to compare the binding affinities of THK5317 and THK5117 (as the former one is the s-enantiomer of THK5117). Even though these two tracers have shown distinct binding profiles in experimental binding assay studies, 20 we consider these as the same in our modeling study. The study is also aimed at explaining the multiple binding sites reported for various tracers as used in in vitro and in vivo studies and to explain the competitive and non-competitive binding of tracers to tau fibrils. We have also addressed the cross interaction of certain tau tracers (namely FDDNP, THK523) with the amyloid beta fibril to investigate the possible ways to optimize the binding specificity towards a specific fibril.

2

Results and Discussion

Results for the binding sites for various tracers within the tau protofibril were analyzed, indicating that certain core and surface sites are associated with high affinity binding. The most important binding sites for the above mentioned tracers are shown in Figure 1. There three core sites are displayed (sites 1, 3 and 4) and one surface site (site 2). The binding affinity values in these sites are presented for various tracers in Table 1. Except the tracer T808, all tracers show significant binding to all four sites examined. Based on these values as calculated from molecular docking it can be deduced that for tracers like FDDNP, THK5105, THK5351, RO6955 and MK6240 the core sites are slightly preferable than the surface site, while the PBB3, T808, THK523 tracers have preference for the surface binding. Certain tracers, like T807, THK5317, JNJ311, have comparable binding affinities to more than one site. The binding affinity values for the tracers are predicted to be in the nM range which is in agreement with certain experimental inhibition constants data reported for these tracers 14–18 . Recently, Lemoine et al. report pM range binding affinity site for tracers like T807, THK5351 and THK5317 in human brain tissues 20,25 which is not correctly captured by the current molecular docking studies. The difference can be easily attributed to the simplicity of the model chosen in the current study which is a tau protofibril while the experimental data is based on binding assay studies in human brain tissue which is highly heterogeneous. To highlight the high affinity binding sites, we used different color code for the values in Table 1. The molecular docking results explain at a qualitative level that there are multiple binding sites available for tracers in the tau fibrils and that they have some preference to certain sites over others. Molecular docking provides aACS first handPlus information Paragon Environment about the binding of ligands to


biological targets. However, the binding affinity calculated in this method are based on certain approximations: (i) the biological target is treated as rigid framework and so the computed affinity is based on a single configuration of the target:ligand complex; (ii) the solvent description is treated implicitly; Given that fibrils in general are highly flexible and known to exist in different polymorphic forms, approximating the fibril as a rigid body may be crude. For this reason, the free energy of binding for tracers have been computed using the integrated molecular dynamics and MM-GBSA approach (refer to methods section). The computed free energies of binding for the tracers in each of the binding sites are given in Table 2. The high binding affinity sites are shown in blue color in the Table. As can be seen the results show some differences when compared to docking results illustrating the importance of sampling for such flexible targets like fibrils. We see significant differences in terms of the preferences for binding sites between the docking and MM-GBSA results, something that has to be attributed to the flexibility and dynamic nature of the binding site environment and its microstructure. Molecular docking only predicted site-1 and site-2 as the preferable sites for all the tracers. However, we see now that preference for site-3 and site-4 is significant for certain tracers. Our data show that the tracers T807 and MK6240 bind to site-1, while FDDNP is shown to have slight preference for site-2 and site-3. Interestingly, PBB3 is shown to have high preference for site-4 while its binding is not insignificant to remaining sites. Among the tracers studied, PBB3 has larger binding affinity with binding free energy values of about

Different PET tau tracers bind to multiple binding sites on the tau fibril

Recommend Documents