- Email: [email protected]
Alzheimer's therapy development: A few points to consider
CHAPTER TEN
Alzheimer’s therapy development: A few points to consider Einar M. Sigurdsson* Departments of Neuroscience and Physiology, and Psychiatry, Neuroscience Institute, New York University School of Medicine, New York, NY, United States *Corresponding author: e-mail address: [email protected]
Contents 1. Introduction 2. Targeting tau intracellularly or extracellularly 3. Drug-screening models: Preventing seeding/spread vs neurotoxicity 4. Strains of Aβ and tau: Influence on therapeutic development 5. Targets other than Aβ and tau 6. Concluding remarks Acknowledgments References
206 207 210 212 214 215 215 216
Abstract Development of therapies for Alzheimer’s disease has only resulted in a few approved drugs that provide some temporary symptomatic relief in certain patients. None of these compounds in clinical use halts or slows the progression of the disease. To date, several drugs targeting the amyloid-β peptide, and some against the tau protein, have failed in clinical trials. While there are various reasons for these failures, considering the following points may aid in improving the outcome of future trials. First, the tau protein should ideally be targeted intracellularly because most of tau pathology is within cells, neurons in particular. Second, an overriding emphasis in recent years has been on implementing drug-screening models that focus on prevention of seeding/spread of aggregates. Much less attention has been paid to identify compounds that inhibit neurotoxicity of these aggregates, which may not necessarily relate to their seeding/spread propensity. Ideally, all these markers should be readouts in the same assay or model. Third, diversity in conformers/strains of aggregates complicates drug development of small molecule aggregation inhibitors but is likely to be less of an issue for antibody-based treatments. Lastly, other more general targets associated with neurodegeneration should continue to be pursued but are in many ways more difficult to address than clearing amyloid-β and tau, the defining hallmarks of AD.
Progress in Molecular Biology and Translational Science, Volume 168 ISSN 1877-1173 https://doi.org/10.1016/bs.pmbts.2019.06.001
#
2019 Elsevier Inc. All rights reserved.
205
206
Einar M. Sigurdsson
1. Introduction Research on Alzheimer’s disease (AD) has expanded greatly in the last few decades and many important advances have been made in clarifying the etiology and pathogenesis of the disease, in the development of diagnostic probes, and in identifying therapeutic targets. However, this progress has been laced with several disappointing outcomes in clinical trials targeting the key hallmarks of the disease, the amyloid-β peptide (Aβ) and the tau protein, as well as in trials for other targets thought to be associated with the disease. Most of the failures have been in trials targeting Aβ since it received earlier and more extensive attention as a target for therapy than the tau protein. The reasons for more drug development effort having been spent on Aβ are logical and relate largely to the discovery of mutations in the amyloid precursor protein (APP) around its enzymatic cleavage sites that generate Aβ. These APP mutations increase Aβ levels or its more pathogenic Aβ42 form, and are associated with a small number of familial cases of AD. In addition, Aβ is mostly found extracellularly in AD and thereby is a more accessible target than tau, which is primarily an intracellular protein. The smaller size of Aβ (40–42 amino acids) compared to tau (352–441 amino acid) also render it easier to study and target with therapies. In hindsight, the failures of the Aβ targeting therapies are not particularly surprising, since when clinical symptoms are evident, Aβ levels have or are about to plateau and Aβ pathology may no longer be closely involved in the progression of the disease. APP is primarily located in synapses and increased generation, or decreased clearance, of Aβ (or other APP fragments) leads to synaptic damage, influx of calcium, activation of protein kinases, hyperphosphorylation of tau and subsequent cytoskeletal instability, which leads to further degeneration of synapses and accelerates their loss. Hence, the source of Aβ diminishes as the disease progresses, and it may be too late to target Aβ when memory impairments have become obvious. Sporadic AD cases have a relatively unknown etiology and no reliable presymptomatic biomarkers, and thus cannot be treated prophylactically before clinical symptoms have manifested. However, ongoing Aβ targeted immunotherapy trials on non-symptomatic individuals with familial forms of AD will hopefully be more positive. A comparable approach on control subjects with high Aβ load detected on PET scans may also provide valuable insight, although it is unclear at this point what percentage of these subjects would develop AD.
