• No results found

Design and Synthesis of Inhibitors Targeting the Aspartic Proteases HIV-1 PR and BACE-1

N/A
N/A
Protected

Academic year: 2021

Share "Design and Synthesis of Inhibitors Targeting the Aspartic Proteases HIV-1 PR and BACE-1"

Copied!
87
0
0

Loading.... (view fulltext now)

Full text

(1)

Design and Synthesis of Inhibitors

Targeting the Aspartic Proteases

HIV-1 PR and BACE-1

(2)

©Jenny Adrian Meredith, Stockholm 2009 Cover: The authors fingerprint, by Jason Meredith ISBN 978-91-7155-940-1

(3)
(4)
(5)

Abstract

This thesis describes the synthesis of molecules designed for inhibition of two aspartic proteases, viral HIV-1 PR and human BACE-1. It also reports on the structure activity relationships of the targeted enzyme inhibitors. It is estimated that currently 33 million people are infected with HIV, the causative agent of AIDS. The virus targets T-lymphocytes and macrophages of the human immune system. The HIV-1 PR plays an important role in the viral replication, and by inhibiting the enzyme the disease progression can be slowed down or even halted.

Herein is reported the design and synthesis of a series of HIV-1 PR inhibi-tors with novel P2 substituents of which several inhibit the enzyme in the nanomolar range. The aim of the second work was to further develop the inhibitors by the introduction of fluorine. Several attempts were performed to fluorinate different P2-substituents.

Alzheimer’s disease (AD) is neurodegenerative, progressive and fatal disor-der of the brain. It is associated with accumulation of plaques and tangles that cause impairment and functional decline of brain tissue which result in loss of memory and cognition. The plaques are mainly constituted of amy-loid-β peptides that are generated in two steps from the amyloid precursor protein (APP). The cleavage sequence is initiated by the aspartic protease BACE-1, which makes the enzyme a key target in the effort of finding a therapy that aim to slow down the progression of AD. Herein are enclosed the development of two series of potent BACE-1 inhibitors. In the first work a synthetic strategy was developed to truncate a previously reported hy-droxyethylene core structure in order to generate more drug-like inhibitors. This generated a series of truncated inhibitors where two amide bonds have been replaced with an ether - or alternatively a secondary amine linkage. A number of these inhibitors show potency against BACE-1. In the second part of the work the aim was investigate the effect of alterations in the P1 posi-tion. Five scaffolds with new P1 substituents were designed, synthesized and coupled with two different P2-P3 substituents. This resulted in a series of potent inhibitors that inhibit BACE-1 in the nanomolar range.

(6)
(7)

List of papers

This thesis is based on the following papers, which will be referred to by Roman numerals I-III and appendix I.

I Design and Synthesis of Novel P2 Substituents in Diol-based HIV Protease Inhibitors

Adrian Meredith, J.; Wallberg, H.; Vrang, L.; Oscarson, S.; Parkes, K.; Hallberg, A.; Samuelsson, B.

Submitted for publication in European Journal of Medicinal Chemistry

II P2’ truncated BACE-1 inhibitors with a novel Hydroxy-ethylene-like core

Adrian Meredith, J.; Björklund,C.; Adolfsson,H.; Hallberg, A.; Rosenquist,Å.; Samuelsson, B

Submitted for publication in European Journal of Medicinal Chemistry

III Design and Synthesis of Potent BACE-1 Inhibitors Con-taining a New Hydroxyethylene (HE) Scaffold: Potent ac-tivities in a cellular assay

Adrian Meredith, J.; Björklund, C.; Adolfsson, H.; Edlund, M.; Jansson, K.; Lindberg, J.; Vrang, L.; Hallberg, A.; Rosenquist, Å.; Samuelsson, B.

Manuscript

(8)
(9)

The modeling in Chapter 4 was performed by Kevin Parkes

The modeling in Chapter 6 and 7 was performed by Katarina Jansson All biological testing was performed at Medivir AB

(10)
(11)

Contents

1 Introduction ...1 1.1 Proteases ...1 1.2 Drug targets ...2 1.3 Nomenclature...2 1.4 Aspartic proteases...2 1.5 Design ...4 2. HIV...5 2.1 Introduction...5

2.2 HIV-1 and HIV-2...5

2.3 The life cycle of the HIV virus ...5

2.4 Anti HIV therapy...7

2.4.1 Approved drugs...7

2.4.2 Protease inhibitors...7

2.5 The HIV-1 protease...8

2.6 Design of HIV-1 PR inhibitors...9

2.7 Why do we need new HIV-1 PR inhibitors? ...10

3. Synthesis of HIV-1 PR inhibitors with new P2 moieties (Paper I) ...11

3.1 Background...11

3.2 Preparation of ligands...12

3.2.1 Synthesis of γ-substituted L-valine and L-homoserine analogues .12 3.2.2 Synthesis of 2-substituted L-valine ethyl amides...14

3.2.3 Oxetan analogues...14

3.2.4 Synthesis of target structures...15

3.3 Biological Results...16

3.4 Modeling and Structure Activity Relationships...18

3.5 Conclusions...21

4 Attempts to prepare fluorinated inhibitors (Appendix I)...23

4.1 Background...23

4.2 Synthesis ...24

4.2.1 Attempts to synthesize γ-fluoro-L-valine analogues...24

4.2.2 Attempts to synthesize γ-fluoro-L-homoserine analogues ...27

(12)

5. BACE-1...29

5.1 Introduction...29

5.2 APP and formation of Aβ ...29

5.3 Aspartic protease BACE-1...31

5.4 AD Therapy...31

5.4.1 Drugs available for AD...31

5.4.2 BACE-1 inhibitors...32

5.5 Design of BACE-1 inhibitors ...32

6. Synthesis of truncated BACE-1 inhibitors (Paper II) ...35

6.1 Background...35

6.2 Synthetic pathway towards truncated inhibitors ...36

6.2.1 Synthesis of primary ethers ...36

6.2.2 Protection of the secondary alcohol ...39

6.2.3 Synthesis of a phenylether and four secondary amines ...39

6.2.4 Synthesis of truncated target structures ...41

6.3 Biological results...43

6.4 Modeling and Structure Activity Relationships...45

6.5 Conclusions...46

7. Synthesis of BACE-1 inhibitors with new P1 substituents (Paper III) ...49

7.1 Introduction...49

7.2 Synthetic pathway towards new P1 substituents ...50

7.2.1 Introduction of new P1 substituents ...50

7.2.2 Synthesis of lactones ...52

7.2.3 Synthesis of target structures...53

7.3 Biological results...54

7.4 Structure Activity Relationships...57

7.5 Conclusions...58

8. Concluding remarks ...59

Appendix...61

Compounds discussed in Chapter 4...61

Acknowledgements ...65

(13)

Abbreviations

Aβ amyloid beta

ACE angiotensin converting enzyme

AcCl acetyl chloride

AD Alzheimer’s disease

AIBN azobisisobutyronitrile

AIDS acquired immune deficiency syndrome

APP amyloid precursor protein

Arg arginine

Asn aspargine

Asp aspartic acid

BACE-1 beta-site APP cleaving enzyme

BBB blood brain barrier

Caco-2 cell line consisting of heterogenous epithelial colorectal adenocarcinoma cells

Cbz benzyloxycarbonyl

CNS central nervous system

DAST diethylamino sulphurtrifluoride

DCE 1,2-dichloroethane

DDQ 2,3-dichloro-5,6-dicyanobenzoquinone

DIAD diisopropyl azodicarboxylate

DIBAL diisobutylaluminium hydride

DIPEA N,N’-diisopropylethylamine

DMAP 4-dimethylaminopyridine

DMF N,N’-dimethylformamide

DPP-IV dipeptidyl peptidase-4

DPPA diphenylphosphoryl azide

EC50 effective concentration, the concentration required to produce

50% of the maximum possible effect in a cell culture system

FDA US food and drug administration

Gly glycine

gp glycoprotein

HAART highly active anti retroviral therapy

HE hydroxyethylene

HEA hydroxyethylamine

HIV human immunodeficiency virus

HMDS hexamethyldisilazane

HPLC high performance liquid chromatography

IC50 inhibitory concentration, the concentration required to inhibit

(14)

