molecules

Review Probing Mechanisms and Therapeutic Potential of γ-Secretase in Alzheimer’s Disease

Michael S. Wolfe

Department of Medicinal Chemistry, University of Kansas, 1567 Irving Hill Road, GLH-2115, Lawrence, KS 66045, USA; [email protected]

Abstract: The membrane-embedded γ-secretase complex carries out hydrolysis within the in proteolyzing nearly 150 different membrane protein substrates. Among these substrates, the precursor protein (APP) has been the most studied, as generation of aggregation-prone amyloid β-protein (Aβ) is a defining feature of Alzheimer’s disease (AD). Mutations in APP and in , the catalytic component of γ-secretase, cause familial AD, strong evidence for a pathogenic role of Aβ. Substrate-based chemical probes—synthetic and peptidomimetics—have been critical to unraveling the complexity of γ-secretase, and small drug-like inhibitors and modulators of γ-secretase activity have been essential for exploring the potential of the as a therapeutic target for Alzheimer’s disease. Such chemical probes and therapeutic prototypes will be reviewed here, with concluding commentary on the future directions in the study of this biologically important protease complex and the translation of basic findings into therapeutics.

Keywords: protease; amyloid; Alzheimer’s disease; inhibitors; modulators

  1. Introduction Citation: Wolfe, M.S. Probing Alzheimer’s disease (AD), the most common form of dementia, is a devastating Mechanisms and Therapeutic neurodegenerative disorder that affects perhaps 30 million people worldwide, with demo- Potential of γ-Secretase in graphic projections suggesting this will increase substantially in the coming decades [1]. Alzheimer’s Disease. Molecules 2021, Cerebral typically takes place first in the hippocampus, a region below 26, 388. https://doi.org/ the neocortex that is critical for consolidating long-term memories. Neuronal loss spreads 10.3390/molecules26020388 to other cortical areas, leading to progressive cognitive decline. By the end stages of the

Academic Editors: Wei Li and disease, patients lose cognitive function to the point of requiring constant care, often in- Andrea Trabocchi stitutionalized. Although a number of risk factors have been associated with AD, disease Received: 21 December 2020 onset correlates best with age, and the large majority of cases occur in the elderly. Among Accepted: 10 January 2021 people over age 85, over a third are afflicted. Published: 13 January 2021 Two types of protein deposits are found in the AD brain: and neurofibrillary tangles [2]. The former are extraneuronal and primarily composed of Publisher’s Note: MDPI stays neu- the 4 kDa amyloid β- (Aβ), whereas the latter are intraneuronal filaments of the tral with regard to jurisdictional clai- normally microtubule-associated protein tau. Neuroinflammation is a third pathological ms in published maps and institutio- feature of AD, in which microglia—phagocytic brain immune cells that release cytokines— nal affiliations. become overactivated [3]. The role of each of these features in AD etiology and pathogenesis are not well understood. However, Aβ aggregation—in the form of oligomers, protofibrils, fibrils, and plaques—is generally observed as the earliest pathology, followed by tau tangle formation and neurodegeneration [4]. For this reason and those mentioned in Copyright: © 2021 by the author. Li- the next section, pathological Aβ is widely considered the initiator of AD, triggering censee MDPI, Basel, Switzerland. β This article is an open access article downstream tau pathology and neuroinflammation, and A has been the primary target distributed under the terms and con- for the development of AD therapeutics for over 25 years [5]. ditions of the Creative Commons At- 2. Familial AD and tribution (CC BY) license (https:// creativecommons.org/licenses/by/ As mentioned above, the “amyloid hypothesis” of AD pathogenesis has reigned for 4.0/). decades, and AD drug development has largely focused on inhibiting Aβ production,

Molecules 2021, 26, 388. https://doi.org/10.3390/molecules26020388 https://www.mdpi.com/journal/molecules Molecules 2021, 26, x 2 of 17

Molecules 2021, 26, 388 2. Familial AD and Genetics 2 of 17 As mentioned above, the “amyloid hypothesis” of AD pathogenesis has reigned for decades, and AD drug development has largely focused on inhibiting Aβ production, blockingblocking Aβ aggregation, Aβ aggregation, or facilitating or facilitating Aβ clearance Aβ clearance from fromthe brain the brain[6]. The [6]. primary The primary basis basis for thisfor dogma this dogma is the isdiscovery the discovery in the in 1990s the 1990sof dominant of dominant genetic genetic mutations mutations associated associated with with early-onsetearly-onset AD [7–10]. AD [7 This–10]. familial This familial AD (FAD) AD (FAD) has a hasdisease a disease onset onsetbefore before age 60 age and 60 can and can occuroccur even evenbefore before age 30. age Other 30. Other than the than mo thenogenetic monogenetic cause causeand mid- andlife mid-life onset, onset, FAD FADis is closelyclosely similar similar to the to sporadic the sporadic AD of AD old of age old with age respect with respect to pathology, to pathology, presentation, presentation, and and progression.progression. The most The mostparsimonious parsimonious explanation explanation is that is similar that similar molecular molecular and cellular and cellular eventsevents are involved are involved in the in pa thethogenesis pathogenesis and progression and progression of both of bothforms forms of the of disease. the disease. The firstThe genetic first genetic mutations mutations associated associated with with FAD FAD were were in the in theamyloid amyloid precursor precursor pro- protein tein (APP)(APP) [7]. [7]. This This gene gene encodes encodes a asingle-pass single-pass membrane membrane protein protein that that is isinitially initially cleaved cleaved by by β-secretase,β-secretase, a membrane-tethered a membrane-tethered aspartyl aspartyl protease protease in the in pepsin the pepsin family, family, to release to release the the large largeAPP ectodomain APP ectodomain [11] (Figure [11] (Figure 1). The1). remnant The remnant C-terminal C-terminal fragment fragment (APP (APP CTF- CTF-β) is β) is then proteolyzedthen proteolyzed within within its transmembrane its transmembrane domain domain (TMD) (TMD) by γ-secretase by γ-secretase to produce to produce Aβ, Aβ, β β whichwhich is then is secreted then secreted from fromthe cell the [12]. cell Most [12]. A Mostβ is 40 A residuesis 40 residues in length in length(Aβ40), (A but40), a but β smalla portion small portion is the ismuch the muchmore moreaggregation-prone aggregation-prone 42-residue 42-residue form form(Aβ42). (A Although42). Although β β β γ Aβ42A is a42 minor is a minor Aβ variant A variant produced produced through through APP CTF- APPβ CTF- processingprocessing by γ-secretase, by -secretase, it is it is the major form deposited in the characteristic cerebral plaques of AD [13]. the major form deposited in the characteristic cerebral plaques of AD [13].

Figure 1. Amyloid precursor protein (APP) processing by β- and γ-secretases. The single-pass

membrane protein APP is proteolyzed just outside the transmembrane domain (TMD) by β-secretase. FigureThe 1. Amyloid remaining precursor membrane-bound protein (APP) C-terminal processing fragment by β-(APP and γ CTF--secretases.β) is then The cleaved single-pass within the TMD membrane protein APP is proteolyzed just outside the transmembrane domain (TMD) by β-secre- to produce the amyloid β-peptide (Aβ) and the APP intracellular domain (AICD). tase. The remaining membrane-bound C-terminal fragment (APP CTF-β) is then cleaved within the TMD toThe produce 27 known the amyloid APP β mutations-peptide (A associatedβ) and the APP with intracellular FAD [14], domain each devastating (AICD). different families, are all missense mutations in and around the small Aβ region of the large APP. The 27 known APP mutations associated with FAD [14], each devastating different These mutations either alter Aβ production or increase the aggregation tendency of the families,peptide are all [15 missense]. A double mutations mutation in and just around outside the the smallN-terminus Aβ region of theof the Aβ largeregion APP. in APP Theseincreases mutations either alter by βA-secretase,β production leading or increase to elevated the APPaggregation CTF-β and tendency therefore of the elevated peptideAβ [15].overall. A double Mutations mutation in the just TMD outside near γ the-secretase N-terminus cleavage of the sites Aβ elevate region A inβ42/A APPβ in-40, and creasesmutations proteolysis within by β the-secretase, Aβ region leading itself to make elevated the peptide APP CTF- moreβ and prone therefore to aggregation. elevated Aβ overall.FAD Mutations mutations in the were TMD then near discovered γ-secretase in cleavage presenilin-1 sites (PSEN1)elevate A [β842/A] andβ40, presenilin-2 and mutations(PSEN2) within [9], the genes Aβ encoding region itself multi-pass make the membrane peptide more proteins prone that to at aggregation. the time had no known FADfunction. mutations These missensewere then mutations—now discovered in pr withesenilin-1 over 200 (PSEN1) known [8] [14 ],and all presenilin-2 but a dozen or so (PSEN2)in PSEN1—are [9], genes encoding located multi-pass throughout membra the sequencene proteins of the that protein at the but time mostly had no within known its nine function.TMDs These [16, 17missense]. Presenilin mutations—now FAD mutations with were over soon 200 foundknown to [14], increase all but A βa 42/Adozenβ 40or [so18 –20], in PSEN1—arefurther strengthening located throughout the idea the that sequence this ratio of isthe critical protein to but pathogenesis. mostly within Moreover, its nine these TMDsfindings [16,17]. indicatedPresenilin thatFAD mutations canwere modulate soon foundγ-secretase to increase cleavage Aβ42/A ofβ40 APP [18–20], substrate, furtheras strengthening the FAD mutations the idea altered that thethis preference ratio is critical for cleavage to pathogenesis. sites by the Moreover, protease. these Soon after findingscame indicated the observation that presenilins that knockout can modulate of PSEN1 γ-secretase dramatically cleavage reduced of APP Asubstrate,β production as at the FADthe levelmutations of γ-secretase altered the [21 ],preference with the remaining for cleavage Aβ sitesproduction by the attributedprotease. Soon to PSEN2 after (later verified [22,23]). Thus, presenilins are required for γ-secretase processing of APP CTF-β to Aβ peptides.

Molecules 2021, 26, x 3 of 17

came the observation that knockout of PSEN1 dramatically reduced Aβ production at the level of γ-secretase [21], with the remaining Aβ production attributed to PSEN2 (later ver- Molecules 2021, 26,ified 388 [22,23]). Thus, presenilins are required for γ-secretase processing of APP CTF-β to 3 of 17 Aβ peptides.