Alzheimer’s therapy development
207
Within the tau field, more than half of the trials are on immunotherapies and are still in the Phase 1–2 stages, which focus on safety and pharmacokinetics rather than efficacy.1,2 Of the 10 trials that have started, only 1 has failed. It occurred early in Phase 1 and it may have been because of a short half-life of the antibody since there were apparently no safety issues. It was being conducted in healthy subjects that likely did not have much if any of the phospho-tau epitope being targeted (P-Ser422). Hence, target engagement could not have been properly analyzed, and lack thereof could then not have been the reason for halting the trial. Beside tau targeting immunotherapies, various tau therapeutic approaches have entered clinical trials and some have failed. Those that have been discontinued include a microtubule stabilizer to counteract presumed loss of function of hyperphosphorylated tau, an aggregation inhibitor to inhibit tau assembly or promote its disassembly, and a GSK-3β enzyme inhibitor to block phosphorylation of tau on certain sites. Similar or other approaches are in Phase 1–2 and have been outlined elsewhere.2 In this overview, I would like to focus on the following three issues relevant to AD therapy and/or pathogenesis: (1) Is it likely to be sufficient to target tau extracellularly or is an intracellular interaction necessary for clinical benefits? (2) In drug-screening models, is prevention of seeding/spread of the protein aggregates the key outcome measure or should the focus be on or include prevention of neurotoxicity? (3) How may diversity in conformers/strains of Aβ and tau influence therapeutic development?
2. Targeting tau intracellularly or extracellularly Tau is primarily an intracellular protein, both in its normal physiological form and in its pathological form as aggregates of various sizes that can form neurofibrillary tangles. Low levels of tau have been measured in brain interstitial fluid and cerebrospinal fluid (CSF) in experimental animals and in CSF in humans. Antibodies have been shown to target tau intra- and extracellularly. The primary form of tau that is being targeted is likely to be soluble, although it presumably consists of small aggregates.1,2 Evidently, secreted tau is soluble and extracellular ghost tangles are relatively few and not a very useful target. Intracellularly, antibodies associate mainly with tau in the endosomal-lysosomal system and the small tau aggregates in these vesicles can be considered rather soluble. Hence, soluble tau is likely the
208
Einar M. Sigurdsson
primary target of the antibodies and thus it is helpful to consider the size of the intra- vs extracellular pools of soluble tau. A recent study assessed these two pools of tau in a group of Alzheimer’s patients and subjects with mild cognitive impairment, compared to age-matched controls.3 Their findings indicate that CSF tau is about 0.001–0.0001% of total brain soluble tau. It is likely that tau released from cells has been largely degraded when it reaches the CSF and the extent of this is not clear in human subjects. However, mouse studies indicate that CSF tau is approximately 10% of tau levels in brain interstitial fluid, although these levels do not correlate in individual animals.4 Assuming a similar relationship in humans, it can be estimated that extracellular tau available to interact with antibodies is then about 0.01–0.001% of intracellular soluble tau. Measurements in a more artificial culture system echo these differences with extracellular tau measuring about 0.1% of intracellular tau.5 Now, in this context, one has to consider that most of soluble tau is in the cytosol and not readily accessible to the antibodies. However, at least some tau antibodies find their way into the cytosol following endocytic uptake, presumably because of leakage or rupture of endosome and lysosomes.6 Once there, the antibodies have access to various forms of tau. These include soluble physiological forms of tau to which antibodies that recognize normal tau epitopes can bind. This would likely initiate the clearance of this complex, possibly via TRIM 21 binding or other proteasomal degradation pathways.7 Such clearance of normal tau could be considered to have detrimental consequences. However, it is unlikely to be extensive and other microtubule-associated proteins can take over tau functions as evident by the relatively limited deficits seen in tau knockout mice. Reducing normal tau levels with an antisense oligonucleotide is also being pursued in clinical trials.8 In diseased neurons, the cytosol will also contain tau aggregates of various sizes that are