Ile isoleucine

Im imidazole

Ki inhibitory constant/dissociation constant for inhibitor binding

calculated as [enzyme][inhibitor]/[enzyme-inhibitor complex]

LC-MS liquid chromatography mass spectroscopy

MeCN acetonitrile Met methionine Ms methanesulphonyl NMM 4-methylmorpholine Ns 2-nitrophenylsulphonyl NS3 non-structural protein 3 P position

Papp apparent permeability

PBSF perfluoro-1-butanesulfonyl fluoride

PI protease inhibitor

PMB p-methoxybenzyl

p-TsOH p-toluenesulphonic acid

PR protease

PyBOP benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate

RT reverse transcriptase

S subsite

SAR structure activity relationship

Ser serine

TBAB tetrabutylammonium bromide

TBAF tetrabutylammonium fluoride

TBAI tetrabutylammonium iodide

TEA triethylamine

TES triethylsilane

TFA trifluoroacetic acid

THF tetrahydrofuran

TMSCl trimethylsilyl chloride TREAT-HF triethylamine trihydrofluride

TS transition state

Tyr tyrosine

Val valine

(15)

1 Introduction

1.1 Proteases

Proteases (also called proteinases, peptidases) are a group of enzymes that catalyze the hydrolytic cleavage of peptide bonds. Proteases are extensively expressed in both mammalian and non-mammalian species, and are involved in a range of functions, such as regulation of essential biological processes

e.g. blood pressure, digestion, blood clotting, cell proliferation and

controlled cell death.1 Proteases can also cause physiological changes that

trigger disease pathologies such as osteoporosis, type II diabetes, hyperten-sion and thrombosis. Furthermore, proteases catalyze important proteolytic steps in tumor invasion and in the infection cycle of a number of pathogenic microorganisms and viruses. Currently there are 570 human proteases in the human degradome database.2-3 Several of these proteases together with a

number of bacterial, viral, and parasitic proteases are potential drug targets. A drug that targets a protease is generally designed as an inhibitor of the catalytic activity of the enzyme in order to reduce or impair its function. Based on the nature of the catalytic site, proteases are divided into five major classes; serine, threonine, cysteine, aspartic and metalloproteases.4-5 In

serine, threonine and cysteine proteases the mechanism is based on a nucleophilic attack on the targeted peptidic bond by a serine, threonine or cysteine, respectively. In aspartic and metalloproteases a water molecule plays this role.

This work presents the design and synthesis of inhibitors for two aspartic proteases, the retroviral HIV-1 PR (Human Immunodeficiency Virus protease) and the endogenous BACE-1 (beta-site APP cleaving enzyme) where APP stands for amyloid precursor protein.

(16)

1.2 Drug targets

As a result of extensive research, today there are commercially available drugs that target a number of endogenous proteases, e.g. ACE (hypertension), thrombin (blood clotting), DPP-IV (type II diabetes), and renin (hypertension) as well as the viral HIV-1 PR. Several drug candidates for proteases are undergoing clinical or preclinical studies e.g. viral hepatitis C NS3 protease, BACE-1 (Alzheimer’s disease), γ-secretase (Alzheimer’s disease) and cathepsin K (osteoporosis).6 Of these proteases apart from

BACE-1 and HIV-1 PR that have already been mentioned, renin and γ-secretase are also aspartic proteases.

1.3 Nomenclature

The amino acid side chains of the peptide substrate are referred to as P1, P1’, P2, P2’ etc (Figure 1), where the scissile bond is located between P1 and P1’.7 These side chains interact with the corresponding protease recognition

subsites S1, S1’, S2, S2’ etc, also numbered starting from the cleavage site.

Figure 1. Substrate binding to protease.

1.4 Aspartic proteases

Aspartic proteases can be found in different organisms ranging from fungi and retroviruses to bacteria and mammals. The best known sources of aspartic proteases are yeast, fungi and the stomach of mammals. Of the 570 human proteases discussed in Chapter 1.1 this is the smallest class with 21 endogenous aspartic proteases reported today. Although small, this class of enzymes comprises a number of important drug targets, as discussed in Chapter 1.2.

(17)

Aspartic proteases in general have two aspartic acid residues in the active site. The most widely accepted mechanism proposed is a general acid-base mechanism that involves coordination of a water molecule between two highly conserved aspartate residues. In the first step of the hydrolysis (Figure 2) the oxygen of the water molecule attacks the carbonyl group of the amide bond to be cleaved. Aspx donates a proton to the carbonyl oxygen

of the substrate, thus acting as an acid, while Aspy acts as a base by

accepting a proton from the water molecule. A highly unstable tetrahedral diol is formed, which collapses into an amine and a carboxylic acid.8-9

Figure 2. Proposed cleavage mechanism for an aspartic protease.

In the majority of aspartic protease inhibitors reported in the literature, the scissile bond is replaced with a stable transition state (TS) isostere motif.5

Characteristically, the aspartate-bound water molecule involved in peptide cleavage is mimicked by an alcohol contained within a hydroxyethylamine (HEA), hydroxyethylene (HE), dihydroxyethyl, statine, or norstatine pseudopeptide backbone framework (Figure 3).

(18)

1.5 Design

Traditionally, protease inhibitors have been developed via a natural substrate screening, which was followed by replacement of the scissile bond by a non-cleavable TS isostere and optimization by truncation of the polypeptide to shorter peptides. This method often proceeds via trial and error, progressively reducing the peptide character of the inhibitor. The design of inhibitors have later been improved by the availability of computer modeling techniques and information derived from NMR-analysis and X-ray crystallography, which provide three-dimensional structural information of both the receptor and the molecules that bind to such a receptor.

When developing a drug such as a protease inhibitor, several requirements should be met. In general, a potential drug is preferably orally bioavailable and administered once or twice daily. Furthermore, it should be safe with few side-effects and selective in its action. These are factors that are hard to predict at the early design stage. Often mentioned in the design of drugs are Lipinski’s rule of five, a rule of thumb that can be used to predict if a compound is likely to be orally bioavailable or not.10-11 Developing potent

and selective inhibitors that can target proteases centrally, i.e. in the brain such as BACE-1 or γ-secretase bears another great challenge in view of the difficulties involved in the passage of a drug across the blood-brain barrier (BBB). If active transport12 cannot be exploited, the potential to penetrate the

BBB is increased if a potential drug has low molecular weight,13 low

polarity, and high lipophilicity and is not a substrate of the P-glycoprotein efflux pumps.14

(19)

2. HIV

2.1 Introduction

The first clinical evidence of AIDS (acquired immune deficiency syndrome) was reported in 1981.15 HIV (human immunodeficiency virus) was identified

in 1983 as the causative agent of AIDS.16-17 Today the AIDS epidemic has

spread over the world, and in August 2008 UNAIDS estimated that 33 million people are living with HIV and AIDS.18 Globally, drug treatments

only reach about 20% of the HIV infected, as the majority of the infected people are found in the developing world where distribution is difficult, compliance low and prices generally not affordable unless heavily subsi-dized.

2.2 HIV-1 and HIV-2

The HIV virus belongs to a group of retroviruses called lentiviruses.19 There

are two types of human viruses causing AIDS, HIV-15,9,20-22 and HIV-2.23

HIV-1 and HIV-2 are closely related to simian immunodeficiency virus, SIV, and are distinguished based on their genomic organization and their evolutionary relationship to other primate lentiviruses.24 HIV-1, which is

responsible for the AIDS pandemic, originates from SIVcpz transmission from chimpanzees to humans. HIV-2 arose via transmission of SIVsm from sooty mangabeys. HIV-2 is prevalent in West Africa but has also spread to Asia and Europe. HIV-1 is more pathogenic than HIV-2.25

2.3 The life cycle of the HIV virus

The single-stranded genome of the HIV virus and enzymes necessary for its replication are packaged inside a protein core particle, and are surrounded by a cover, in which the outer envelope proteins (gp120 and gp41) are embed-ded.26 Two important targets for HIV are T-lymphocytes and macrophages,

(20)

parti-cle binds to CD4 on the host cell, which allows fusion of the viral envelope to the host membrane (Figure 4).