3. Presenilin and the γ-Secretase Complex 3. Presenilin and the γ-Secretase Complex Meanwhile, the design of substrate-based peptidomimetic inhibitors suggested that Meanwhile, the design of substrate-based peptidomimetic inhibitors suggested that γ-secretase is an aspartyl protease [24–26]. Peptide analogs, based on the γ-secretase γ-secretase is an aspartyl protease [24–26]. Peptide analogs, based on the γ-secretase cleavage site in the APP TMD leading to Aβ production and containing difluoroketone or cleavage site in the APP TMD leading to Aβ production and containing difluoroketone or difluoroalcohol moieties—mimetics of the transition state of aspartyl protease catalysis— difluoroalcohol moieties—mimetics of the transition state of aspartyl protease catalysis— were effective inhibitors of γ-secretase activity in cell-based assays. Given the requirement were effective inhibitors of γ-secretase activity in cell-based assays. Given the requirement of presenilin for γ-secretase activity, the site of proteolysis of APP within its TMD, the of presenilin for γ-secretase activity, the site of proteolysis of APP within its TMD, the multi-TMD naturemulti-TMD of presenilin, nature and of presenilin,the evidence and that the γ evidence-secretase that is anγ -secretaseaspartyl protease, is an aspartyl protease, the possibility wasthe possibilityraised that waspresenilin raised could that presenilin be a novel could membrane-embedded be a novel membrane-embedded protease. protease. Indeed, two conservedIndeed, two TMD conserved aspartates TMD were aspartates found in were presenilins, found in presenilins, and both aspartates and both aspartates were were required forrequired γ-secretase for γ-secretase activity [27]. activity Subsequent [27]. Subsequent reports that reports affinity-labeling that affinity-labeling rea- reagents gents based on basedthe transition-state on the transition-state analog inhibitors analog inhibitors of γ-secretase of γ-secretase bound directly bound to directly pre- to presenilin senilin cementedcemented the idea the that idea presenilin that presenilin is an unprecedented is an unprecedented aspartyl aspartyl protease protease with withits its active site locatedlocated within within the lipid the lipidbilayer bilayer [28,29]. [28 ,29]. Although presenilinsAlthough appeared presenilins to be appeared unusual aspartyl to be unusual , aspartyl it was proteases, clear that it was clear that they did not havethey this did activity not have on thistheir activity own. Presenilins on their own. themselves Presenilins undergo themselves proteolysis undergo proteolysis within the largewithin loop between the large TMD6 loop between and TMD7 TMD6 to form and TMD7an N-terminal to form anfragment N-terminal (NTF) fragment (NTF) and C-terminaland fragment C-terminal (CTF) fragment [30] (Figure (CTF) 2). [The30] (Figureformation2). Theof PSEN formation NTF and of PSEN CTF NTFis and CTF is gated by limitinggated cellular by limiting factors cellular[31], and factors these [two31], presenilin and these twosubunits presenilin assemble subunits into a assemble into a larger complexlarger [32,33]. complex Biochemical [32,33 analysis]. Biochemical and genetic analysis screening and genetic ultimately screening identified ultimately identified three other componentsthree other of componentswhat became of known what becameas the γ known-secretase as thecomplexγ-secretase [34–36]. complex These [34–36]. These three components,three membrane components, proteins membrane , proteins Aph-1, nicastrin, and Pen-2, Aph-1, assemble and with Pen-2, prese- assemble with pre- nilin, activatingsenilin, an autoproteolytic activating an function autoproteolytic of presenilin function to form of presenilin PSEN NTF to form and PSENCTF. NTF and CTF. [The two essential[The TMD two essentialaspartates TMD of presenilins aspartates are of presenilinsalso required are for also PSEN required NTF/CTF for PSEN NTF/CTF formation [27].]formation This assembly [27].] with This assemblycleaved presenilin with cleaved is the presenilin active form is theof γ active-secretase. form of γ-secretase. Indeed, the transition-stateIndeed, the transition-state analog affinity analog labeling affinity reagents labeling that reagents tag presenilins that tag specifi- presenilins specifically cally bound to PSENbound NTF to PSEN and NTFCTF and[28,29], CTF suggesting [28,29], suggesting that the active that the site active of the site protease of the protease resides resides at the interfaceat the interface betweenbetween these two these presenilin two presenilin subunits. subunits.This idea is This consistent idea is with consistent with the the observationobservation that one of thatthe essential one of the aspart essentialates aspartatesis in TMD6 is in in the TMD6 PSEN in theNTF, PSEN and NTF,the and the other other is in TMD7is inin TMD7the PSEN in the CTF. PSEN CTF.

Figure 2. PresenilinFigure 2. Presenilin and other and components other components of the γ-secretase of the γ-secretase complex. complex. Presenilin Presenilin is a 9-TMD is a protein9-TMD thatpro- contains two TMD aspartatestein that (D) contains essential two to catalysis. TMD aspartates Assembly (D) of essentia presenilinl to withcatalysis. the three Assembly other of components—nicastrin, presenilin with the Aph-1, and Pen-2—triggersthreeautoproteolysis other components—nicastrin of presenilin into, Aph-1, an N-terminal and Pen-2—triggers fragment (NTF) autoproteolysis and C-terminal of presenilin fragment (CTF)into to form the active γ-secretasean N-terminal complex. fragment (NTF) and C-terminal fragment (CTF) to form the active γ-secretase com- plex. Soon after the discovery of presenilin as the catalytic component of γ-secretase, analy- sis of the other proteolytic product of γ-secretase cleavage of APP (AICD), revealed that the APP TMD was proteolyzed at two different sites [37–41]. Cleavage at the second (ε) site

releases AICD products composed of residues 49–99 or 50–99 of the 99-residue APP CTF-β substrate for γ-secretase (Figure3). With secreted A β peptides ranging from 38–43 residues Molecules 2021, 26, x 4 of 17

Soon after the discovery of presenilin as the catalytic component of γ-secretase, anal- Molecules 2021, 26, 388 ysis of the other proteolytic product of γ-secretase cleavage of APP (AICD), revealed that4 of 17 the APP TMD was proteolyzed at two different sites [37–41]. Cleavage at the second (ε) site releases AICD products composed of residues 49–99 or 50–99 of the 99-residue APP CTF-β substrate for γ-secretase (Figure 3). With secreted Aβ peptides ranging from 38-43 residues(Aβ 38-A(Aβ38-Aβ43),β this43), leftthis 5left to 115 to APP 11 TMDAPP TMD residues residues unaccounted unaccounted for. Subsequent for. Subsequent discovery discoveryof Aβ of45, A Aββ45,46, A Aββ46,48, A andβ48, A andβ49, A butβ49, no but N-terminally no N-terminally extended extended AICD AICD peptides, peptides, led to the led tohypothesis the hypothesis that ε thatproteolysis ε proteolysis occurs occurs first [ 42first–44 [42–44].].

γ FigureFigure 3. Processive 3. Processive proteolysis proteolysis of APP substrate of APP substrateby γ-secretase. by Initial-secretase. endoproteolysis Initial endoproteolysis at the ε at cleavagethe siteε cleavage of APP substrate site of APP C99 substrate results in C99 Aβ48 results or Aβ in49 and Aβ48 corresponding or Aβ49 and APP corresponding intracellular APP in- domaintracellular fragments domain AICD49-99 fragments and AICD50-99. AICD49-99 The and AICD50-99. The activity carboxypeptidase of γ-secretase activity then of γ- trimssecretase the initial then Aβ peptides trims the along initial two Aβ peptidespathways: along Aβ49 two→A pathways:β46→Aβ43 A→βA49β→40A andβ46 →Aβ43→Aβ40 and Aβ48A→βA48β→45A→βA45β42→→AβA42β38.→ Aβ38.

The generatedThe generated Aβ48 A andβ48 A andβ49 A (counterpartsβ49 (counterparts of AICD49-99 of AICD49-99 and AICD50-99, and AICD50-99, respec- respec- tively)tively) were were postulated postulated to undergoundergo tripeptide tripeptide trimming trimming along twoalong pathways: two pathways: Aβ49→A β46 Aβ49→→AAββ4346→→Aβ4043→ andAβ40 Aβ and48→ AAββ4845→→AAββ4542→→AAβ42β38→A (thisβ38 last(this cleavage last cleavage step step generating gen- a eratingtetrapeptide a tetrapeptide coproduct). coproduct). Mass Mass spectrometric spectrometric analysis analysis of the of smallthe small peptide peptide products prod- sup- ucts portedsupported this this notion notion [45], [45], as did as did the findingthe finding that that synthetic synthetic Aβ49 Aβ is49 primarily is primarily processed pro- to cessedAβ to40 A andβ40 A andβ48 A isβ primarily48 is primarily trimmed trimmed to Aβ42 to byAβ purified42 by purifiedγ-secretase γ-secretase [46]. Kinetic [46]. analysisKi- neticof analysis trimming of tri ofmming synthetic of Asyntheticβ48 and A Aββ4849 and by fiveAβ49 different by five FAD-mutantdifferent FAD-mutantγ-secretase γ- com- secretaseplexes complexes revealed thatrevealed all five that were all dramaticallyfive were dramatically deficient in deficient this carboxypeptidase in this carboxypep- trimming tidaseactivity trimming [46]. activity [46]. PresenilinPresenilin and the and γ the-secretaseγ-secretase complex complex have havemany many more more substrates substrates besides besides APP APP[47]. [47]. Indeed,Indeed, so many so many substrates substrates have havebeen beenidentified identified that thatthe γ the-secretaseγ-secretase complex complex has been has been calledcalled the proteasome the proteasome of the of membrane the membrane [48], [im48],plying implying that thatone of one its of major its major functions functions is to is to clearclear out outmembrane membrane protein protein stubs stubs that that remain remain after after ectodomain releaserelease by by . sheddases. While γ Whilemembrane membrane protein protein clearance clearance may may be be an an important important functionfunction ofof γ-secretase,-secretase, the pro- protease teasealso also plays plays essential essential roles roles in in certain certain cell signaling pathways. pathways. The The most most important important of of these theseis is signaling signaling fromfrom the Notch Notch family family of of receptors receptors [49]. [49 Notch]. Notch receptors receptors are single-pass are single-pass membrane proteins like APP, and proteolytic processing of Notch, triggered by interaction membrane proteins like APP, and proteolytic processing of Notch, triggered by interac- with cognate ligands on neighboring cells, leads to release of its Notch intracellular domain tion with cognate ligands on neighboring cells, leads to release of its Notch intracellular (NICD) (Figure4). The NICD translocates to the nucleus and interacts with specific domain (NICD) (Figure 4). The NICD translocates to the nucleus and interacts with spe- transcription factors that control the expression of genes involved in cell differentiation. cific transcription factors that control the expression of genes involved in cell differentia- These signaling pathways, particularly from Notch1 receptors, are essential to proper tion. These signaling pathways, particularly from Notch1 receptors, are essential to proper development in all multi-cellular animals. Knockout of presenilin genes in mice is lethal development in all multi-cellular animals. Knockout of presenilin genes in mice is lethal and leads to phenotypes that are virtually identical with those observed upon knockout of and leads to phenotypes that are virtually identical with those observed upon knockout the Notch1 gene [50,51], findings that, as explained later, have major implications for the of the Notch1 gene [50,51], findings that, as explained later, have major implications for potential of γ-secretase inhibitors as AD therapeutics. the potential of γ-secretase inhibitors as AD therapeutics.

Molecules 2021, 26, 388 5 of 17 Molecules 2021, 26, x 5 of 17

Figure 4. Notch receptor processing and signaling. Interaction with a cognate ligand on a neighboring Figurecell triggers 4. Notch Notch receptor ectodomain processing shedding and signaling. by the metalloprotease Interaction with ADAM-10 a cognate and ligand then on cleavage a neigh- of the boringNotch cell extracellular triggers Notch truncation ectodomain (NEXT) shedding fragment by withthe metalloprotease its single TMD byADAM-10γ-secretase. and Releasethen cleav- of the ageNotch of the intracellular Notch extracellula domainr truncation (NICD) leads (NEXT) to translocation fragment with to the its nucleus,single TMD activation by γ-secretase. of transcriptor Release of the Notch intracellular domain (NICD) leads to translocation to the nucleus, activation factors, and gene expression that controls cell differentiation. of transcriptor factors, and gene expression that controls cell differentiation. 4. Chemical Probes 4. ChemicalSmall-molecule Probes substrate-based peptidomimetics have been valuable chemical probes in elucidatingSmall-moleculeγ-secretase substrate-based biochemistry peptidomimetics and biology. Ashave mentioned been valuable earlier, chemical transition- probesstate analogin elucidating inhibitors γ-secretase (TSAs; see biochemistry examplesin and Figure biolo5gy.) suggested As mentioned that γ earlier,-secretase transi- is an tion-stateaspartyl analog protease inhibitors [24–26], (TSAs; leading see to examples the identification in Figure of 5) two suggested conserved that TMD γ-secretase aspartates is anin aspartyl presenilin protease essential [24–26], to γ-secretase leading to activity the identification [27], and TSA of two affinity conserved probes TMD labeled aspar- PSEN tatesNTF in and presenilin CTF [28 essential,29]. Panels to γ of-secretase systematically activity varied [27], and TSAs TSA also affinity helped probes characterize labeled the PSENnature NTF of and the activeCTF [28,29]. site, particularly Panels of systematically the pockets that varied accommodate TSAs also helped substrate characterize theside nature chains of the [25 ,52active–54]. site, In proteaseparticularly terminology, the pockets these that substrateaccommodate residues substrate are P1, amino P2, P3 acidand side so onchains moving [25,52–54]. in the N In-terminal protease direction terminology, from these the scissile substrate amide residues bond andare P1, P1 0P2,, P2 0, P3P3 and0 etc. so movingon moving in thein the C-terminal N-terminal direction. direction Corresponding from the scissile pockets amide onbond the and protease P1′, P2 are′, P3termed′ etc. moving S1, S2, in S3 the etc. C-terminal and S10, S2 direction.0, S30 etc. Corresponding TSA peptidomimetic pockets probes on the forproteaseγ-secretase are termedsuggested S1, S2, (1) S3 relatively etc. and looseS1′, S2 sequence′, S3′ etc. specificityTSA peptidomimetic of the protease, probes consistent for γ-secretase with its nearlysug- 0 0 gested150 other (1) relatively TMD substrates loose sequence besides specificity APP [55]; (2)of the pockets protease, S2, S1, consistent S1 and S3withare its relatively nearly 0 150large, other accommodating TMD substrates the besides bulky aromaticAPP [55]; phenylalanine (2) pockets S2, side S1, chain,S1′ and while S3′ are the relatively S2 pocket large,is relatively accommodating shallow andthe bulky does notaromatic tolerate phenylalanine phenylalanine; side and chain, (3) thewhile protease the S2′ has pocket three 0 0 0 0 is pocketsrelatively S1 shallow, S2 , and and S3 doesbut not apparently tolerate nophenylalanine; S4 pocket. Thisand (3) last the point protease has implications has three pocketsfor how S1′γ, -secretaseS2′, and S3 processes′ but apparently APP substrate: no S4′ pocket. As described This last point earlier, has it implications is now clear for that howthe γ -secretase carries processes out multiple APP substrate: proteolytic As descri eventsbed in earlier, the TMD, it is trimming now clear down that the initially en- zymeformed carries long out A βmultiplepeptides proteolytic of 48 or 49 events residues, in the generally TMD, trimming in intervals down of three initially amino formed acids, 0 0 0 longto shorterAβ peptides secreted of 48 variants or 49 residues, such as generally Aβ40 and in Aintervalsβ42. The of threethree pocketsamino acids, S1 , S2 to shorter, and S3 secretedapparently variants dictate such tripeptide as Aβ40 carboxypeptidaseand Aβ42. The three trimming pockets of S1 long′, S2 to′, and short S3 A′ βapparently. Moreover, the intolerance of Phe in the S20 pocket was confirmed by systematic Phe mutagenesis dictate tripeptide carboxypeptidase trimming of long to short Aβ. Moreover,0 the intoler- ancewithin of Phe the APPin the TMD S2′ pocket substrate: was In confirmed every case by where system Pheatic was Phe in themutagenesis P2 position within relative the to γ APPa specific TMD substrate: cleavage site,In every proteolysis case where of that Phe site was by in-secretase the P2′ position was blocked relative [56 to]. a specific cleavage site, proteolysis of that site by γ-secretase was blocked [56].