derived from ruptured lysosomes. Neuronal accumulation of lysosomes filled with tau aggregates is a well-known phenomenon in AD and related transgenic tauopathy mouse models.9 Presumably, with advancing tau pathology, the lysosomal clearance system exceeds its capacity, resulting in accumulation of lysosomes with tau aggregates that eventually leak into the cytosol. There, these small rather soluble aggregates then seed the formation of larger aggregates, which eventually results in the formation of neurofibrillary tangles. The tangles may be rather inert but eventually these neurons die, perhaps in part because the accumulating fibrils affect the function of various cellular organs. Antibody interaction with these larger aggregates may in some instances, depending on the epitope, prevent further aggregation and/or lead to the disassembly of the aggregates. Overall,
209
Alzheimer’s therapy development
this may be beneficial or detrimental, depending on if these aggregates cause dysfunction or have been assembled to prevent toxicity caused by the smaller assemblies. The stage of the disease may also matter in this context. Regardless of the exact mechanism of action of antibody-mediated clearance of tau pathology, it is very clear that intracellular tau levels exceed extracellular tau levels by several orders of magnitude (Fig. 1A).
A
B
Extracellular Tau
Neurotoxicity
Intracellular
Tau
seeding/ sequestration
propagation
Conformers / Strains
C
Small molecules
Antibodies
Fig. 1 (A) The tau protein is primarily found intracellularly so the most efficacious therapies targeting tau need to be able to enter the cell to access its largest pool. (B) The seeding/sequestration and propagation pathway may be less neurotoxic (broken arrows) than other misfolding/aggregation pathway (solid arrows). (C) Small molecules may be less likely to recognize different conformers/strains than antibodies. To make this point, the panel depicts three different small molecules that are specific for each of the three conformers whereas one of the two antibodies shown recognizes two out of three conformers equally well.
210
Einar M. Sigurdsson
Considering these large differences in intra- vs extracellular pools of tau, an antibody that gets into neurons is likely to be more efficacious than an antibody that can only work extracellularly. Our findings indicate that antibody charge has a great influence on antibody uptake and that antibody chimerization/humanization can greatly affect its charge and thereby uptake.10 This needs to be carefully examined for clinical candidate antibodies. However, the extent of animal studies that can be conducted on humanized antibodies is rather limited because of species related immunogenicity issues. It is conceivable to use mice with human-like immune system for this type of studies but it is unclear how such a drastic alteration affects tau pathology in transgenic models. However, the efficacy of these humanized antibodies can be easily examined in human-derived culture systems. We have shown in such a model that a partial humanization of an antitau mouse monoclonal antibody, with well-established therapeutic potential in various models, prevents its intracellular efficacy in blocking tau toxicity and promoting tau clearance.6,10–13 The reason for this loss of efficacy is that the chimeric antibody is not taken up into neurons because of substantial change in antibody charge. It is not clear how well the current humanized antibodies have been re-examined for efficacy prior to clinical trials and it raises concerns that some of these may fail in clinical trials for reasons unrelated to their direct interaction with the target. With regard to binding itself, we have reported that at least for one prominent tau epitope, the phospho-serine 396,404 region, stronger binding does not convey improved efficacy.10,12,13 This important finding should also be kept in mind as the humanization may also influence binding, even though the binding regions remain the same, as we have reported.10 Lastly, tau levels in the CSF are not altered in non-AD tauopathies, suggesting that tau lesions in these subjects may spread in a cell autonomous manner, not involving extracellular tau. Therefore, in those subjects, targeting tau intracellularly is even more important than in AD.
3. Drug-screening models: Preventing seeding/spread vs neurotoxicity It is now well-established that various peptide/protein aggregates can be seeded in vitro and in vivo, and that these seeds can spread through anatomically connected pathways.14–16 Incorporating this phenomenon into antibody efficacy studies has been popular and is in many ways appropriate.