Figure 4. Life cycle of the virus.27-28 a: attachment; b: fusion; c: reverse transcript-

tion; d: integration; e: transcription; f: translation; g: post-translation processing; h: assembly; i: budding; j: maturation.

After fusion, viral RNA and viral enzymes are released into the cytoplasm of the cell. As reverse transcriptase (RT) becomes active, it acts as a DNA po-lymerase, and synthesizes a double-stranded DNA from the viral RNA. After entering the nucleus of the host cell, the virally coded DNA is integrated into the genetic material of the host, a process that is catalyzed by the viral en-zyme integrase (IN). Viral messenger RNA (mRNA) is transcribed and leaves the nucleus. The mRNA, employing the host’s cellular mechanisms, is then translated into the virus proteins gag and gag-pol polyproteins, which are multi-protein molecules that are further processed into smaller functional proteins during viral maturation. These polyproteins and viral RNA are transported to the cell membrane in preparation for assembly and viral bud-ding. The viral proteins assemble at the host cell membrane, and the virus buds from the cell surface and is then released, bringing all the viral proteins and RNA needed to form virions. During the budding of the new HIV

(21)

parti-cle from the host cell membrane, the viral protease (PR) is activated, result-ing in the proteolytic cleavage of polyproteins into various subunits, thus generating mature HIV. Some of these cleavages might also occur after bud-ding. This processing of polypeptides by the HIV PR is essential for the generation of infectious virions.

2.4 Anti HIV therapy

2.4.1 Approved drugs

Almost every step of the life cycle of the HIV virus is a potential target for antiviral chemotherapy.29-30 Today there are four types of drugs that have

been approved by the US Food and Drug Administration, FDA. Reverse transcriptase inhibitors were the first drugs to enter the market, and currently both nucleoside reverse transcriptase inhibitors (NRTIs), nucleotide reverse transcriptase inhibitors (NtRTIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) are available.31 Reverse transcriptase inhibitors were

followed by HIV-1 PR inhibitors (PIs), which have proven to be valuable therapeutics in combination with NRTIs for the treatment of AIDS. More recently FDA has approved two entry inhibitors that prevent the HIV from adhering to the endogenous immune system cells.32-33 The most recent type

of chemotherapy approved by the FDA for HIV treatment is an integrase inhibitor that hinders integration of the proviral DNA into the host cell chromosomal DNA.34

2.4.2 Protease inhibitors

Interruption of the viral lifecycle through HIV-1 PR inhibitors in theory results in the production of immature and non-infectious virions.35-36 This

mechanism of action has been remarkably effective in the clinic in suppressing the viral replication of HIV-1 infected patients. There are currently nine FDA-approved HIV-1 PI’s, saquinavir37 (Figure 5),

ritonavir,38 indinavir,39-40 nelfinavir,41 amprenavir,42 lopinavir,43 atazanavir,44

tipranavir45 and darunavir46 as well as fosamprenavir,47 a produg to

amprenavir. With the exception of tipranavir, all commercial inhibitors are more or less substrate analogues where the scissile P1-P1’ amide bond has been replaced with a hydroxyethylamine (HEA) or a hydroxyethylene (HE) TS isostere (Figure 3, Chapter 1).

(22)

N N H O H N O OH N H H O N H Ph saquinavir IC500.23 nM MW 679 Da N N N N H O OH Ph H N O OH indinavir O HN N H H N N N S O OH Ph Ph O O N S ritonavir H N O OH N H H O N H S Ph HO nelfinavir O N H H N N N H H N O O O Ph OH O O N atazanavir O H N S N CF3 O O OH O tipranavir O H N N H N O HN N S O OH Ph Ph O NH O lopinavir O O Ph OH NH2 O O amprenavir Ki0.6 nM MW 505 Da O O O HN N O Ph OH S NH2 O O H H darunavir Ki 0.016 nM MW 547 Da CONH2

Figure 5. FDA approved HIV-1 PR inhibitors

2.5 The HIV-1 protease

The HIV-1 PR consists of two monomers that form a C2-symmetrical

dimer.48-50 The monomers contain 99 amino acid residues, and contribute

with one aspartic acid residue each (Asp25 and Asp125) to the catalytic site of the protease. The protease is not very specific, but has some preference for cleavage sites in the viral gag and gag-pol polyproteins, from which it produces essential structural and functional viral proteins.51-52

The X-ray crystal structure of the HIV-1 protease in complex with an inhibi-tor reveals two flexible flaps in which Ile50 and Ile150 are coordinating a structural water molecule, which is also forming two hydrogen bonds to the inhibitor.52-53 The flaps, by opening and folding, allow the substrate or

(23)

in-hibitor to access the active site, and close around the substrate or inin-hibitor when it is bound to the active site.

2.6 Design of HIV-1 PR inhibitors

HIV-1 PR does not carry out a single reaction, but is an enzyme which cleaves the viral precursors in a specific order. The quite loose specificity in HIV-1 PR makes the design of an inhibitor challenging. The design of early inhibitors was based on natural substrates, with the scissile amide bond replaced by a TS analogue. The step from substrate screening to X-ray structure of an inhibitor co-crystallized with HIV-1 PR was important for the design of PI’s. In later inhibitors the use of a central hydroxyl TS analogue was still employed but the varying functionality that was systematically added, could be optimized through iterative synthesis, co-crystallization of inhibitor and enzyme, molecular modeling and redesigning the inhibitor until the required balance between potency, oral absorption and in vivo activity was reached.54

An example of such optimization is the development of amprenavir and darunavir.55 Structure activity relationship (SAR) studies of saquinavir

(Figure 5) revealed that the P2-aspargine forms a hydrogen bond with the Asp30 N-H of the protease backbone. Furthermore the carbonyl oxygen of the P3-quinaldic amide forms a hydrogen bond with Asp29 also located in the backbone of the protease. The backbone of the protease is less sensitive to mutations, and interactions between a potential inhibitor and residues of the main chains of the enzyme is therefore of importance to maintain potency against multi-drug-resistant HIV-1 PR. Truncation of the nonprime region and switching the P2-aspargine to a tetrahydrofuranylurethane resulted in a drop in potency, but the results were still encouraging since three amide linkages had been removed and the molecular weight reduced. Further modification in the prime side where a sulfonamide was introduced resulted in amprenavir, a very potent inhibitor of HIV-1 PR with low molecular weight and removal of the final amide linkage.

Examination of an X-ray structure of amprenavir co-crystallized with HIV-1 PR show that the ring oxygen of the THF moiety makes weak interactions with the N-Hs of Asp29 and Asp30. Optimization resulted in the design of several bicyclic ligands that effectively fills the S2 pocket and has ring oxy-gens positioned to form hydrogen bonds with the N-H’s of Asp29 and

(24)

Asp30. One of the most potent inhibitors developed was darunavir, in which both the oxygens of the bis-THF hydrogen form hydrogen-bonds to the backbone of Asp29 and Asp30 and the P2’ amine form hydrogen bonds to the carbonyl oxygen and carboxylate of backbone Asp130.56 These

interac-tions are proposed to be responsible for the potent activity of darunavir against multi-PI resistant strains.57

2.7 Why do we need new HIV-1 PR inhibitors?

Due to the growing number of drug-resistant HIV-1 PR mutants, there is a limit to the efficacy of the approved HIV-1 PR inhibitors.58 Other drawbacks

with the FDA-approved first generation inhibitors are severe side effects, toxicities, high dosage and high treatment costs.59 This urges for new HIV-1

PR inhibitors that need to demonstrate high potency, high genetic barrier, lack of cross-resistance towards the resistant mutant strains selected by the currently approved PI’s, good tolerability and high safety. Convenient once daily dosing is today highly desirable as well as lower treatment costs.