Molecules 2021, 26, 388 6 of 17 Molecules 2021, 26, x 6 of 17

Figure 5. Representative transition-state analog inhibitorsinhibitors (TSAs) and affinityaffinity labeling reagents for γ-secretase. Functional groups that mimic the transition state of aspartyl protease catalysis are highlighted in red. Reactive groups leading to groups that mimic the transition state of aspartyl protease catalysis are highlighted in red. Reactive groups leading to covalent attachment are blue, and linker-biotin moieties are in green. Boxed is the general structure of the tetrahedral covalent attachment are blue, and linker-biotin moieties are in green. Boxed is the general structure of the tetrahedral intermediate formed upon addition of to the scissile amide bond during aspartyl protease catalysis. Residues P1, intermediateP2, P3 and P1 formed′, P2′, P3 upon′ relative addition to the of scissile water toamide the scissile bond are amide indicated. bond during aspartyl protease catalysis. Residues P1, P2, P3 and P10, P20, P30 relative to the scissile amide bond are indicated. Immobilization of a TSA inhibitor allowed isolation [57] and ultimately purification Immobilization of a TSA inhibitor allowed isolation [57] and ultimately purifica- [58] of the γ-secretase complex, but this affinity purification approach also revealed some- tion [58] of the γ-secretase complex, but this affinity purification approach also revealed thing important about how a membrane-embedded protease recognizes substrates. TSA something important about how a membrane-embedded protease recognizes substrates. affinity purification of γ-secretase from solubilized cell membranes resulted in co-purifi- TSA affinity purification of γ-secretase from solubilized cell membranes resulted in co- cation of APP substrate [57]. This was initially surprising, as it was not immediately obvi- purification of APP substrate [57]. This was initially surprising, as it was not immediately ous how substrate could be bound to the enzyme when the active site was occupied by obvious how substrate could be bound to the enzyme when the active site was occupied immobilized TSA. However, this makes sense considering that both enzyme active site by immobilized TSA. However, this makes sense considering that both enzyme active and substrate cleavage site reside in the membrane. The active site contains two catalytic site and substrate cleavage site reside in the membrane. The active site contains two aspartates that activate a water molecule for hydrolysis of an amide bond. The hydrophilic catalytic aspartates that activate a water molecule for hydrolysis of an amide bond. The aspartateshydrophilic and aspartates water should and water be inside should the be pres insideenilin the protein, presenilin sequestered protein, sequestered from the hydro- from phobicthe hydrophobic environment environment of the lipid of bilayer. the lipid Th bilayer.e substrate The is substrate also embedded is also embeddedin the membrane in the andmembrane can only and diffuse can only in two diffuse dimensions. in two dimensions. To gain access To gain to the access internal to the active internal site, active substrate site, TMDsubstrate must TMD first dock must on first the dock outer on surface the outer of the surface protease of the complex. protease Apparently, complex. Apparently, during the TSAduring affinity the TSA purification, affinity purification, enzyme was enzyme isolated was with isolated substrate with bound substrate to this bound initial to sub- this initial substrate docking exosite. With the immobilized TSA occupying the active site,

Molecules 2021, 26, x 7 of 17

Molecules 2021, 26, 388 7 of 17

strate docking exosite. With the immobilized TSA occupying the active site, substrate re- mains boundsubstrate to the remains external bound docking to the site external and cannot docking enter site the and internal cannot active enter site the for internal cat- active alytic conversion.site for catalytic conversion. Other substrate-basedOther substrate-based peptidomimetics peptidomimetics were designed were designed to target tothis target docking this dockingexosite. exosite. With theWith reasoning the reasoning that the thatinitial the contact initial with contact this with docking this docking site would site involve would involvea classical a classical α-helicalα conformation-helical conformation of the substrate of the substrate TMD, peptides TMD, peptides based on based the onAPP the TMD APP TMDwere syn- were synthe- thesized,sized, incorporating incorporating the helix-inducing the helix-inducing α-aminoisobutyricα-aminoisobutyric acid acid (Aib; (Aib; α-methylalanine)α-methylalanine) [59]. [59]. In theIn the design, design, Aib Aib residues residues were were spaced spaced apart apart every every 3–4 3–4 residues, residues, soso thatthat thetheAib Aib residues residueswould would lie lie along along one one face face of of the the helix, helix with, with APP APP residues residues arrayed arrayed along along the the rest rest of the helix of the helixand and available available for for binding binding to theto the docking docking exosite exosite (Figure (Figure6). 6). A setA set of of APP-based APP-based peptides peptideswas was synthesized synthesized with with the the placement placement of the of Aibthe residuesAib residues staggered staggered to present to present different faces differentof faces the TMDof the to TMD the protease to the protea complex.se complex. Additionally, Additionally, 10-residue 10-residueL-peptides L were-peptides identified as were identifiedlow micromolar as low micromolar inhibitors ofinhibitors Aβ production of Aβ production at the γ-secretase at the γ level-secretase in human level cells in stably human cellsexpressing stably APP.expressing Surprisingly, APP. Surprisingly, the corresponding the correspondingD-peptides, synthesized D-peptides, as synthe- controls, were sized as morecontrols, potent were that more their potent mirror-image that their counterparts. mirror-image Inverting counterparts. two internalInverting stereocenters, two internal however,stereocenters, disrupted however, helicity disrupted and dramatically helicity and decreased dramatically inhibitory decreased potency. inhibitory Phenylalanine potency.scanning Phenylalanine through scanning the ten-residue through helicalthe ten-residue peptide inhibitors helical peptide (HPIs) inhibitors led to identification (HPIs) of L- led to identificationand D-peptides of withL- and IC50 D-peptides values in with the mid-nanomolar IC50 values in the range. mid-nanomolar Exploring L- andrange.D-peptides Exploringup L to- and 16 residuesD-peptides in lengthup to 16 led residues to discovery in length of aled 13-residue to discovery D-HPI of a with 13-residue subnanomolar D-HPI withpotency subnanomolar [60]. potency [60].

Figure 6. Representative helical peptide inhibitors (HPIs) and affinity labeling reagents for γ-secretase. Each depicted HPI is a D-peptide. Helix-inducing aminoisobutyric acid (Aib) residues are highlighted Figure 6. Representative helical peptide inhibitors (HPIs) and affinity labeling reagents for γ-secre- in red. Photoreactive 4-benzoyl-D-phenylalanine residues are blue, and linker-biotin moieties are tase. Each depicted HPI is a D-peptide. Helix-inducing aminoisobutyric acid (Aib) residues are highlightedin green.in red. Photoreactive 4-benzoyl-D-phenylalanine residues are blue, and linker-biotin moieties are in green. The increased inhibitory potencies of HPIs with phenylalanine in specific positions en- Thecouraged increased the inhibitory replacement potencies of this residueof HPIs with with the phenylalanine photoactivatable in specific 4-benzoyl-phenylalanine positions encouragedand installationthe replacement of a linker-biotin of this residue on thewith N-terminus the photoactivaTable to provide affinity 4-benzoyl-phenyl- labeling reagents [61]. alanine andSuch installation modifications of a of linker-biotin the potent 10- onand the 13-residueN-terminusD to-HPIs provide led to affinity some loss labeling of inhibitory reagentsactivity [61]. Such but modifications still with low-to-mid-nanomolar of the potent 10- and potencies. 13-residue Interestingly, D-HPIs led to photoactivation some loss of these modified D-HPIs in the presence of HeLa cell lysates resulted in labeling of PSEN1 of inhibitory activity but still with low-to-mid-nanomolar potencies. Interestingly, photo- NTF and CTF. As TSA inhibitors also bind PSEN1 NTF and CTF, competition experiments activation of these modified D-HPIs in the presence of HeLa cell lysates resulted in label- were run and showed that TSA inhibitor did not compete with the 10-residue D-HPI ing of PSEN1 NTF and CTF. As TSA inhibitors also bind PSEN1 NTF and CTF, competi- photoprobe, and the 10-residue D-HPI parent compound did not compete with the TSA tion experiments were run and showed that TSA inhibitor did not compete with the 10- photoprobe. This demonstrated that the HPI and TSA compounds, while they both interact residue D-HPI photoprobe, and the 10-residue D-HPI parent compound did not compete with PSEN1 NTF and CTF, bind to distinct sites. This is consistent with the active site with the TSA photoprobe. This demonstrated that the HPI and TSA compounds, while residing inside PSEN1 and the docking site residing outside PSEN1, but with both sites at they both interact with PSEN1 NTF and CTF, bind to distinct sites. This is consistent with the NTF/CTF interface. Thus, substrate TMD apparently docks and then moves in whole the active site residing inside PSEN1 and the docking site residing outside PSEN1, but or in part between the NTF and CTF subunits of PSEN1 to access the internal active site. with both sites at the NTF/CTF interface. Thus, substrate TMD apparently docks and then In contrast to the 10-residue D-HPI, the 13-residue D-HPI could compete with the TSA moves in whole or in part between the NTF and CTF subunits of PSEN1 to access the protoprobe, and TSA inhibitor could compete with the 13-residue D-HPI photoprobe. The internal differenceactive site. in In abilities contrast of to the the 10- 10-residue and 13-residue D-HPI, D-HPIs the 13-residue to compete D-HPI with could TSAs com- suggests that pete withthe the docking TSA protoprobe, and active sitesand TSA are proximal, inhibitor withincould compete three amino with acids the 13-residue of each other. D- HPI photoprobe. The difference in abilities of the 10- and 13-residue D-HPIs to compete