Alzheimer’s therapy development
211
However, we have observed that antibodies’ ability to prevent tau seeding and block tau toxicity does not always go hand in hand.10 For example, one antibody that we developed against the P-[Thr212, Ser214] tau region, prevented tau seeding but not tau toxicity.10 This can be explained by the tau species to which it binds, which may promote seeding without influencing toxicity. Antibodies against some other regions can prevent both seeding and toxicity.10 With toxicity, we mean loss of neurons as detected by LDH ELISA assay or retraction of neuronal processes as detected by reduced NeuN signal on Western blots. At least in culture, the toxicity pathway appears to be more acute whereas the seeding pathway is more gradual. It can be envisioned in vivo that certain forms of tau assemblies may have an acute toxic effect, whereas other tau forms may be more prone to initiate tau seeding. Such seeding is eventually likely to have detrimental effects but may to some extent be an effort of the cell to sequester toxic forms of tau (Fig. 1B). Hence, prevention of seeding may unleash some toxic forms of tau that are better kept sequestered in aggregates. Reports from the prion field supports this view, showing that prion replication can be separated from toxicity and that some forms of prion aggregates may actually delay the onset of clinical symptoms.17,18 Specifically, levels of the infectious scrapie form of the prion protein, PrPSc, have been shown to plateau before clinical onset,17 analogous to Aβ levels in AD. While this may also relate to neuronal reserve that can be differentially affected by PrPSc and other perhaps more toxic forms of PrP in different models, it is likely that different species of the amyloidogenic peptide/protein in amyloid diseases can have varying effects on seeding/propagation and toxicity. However, this is not always the case, at least in vitro. A culture study on tau strains reported that seeding efficiency correlated with inhibition of growth.19 How this plays out in vivo is less clear as in that particular study, tau strain inoculations led to robust tau pathology without overt neuronal loss.19 As pointed out above, using antibodies as tools to neutralize certain tau species provides an insight into this matter, showing some overlap between seeding and toxicity but also that these can proceed via separate pathways. To reiterate, how antibodies affect the different pathways of acute toxicity vs chronic spreading is likely to be epitope-dependent, as we have observed for some of our antibodies.10 This and their efficacy in general may also vary depending on subtle binding differences within the same epitope region.12 It may also depend on the exact tauopathy under study. To choose the best antibodies to advance to clinical trials, it is important to include various assays that can distinguish between these two pathways,
212
Einar M. Sigurdsson
the acute toxicity vs chronic seeding/toxicity. Ideally, a clinical candidate antibody should perform well in both types of assays, assuming that preventing the latter does not exacerbate the former. Hence, each model should ideally allow examining both pathways at the same time as we have done in culture models.10,12 Subsequent in vivo studies should likewise incorporate both pathways into data analysis. Based on publications in this field, it appears that much more attention is being paid to interfering with the seeding/aggregation/propagation pathway (protein-protein interaction) as reflected by reports on the development of and testing of drug candidates in high throughput in vitro assays. Prevention of toxicity is typically a secondary measure conducted on key hits from the protein-protein interaction assay. Hence, these assays miss compounds that may be preferable for clinical benefits. This is also evident in in vivo assays that focus on the ability of drug candidates to prevent seeding and spread in animal models. Neurotoxicity/ neuronal loss may not be very evident in such injection models and is more time consuming to analyze. The neurotoxicity pathway may be more relevant for identifying therapies that more quickly provide cognitive improvements, whereas the functional benefits of targeting the more chronic seeding pathway may take longer to manifest themselves.