(25)

3. Synthesis of HIV-1 PR inhibitors with new

P2 moieties (Paper I)

3.1 Background

This chapter discusses the design and synthesis of a series of inhibitors for the HIV-1 protease. C2-symmetric inhibitor 1

60 (Figure 6) was previously

presented by our group and is a highly efficient inhibitor of the HIV-1 PR. Inhibitor 1 was synthesized in 3 steps, starting from L-mannaric acid

γ-lactone and has a dihydroxyethylene as a TS isostere, benzyloxy groups as P1/P1’ substituents and (1S,2R)-1-amino-indanol in the P2/P2’ positions. Amino-indanol can also be found in indinavir (Figure 5), and as a P2/P2’-substituent amino-indanol occupies the lipophilic S2/S2’ pocket with a neat fit, and also forms a hydrogen bond to the Asp29/129 of the HIV-1 PR backbone.40,61

Figure 6. Lead compounds 1 and 2.

The greatest drawback of the amino-indanol group is that it is readily oxidized by the 3A4 isozyme of the hepatic cytochrome P-450 system.62 The

consequence is rapid degradation which in turn results in poor oral bioavailability. This liability has prompted the development of new inhibitors which resulted in, e.g., the asymmetrical inhibitor 263 (Figure 6).

Inhibitor 2 binds to the HIV-1 protease with high affinity, but the L-valine

(26)

As discussed in Chapter 2.6, it has been shown that darunavir46 (Figure 5),

that have a tight binding between the P2 residue and Asp29 and Asp30 display broad-spectrum activity against multi-PI resistant HIV-1 variants. In order to optimize the L-valine methyl amide of inhibitor 2, new P2 ligands

were designed to explore the effects on potency from establishing hydrogen bonds or hydrogen interactions to the backbone N-H of Asp29 or Asp30 of the HIV-1 protease. This was performed by introducing extensions from the methyl amide, as well as from the methyl substituents of the γ-position of the

L-valine of inhibitor 2 (Figure 7). In addition two oxetanylglycine analogues

were synthesized. The unsymmetrical inhibitors have been evaluated for their potency against the HIV-1 protease and antiviral activity in a MT4 cell-based assay.These inhibitors were also tested against an HIV-1 protease inhibitor resistant strain carrying the M46I, V82F, and I84V mutations.

Figure 7. Outline of target compounds.

3.2 Preparation of ligands

3.2.1 Synthesis of γ-substituted L-valine and L-homoserine analogues

Commercially available Cbz-Asp-OMe 3 (Scheme 1) was treated with isobu-tylene64 and a catalytic amount of H

2SO4 to generate tert-butyl ester 465 in

82% yield.β-Methylation66-67 of compound 4 was performed using LiHMDS

and iodomethane furnishing compounds 5a/b65 in 89% yield as a 3:1 mixture

of inseparable diastereomers. After selective hydrolysis of the methyl ester with 1 M NaOH the β-acids 6a/b were obtained in 95% yield. The mixture of 6a/b was treated with ethyl chloroformate, and then reduced with sodium

(27)

borohydride.67 The diastereomers 7a68 and 7b68 could be separated by

col-umn chromatography and were obtained in 53% and 19% yield, respectively.

Scheme 1. Reagents and conditions: i: Isobutylene, H2SO4 (cat), CH2Cl2; ii: HMDS,

BuLi, MeI, THF, -78 ºC; iii: NaOH 1M, MeOH; iv: 1. Ethyl chloroformate, NMM, THF; 2. NaBH4, H2O.

Commercially available L-homoserine 8 was Cbz protected and converted

into the corresponding benzyl ester69 to obtain 7c in 50% yield over two

steps (Scheme 2).

Scheme 2. Reagents and conditions: i: Benzyl chloroformate, NaHCO3; ii: NaOH,

BnBr.

When treating 7a-c with silver oxide and iodomethane the only products that could be isolated were the corresponding lactones. Instead, methyl triflate and 2,6-di-tert-butyl-4-methylpyridine was used to furnish methyl ethers

9a-c (Scheme 3) in 44-56% yield.70 The yields are moderate due to the

forma-tion of lactone. Cleavage of the tert-butyl ester of 9a and 9b with TFA gen-erated the acids 10a and 10b in 88% and 74% yield, respectively.71

(28)

carbox-ylic acid 10c in 58% yield. After treatment with PyBOP,73 DIPEA and

me-thylamine, the methyl amides 11a-c were obtained in 62-88% yield.

Scheme 3. Reagents and conditions: i: MeOTf, 2,6-di-tert-butyl-4-methylpyridine,

CH2Cl2; ii: TFA, TES, CH2Cl2; iii: NaOH(aq), THF/MeOH (13:10); iv: PyBOP,

DIPEA, MeNH2, CH2Cl2.

3.2.2 Synthesis of 2-substituted L-valine ethyl amides

Commercially available ethylamines 12a-c (Scheme 4) were coupled with commercially available Cbz-L-valine-O-succinimide 1363 to obtain 14a, 14b

and 14c74 in 83%, 84% and 25% yield, respectively.

Scheme 4. Reagents and conditions: i: NMM, 12a, 12b or 12c, THF.

3.2.3 Oxetan analogues

For the synthesis of the Cbz protected amines 15a and 15b (Scheme 5) a racemic mixture of oxetanylglycine methyl ester 16a75 and 16b75 was Cbz

protected and separated by chiral HPLC to generate the protected amines

17a and 17b. The methyl ester was converted to the corresponding methyl

amide using methylamine in ethanol (33% solution) to generate 15a and 15b in 86% and 49% yield respectively.

(29)

Scheme 5. Reagents and conditions: i: Benzyl chloroformate, NaHCO3; ii: MeNH2

33% in ethanol.

3.2.4 Synthesis of target structures

Lactone 1876 (Scheme 6) was synthesized in three steps from L-mannaric

acid γ-lactone 19.76

Scheme 6. Reagents and conditions: i: HNO3, H2O; ii: benzyl-2,2,2-trichloro-

acetimidate, TfOH, 1,4-dioxane; iii: (1S,2R)-1-amino-2-indanol, 2-hydroxypyridine, CH2Cl2.

The Cbz group of 11a, 11c and 14a-c was removed by hydrogenolytic cleavage employing palladium on charcoal generating the free amines, which were refluxed with 18 in DCE to generate the target compounds 21a,

21c and 21d-f (Scheme 7) 15-28% yield.76 The low yield is due to the

formation of symmetric compound 1 (Figure 6), which was difficult to separate from the product by column chromatography. Some unreacted starting material was also recovered. In order to improve the yields a different method was introduced. The Cbz-group of 11b and 15a-b was removed using the same protocol as for 11a, 11c and 14a-c, vide supra. This was followed by treatment with 18 in neat DIPEA with 2-hydroxypyridine to obtain target compounds 21b, 21g and 21h in 32-61% yield. Symmetric compound 1 was still a major byproduct but more starting material had re-acted, which led to an increase in yield.

(30)

H N N H O R3 Cbz R3 O N H O Ph Ph HO O H N OH OH O O H N 21d: R3= CH2F, 28% 21e: R3= CF3, 15% 21f: R3= CH2OCH3, 18% 14a: R3= CH2F 14b: R3= CF3 14c: R3= CH2OCH3 i, ii H N N H O O Cbz O N H O Ph Ph HO O H N OH OH O O H N O 21g: R, 61% 21h: S, 35% 15a: R 15b: S i, iii H N N H O O R2 R1 Cbz Ph HO O H N OH OH O O H N O N H O Ph R2 R1 O 21a: R1= CH3, R2= H, 15% 21b: R1= H, R2= CH3, 32% 21c: R1= R2= H, 28% 11a: R1= CH3, R2= H 11b: R1= H, R2= CH3 11c: R1= R2= H i, ii or iii * *

Scheme 7. Reagents and conditions: i: H2, Pd/C, MeOH; ii: 18, DCE, reflux, 16 h;

iii: 18, DIPEA, 2-hydroxypyridine, 16 h, 70 ºC.