Molecules 2021, 26, 388 8 of 17

5. Structural Probes More recently, HPI and TSA inhibitors have been combined to create substrate-based structural probes for the γ-secretase complex [62]. The structure of the γ-secretase complex was elucidated by cryo-electron microscopy (cryo-EM) in 2015 [63]. However, certain regions of PSEN1 were not resolved, and the catalytic aspartates were close but not aligned properly for catalysis. Subsequently, cryo-EM structures of γ-secretase bound to APP and Notch1 substrates were elucidated [64,65], with previously visible regions of PSEN1 now resolved. Both substrates were found enveloped by PSEN1, between NTF and CTF. However, determination of these substrate-bound structures required mutation of one of the active site aspartates and cysteine mutagenesis with disulfide crosslinking between substrate and PSEN1. Thus, the enzyme was catalytically inactive, and the Cys mutations with crosslinking raised the possibility of artifacts. To trap the active enzyme without crosslinking, full TMD substrate-based mimetics were recently developed as tight-binding inhibitors of γ-secretase [62]. In the design of these structural probes, linking 10-residue helical peptide inhibitors (HPIs) to transition-state analog inhibitors (TSA) was envisioned to provide potent inhibitors that would simultaneously bind to both the docking site and active site. Another thing critical to the design was use of an HPI composed of L-amino acids to more closely mimic substrate, so the TMD mimetic would not only bind the docking site but potentially enter the interior of PSEN1 through lateral gating, as seen with substrates in the new cryo-EM studies. In this way, active enzyme would tightly bind to these TMD mimetics and trap the enzyme at the transition state, poised to carry out intramembrane proteolysis. Cryo-EM analysis of these TMD mimetics bound to γ- secretase could be carried out with proteolytically active enzyme and without the need for chemical crosslinking. TSA 10 and HPI 11 were selected for these studies (Figure7)[ 62]. TSA 10, a pen- tapeptide analog spanning residues P2 through P30, was among the most potent com- pounds to emerge from systematic variation of hydroxyethylurea peptidomimetics [54]. Reanalysis with purified enzyme showed 10 inhibited γ-secretase with an IC50 of 41 nM. Aib-containing HPI 11 [59], based on the TMD of APP and composed of L-amino acids, inhibited purified γ-secretase with an IC50 of 58 nM. Synthesis of the HPI–TSA conjugate with the C-terminus of 10 directly connected to the N-terminus of 11 (i.e., with no linker), did not improve the inhibitory potency. However, insertion of ω-aminoalkanoyl linkers of varying lengths improved potency and resulted in discovery of 15 (Figure7) with an IC50 of 0.8 nM. This compound, with a 10-atom spacer, was essentially a stoichiometric inhibitor, as the concentration of γ-secretase used in the assay was 1 nM. The linker of 15 apparently contributed to potency, as the linker-10 analog displayed an IC50 of 16 nM (cf. IC50 of 10, 41 nM). Replacement of the hydroyxyethylurea moiety in the TSA component of 15 with a peptide bond resulted in substantial loss of potency (IC50 of 18 nM), and mass spectrometric analysis revealed that this peptide was cleaved by γ-secretase between the two Phe residues, validating the correct placement of the transition-state mimicking hydroxyethylurea moiety in the stoichiometric inhibitor 15. Finally, disrupting the helicity of the HPI component of 15 by inverting two internal stereocenters resulted in an 8-fold loss of potency (IC50 of 6 nM). Thus, each component of 15—the HPI, the linker, and the TSA—contributed to its high inhibitory potency. Enzyme inhibition kinetics demonstrated that 15 had a Ki of 0.42 nM toward γ- secretase [62]. Moreover, cross-competition kinetic experiments revealed that while TSA and HPI compounds were noncompetitive with respect to each other, 15 was competitive with both TSA and HPI, consistent with interaction of 15 with both docking site and active site. This was confirmed using the biotin-tagged photoaffinity probes of TSA and HPI described earlier: TSA but not HPI prevented labeling by the TSA photoprobe, and HPI but not TSA prevented labeling by the HPI photoprobe, while 15 prevented labeling by both probes. Provocatively, conformational analysis of 15 via 2D NMR experiments showed that the HPI component is indeed in a helical conformation, while the linker and TSA components are more flexible. Among the 10 lowest energy conformations is one that Molecules 2021, 26, 388 9 of 17

overlaps well with bound APP substrate in the recently reported cryo-EM structure of γ-secretase, with the hydroxyl group of the transition-state-mimicking moiety overlapping with the scissile amide bond in the APP substrate. Taken together, these findings suggest that 15 is a TMD mimetic pre-organized for optimal binding to γ-secretase and a suitable Molecules 2021, 26, x structural probe to trap the protease complex at the transition state of intramembrane9 of 17

proteolysis for analysis by cryo-EM. Such studies are underway.

Figure 7. Discovery of structural probes for γ-secretase. Schematic shows the design concept. Helical peptide inhibitors Figure 7. Discovery of structural probes for γ-secretase. Schematic shows the design concept. Helical peptide inhibitors (HPIs)(HPIs) directed directed to to the the substrate substrate docking docking exosite exosite were were conjugated conjugated through through a variablea variable linker linker to to transition-state transition-state analogue analogue inhibitorsinhibitors (TSAs) (TSAs) directed directed to to the the active active site.site. Presenilin Presenilin (blue-grey) (blue-grey) and and other other components components of ofthe the γ-secretaseγ-secretase complex complex (out- (outlined)lined) are are shown shown schematically schematically in in the the abse absencence and presence ofof aa hybridhybrid HPI-TSA HPI-TSA inhibitor. inhibitor. The The active active site site contains contains two two catalyticcatalytic transmembrane transmembrane aspartates aspartates and and three three hydrophobic hydrophobic pockets pockets S10, S2 S10 and′, S2′ S3 and0 that S3 must′ that bemust engaged be engaged for intramembrane for intramem- proteolysis.brane proteolysis. Linking TSALinking10 to TSA HPI 1011 tothrough HPI 11 athrough 10-atom a linker10-atom led linker to stoichiometric led to stoichiometric inhibitor inhibitor15. 15.

Enzyme inhibition kinetics demonstrated that 15 had a Ki of 0.42 nM toward γ-secre- 6.tase Functional [62]. Moreover, Probes cross-competition kinetic experiments revealed that while TSA and HPIBased compounds on the findingwere noncompetitive that replacement with of respect the hydroxyethylurea to each other, 15 was of 15 competitiveby a peptide with bondboth resulted TSA and in HPI, proteolysis consistent by γwith-secretase, interaction new of helical 15 with peptides both docking were designed site and basedactive on site. theThis APP was TMD confirmed as substrates using for theγ-secretase biotin-tagged [66]. ThephotoaffinityN-terminal probes Aβ-like of cleavage TSA and products HPI de- containscribed Aib earlier: residues, TSA whichbut not should HPI prevented induce helicity labeling and by therebythe TSA improve photoprobe, their and solubility HPI but andnot detectability. TSA prevented Thus, labeling these syntheticby the HPI full photoprobe, TMD functional while 15 probes prevented for γ-secretase labeling by were both developedprobes. Provocatively, to allow ready conformational analysis of all analysis the proteolytic of 15 via products 2D NMR by experiments LC-MS. Peptides showed rangingthat the from HPI Gly29 component to Lys55 is of indeed APP substrate in a helical (Aβ conformation,numbering) were while generated, the linker installing and TSA helix-inducingcomponents are Aib more residues flexible. in the AmongN-terminal the 10 half lowest of the energy TMD-based conformations peptides (Figure is one8 that). Theoverlaps C-terminal well with half ofbound the peptides APP substrate contained in the only recently APP reported residues, cryo-EM potentially structure allowing of γ- ε unwindingsecretase, andwith bindingthe hydroxyl to the group active of site thefor transition-state-mimicking-like proteolysis followed moiety by processiveoverlapping proteolysis.with the scissile Initially, amide three bond such peptidesin the APP (19 substrate.–21) were Taken designed, together, placing these the Aibfindings residues suggest in a staggeredthat 15 is a manner TMD mimetic in order pre-organized to present different for optimal faces of binding the APP to TMD γ-secretase helix to theand protease. a suitable structural probe to trap the protease complex at the transition state of intramembrane pro- teolysis for analysis by cryo-EM. Such studies are underway.

6. Functional Probes Based on the finding that replacement of the hydroxyethylurea of 15 by a peptide bond resulted in proteolysis by γ-secretase, new helical peptides were designed based on the APP TMD as substrates for γ-secretase [66]. The N-terminal Aβ-like cleavage products contain Aib residues, which should induce helicity and thereby improve their solubility and detectability. Thus, these synthetic full TMD functional probes for γ-secretase were developed to allow ready analysis of all the proteolytic products by LC-MS. Peptides

Molecules 2021, 26, x 10 of 17

ranging from Gly29 to Lys55 of APP substrate (Aβ numbering) were generated, installing helix-inducing Aib residues in the N-terminal half of the TMD-based peptides (Figure 8). The C-terminal half of the peptides contained only APP residues, potentially allowing unwinding and binding to the active site for ε-like proteolysis followed by processive pro- Molecules 2021, 26, 388 teolysis. Initially, three such peptides (19–21) were designed, placing the Aib residues in 10 of 17 a staggered manner in order to present different faces of the APP TMD helix to the prote- ase.

Figure 8. Functional probes for γ-secretase. Peptides based on the TMD of APP were designed with Figure 8. Functional probes for γ-secretase. Peptides based on the TMD of APP were designed helix-inducing α-aminoisobutyric acid (Aib) installed in the N-terminal region. Proteolytic products with helix-inducing α-aminoisobutyric acid (Aib) installed in the N-terminal region. Proteolytic were detected by LC-MS. Endoproteolytic (ε) sites, indicated by the arrowheads below the peptide products were detected by LC-MS. Endoproteolytic (ε) sites, indicated by the arrowheads below the peptidesequence, sequence, were were determined determined by by analysis analysis of of AICD-like AICD-like products. products. TrimmingTrimming sites,sites, indicatedindi- by the cated by thearrows arrows above above the the peptide peptide sequence, sequence, were were determined determined by by analysis analysis of of A βA-likeβ-like products. products. Note that Note thatreplacement replacement of of the the most most C-terminal C-terminal Aib Aib in in peptides peptides19 19–21–21with with the the corresponding corresponding APP APP TMD residue TMD residue(peptides (peptides22–24 22) leads–24) leads to most to most effective effective trimming, trimming, producing producing shorter shorter Aβ-like Aβ products.-like prod- ucts. Analysis of the proteolytic products by LC-MS revealed that ε proteolysis took place Analysisonly at of the the two proteolytic sites that products occur naturally by LC-MS with revealed APP: both that Aε βproteolysis-like peptides took corresponding place only at theto Atwoβ48 sites and that Aβ49 occur were naturally observed with along APP: with both the A correspondingβ-like peptides AICD-like corresponding peptide prod- to Aβ48 uctsand A [66β].49 Moreover, were observed proteolytic along with products the corresponding corresponding AICD-like to Aβ43, Apeptideβ45, and prod- Aβ46 were ucts [66].also Moreover, observed, proteolytic suggesting products that normal corresponding tripeptide to trimming Aβ43, Aβ of45, the and Aβ A48-β46 and were Aβ 49-like also observed,peptides suggesting occurred. that This normal was validatedtripeptide in trimming two ways: of the first, A longerβ48- and AICD-like Aβ49-like products peptidescorresponding occurred. This to was the validated shorter A βin-like two productsways: first, were longer not observed,AICD-like and products second, cor- incubation respondingof Phe-mutant to the shorter versions Aβ-like of products peptide were19 with not γobserved,-secretase and demonstrated second, incubation that Phe of blocked 0 Phe-mutantcleavage versions wherever of peptide it was 19 placed with γ in-secretase the P2 position,demonstrated just as that seen Phe with blocked APP substrate. cleav- age whereverBecause it was placed these initialin the P2 functional′ position, probes just as were seen onlywith trimmedAPP substrate. up to the Aβ43-like site, Becausea second these round initial offunctional peptides prob wases designed were only in trimmed which the up mostto the C-terminal Aβ43-like site, Aib a in 19, 20, second roundand 21 ofwas peptides replaced was with designed the corresponding in which the most residue C-terminal in the APP Aib TMDin 19,(peptides 20, and 22–24, 21 was replacedFigure8)[ with66]. the The corresponding suspicion was residue that the in the most APP C-terminal TMD (peptides Aib was 22– preventing24, Figure further β β 8) [66]. Thetrimming suspicion to A was40- that and the A most42-like C-terminal peptides Aib and was that preventing extending thefurther natural trimming APP sequence would allow another round of tripeptide trimming. Indeed, while 22–24 were still cleaved to Aβ40- and Aβ42-like peptides and that extending the natural APP sequence would al- exclusively at the normal ε sites, generating only the expected AICD-like products, the low another round of tripeptide trimming. Indeed, while 22–24 were still cleaved exclu- Aβ-like products now ranged from “Aβ40” to “Aβ45”. Thus, trimming is more efficient sively at the normal ε sites, generating only the expected AICD-like products, the Aβ-like with the second-round peptides, as no “Aβ46”, “Aβ48” or “Aβ49” were detected and products now ranged from “Aβ40” to “Aβ45”. Thus, trimming is more efficient with the trimming to “Aβ40” and “Aβ42” occurred with peptide 22. Replacement of an additional second-round peptides, as no “Aβ46”, “Aβ48” or “Aβ49” were detected and trimming to Aib residue of 22 with the corresponding APP TMD residue led to even further trimming “Aβ40” and “Aβ42” occurred with peptide 22. Replacement of an additional Aib residue to “Aβ37” and “Aβ38”, as seen with APP substrate. of 22 with the corresponding APP TMD residue led to even further trimming to “Aβ37” These helical peptide functional probes were further validated as surrogate substrates and “Aβ38”, as seen with APP substrate. by synthesizing peptides based on 22–24 containing APP TMD FAD mutations V44A These helical peptide functional probes were further validated as surrogate sub- and I45F [66]. In APP substrate, these disease-causing mutations have opposite effects strates by synthesizing peptides based on 22–24 containing APP TMD FAD mutations on ε cleavage site preference: V44A skews cleavage toward the Aβ48-producing site, V44A andand I45F I45F [66]. favors In APP the Asubstrate,β49-producing these disease-causing site compared mutations to what is seenhave withopposite wild-type ef- APP. fects on Afterε cleavage incubation site preference: with γ-secretase, V44A skews the mutant cleavage versions toward of the22– A24βshowed48-producing these samesite, trends, and I45Fespecially favors the the Aβ V44A49-producing and I45F site mutants compared of 22 to. what Provocatively, is seen with the wild-type FAD-mutant APP. peptides After incubationwere all deficientwith γ-secretase, in trimming the mutant compared versions to their of 22 wild-type–24 showed counterparts, these same with trends, longer Aβ- like peptides up to “Aβ49” being detectable. These findings are consistent with earlier studies with wild-type APP substrate and PSEN1 FAD-mutant γ-secretases, which showed deficient trimming and increased levels of long Aβ peptides [46]. Taken together, these observations raise the possibility that Aβ peptides ranging from 45 to 49 residues, which contain most of the APP TMD and are membrane-anchored, are pathogenic in FAD. Molecules 2021, 26, x 11 of 17