4. Strains of Aβ and tau: Influence on therapeutic development In recent years, it has become popular to refer to strains of Aβ and tau, analogous to the well-established concept of prion strains. Within the prion field, a definition of a strain is that it retains its properties following multiple passages, ideally between animals. There is some indication that this may occur for certain Aβ and tau preparations but the fidelity of this propagation is not yet entirely clear, at least not in animal models. Changes in tau profiles with continued passages can indicate mixtures of strains and the presence of substrains.20 It can also be interpreted to reveal certain conformational variability that does not clearly support the presence of a strain. A key reason why the presence of Aβ and tau strains is not as clear as for PrP prions is that unlike PrPSc, Aβ and tau aggregates are much less transmissible, presumably because they are more easily degraded. Hence, because of subtle differences in degradation profiles, some diversity of Aβ or tau profiles can be expected. Regarding how possible strain/conformer presence may affect therapeutic development, this is likely to differ for small molecules vs antibodies. Generally speaking, development of small molecule aggregation inhibitors
Alzheimer’s therapy development
213
targeting Aβ or tau has been disappointing. Several classes of small molecules have been identified that recognize the characteristic β-sheet conformation, which defines amyloids in general. These have been effective in various in vitro assays and some in vivo models but very few have advanced to clinical trials, presumably because of low blood-brain barrier permeability and/or toxicity. Those that have advanced to late stage clinical trials have either failed to show efficacy or are still under study, in what appears to be limited capacity as defined by small trial size.21 Some insight into how this may play out can be derived from related positron emission tomography (PET) studies on compounds targeting Aβ or tau. Approved PET ligands targeting Aβ are comparable in many ways, which may relate to the fact that Aβ deposits are large and extracellular. Even if the different Aβ ligands may have different affinities for certain β-sheet subtypes of Aβ, each deposit likely comprises a diversity of β-sheet structures because it is formed by accretion of many different types of large Aβ assemblies. However, at least some of the Aβ PET ligands, like PiB, do not bind well to plaques in most mouse models, which can be explained by differences in binding sites within Aβ in humans vs mice.22 Hence, potential therapeutic β-sheet binders that are effective in the mouse may not work in humans, and vice versa. A similar scenario can be envisioned in the tau field, but it is not as clearcut. Tau aggregates are primarily found intracellularly and the various small molecule tau β-sheet ligands show some diversity in their binding profile.23 Many of the first tau ligands that were examined only recognized tau in Alzheimer’s patients but not in other tauopathies. Those early findings suggest differences in tau conformations in the different tauopathies, as recently shown more clearly in cryo-electron microscopy studies.24,25 In addition, some of the small molecule tau PET ligands turned out to bind strongly to unrelated proteins.23 Even if all the tau deposits from different tauopathies contain β-sheets and all the ligands are defined as β-sheet binders, subtle conformational differences likely explain why the ligands bind differently to tau from different tauopathies. However, to date, at least some of the ligands have been reported to give a positive PET signal in tauopathies other than AD.23 Future studies will likely reveal the molecular explanation for these differences in binding profile. In summary, for these small molecule β-sheet binders to work as a therapy, it may be necessary to develop strain/conformer specific inhibitors, at least for the tau protein (Fig. 1C). Because of its size, disordered structure, different isoforms and diversity of posttranslational modifications, its
214
Einar M. Sigurdsson
secondary structure is much more variable than for Aβ. Beside these issues, there are some concerns that chronic targeting of β-sheets is likely to result in adverse reactions. Various proteins have at least some segments of β-sheets, and β-sheet assemblies are used, for example, to store hormones for later use.26 In addition, chronic targeting of a subgroup of tau strains/conformers, may lead to an expansion of other tau subgroups, analogous to changes in bacterial flora following antibiotic treatment, which can have unforeseen detrimental consequences. For antibody therapies, this is likely to be less of an issue as many antibodies against a particular epitope will recognize different strains/conformers (Fig. 1C). For example, most of tau antibodies will recognize tau deposits in all tauopathies. The exception may be certain antibodies against discontinuous epitopes (conformational) and antibodies against regions that are buried in some strains/conformers. For example, for tau, an antibody directed at P-Ser262 binds well to tau aggregates in AD but not in Pick’s disease, in which this epitope is buried.25 Typically, the N- and C-terminal regions of full-length tau are accessible to antibodies, as well as large segments of the middle region. However, tau exists in various fragments in the brain and it is unclear if some of these may constitute certain strains/conformers of tau, which would then be differentially affected by antibodies against different epitopes. Overall, it is unlikely that subtle structural differences between strains/ subtypes of tau will greatly affect therapeutic efficacy of most antibodies. This ability of most antibodies to recognize all the different tauopathies bodes also well for their diagnostic potential. Whole antibodies are likely too large to be used as PET ligands for brain imaging but small antibody derivatives have great potential as imaging agents because of better brain penetration.27 Compared to β-sheet small molecule dye based ligands, the antibody derivatives should be more specific and can provide important information on the epitope profile of tau aggregates. That information can then be used to personalize antibody treatment to target these aggregates.