3.3 Biological Results

Investigation of the X-ray crystal structure of inhibitor 2 (Figure 8) in complex with the HIV-1 protease77 revealed two possible ways to increase

the interactions with the P2 substituent and Asp29 or Asp30 of the enzyme. The first was to build from the γ-carbon of the isopropyl group that is at 4.27 Å distance from the backbone nitrogen of Asp29 and 4.05 Å from that of Asp30. The second was to extend the methyl carbon of the methyl amide that is 5.07 Å from the backbone nitrogen of Asp29 and significantly further, 6.17 Å, from that of Asp30.

(31)

Figure 8. X-ray crystal structure of inhibitor 2 bound to the HIV-1 protease.77 The

flap residues (Met46 – Ile54 and Met146 – Ile154) have been deleted to improve visibility of the inhibitor binding, and the surface is color coded by lipophilic potential.

Based on these suggestions a series of P2 substituents, containing the potentially hydrogen interacting elements fluorine and oxygen, was designed, synthesized and introduced into our inhibitor scaffold. 21d-f (Scheme 7) were designed to interact with Asp29 and 21a-c and 2h-i with Asp30. Compounds 21a-h (Scheme 7) were evaluated for their antiviral potency against the HIV-1 protease (Ki)78-79 and for anti-HIV activity in a

MT4 cell-based assay (EC50)80 against both wild-type virus and a HIV-1

protease inhibitor-resistant strain carrying the M46I, V82F, and I84V mutations (Table 1). Inhibitors 21d and 21e were also evaluated for their apparent permeability (Papp) in a Caco-2 cell assay, this is further discussed

in Chapter 3.4 vide infra. Topological polar surface area (PSA) and logP (Table 1) were calculated for all inhibitors (Table 1).

(32)

Table 1. HIV-1 protease inhibitory activities, calculated logP and calculate

topologi-cal polar surface areas for targets compounds 21a-h.

Cpd. R= Ki (nM) EC50 (µM) EC50 R.C.a (µM) EC50

ratio logPb PSAc

21a 18 3.5 >10 >2 1.86 176 21b 0.7 0.33 1.3 4 1.86 176 21c 10 1.8 6.9 3 1.36 176 21d 1.0 0.17 0.44 3 2.84 166 21e 1.9 0.15 0.92 5 3.66 166 21f 2.3 0.25 1.2 5 2.52 176 21g 99 >10 >10 - 1.22 176 21h 10 5.4 >10 >1.6 1.22 176 2 0.2 0.14 1.2 9 2.68 166

aR.C.: Resistant clone. bCalculated log P. cCalculated polar surface area

3.4 Modeling and Structure Activity Relationships

Four inhibitors, 21b, 21d, 21e and 21f, displayed potencies relatively close to that of lead compound 2 (Table 1), with Ki values of 0.7 nM, 1.0 nM, 1.9

(33)

as-tively, were quite close to that of inhibitor 2. The introduction of fluorine in

21d lead to a slight increase in the calculated logP (Table 1) which is

re-flected in the minor improvements in apparent permeability (Papp) in a

Caco-2 assay with a value of 1.6 x10-6 cm/s. Inhibitor 2 displayed a P

app (Caco-2)

value of 1.4 x10-6 cm/s. Thus, the effect of fluorination on permeability is

more marked in the trifluorinated analogue 21e with a Papp (Caco-2) value of

2.9 x10-6 cm/s. The increase in calculated logP for compounds 21d and 21e

supports the observed increase in permeability.

In order to further investigate the SAR, modeling of two compounds, 21b and 21d, was performed using data from the X-ray crystal structure of 2 co-crystallized HIV-177 (Figure 9 and Figure 10) as a starting point. Docking of

the isopropyl substituted inhibitor 21b (Figure 9) indicates that a weak H-bond can be formed with Asp30, and possibly a more tenuous one with Asp29. However these interactions are counterbalanced by a quite tight fit of the substituent in the S2 pocket, which may explains why the activity is not improved compared to that of lead compound 2.

Figure 9. Docked pose of inhibitor 21b (magenta) overlaid with that of inhibitor 2

(orange) from the crystal structure prepared as in Figure 8.77

The decrease in activity of diastereomeric inhibitor 21a can be explained by these steric constraints of the S2 pocket, since 21a need to turn the

(34)

unsubsti-tuted methyl towards the flap to achieve the desired H-bonding interactions. The lower activity of 21c, which was expected to adopt a conformation mak-ing interactions similar to that of 21b, may be due to the entropic cost of constraining the highly mobile side-chain.

Docking of 21d indicates that the ethyl fluorine substituent is generally along the ligand binding channel towards the solvent and therefore can not achieve the envisioned polar interactions with the enzyme (Figure 10). This can explain the somewhat reduced potency in inhibitors 21a and 21b relative to that of the parent inhibitor 2.

Figure 10. Docked structure of inhibitor 21d (magenta) overlaid with that of

inhibitor 2 (orange) from the crystal structure prepared as in Figure 8.77

The low activity of the oxetanylglycines 21g and 21h (Table 1) was expected for the “unnatural” enantiomer 21g, but more surprising for 21h. This result might be due to the effect of the ring system that “ties back” the oxygen atom, effectively keeping it away from the Asp29 and Asp30 N-H’s. Inhibitors 21a-h were also tested against an HIV-1 protease inhibitor-resistant strain carrying the M46I, V82F, and I84V mutations (Table 1). Inhibitors 21d and 21b in particular displayed substantially lower loss of

(35)

whilst the I84V mutation is on the edge of the S2 pocket and might influence the binding, there is no clear overall pattern in the fold changes observed for these inhibitors against the mutant clone.

3.5 Conclusions

In summary eight novel inhibitors have been designed and synthesized. Inhibitor 21b and 21d showed the most potent inhibition of the HIV-1 protease. Inhibitors 21d and 21e were roughly equipotent to lead 2 with regard to anti-HIV activity in the MT4 cell-based assay. The introduction of one or more fluorines, as in inhibitors 21d and 21e, appears to contribute to the increase in cell permeability, which was expected. This observation is supported by Papp (Caco-2) data and calculated logP data. The increase in

cell permeability may contribute to the overall favorable anti-HIV activity. Extensions from the methyl group, i.e. inhibitors 21d-f appear to be generally well tolerated in the lipophilic S2 pocket although modeling indicates that the substituents fail to achieve the interactions with Asp29 that was envisaged. With the exception of inhibitor 21b extensions with polar groups from the isopropyl side chain of the P2 Val appear to be less well accommodated in the S2 pocket. When tested against a multi resistant clone, two of the analogues, 21d and 21e, display favorable lower fold changes compared to lead inhibitor 2.

The advantageous cellular anti-HIV activities for the fluorinated compounds

21d and 21e, as well as the short synthesis routes from the appropriate P2

(36)
(37)

4 Attempts to prepare fluorinated inhibitors

(Appendix I)

4.1 Background

Fluorine is the smallest substitute possible for hydrogen, although the C-F van der Waals radius (1.47 Å) is closer to that of C-O (1.57 Å) than to C-H (1.20 Å). The electronegativity of fluorine introduces a significant dipole moment into the bond to carbon. If fluorine is a hydrogen bond acceptor that forms fluorine-hydrogen bonds has been up for debate in the literature, but due to the low polarizability of fluorine it is probably more suitable to discuss these interactions in terms of weak polar interactions.81 The presence

of a fluorine atom in a compound can result in an increase of bioavailability due to enhanced metabolic stability, reduction of pKa and increase in

lipophilicity.82-83 This increase in lipophilicity can in turn result in improved

cell permeability.

As discussed in Chapter 3, fluorinated compounds 21d and 21e showed cellular antiviral activities close to those of lead compound 2, as well as improved cell permeability and increased lipophilicity (LogP, Table 1). It was hypothesized that the introduction of a fluorine in the γ-position of L

-valine and L-homoserine could result in potent inhibitors which also would

be better accommodated in the S2 pocket compared to 21a-c. Compounds

22a-c (Figure 11), which are close analogues of Cbz protected compounds 11a-c (Scheme 3) were designed as new P2 substituents based on our

(38)

Figure 11. P2 substituents 22a-c.