especially the V44A and I45F mutants of 22. Provocatively, the FAD-mutant peptides were all deficient in trimming compared to their wild-type counterparts, with longer Aβ-like peptides up to “Aβ49” being detectable. These findings are consistent with earlier studies with wild-type APP substrate and PSEN1 FAD-mutant γ-secretases, which showed defi- cient trimming and increased levels of long Aβ peptides [46]. Taken together, these obser- Molecules 2021, 26, 388 vations raise the possibility that Aβ peptides ranging from 45 to 49 residues, which con- 11 of 17 tain most of the APP TMD and are membrane-anchored, are pathogenic in FAD.

7. γ-Secretase Inhibitors: Therapeutic Potential 7. γ-Secretase Inhibitors: Therapeutic Potential The search for γ-secretase inhibitors (GSIs) as potential therapeutics for AD had gone The search for γ-secretase inhibitors (GSIs) as potential therapeutics for AD had gone on for over two decades, ever since Aβ was discovered to be a normally secreted peptide on for over two decades, ever since Aβ was discovered to be a normally secreted peptide produced from a variety of cell types in culture [67]. Such inhibitors were identified even produced from a variety of cell types in culture [67]. Such inhibitors were identified even before the components of the protease were known and ultimately served as critical tools before the components of the protease were known and ultimately served as critical tools for discovery of presenilin as the catalytic component [28,29], as described earlier. Initially, for discovery of presenilin as the catalytic component [28,29], as described earlier. Initially, these inhibitorsthese inhibitors were simple were peptidomimetics simple peptidomimetics (e.g., TSAs), (e.g., but TSAs), pharmaceutical but pharmaceutical companies companies quickly quicklydeveloped developed compounds compounds with much with better much drug-like better drug-like properties properties that allowed that allowed in vivo in vivo testing fortesting the ability for the to ability lower to A lowerβ in the Aβ brainsin the of brains transgenic of transgenic AD mice AD (e.g., mice expressing (e.g., expressing FAD-mutantFAD-mutant APP and APP presenilin). and presenilin). Acute treatmentAcute treatment with GSIs with did GSIsshow did such show proof such of principle proof of for principle these drug for candidates these drug candi- [68,69]; dateshowever, [68,69 chronic]; however, treatm chronicent revealed treatment serious revealed peripheral serious toxicities peripheral [70,71], toxicities such [70 as,71 ], such gastrointestinalas gastrointestinal bleeding, bleeding,immunosuppression, immunosuppression, and skin andlesions, skin all lesions, effects all that effects could that be could be traced totraced inhibition to inhibition of Notch of proteolysis Notch proteolysis and signaling. and signaling. As AD patients As AD patients would be would required be required to take GSIsto take for GSIs years for and years perhaps and perhaps decades, decades, the severe the severe toxic consequences toxic consequences of γ-secretase of γ-secretase in- inhibitionhibition caused caused great great concern. concern. While While ther theree were were hopes hopes for for a therapeutic a therapeutic window window that that would would allowallow lowering lowering brain brain A Aββ levelslevels without without the the peripheral peripheral Notch-deficient Notch-deficient toxicity, toxicity, the the failure failure ofof one one GSI, (Figure(Figure9 ),9), in in phase phase III III clinical clinical trials, trials, dashed dashed these these hopes hopes [ 72 ]. The [72]. Thetrial trial resulted resulted in in unacceptable unacceptable peripheral peripheral toxicities toxicities and—more and—more worrisome—cognition worrisome—cog- that nition thatwas was worse worse than than the the placebo placebo control control groups. groups.

Figure 9. Clinical candidate γ-secretase inhibitors (GSIs). Semagacestat (left) is nonselective for APP vis-à-vis Notch, while avagacestat (right) is reported as selective for blocking γ-secretase proteolysis Figure 9.of Clinical APP over candidate Notch. γ-secretase inhibitors (GSIs). Semagacestat (left) is nonselective for APP vis-à-vis Notch, while avagacestat (right) is reported as selective for blocking γ-secretase pro- teolysis of APPBecause over Notch. all the serious toxic effects were apparently caused by inhibition of Notch signaling, the focus then went toward finding GSIs that could selectively inhibit the prote- Becauseolysis all of APPthe serious by γ-secretase toxic effects without were affecting apparently Notch1 caused proteolysis. by inhibition This led of to Notch the discovery signaling,of so-calledthe focus “Notch-sparing”then went toward GSIs finding [73– 76GSIs], a that misleading could selectively term, as these inhibit compounds the pro- show teolysisAPP/Notch of APP by γ selectivity-secretase without and not completeaffecting Notch1 lack of effectproteolysis. on Notch This proteolysis. led to the dis- Moreover, covery ofthe so-called degree of“Notch-sparing” selectivity was GSIs a matter [73–76], of debate,a misleading with term, some as reports these compounds of a lack of any se- show APP/Notchlectivity for selectivity APP [77 and,78]. not One complete such compound, lack of effect avagacestat on Notch proteolysis. (Figure9), wentMoreo- as far as ver, thephase degree II of clinical selectivity trials, was and a like matter semagacestat of debate, causedwith some Notch-deficient reports of a toxicitieslack of any at higher selectivitydoses, for APP equivocal [77,78]. Aβ Onelowering such compound, in cerebrospinal avagacestat fluid at (Figure lower doses,9), went and as worseningfar as of cognition [79]. Evidence from mouse models suggest that the cognitive worsening may be due to increased γ-secretase substrates [80], although elevation of total Aβ, seen in plasma at low inhibitor concentrations, may be responsible [81,82]. These findings have effectively

halted further development of GSIs for AD. Interestingly though, these compounds may be repurposed for oncology, for the treatment of various cancers that involve overactive Notch signaling [83]. Molecules 2021, 26, x 12 of 17

phase II clinical trials, and like semagacestat caused Notch-deficient toxicities at higher doses, equivocal Aβ lowering in cerebrospinal fluid at lower doses, and worsening of cog- nition [79]. Evidence from mouse models suggest that the cognitive worsening may be due to increased γ-secretase substrates [80], although elevation of total Aβ, seen in plasma at low inhibitor concentrations, may be responsible [81,82]. These findings have effectively halted further development of GSIs for AD. Interestingly though, these compounds may Molecules 2021, 26, 388 12 of 17 be repurposed for oncology, for the treatment of various cancers that involve overactive Notch signaling [83].

8. γ-Secretase Modulators:8. γ-Secretase Therapeutic Modulators: Potential Therapeutic Potential While GSIs areWhile out of GSIs further are consideration out of further considerationfor AD therapeutics, for AD γ therapeutics,-secretase modu-γ-secretase modula- lators (GSMs) aretors still (GSMs) of keen are interest still of [84] keen. These interest compounds [84]. These (see compounds Figure 10 for (see examples) Figure 10 for examples) have the effect haveof lowering the effect Aβ of42 loweringlevels without Aβ42 decreasing levels without overall decreasing Aβ levels overall or otherwise Aβ levels or otherwise inhibiting generalinhibiting γ-secretase general activityγ-secretase [85,86]. activityThe decrease [85,86 in]. A Theβ42 decrease is correlated in A βwith42 is an correlated with increase in Aβ38,an thereby increase replacing in Aβ38, a thereby highly aggregation-prone replacing a highly form aggregation-prone of Aβ with a much form of Aβ with a more soluble form.much Thus, more these soluble compounds form. Thus, can these prevent compounds the formation can prevent of plaques the formation and of plaques other higher-orderand assembly other higher-order states of A assemblyβ42 in the states brain. of GSMs, Aβ42 however, in the brain. have GSMs, no effect, however, have no even at very higheffect, concentrations, even at very on high Notch concentrations, proteolysis and on Notch signaling, proteolysis nor do they and signaling,elevate nor do they γ-secretase substrates.elevate γ Presumably-secretase substrates. for these Presumablyreasons, these for thesecompounds reasons, have these shown compounds ex- have shown cellent safety profiles,excellent both safety in animal profiles, models both in and animal in human models trials. and in human trials.