5. Targets other than Aβ and tau It makes sense to focus on Aβ and tau aggregates for AD therapy as they are the defining hallmarks of AD. However, there has always been interest in other more general targets related to neurodegeneration, such as promoting synaptic health, manipulating inflammation/microglial
Alzheimer’s therapy development
215
activation by decreasing or enhancing it, influencing lipid metabolism via ApoE or by other means, and enhancing intracellular protein degradation via proteasomal or lysosomal degradation. While these are certainly valid targets, they are fraught with their own difficulties that are more formidable than targeting Aβ and tau. Cognitive enhancers designed to improve synaptic health have so far not worked in the clinic. The benefits of inhibiting or promoting inflammation/microglial activation may depend on the severity of the pathology, which varies between brain regions as the disease progresses. Therefore, what is beneficial for the entorhinal cortex at some stage of the disease may be detrimental for the neocortex at the same time. Lipid metabolism is complex and not well understood. ApoE targeted approaches have not yet made it into the clinic over 25 years after its association with AD was discovered. Promoting intracellular protein degradation by activating the pathways involved is a double-edged sword that can be very toxic and is difficult to fine-tune, which explains why some of the experimental compounds that have been used to study these pathways have not made it into the clinic.
6. Concluding remarks While there are many explanations for the failures of drugs to attenuate progression of AD, there are reasons to be optimistic about the future of therapeutic development in this field. In particular, targeting the tau protein is likely to provide therapeutic benefits later in the disease process than Aβ since tau pathology correlates better with the degree of dementia than Aβ lesions. However, as outlined in this brief overview, several uncertainties remain about the ideal preclinical features of a drug that is likely to succeed in clinical trials. I hope that the points raised here will improve the selection process and thereby accelerate the development of effective therapies. Other more general targets associated with neurodegeneration should continue to be pursued but are in many ways more difficult to address than targeting the defining hallmarks of AD.
Acknowledgments E.M.S. is currently funded by NIH grants R01 AG032611, R01 NS077239, R21 AG058282, and R21 AG059391. He is an inventor on several patents that are assigned to New York University. Some of the patents that focus on tau immunotherapy and related diagnostics are licensed to and are being co-developed with H. Lundbeck A/S.