4.2 Synthesis

4.2.1 Attempts to synthesize γ-fluoro-L-valine analogues

Deoxyfluorination was attempted on both 7a and 7b (Scheme 8) using one of two systems: (1) DAST84 at -78 ºC in CH

2Cl2 (entry 1, Table 2), or (2)

perfluoro-1-butanesulfonyl fluoride (PBSF) together with triethylamine trihydrofluoride (TREAT) and TEA85 at room temperature with THF (entry

2 and 5, Table 2) or CH2Cl2 (entry 3 and 4, Table 2) as solvent. This resulted

in the formation of lactones 23a and 23b.86-87 At one point, when using

DAST, 24a could be isolated as a byproduct in 3% yield, which served as a reference in further experiments.

Scheme 8. Reagents and conditions: i: DAST, CH2Cl2, ii: PBSF, THF, TREAT, THF

(39)

Table 2. Conditions for attempts to deoxyfluorinate compounds 7a and 7b.

entry substrate Fluorinating agent

(equivalents) solvent T (ºC) t (h) Product 1 7a DAST:TEA (6:6) CH2Cl2 -78 → r t 1.5 23a+24aa 2 7a PBSF:TEA:TREAT (2:2:2) THF r t 16 23a 3 7a PBSF:TEA:TREAT (2:2:2) CH2Cl2 r t 16 23a 4 7b PBSF:TEA:TREAT (2:6:2) CH2Cl2 r t 16 23b 5 7b PBSF:TEA:TREAT (2:6:2) THF r t 16 23b a

24a was isolated in 3% yield

With DAST the lactone was formed rapidly, but the reactions with PBSF and TREAT required quite a long time before the starting material was completely converted into lactone even at room temperature. In order to examine if heating could promote the formation of a fluorinated product in favor of the lactone several experiments were performed with 7b using PBSF and TREAT. Heating was carried out via microwave irradiation in different aprotic solvents (THF, CH2Cl2 and MeCN) and temperatures

(120-180 ºC), but neither the fluorinated product nor lactone 23b were formed (Table 3), only starting material 7b could be detected.

Table 3. Microwave conditions for attempts to deoxyfluorinate compound 7b.

entry substrate PBSF:TEA:TREAT

(equivalents)

solvent T (ºC) t (min) product

1 7b 1:3:1 CH2Cl2 120a 5-30 n.d.b

2 7b 2:6:2 CH2Cl2 120 a 5-20 n.d.b

3 7b 2:6:2 THF 150 a 5-30 n.d.b

4 7b 4:12:4 THF 150 a 5-30 n.d.b

5 7b 2:6:2 MeCN 180 a 20 n.d.b

aThe reaction was run in a microwave reactor. bNot detected, only starting material

7b could be recovered.

To overcome the problem with lactone formation alcohol 7a was converted into its corresponding mesylate 25a (Scheme 9).88 Compound 25a was then

reacted with different nucleophilic fluorides such as TBAF, CsF and KF with 18-crown-6 (Table 4). Different aprotic solvents were evaluated, as well as different reaction times and temperatures, and microwave irradiation.

(40)

Scheme 9. Reagents and conditions: i: MsCl, TEA, DMAP, CH2Cl2, 0 ºC.

Table 4. Conditions for attempts to fluorinate compound 25a.

entry substrate Fluorination agent (equivalents)

solvent T (ºC) t (min) Product

1 25a CsF (2) MeCN 100b 2-100 26a

2 25a CsF (4) DMF 100b 5-30 26a

3 25a KF (2)a Acetone 85b 5-30 n.d.c

4 25a TBAF (6) MeCN 150b 15-360 n.d.c

5 25a TBAF (1.2) THF 90b 90-130 n.d.c

6 25a TBAF (2) THF reflux 75 n.d.c

a0.015 equivalents of 18-crown-6. bThe reaction was run in a microwave reactor. cNot detected, only starting material 25a could be recovered.

However, none of these reaction conditions yielded the desired product. Either the 25a had not reacted, or if consumed, the only isolated material was the cyclized compound 26a, (Figure 12), presumably resulting from a nucleophilic attack on the mesylate group by the Cbz carbonyl oxygen.

(41)

4.2.2 Attempts to synthesize γ-fluoro-L-homoserine analogues

In order to avoid intramolecular ring closure, the Cbz protection was replaced by benzyl and phenylfluorenyl (PhFl) groups (Scheme 10). At this point it was decided to use L-homoserine as the starting material, since it was more readily available than 7a and 7b.

The Cbz group in compound 4 (Scheme 10) was cleaved by hydrogenolysis using a catalytic amount of palladium on charcoal to generate the free amine89 in quantitative yield. The amine was benzylated via reductive

amination using benzaldehyde and sodium cyanoborohydride to generate

2768 in 53% yield. Compound 27 was treated with

9-bromo-9-phenylfluorene, in the presence of PbNO3 and K3PO4 to deliver 2868in 94%

yield. The methyl ester was reduced using DIBAL, resulting in the protected

L-homoserine 29 in 67% yield.90 The hydroxyl group of 29 was converted

into its mesylate using methanesulfonyl chloride and TEA and a catalytic amount of DMAP to furnish 30 in quantitative yield.

Scheme 10. Reagents and conditions: i: H2, Pd/C, MeOH; ii: 1. benzaldehyde, acetic

acid, MeOH, 2. NaCNBH4; iii: 9-bromo-9-phenylfluorene, PbNO3, K3PO4,

MeCN; iv: DIBAL, THF, -40 ºC; v: MsCl, TEA, DMAP, CH2Cl2, 0 ºC.

Compound 30 was treated with TBAF in MeCN or KF together with 18-crown-6 in DMF using microwave irradiation. Neither method gave any product at all, only starting material could be isolated (entries 1 and 2, Table

5). When compound 30 was heated at reflux with TBAF in THF overnight,

(42)

Table 5. Conditions for attempts to fluorinate compound 30.

entry substrate Fluorination agent (equivalents) solvent T (ºC) t (h) Product 1 30 TBAF (6) CH2Cl2 150b 0.25-6 n.d. c 2 30 KF (5)a DMF 150b 0.25-6 n.d. c 3 30 TBAF (2) THF Reflux 16 n.d. c 4 30 TBAF (6) THF Reflux 16 n.d. c

a2 Equivalents of 18-crown-6 were used. bThe reaction was run in a microwave

reactor. cNot detected, only starting material 30 could be recovered.

4.3 Conclusions

Cellular antiviral activities for the fluorine containing compounds discussed in Chapter 3 are close to that of the lead compounds 1 and 2, and this urged us to further explore the effect of fluorinated P2-substituents. Starting from

L-aspartic acid, different strategies towards the target structures were

explored. Lactone is the major undesired product when using DAST or PBSF for the γ-deoxyfluorination of compound 7a or 7b. If nucleophilic substitution of the corresponding mesylate is attempted instead, neither fluorination nor lactonization occurs. The fluorine is a too weak nucleophile. Instead, competing side reactions with intramolecular nucleophiles can take place. Our attempts to synthesize γ-fluorinated L-valine analogues have so

(43)

5. BACE-1

5.1 Introduction

Alzheimer’s disease (AD) was named after Alois Alzheimer, who in 1907 described rare structures that are now referred to as senile plaques, in the cerebral cortex of a 51 year-old woman with progressive dementia. AD is the most common cause of dementia in the elderly worldwide. This disabling, progressive and ultimately fatal form of dementia affects approximately 40 percent of the population over 80 years, with over 24 million people suffering from AD worldwide.91 In addition to the disease burden of the

individuals, and social impact on families and society, the financial costs to society are staggering, estimated at over 100 billion USD per year in the US alone.92-93

AD has two distinct pathological features – extracellular amyloid plaques and intracellular neurofibrillary tangles in the brain.94 The mechanism that is

believed to be responsible for the formation of amyloid plaques is referred to as the amyloid hypothesis,95 and will be discussed further in Chapter 5.2.