Figure 10. Example γ-secretase modulators (GSMs). Figure 10. Example γ-secretase modulators (GSMs). The mechanism of action of these compounds is not entirely clear, although the corre- The mechanismlation betweenof action Aofβ these42 lowering compounds and Aisβ not38 elevationentirely clear, is relevant, although as Atheβ42 cor- is a precursor to relation betweenAβ A38.β42γ -Secretaselowering and cleaves Aβ38 the elevation Aβ42 C-terminus is relevant, to as release Aβ42 ais tetrapeptide a precursor [to45 ], and isolated Aβ38. γ-Secretaseγ-secretase cleaves the converts Aβ42 C-terminus synthetic A βto42 release to Aβ a38 tetrapeptide with release [45], of this and tetrapeptide isolated [87]. More- γ-secretase convertsover, presenilinsynthetic A mutationsβ42 to Aβ decrease38 with release the Aβ of42-to-A this tetrapeptideβ38 conversion [87]. whileMoreo- GSMs stimulate ver, presenilin mutationsit. Thus, GSMs decrease appear the A toβ decrease42-to-Aβ38 Aβ conversion42 by enhancing while GSMs the carboxypeptidase stimulate it. activity of Thus, GSMs appearγ-secretase to decrease that converts Aβ42 by this enhancing aggregation-prone the carboxypeptidase peptide to A βactivity38. of γ- secretase that convertsA critical this aggregation-prone issue with GSMs, peptide however, to A likeβ38. all anti-Aβ therapeutic strategies, is the A critical designissue with of clinical GSMs, trials however, [88]. So like far, all all reportedanti-Aβ therapeutic clinical trials strategies, with candidate is the AD therapeutic design of clinicalagents, trials including[88]. So far, GSIs, all repo GSMsrted and clinical anti-A trialsβ immunotherapy, with candidate AD have therapeu- been with individuals tic agents, includingwho already GSIs, GSMs have and AD anti-A or a pre-ADβ immunotherapy, condition called have mildbeen with cognitive individu- impairment (MCI). als who alreadyEven have in AD those or a with pre-AD MCI, condition substantial called neurodegeneration mild cognitive hasimpairment occurred, (MCI). and there are serious Even in those withconcerns MCI, thatsubstantial targeting neurod Aβ afteregeneration the onset has of symptomsoccurred, and is too there late. are A seri-β pathology in the ous concerns thatbrain targeting may appear Aβ after more the than onset 10 of years symptoms before is the too clinical late. A manifestationβ pathology in of AD [89]. As the brain may appearAβ pathology more than apparently 10 years before precedes the tau clinical pathology manifestation [4,90], andof AD tau [89]. pathology As may then Aβ pathology propagateapparently from precedes neuron tau to neuronpathology [91 ,92[4,90],], blocking and tau Aβ pathologyafter tau pathology may then is initiated may propagate fromnot neuron prevent to neuron or slow [91,92], the progression blocking of A ADβ after and tau may pathology not even is prevent initiated or may delay disease onset. not prevent or Forslow anti-A the progressionβ strategies—including of AD and GSMs—tomay not even succeed, prevent clearer or knowledgedelay disease of the pathogenic onset. For anti-Aprocessβ strategies—including and timing is needed, GSMs—to as are convenient succeed, andclearer reliable knowledge biomarkers of the and diagnostics. pathogenic process Moreover,and timing the is needed, clinical successas are convenient of GSMs isand completely reliable biomarkers dependent and on whether Aβ42 diagnostics. is indeed the pathogenic entity in AD. The reasons for the focus on Aβ42 are arguably more historic than a result of an objective search with no preconceptions. Over 100 years ago, Alois Alzheimer described extraneuronal amyloid plaques as a signature pathological characteristic of the disease, and the discovery in the late 1980s and early 1990s that the primary protein component of the plaques is Aβ [93,94], with Aβ42 being the predominant species [13] and most aggregation-prone [95], led to the assumption that Aβ42 is the likely pathogenic species. Subsequent studies ranging from effects of APP and presenilin mutations to the neurotoxicity of various Aβ assemblies would seem to confirm this hypothesis. However, selectivity in the reporting of findings (negative results are more difficult to publish) in combination with incomplete knowledge of all forms of Aβ could Molecules 2021, 26, 388 13 of 17

result in mistaking correlation for causality. In light of more recent findings, mentioned earlier, that FAD mutations in PSEN1 increase Aβ peptide intermediates of 45 residues and longer, it would seem worthwhile to determine the effects of GSMs on the first or second trimming events of APP substrate by γ-secretase (e.g., Aβ49→46, Aβ45→42).

9. Summary and Perspective The devastation of AD is severe, the numbers afflicted are large, and the need for effective therapeutics is great. Despite decades of intense efforts by laboratories around the world, in academia, government, and industry, only symptomatic treatments for AD are available, and no drugs of any kind have been approved since 2003. The large majority of clinical trials have involved agents that target Aβ in some way, to block its production or aggregation or to stimulate its clearance. γ-Secretase remains a top target, especially as altered processing of APP substrate by this protease complex is clearly involved in the pathogenesis of FAD. Dominant missense mutations in presenilins—the catalytic component of the γ-secretase complex—and APP—the γ-secretase substrate precursor to Aβ—cause FAD with virtually 100% penetrance. A single mutant allele in either the enzyme or substrate that produce Aβ fates the carrier to AD in midlife. The pathology, presentation, and progression of this genetic form of the disease is essentially indistinguishable from the much more common form that strikes in late life. Thus, it seems likely that Aβ in some form similarly plays a key role in the etiology of sporadic late-onset AD. Given the close similarities between the familial and sporadic forms of the disease, elucidating the pathogenic mechanisms of the former—a simpler problem—is likely to be illuminating for the latter. Early work had focused on the aggregation-prone Aβ42, as this Aβ variant is the principal component of the characteristic cerebral plaques of AD. In many cases, FAD mutations can elevate the ratio of Aβ42 to the more soluble Aβ40. However, processing of APP substrate by γ-secretase is complex, involving carboxypeptidase trim- ming of initially formed Aβ48 or Aβ49, and FAD mutations can elevate longer forms of Aβ that contain most of the APP TMD and remain membrane-bound. A critical question is whether these long Aβ peptides play an important role in the pathogenic process. A variety of substrate-based probes for γ-secretase have been developed. Transition- state analog inhibitors (TSAs) provided the first clue that the enzyme is an aspartyl protease, and affinity labeling reagents based on TSAs covalently bind to presenilin NTF and CTF, early evidence that the active site resides between these two subunits. TSAs also helped characterize active site pockets and suggested the existence of a separate initial substrate docking exosite. Helical peptide inhibitors (HPIs) designed to interact with this docking site identified the presenilin NTF/CTF interface as its location and close proximity to the active site. More recently, linking HPI to TSA has provided full substrate TMD-based stoichiometric inhibitors as structural probes for cryo-EM analysis to trap the protease complex in its transition state for intramembrane proteolysis. Full substrate TMD-based functional probes have also been developed to facilitate analysis of all proteolytic prod- ucts generated during the complex processing of APP substrate by γ-secretase. Toward therapeutics targeting γ-secretase, a wide variety of small-molecule inhibitors (GSIs) and modulators (GSMs) have been reported, with some going through clinical trials. GSIs can effectively lower Aβ production in vivo. However, GSIs also interfere with critical Notch signaling, causing severe side effects. More concerning, GSIs cause cognitive worsening. These devastating failures of GSIs in late-stage clinical trials has turned the field toward GSMs. Modulation rather than inhibition appears to be safe and allows specific targeting of Aβ42. These compounds apparently stimulate the Aβ42→38 trimming step, to lower the aggregation-prone peptide. Whether GSMs will prevent AD remains to be determined, however. If they are efficacious, this would be a strong argument for Aβ42 being the pathogenic entity. If they are not, exploring the effects of GSMs on longer forms of Aβ and testing potential pathogenic roles for these membrane-associated forms of the peptide would seem to be worthwhile. Molecules 2021, 26, 388 14 of 17

Funding: This research received no external funding. The APC was funded by MDPI. Conflicts of Interest: The author declares no conflict of interest.

References 1. Querfurth, H.W.; LaFerla, F.M. Alzheimer’s disease. N. Eng. J. Med. 2010, 362, 329–344. [CrossRef][PubMed] 2. Vinters, H.V. Emerging concepts in Alzheimer’s disease. Ann. Rev. Pathol. 2015, 10, 291–319. [CrossRef][PubMed] 3. Marttinen, M.; Takalo, M.; Natunen, T.; Wittrahm, R.; Gabbouj, S.; Kemppainen, S.; Leinonen, V.; Tanila, H.; Haapasalo, A.; Hiltunen, M. Molecular Mechanisms of Synaptotoxicity and Neuroinflammation in Alzheimer’s Disease. Front. Neurosci. 2018, 12, 963. [CrossRef][PubMed] 4. Jack, C.R., Jr.; Knopman, D.S.; Jagust, W.J.; Shaw, L.M.; Aisen, P.S.; Weiner, M.W.; Petersen, R.C.; Trojanowski, J.Q. Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol. 2010, 9, 119–128. [CrossRef] 5. Selkoe, D.J.; Hardy, J. The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol. Med. 2016, 8, 595–608. [CrossRef] 6. Karran, E.; Mercken, M.; De Strooper, B. The amyloid cascade hypothesis for Alzheimer’s disease: An appraisal for the development of therapeutics. Nat. Rev. Drug Discov. 2011, 10, 698–712. [CrossRef] 7. Chartier-Harlin, M.C.; Crawford, F.; Houlden, H.; Warren, A.; Hughes, D.; Fidani, L.; Goate, A.; Rossor, M.; Roques, P.; Hardy, J.; et al. Early-onset Alzheimer’s disease caused by mutations at codon 717 of the β-amyloid precursor protein gene. Nature 1991, 353, 844–846. [CrossRef] 8. Sherrington, R.; Rogaev, E.I.; Liang, Y.; Rogaeva, E.A.; Levesque, G.; Ikeda, M.; Chi, H.; Lin, C.; Li, G.; Holman, K.; et al. Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease. Nature 1995, 375, 754–760. [CrossRef] 9. Levy-Lahad, E.; Wasco, W.; Poorkaj, P.; Romano, D.M.; Oshima, J.; Pettingell, W.H.; Yu, C.E.; Jondro, P.D.; Schmidt, S.D.; Wang, K.; et al. Candidate gene for the chromosome 1 familial Alzheimer’s disease locus. Science 1995, 269, 973–977. [CrossRef] 10. Tanzi, R.E.; Bertram, L. Twenty years of the Alzheimer’s disease amyloid hypothesis: A genetic perspective. Cell 2005, 120, 545–555. [CrossRef] 11. Cole, S.L.; Vassar, R. The role of APP processing by BACE1, the β-secretase, in Alzheimer’s disease pathophysiology. J. Biol. Chem. 2008, 283, 29621–29625. [CrossRef] 12. Wolfe, M.S. γ-Secretase in biology and medicine. Sem. Cell Dev. Biol. 2009, 20, 219–224. [CrossRef][PubMed] 13. Iwatsubo, T.; Odaka, A.; Suzuki, N.; Mizusawa, H.; Nukina, N.; Ihara, Y. Visualization of Aβ42(43) and Aβ40 in senile plaques with end-specific Aβ monoclonals: Evidence that an initially deposited species is Aβ42(43). Neuron 1994, 13, 45–53. [CrossRef] 14. Available online: http://www.alzforum.org/mutations (accessed on 12 January 2021). 15. Selkoe, D.J. Cell biology of the amyloid β-protein precursor and the mechanism of Alzheimer’s disease. Annu. Rev. Cell Biol. 1994, 10, 373–403. [CrossRef][PubMed] 16. Laudon, H.; Hansson, E.M.; Melen, K.; Bergman, A.; Farmery, M.R.; Winblad, B.; Lendahl, U.; von Heijne, G.; Naslund, J. A nine-transmembrane domain topology for presenilin 1. J. Biol. Chem. 2005, 280, 35352–35360. [CrossRef] 17. Spasic, D.; Tolia, A.; Dillen, K.; Baert, V.; De Strooper, B.; Vrijens, S.; Annaert, W. Presenilin-1 maintains a nine-transmembrane topology throughout the secretory pathway. J. Biol. Chem. 2006, 281, 26569–26577. [CrossRef] 18. Citron, M.; Westaway, D.; Xia, W.; Carlson, G.; Diehl, T.; Levesque, G.; Johnson-Wood, K.; Lee, M.; Seubert, P.; Davis, A.; et al. Mutant presenilins of Alzheimer’s disease increase production of 42-residue -protein in both transfected cells and transgenic mice. Nat. Med. 1997, 3, 67–72. [CrossRef] 19. Scheuner, D.; Eckman, C.; Jensen, M.; Song, X.; Citron, M.; Suzuki, N.; Bird, T.D.; Hardy, J.; Hutton, M.; Kukull, W.; et al. Secreted amyloid β-protein similar to that in the senile plaques of Alzheimer’s disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer’s disease. Nat. Med. 1996, 2, 864–870. [CrossRef] 20. Borchelt, D.R.; Thinakaran, G.; Eckman, C.B.; Lee, M.K.; Davenport, F.; Ratovitsky, T.; Prada, C.M.; Kim, G.; Seekins, S.; Yager, D.; et al. Familial Alzheimer’s disease-linked presenilin 1 variants elevate Aβ1-42/1-40 ratio in vitro and in vivo. Neuron 1996, 17, 1005–1013. [CrossRef] 21. De Strooper, B.; Saftig, P.; Craessaerts, K.; Vanderstichele, H.; Guhde, G.; Annaert, W.; Von Figura, K.; Van Leuven, F. Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature 1998, 391, 387–390. [CrossRef] 22. Herreman, A.; Serneels, L.; Annaert, W.; Collen, D.; Schoonjans, L.; De Strooper, B. Total inactivation of γ-secretase activity in presenilin-deficient embryonic stem cells. Nat. Cell Biol. 2000, 2, 461–462. [CrossRef][PubMed] 23. Zhang, Z.; Nadeau, P.; Song, W.; Donoviel, D.; Yuan, M.; Bernstein, A.; Yankner, B.A. Presenilins are required for gamma-secretase cleavage of beta-APP and transmembrane cleavage of Notch-1. Nat. Cell. Biol. 2000, 2, 463–465. [CrossRef] 24. Wolfe, M.S.; Citron, M.; Diehl, T.S.; Xia, W.; Donkor, I.O.; Selkoe, D.J. A substrate-based difluoro ketone selectively inhibits Alzheimer’s γ-secretase activity. J. Med. Chem. 1998, 41, 6–9. [CrossRef] 25. Wolfe, M.S.; Xia, W.; Moore, C.L.; Leatherwood, D.D.; Ostaszewski, B.; Donkor, I.O.; Selkoe, D.J. Peptidomimetic probes and molecular modeling suggest Alzheimer’s γ-secretases are intramembrane-cleaving aspartyl proteases. Biochemistry 1999, 38, 4720–4727. [CrossRef][PubMed] 26. Shearman, M.S.; Beher, D.; Clarke, E.E.; Lewis, H.D.; Harrison, T.; Hunt, P.; Nadin, A.; Smith, A.L.; Stevenson, G.; Castro, J.L. L-685,458, an aspartyl protease transition state mimic, is a potent inhibitor of amyloid β-protein precursor γ-secretase activity. Biochemistry 2000, 39, 8698–8704. [CrossRef][PubMed] Molecules 2021, 26, 388 15 of 17