The formation of neurofibrillary tangles is believed to have its origins in abnormal hyperphosphorylation of the tau protein.91 The reason for this is

not well understood, and the tau hypothesis will not be further discussed in this thesis.

5.2 APP and formation of Aβ

According to the amyloid hypothesis extracellular plaques consist of aggregated polypeptides referred to as amyloid-β (Aβ), which are generated from the proteolytic cleavages of the amyloid precursor protein (APP).95

APP is a large type I membrane-bound protein that is expressed ubiquitously, and Aβ, unless it aggregates in brain tissue, is a normal catabolic product of the APP metabolism that take place in most cells.96-97

(44)

There are two pathways for APP-metabolism. The non-amyloidogenic pathway is initiated by α-secretase cleavage of APP (Figure 13) and predominates in most cell types.98-99 This results in a α-APPs (where s stands

for soluble) fragment that is believed to be neuroprotective, and a 83 amino-acid membrane-bound C-terminal fragment (C83), which in turn is cleaved by γ-secretase to generate a fragment named p3.100-102

The amyloidogenic pathway starts with APP being cleaved by the protease β-secretase (also referred to as BACE-1, Asp1 or memapsin-2) (Figure 13), this results in two products: a secreted ectodomain of APP named β-APPs,103

and a membrane-bound 99-amino-acid fragment of APP (C99).104 C99 is the

substrate of a second protease, γ-secretase, which cleaves C99 to generate the mature peptide Aβ, which is released from the cell. γ-Secretase cleavage is not precise, but the majority of Aβ are 40 or 42 amino acids in length, termed Aβ40 and Aβ42.105-106 In neuritic plaque found in the brain of AD

patients Aβ42 dominates,95,107 and in vitro the aggregation state of Aβ has

been show to influence neurotoxicity.108-109

Figure 13. The two routes by which APP is processed: the α-secretase (or non-

(45)

5.3 Aspartic protease BACE-1

In this thesis β-secretase will from now on be referred to as BACE-1. BACE-1 was identified as a type I transmembrane aspartic protease in 1999.96,111-114 BACE-1 is constituted of 501 amino acid residues, and is

expressed ubiquitously but predominates in the brain. The two aspartic residues involved in the cleavage of APP, Asp32 and Asp228, are located in the large extracellular domain of the enzyme. BACE-1 also has a short intracellular domain containing a sorting sequence, which has been shown to be involved in the trafficking of the protein.115 The X-ray structure of

BACE-1 co-crystallized with transition state inhibitor OM99-2116 (A1,

Figure 14) shows that the BACE-1 catalytic domain is similar in structure to

that of pepsin and other aspartic proteases. The active site is constitutes by a long cleft with eight main subsites spanning from S4 to S4’. Like many other aspartic proteases BACE-1 has a flap that closes down on the substrate.

5.4 AD Therapy

5.4.1 Drugs available for AD

There are currently two types of therapies available for the treatment of AD, cholinesterase inhibitors117 and memantine.91 Cholinesterase inhibitors do

not target the underlying pathology of the disease but only provide a symptomatic treatment addressing the cognitive impairment of the patients. Memantine belongs to a class of N-methyl-D-aspartate (NMDA) inhibitors

and is often administrated in combination with cholinesterase inhibitors. Memantine is believed to target the neuron-damaging excitotoxic activities in the brain, though it is not clear if the benefits are lasting.91 The need for

disease-modifying treatment for AD is enormous, and although many research groups have put a great effort into this area of research, currently effective drugs for the treatment of AD are not available. In theory there are several possible targets for disease-modifying treatment, i.e. immunothera-peutic approaches, inhibitors and modulators of BACE-1 and γ-secretase, inhibitors of amyloid aggregation, and compounds which seek to increase clearance of Aβ peptides.

(46)

5.4.2 BACE-1 inhibitors

The amyloid hypothesis is under debate in the literature, but it is still one of the main discovery approaches in the search for efficacious AD treatments. In support of the amyloid hypothesis, BACE-1 cleavage of APP is the first step in the production of Aβ and any factor that might have an impact on Aβ production would be likely to have an effect on APP processing and Aβ clearance. Another factor is that knockout mice, transgenic for human APP show no Aβ build-up in the brain and are healthy and fertile.118-119 Third,

BACE-1 is an aspartic protease, for which the mechanism of inhibition and design of TS analogues is well known.120 It is also possible to take advantage

of the knowledge that has been derived from extensive research on inhibitors for two other aspartic proteases, renin and HIV-1 PR.

An intense search to identify safe and efficacious drugs based on selective BACE-1 inhibition has become a significant goal for industrial as well as academic research efforts.121-124 There are no approved BACE-1 inhibitors on

the market today, although one, CTS 21166 is in clinical trials (phase I), and at least five more inhibitors are currently in preclinical studies.6 The

structure of CTS 21166 is yet to be published.

5.5 Design of BACE-1 inhibitors

As discussed in Chapter 2, targeting APP of the brain, the design of an in-hibitor for BACE-1 provides a great challenge, since such an inin-hibitor needs to be safe, selective and centrally active, i.e. being able to permeate the BBB by oral administration. The earliest attempts to develop inhibitors for BACE-1 were based on natural substrate screening. Swedish mutant APP is associ-ated with an increase in the production of Aβ and was used by Ghosh and co-workers in the development of A1116 (OM99-2) (Figure 14). A1 was

based on P4 to P4’ of the Swedish mutant APP but with some modifications, Asp in P1’ was exchanged for Ala, and the scissile bond was replaced with a HE transition state isostere. The X-ray structure of A1 co-crystallized with BACE-1, as discussed in Chapter 5.3, revealed that A1 interacted mainly with S4-S2’ of BACE-1. Truncation of A1 resulted in A2125 (Figure 14), an

inhibitor with fewer amide linkages and of lower molecular weight. Further truncation in the prime-side and the introduction of a five-substituted isophthalamide126 as a P2-P3 substituent resulted in A3127 a potent inhibitor

(47)

bonds to amino acids of S2 pocket of the enzyme.128 Furthermore it has been

shown that the R-configuration of the isophthalamide substituent is impor-tant for the activity against BACE-1. The five-substituted isophthalamide can also be found in A4129 (GRL 8234), which furthermore has a HEA as a

TS analogue. The HEA TS analogue requires inversion of the hydroxyl func-tionality to achieve the best interactions with BACE-1.55 As a P2’ substituent

A4 has a p-methoxybenzyl, a lipophilic ligand that is important for potency.

Other nonpeptidic P2-P3 substituents are the isonicotinamide in A5130-131

(Merck) and the tricyclic indolecarboxamide in A6132 (Glaxo Smith Kline).

Macrocyclic A7133 is potent against BACE-1 and also displays good in vivo

properties.

(48)
(49)

6. Synthesis of truncated BACE-1 inhibitors

(Paper II)

6.1 Background

This chapter describes the design and synthesis of a series truncated inhibitors for BACE-1. Previously our research group has reported several potent inhibitors with a novel hydroxyethylene (HE) core.134 The most potent

compound 31134 (Figure 15) in this series has a

3,5-di-fluorophenyloxymethyl group as P1, a methoxy group as a P1’, a five-substituted isophthalamide126 as the P2-P3 substituent and a L

-Val-benzylamine125 as P2’-P3’ substituent.

Figure 15. Lead compound 31.

Although very potent, lead compound 31 (Figure 15) has high molecular weight and a peptide-like character, drawbacks that warrants for optimization. The work that describe compound 31 also show that it is possible to truncate between P2’ and P3’ and still maintain activity against BACE-1, although significantly lower than that of 31.

(50)

Based on these findings a prime side truncated template was designed and synthesized. Two amide linkages were removed, and the new scaffold could be substituted with several ethers and secondary amines in the P2’ position (Figure 16). The template also allows for different P2-P3 substituents to be introduced in the last step.

Figure 16. Outline of target compounds.