27. Wolfe, M.S.; Xia, W.; Ostaszewski, B.L.; Diehl, T.S.; Kimberly, W.T.; Selkoe, D.J. Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and γ-secretase activity. Nature 1999, 398, 513–517. [CrossRef] 28. Esler, W.P.; Kimberly, W.T.; Ostaszewski, B.L.; Diehl, T.S.; Moore, C.L.; Tsai, J.-Y.; Rahmati, T.; Xia, W.; Selkoe, D.J.; Wolfe, M.S. Transition-state analogue inhibitors of γ-secretase bind directly to presenilin-1. Nat. Cell Biol. 2000, 2, 428–434. [CrossRef] 29. Li, Y.M.; Xu, M.; Lai, M.T.; Huang, Q.; Castro, J.L.; DiMuzio-Mower, J.; Harrison, T.; Lellis, C.; Nadin, A.; Neduvelil, J.G.; et al. Photoactivated γ-secretase inhibitors directed to the active site covalently label presenilin 1. Nature 2000, 405, 689–694. [CrossRef] 30. Thinakaran, G.; Borchelt, D.R.; Lee, M.K.; Slunt, H.H.; Spitzer, L.; Kim, G.; Ratovitsky, T.; Davenport, F.; Nordstedt, C.; Seeger, M.; et al. Endoproteolysis of presenilin 1 and accumulation of processed derivatives in vivo. Neuron 1996, 17, 181–190. [CrossRef] 31. Thinakaran, G.; Harris, C.L.; Ratovitski, T.; Davenport, F.; Slunt, H.H.; Price, D.L.; Borchelt, D.R.; Sisodia, S.S. Evidence that levels of presenilins (PS1 and PS2) are coordinately regulated by competition for limiting cellular factors. J. Biol. Chem. 1997, 272, 28415–28422. [CrossRef] 32. Capell, A.; Grunberg, J.; Pesold, B.; Diehlmann, A.; Citron, M.; Nixon, R.; Beyreuther, K.; Selkoe, D.J.; Haass, C. The proteolytic fragments of the Alzheimer’s disease-associated presenilin-1 form heterodimers and occur as a 100-150-kDa molecular mass complex. J. Biol. Chem. 1998, 273, 3205–3211. [CrossRef][PubMed] 33. Yu, G.; Chen, F.; Levesque, G.; Nishimura, M.; Zhang, D.M.; Levesque, L.; Rogaeva, E.; Xu, D.; Liang, Y.; Duthie, M.; et al. The presenilin 1 protein is a component of a high molecular weight intracellular complex that contains β-catenin. J. Biol. Chem. 1998, 273, 16470–16475. [CrossRef][PubMed] 34. Takasugi, N.; Tomita, T.; Hayashi, I.; Tsuruoka, M.; Niimura, M.; Takahashi, Y.; Thinakaran, G.; Iwatsubo, T. The role of presenilin cofactors in the γ-secretase complex. Nature 2003, 422, 438–441. [CrossRef][PubMed] 35. Kimberly, W.T.; LaVoie, M.J.; Ostaszewski, B.L.; Ye, W.; Wolfe, M.S.; Selkoe, D.J. γ-Secretase is a membrane protein complex comprised of presenilin, nicastrin, aph-1, and pen-2. Proc. Natl. Acad. Sci. USA 2003, 100, 6382–6387. [CrossRef][PubMed] 36. Edbauer, D.; Winkler, E.; Regula, J.T.; Pesold, B.; Steiner, H.; Haass, C. Reconstitution of γ-secretase activity. Nat. Cell Biol. 2003, 5, 486–488. [CrossRef] 37. Gu, Y.; Misonou, H.; Sato, T.; Dohmae, N.; Takio, K.; Ihara, Y. Distinct intramembrane cleavage of the β-amyloid precursor resembling γ-secretase-like cleavage of Notch. J. Biol. Chem. 2001, 276, 35235–35238. [CrossRef] 38. Yu, C.; Kim, S.H.; Ikeuchi, T.; Xu, H.; Gasparini, L.; Wang, R.; Sisodia, S.S. Characterization of a presenilin-mediated amyloid precursor protein carboxyl-terminal fragment gamma. Evidence for distinct mechanisms involved in γ-secretase processing of the APP and Notch1 transmembrane domains. J. Biol. Chem. 2001, 276, 43756–43760. [CrossRef] 39. Weidemann, A.; Eggert, S.; Reinhard, F.B.; Vogel, M.; Paliga, K.; Baier, G.; Masters, C.L.; Beyreuther, K.; Evin, G. A novel var epsilon-cleavage within the transmembrane domain of the Alzheimer amyloid precursor protein demonstrates homology with Notch processing. Biochemistry 2002, 41, 2825–2835. [CrossRef] 40. Sato, T.; Dohmae, N.; Qi, Y.; Kakuda, N.; Misonou, H.; Mitsumori, R.; Maruyama, H.; Koo, E.H.; Haass, C.; Takio, K.; et al. Potential link between amyloid β-protein 42 and C-terminal fragment gamma 49–99 of β-amyloid precursor protein. J. Biol. Chem. 2003, 278, 24294–24301. [CrossRef] 41. Sastre, M.; Steiner, H.; Fuchs, K.; Capell, A.; Multhaup, G.; Condron, M.M.; Teplow, D.B.; Haass, C. Presenilin-dependent γ-secretase processing of β-amyloid precursor protein at a site corresponding to the S3 cleavage of Notch. EMBO Rep. 2001, 2, 835–841. [CrossRef] 42. Qi-Takahara, Y.; Morishima-Kawashima, M.; Tanimura, Y.; Dolios, G.; Hirotani, N.; Horikoshi, Y.; Kametani, F.; Maeda, M.; Saido, T.C.; Wang, R.; et al. Longer forms of amyloid β protein: Implications for the mechanism of intramembrane cleavage by γ-secretase. J. Neurosci. 2005, 25, 436–445. [CrossRef][PubMed] 43. Yagishita, S.; Morishima-Kawashima, M.; Ishiura, S.; Ihara, Y. Aβ46 is processed to Aβ40 and Aβ43, but not to Aβ42, in the low density membrane domains. J. Biol. Chem. 2008, 283, 733–738. [CrossRef] 44. Zhao, G.; Cui, M.Z.; Mao, G.; Dong, Y.; Tan, J.; Sun, L.; Xu, X. γ-Cleavage is dependent on ζ-cleavage during the proteolytic processing of amyloid precursor protein within its transmembrane domain. J. Biol. Chem. 2005, 280, 37689–37697. [CrossRef] [PubMed] 45. Takami, M.; Nagashima, Y.; Sano, Y.; Ishihara, S.; Morishima-Kawashima, M.; Funamoto, S.; Ihara, Y. γ-Secretase: Successive tripeptide and tetrapeptide release from the transmembrane domain of β-carboxyl terminal fragment. J. Neurosci. 2009, 29, 13042–13052. [CrossRef][PubMed] 46. Fernandez, M.A.; Klutkowski, J.A.; Freret, T.; Wolfe, M.S. Alzheimer presenilin-1 mutations dramatically reduce trimming of long amyloid β-peptides (Aβ) by γ-secretase to increase 42-to-40-residue Aβ. J. Biol. Chem. 2014, 289, 31043–31052. [CrossRef] 47. Haapasalo, A.; Kovacs, D.M. The many substrates of presenilin/γ-secretase. J. Alzheimer Dis. 2011, 25, 3–28. [CrossRef] 48. Kopan, R.; Ilagan, M.X. γ-Secretase: Proteasome of the membrane? Nat. Rev. Cell Mol. Biol. 2004, 5, 499–504. [CrossRef] 49. Selkoe, D.; Kopan, R. Notch and Presenilin: Regulated intramembrane proteolysis links development and degeneration. Annu. Rev. Neurosci. 2003, 26, 565–597. [CrossRef] 50. Wong, P.C.; Zheng, H.; Chen, H.; Becher, M.W.; Sirinathsinghji, D.J.; Trumbauer, M.E.; Chen, H.Y.; Price, D.L.; Van der Ploeg, L.H.; Sisodia, S.S. Presenilin 1 is required for Notch1 and DII1 expression in the paraxial mesoderm. Nature 1997, 387, 288–292. [CrossRef] 51. Shen, J.; Bronson, R.T.; Chen, D.F.; Xia, W.; Selkoe, D.J.; Tonegawa, S. Skeletal and CNS defects in Presenilin-1-deficient mice. Cell 1997, 89, 629–639. [CrossRef] Molecules 2021, 26, 388 16 of 17

52. Moore, C.L.; Leatherwood, D.D.; Diehl, T.S.; Selkoe, D.J.; Wolfe, M.S. Difluoro Ketone Peptidomimetics Suggest a Large S1 Pocket for Alzheimer’s γ-Secretase: Implications for Inhibitor Design. J. Med. Chem. 2000, 43, 3434–3442. [CrossRef][PubMed] 53. Bakshi, P.; Wolfe, M.S. Stereochemical analysis of (hydroxyethyl)urea peptidomimetic inhibitors of γ-secretase. J. Med. Chem. 2004, 47, 6485–6489. [CrossRef][PubMed] 54. Esler, W.P.; Das, C.; Wolfe, M.S. Probing pockets S2-S40 of the gamma-secretase active site with (hydroxyethyl)urea peptidomimet- ics. Bioorg. Med. Chem. Lett. 2004, 14, 1935–1938. [CrossRef][PubMed] 55. Güner, G.; Lichtenthaler, S.F. The substrate repertoire of γ-secretase/presenilin. Sem. Cell Dev. Biol. 2020, 105, 27–42. [CrossRef] 56. Bolduc, D.M.; Montagna, D.R.; Seghers, M.C.; Wolfe, M.S.; Selkoe, D.J. The amyloid-βforming tripeptide cleavage mechanism of γ-secretase. eLife 2016, 5, e17578. [CrossRef] 57. Esler, W.P.; Kimberly, W.T.; Ostaszewski, B.L.; Ye, W.; Diehl, T.S.; Selkoe, D.J.; Wolfe, M.S. Activity-dependent isolation of the presenilin/γ-secretase complex reveals nicastrin and a γ substrate. Proc. Natl. Acad. Sci. USA 2002, 99, 2720–2725. [CrossRef] 58. Fraering, P.C.; Ye, W.; Strub, J.M.; Dolios, G.; LaVoie, M.J.; Ostaszewski, B.L.; Van Dorsselaer, A.; Wang, R.; Selkoe, D.J.; Wolfe, M.S. Purification and Characterization of the Human γ-Secretase Complex. Biochemistry 2004, 43, 9774–9789. [CrossRef] 59. Das, C.; Berezovska, O.; Diehl, T.S.; Genet, C.; Buldyrev, I.; Tsai, J.Y.; Hyman, B.T.; Wolfe, M.S. Designed helical peptides inhibit an . J. Am. Chem. Soc. 2003, 125, 11794–11795. [CrossRef] 60. Bihel, F.; Das, C.; Bowman, M.J.; Wolfe, M.S. Discovery of a subnanomolar helical D-tridecapeptide inhibitor of γ-secretase. J. Med. Chem. 2004, 47, 3931–3933. [CrossRef] 61. Kornilova, A.Y.; Bihel, F.; Das, C.; Wolfe, M.S. The initial substrate- of γ-secretase is located on presenilin near the active site. Proc. Natl. Acad. Sci. USA 2005, 102, 3230–3235. [CrossRef] 62. Bhattarai, S.; Devkota, S.; Meneely, K.M.; Xing, M.; Douglas, J.T.; Wolfe, M.S. Design of Substrate Transmembrane Mimetics as Structural Probes for γ-Secretase. J. Am. Chem. Soc. 2020, 142, 3351–3355. [CrossRef] 63. Bai, X.C.; Yan, C.; Yang, G.; Lu, P.; Ma, D.; Sun, L.; Zhou, R.; Scheres, S.H.; Shi, Y. An atomic structure of human γ-secretase. Nature 2015, 525, 212–217. [CrossRef][PubMed] 64. Yang, G.; Zhou, R.; Zhou, Q.; Guo, X.; Yan, C.; Ke, M.; Lei, J.; Shi, Y. Structural basis of Notch recognition by human γ-secretase. Nature 2019, 565, 192–197. [CrossRef][PubMed] 65. Zhou, R.; Yang, G.; Guo, X.; Zhou, Q.; Lei, J.; Shi, Y. Recognition of the amyloid precursor protein by human γ-secretase. Science 2019, 363, eaaw0930. [CrossRef][PubMed] 66. Philip, A.T.; Devkota, S.; Malvankar, S.; Bhattarai, S.; Meneely, K.M.; Williams, T.D.; Wolfe, M.S. Designed Helical Peptides as Functional Probes for γ-Secretase. Biochemistry 2019, 58, 4398–4407. [CrossRef] 67. Haass, C.; Schlossmacher, M.G.; Hung, A.Y.; Vigo-Pelfrey, C.; Mellon, A.; Ostaszewski, B.L.; Lieberburg, I.; Koo, E.H.; Schenk, D.; Teplow, D.B.; et al. Amyloid β-peptide is produced by cultured cells during normal metabolism. Nature 1992, 359, 322–325. [CrossRef] 68. Dovey, H.F.; John, V.; Anderson, J.P.; Chen, L.Z.; de Saint Andrieu, P.; Fang, L.Y.; Freedman, S.B.; Folmer, B.; Goldbach, E.; Holsztynska, E.J.; et al. Functional γ-secretase inhibitors reduce β-amyloid peptide levels in brain. J. Neurochem. 2001, 76, 173–181. [CrossRef] 69. Barten, D.M.; Guss, V.L.; Corsa, J.A.; Loo, A.T.; Hansel, S.B.; Zheng, M.; Munoz, B.; Srinivasan, K.; Wang, B.; Robertson, B.J.; et al. Dynamics of β-Amyloid Reductions in Brain, Cerebrospinal Fluid and Plasma of β-Amyloid Precursor Protein Transgenic Mice Treated with a γ-Secretase Inhibitor. J. Pharmacol. Exp. Ther. 2004, 27, 27. [CrossRef] 70. Searfoss, G.H.; Jordan, W.H.; Calligaro, D.O.; Galbreath, E.J.; Schirtzinger, L.M.; Berridge, B.R.; Gao, H.; Higgins, M.A.; May, P.C.; Ryan, T.P. Adipsin: A biomarker of gastrointestinal toxicity mediated by a functional γsecretase inhibitor. J. Biol. Chem. 2003, 278, 46107–46116. [CrossRef] 71. Wong, G.T.; Manfra, D.; Poulet, F.M.; Zhang, Q.; Josien, H.; Bara, T.; Engstrom, L.; Pinzon-Ortiz, M.; Fine, J.S.; Lee, H.J.; et al. Chronic treatment with the γ-secretase inhibitor LY-411,575 inhibits β-amyloid peptide production and alters lymphopoiesis and intestinal cell differentiation. J. Biol. Chem. 2004, 279, 12876–12882. [CrossRef] 72. Doody, R.S.; Raman, R.; Farlow, M.; Iwatsubo, T.; Vellas, B.; Joffe, S.; Kieburtz, K.; He, F.; Sun, X.; Thomas, R.G.; et al. A phase 3 trial of semagacestat for treatment of Alzheimer’s disease. N. Eng. J. Med. 2013, 369, 341–350. [CrossRef][PubMed] 73. Netzer, W.J.; Dou, F.; Cai, D.; Veach, D.; Jean, S.; Li, Y.; Bornmann, W.G.; Clarkson, B.; Xu, H.; Greengard, P. Gleevec inhibits β-amyloid production but not Notch cleavage. Proc. Natl. Acad. Sci. USA 2003, 100, 12444–12449. [CrossRef][PubMed] 74. Fraering, P.C.; Ye, W.; Lavoie, M.J.; Ostaszewski, B.L.; Selkoe, D.J.; Wolfe, M.S. γ-Secretase substrate selectivity can be modulated directly via interaction with a nucleotide binding site. J. Biol. Chem. 2005, 280, 41987–41996. [CrossRef][PubMed] 75. Mayer, S.C.; Kreft, A.F.; Harrison, B.; Abou-Gharbia, M.; Antane, M.; Aschmies, S.; Atchison, K.; Chlenov, M.; Cole, D.C.; Comery, T.; et al. Discovery of begacestat, a Notch-1-sparing γ-secretase inhibitor for the treatment of Alzheimer’s disease. J. Med. Chem. 2008, 51, 7348–7351. [CrossRef] 76. Gillman, K.W.; Starrett, J.E.; Parker, M.F.; Xie, K.; Bronson, J.J.; Marcin, L.R.; McElhone, K.E.; Bergstrom, C.P.; Mate, R.A.; Williams, R.; et al. Discovery and Evaluation of BMS-708163, a Potent, Selective and Orally Bioavailable γ-Secretase Inhibitor. ACS Med. Chem. Lett. 2010, 1, 120–124. [CrossRef] 77. Crump, C.J.; Castro, S.V.; Wang, F.; Pozdnyakov, N.; Ballard, T.E.; Sisodia, S.S.; Bales, K.R.; Johnson, D.S.; Li, Y.M. BMS-708,163 targets presenilin and lacks notch-sparing activity. Biochemistry 2012, 51, 7209–7211. [CrossRef] Molecules 2021, 26, 388 17 of 17

78. Chavez-Gutierrez, L.; Bammens, L.; Benilova, I.; Vandersteen, A.; Benurwar, M.; Borgers, M.; Lismont, S.; Zhou, L.; Van Cleynenbreugel, S.; Esselmann, H.; et al. The mechanism of γ-secretase dysfunction in familial Alzheimer disease. EMBO J. 2012, 31, 2261–2274. [CrossRef] 79. Coric, V.; van Dyck, C.H.; Salloway, S.; Andreasen, N.; Brody, M.; Richter, R.W.; Soininen, H.; Thein, S.; Shiovitz, T.; Pilcher, G.; et al. Safety and tolerability of the γ-secretase inhibitor avagacestat in a phase 2 study of mild to moderate Alzheimer disease. Arch. Neurol. 2012, 69, 1430–1440. [CrossRef] 80. Mitani, Y.; Yarimizu, J.; Saita, K.; Uchino, H.; Akashiba, H.; Shitaka, Y.; Ni, K.; Matsuoka, N. Differential Effects between γ-Secretase Inhibitors and Modulators on Cognitive Function in Amyloid Precursor Protein-Transgenic and Nontransgenic Mice. J. Neurosci. 2012, 32, 2037–2050. [CrossRef] 81. Lanz, T.A.; Karmilowicz, M.J.; Wood, K.M.; Pozdnyakov, N.; Du, P.; Piotrowski, M.A.; Brown, T.M.; Nolan, C.E.; Richter, K.E.; Finley, J.E.; et al. Concentration-dependent modulation of amyloid-β in vivo and in vitro using the γ-secretase inhibitor, LY-450139. J. Pharmacol. Exp. Ther. 2006, 319, 924–933. [CrossRef] 82. Barnwell, E.; Padmaraju, V.; Baranello, R.; Pacheco-Quinto, J.; Crosson, C.; Ablonczy, Z.; Eckman, E.; Eckman, C.B.; Ramakrishnan, V.; Greig, N.H.; et al. Evidence of a Novel Mechanism for Partial γ-Secretase Inhibition Induced Paradoxical Increase in Secreted Amyloid βProtein. PLoS ONE 2014, 9, e91531. [CrossRef][PubMed] 83. Takebe, N.; Nguyen, D.; Yang, S.X. Targeting in cancer: Clinical development advances and challenges. Pharmacol. Ther. 2014, 141, 140–149. [CrossRef] 84. Bursavich, M.G.; Harrison, B.A.; Blain, J.F. γ-Secretase Modulators: New Alzheimer’s Drugs on the Horizon? J. Med. Chem. 2016, 59, 7389–7409. [CrossRef][PubMed] 85. Weggen, S.; Eriksen, J.L.; Das, P.; Sagi, S.A.; Wang, R.; Pietrzik, C.U.; Findlay, K.A.; Smith, T.E.; Murphy, M.P.; Bulter, T.; et al. A subset of NSAIDs lower amyloidogenic Aβ42 independently of cyclooxygenase activity. Nature 2001, 414, 212–216. [CrossRef] [PubMed] 86. Golde, T.E.; Koo, E.H.; Felsenstein, K.M.; Osborne, B.A.; Miele, L. γ-Secretase inhibitors and modulators. Biochim. Biophys. Acta 2013, 1828, 2898–2907. [CrossRef][PubMed] 87. Okochi, M.; Tagami, S.; Yanagida, K.; Takami, M.; Kodama, T.S.; Mori, K.; Nakayama, T.; Ihara, Y.; Takeda, M. γ-Secretase modulators and presenilin 1 mutants act differently on presenilin/γ-secretase function to cleave Aβ42 and Aβ43. Cell Rep. 2013, 3, 42–51. [CrossRef] 88. Golde, T.E.; Schneider, L.S.; Koo, E.H. Anti-Aβ therapeutics in Alzheimer’s disease: The need for a paradigm shift. Neuron 2011, 69, 203–213. [CrossRef] 89. Bateman, R.J.; Xiong, C.; Benzinger, T.L.; Fagan, A.M.; Goate, A.; Fox, N.C.; Marcus, D.S.; Cairns, N.J.; Xie, X.; Blazey, T.M.; et al. Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N. Eng. J. Med. 2012, 367, 795–804. [CrossRef] 90. Hardy, J.; Allsop, D. Amyloid deposition as the central event in the aetiology of Alzheimer’s disease. Trends Pharmacol. Sci. 1991, 12, 383–388. [CrossRef] 91. de Calignon, A.; Polydoro, M.; Suarez-Calvet, M.; William, C.; Adamowicz, D.H.; Kopeikina, K.J.; Pitstick, R.; Sahara, N.; Ashe, K.H.; Carlson, G.A.; et al. Propagation of tau pathology in a model of early Alzheimer’s disease. Neuron 2012, 73, 685–697. [CrossRef] 92. Liu, L.; Drouet, V.; Wu, J.W.; Witter, M.P.; Small, S.A.; Clelland, C.; Duff, K. Trans-synaptic spread of tau pathology in vivo. PLoS ONE 2012, 7, e31302. [CrossRef][PubMed] 93. Glenner, G.G.; Wong, C.W. Alzheimer’s disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem. Biophys. Res. Comm. 1984, 120, 885–890. [CrossRef] 94. Masters, C.L.; Simms, G.; Weinman, N.A.; Multhaup, G.; McDonald, B.L.; Beyreuther, K. Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc. Natl. Acad. Sci. USA 1985, 82, 4245–4249. [CrossRef][PubMed] 95. Jarrett, J.T.; Berger, E.P.; Lansbury, P.T., Jr. The carboxy terminus of the βamyloid protein is critical for the seeding of amyloid formation: Implications for the pathogenesis of Alzheimer’s disease. Biochemistry 1993, 32, 4693–4697. [CrossRef]