6.2 Synthetic pathway towards truncated inhibitors

6.2.1 Synthesis of primary ethers

Compound 32 was synthesized in three steps starting from D-glucose 33

according to the literature (Scheme 11).76,135-136 Compound 33 was treated

with thiocarbonyldiimidazol to obtain thioester 34 in 97% yield. Subsequent deoxygenation of compound 34 was performed using 2,2’-azobisisobutyronitrile (AIBN) and tributyl tinhydride to furnish monodeoxy sugar 35 in 98% yield. The 5,6-O-isopropyliden was cleaved using acetic acid (70%, aq) to yield 32 in 92% yield.

(51)

Scheme 11. Reagents and conditions: i: thiocarbonyldiimidazol, THF; ii: tributyltin

hydride, AIBN, toluene; iii: 70% acetic acid (aq).

Starting from compound 32, lactol 36 was synthesized in a six-step synthetic route previously described by our research group (Scheme 12).134 Compound

32 was treated with triphenylphosphine (Ph3P) and DIAD to obtain epoxide

37 (scheme 12) in 80% yield.137 Compound 37 was selectively opened with

difluorophenoxide and potassium carbonate to generate 3,5-difluorophenoxy ether 38 in 71% yield.138 Compound 38 was converted to

the corresponding 5-azide 39 with inversion of configuration in quantitative yield using Mitsunobu conditions i.e. Ph3P, DIAD and diphenylphosphoryl

azide (DPPA).137 Compound 39 was treated with acetyl chloride in

methanol, to obtain compound 40 in 89% yield. Subsequent methylation of the 2-OH-position using silver oxide and iodomethane generated compound

41 in 79% yield. Deglycosylation of compound 41 was performed with

H2SO4 in 1,4-dioxane to furnish lactol 36 in 78% yield.

Scheme 12. Reagents and conditions: i: Ph3P, DIAD, DCE, reflux; ii:

3,5-difluoro-phenol, K2CO3, DMF, 120 ˚C; iii: Ph3P, DIAD, DPPA, THF, r t; iv: AcCl, MeOH, r

(52)

Compound 36 was initially treated with NaBH4 to obtain diol 42 (Scheme

13) but only unreacted 36 could be isolated. Attempts were also performed

with LiAH4. In these reactions the reduced compound, where also the azide

had been reduced into a free amine, could be detected by mass spectrometry but was not possible to isolate. Reductions with NaCNBH3 on a small scale

showed that some reductively opened product 42 was formed, and to optimize the reaction LiBH4 was used as the reducing agent, generating diol

42 in 95% yield. Compound 42 was converted to its corresponding tinacetal

with dibutyl tinoxide. The tinacetal was treated with benzyl bromide and TBAI to generate selectively opened 43a in 38% yield.

Scheme 13. Reagents and conditions: i: LiBH4, THF, 0 ºC; ii: dibutyl tinoxide, tolu-

ene, reflux; iii: BnBr, TBAI, 90 ºC.

When using the same protocol with iodomethane and allyl bromide the yields were unacceptably low (<10%), and an alternative strategy was developed. After treatment with sodium hydride, 42 was reacted with various organohalides preferentially alkylating at the primary alcohol, fur-nishing ethers 43b-h (Scheme 14) in 14 – 51% yield. The dialkylated and secondary alkylated products were formed in low yields (2-14%), and were easily separated from the target compounds by column chromatography.

(53)

6.2.2 Protection of the secondary alcohol

To convert compound 42 to its corresponding phenylether and phenyl- and benzylamines, protection of the secondary alcohol was necessary. The introduction of p-methoxybenzylidene acetal into 42 (Scheme 15) was performed using anisaldehyde dimehtylacetal and p-TsOH in DMF under reduced pressure to furnish 44 in 98% yield.139 The p-methoxybenzylidene

acetal was reductively opened employing NaCNBH3 and TMSCl in MeCN

to generate the two PMB ethers 45 and 46 in 43% and 42% yield respectively. The yield of 46 was not as high as hoped for based on reports in the literature. This is probably due to the difference in a seven membered acetalring compared to the six membered acetalring reported.139

Scheme 15. Reagents and conditions: i: Anisaldehyde dimehtylacetal, p-TsOH,

DMF, reduced pressure, 50 ºC; ii: NaCNBH3, TMSCl, MeCN, 0 ºC.

When using the method described to generate the corresponding secondary PMB-protected hydroxyl group (i.e. NaCNBH3, TFA, DMF) our results

were in accordance with those reported and product 45 was generated almost exclusively.139 Compound 45 could be used as a building block in the

synthesis of three targets, vide infra, as well as to regenerate 42 in 90% yield when treated with 10% TFA in CH2Cl2.

6.2.3 Synthesis of a phenylether and four secondary amines

Compound 46 was treated with phenol, Ph3P and DIAD to obtain

phenylether 47 (Scheme 16) in 63% yield. Subsequent cleavage of the PMB-group with DDQ generated 48 in 89% yield.

(54)

Scheme 16. Reagents and conditions: i: Phenol, Ph3P, DIAD, CH2Cl2; ii: DDQ,

CH2Cl2/H2O 19:1.

Compound 46 was treated with Ph3P, DIAD and

2-nitro-N-phenyl-benzenesulfonamide140 or N-(4-fluorophenyl)-2-nitro-benzenesulfonamide

(Scheme 17) to render the crude nosylated phenylamine and 4-fluorophenylamine, which were used in the next step without further purification. The nosyl group was cleaved using thiophenol and potassium carbonate to generate phenylamine 49a and 4-fluorophenylamine 49b in 74% and 50% yield over two steps, respectively. The PMB-group was cleaved using 10% TFA in CH2Cl2 to obtain 50a and 50b in 84% and 99%

yield, respectively.

Scheme 17. Reagents and conditions: i: 2-Nitro-N-phenyl-benzenesulfonamide or N-

(4-fluorophenyl)-2-nitro-benzenesulfonamide, DIAD, Ph3P, CH2Cl2; ii: PhSH,

K2CO3, MeCN, 50 ºC; iii: TFA 10% in CH2Cl2, 30 min.

Compound 46 was converted into nosylated benzylamine 51 in 61% yield

(Scheme 18) using Ph3P, DIAD and N

-(benzyl)-2-nitro-benzenesulfonamide.141 In order to avoid coupling of the benzylamine to the

carboxylic acid in one of the last steps of the synthesis, the nosylate was maintained as a protective group to be removed after the coupling. The PMB-group of 51 was cleaved using DDQ to generate 52 in 72% yield. Inversion of the hydroxyl group of 52 was performed under Mitsunobu conditions, i.e. Ph3P, DIAD and p-nitrobenzoic acid, followed by treatment

Figure

Figure 1. Substrate binding to protease.
Figure 3. Examples of stable TS isosteres.
Figure 4. Life cycle of the virus. 27-28  a: attachment; b: fusion; c: reverse transcript-  tion; d: integration; e: transcription; f: translation; g: post-translation processing; h:
Figure 5. FDA approved HIV-1 PR inhibitors
+7

References

Related documents

In addition, development of methodologies compatible with solid-phase synthesis is required since the conditions are generally different from those for solution-phase chemistry

Keywords: Aminoacyl-tRNA synthetases, Aminoacyl-AMP, Bioisosteres, Amino Acids, Solution-Phase Chemistry, Protective Groups, Solid-Phase Chemistry, Biological Evaluation

Cover Art: (top) compound 20 (Paper II) modeled in the active site of the HCV NS3 serine protease, (middle) model of compound 18 (Paper II) in the same protease (not used in

Department of Physics, Chemistry and Biology Linköping University, SE-58183 Linköping, Sweden.

(Bedford et al, 2020) In order to fully understand the expressed reactions amongst Swedish people living in Spain, a brief summary of the two countries regulations during the

A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series

Moreover, several inhibitors displayed improved cell-based permeability compared to other inhibitors previously synthesized by our group, but not unexpectedly these

In 2006, she began her postgraduate studies in Organic Chemistry under Professor Ingemar Kvarnström and Professor Bertil Samuelsson... Design and Synthesis