216
Einar M. Sigurdsson
References 1. Sigurdsson EM. Tau immunotherapies for Alzheimer’s disease and related tauopathies: progress and potential pitfalls. J Alzheimers Dis. 2018;64:S555–S565. 2. Congdon EE, Sigurdsson EM. Tau-targeting therapies for Alzheimer disease. Nat Rev Neurol. 2018;14:399–415. 3. Han P, Serrano G, Beach TG, et al. A quantitative analysis of brain soluble tau and the tau secretion factor. J Neuropathol Exp Neurol. 2017;76:44–51. 4. Yamada K, Cirrito JR, Stewart FR, et al. In vivo microdialysis reveals age-dependent decrease of brain interstitial fluid tau levels in P301S human tau transgenic mice. J Neurosci. 2011;31:13110–13117. 5. Chai X, Dage JL, Citron M. Constitutive secretion of tau protein by an unconventional mechanism. Neurobiol Dis. 2012;48:356–366. 6. Gu J, Congdon EE, Sigurdsson EM. Two novel tau antibodies targeting the 396/404 region are primarily taken up by neurons and reduce tau protein pathology. J Biol Chem. 2013;288:33081–33095. 7. McEwan WA, Falcon B, Vaysburd M, et al. Cytosolic Fc receptor TRIM21 inhibits seeded tau aggregation. Proc Natl Acad Sci U S A. 2017;114:574–579. 8. ClinicalTrials gov. Safety, Tolerability, Pharmacokinetics, and Pharmacodynamics of IONISMAPTRx in Patients With Mild Alzheimer’s Disease. https://www.clinicaltrials.gov/ct2/ show/NCT03186989 US National Library of Medicine, ClinicalTrials gov; 2019. 9. Lin WL, Lewis J, Yen SH, Hutton M, Dickson DW. Ultrastructural neuronal pathology in transgenic mice expressing mutant (P301L) human tau. J Neurocytol. 2003;32:1091–1105. 10. Congdon EE, Chukwu JE, Shamir DB, et al. Tau antibody chimerization alters its charge and binding, thereby reducing its cellular uptake and efficacy. EBioMedicine. 2019;42:157–173. 11. Congdon EE, Gu J, Sait HB, Sigurdsson EM. Antibody uptake into neurons occurs primarily via clathrin-dependent fcgamma receptor endocytosis and is a prerequisite for acute tau protein clearance. J Biol Chem. 2013;288:35452–35465. 12. Congdon EE, Lin Y, Rajamohamedsait HB, et al. Affinity of tau antibodies for solubilized pathological tau species but not their immunogen or insoluble tau aggregates predicts in vivo and ex vivo efficacy. Mol Neurodegener. 2016;11:62–85. 13. Wu Q, Lin Y, Gu J, Sigurdsson EM. Dynamic assessment of tau immunotherapies in the brains of live animals by two-photon imaging. EBioMedicine. 2018;35:270–278. 14. Sigurdsson EM, Wisniewski T, Frangione B. Infectivity of amyloid diseases. Trends Mol Med. 2002;8:411–413. 15. Goedert M, Falcon B, Clavaguera F, Tolnay M. Prion-like mechanisms in the pathogenesis of tauopathies and synucleinopathies. Curr Neurol Neurosci Rep. 2014;14:495. 16. Walker LC, Diamond MI, Duff KE, Hyman BT. Mechanisms of protein seeding in neurodegenerative diseases. JAMA Neurol. 2013;70:304–310. 17. Sandberg MK, Al-Doujaily H, Sharps B, Clarke AR, Collinge J. Prion propagation and toxicity in vivo occur in two distinct mechanistic phases. Nature. 2011;470:540–542. 18. Diaz-Espinoza R, Morales R, Concha-Marambio L, Moreno-Gonzalez I, Moda F, Soto C. Treatment with a non-toxic, self-replicating anti-prion delays or prevents prion disease in vivo. Mol Psychiatry. 2018;23:777–788. 19. Kaufman SK, Sanders DW, Thomas TL, et al. Tau prion strains dictate patterns of cell pathology, progression rate, and regional vulnerability in vivo. Neuron. 2016;92:796–812. 20. Sharma AM, Thomas TL, Woodard DR, Kashmer OM, Diamond MI. Tau monomer encodes strains. Elife. 2018;7:e37813.
Alzheimer’s therapy development
217
21. Alzforum Therapeutics Database, https://www.alzforum.org/therapeutics/search?fda_ statuses¼&target_types%5B%5D¼170&therapy_types¼&conditions¼&keywords-entry¼ &keywords¼#results. Accessed 05/08/19 2019. 22. Klunk WE, Lopresti BJ, Ikonomovic MD, et al. Binding of the positron emission tomography tracer Pittsburgh compound-B reflects the amount of amyloid-beta in Alzheimer’s disease brain but not in transgenic mouse brain. J Neurosci. 2005;25: 10598–10606. 23. Leuzy A, Chiotis K, Lemoine L, et al. Tau PET imaging in neurodegenerative tauopathies-still a challenge. Mol Psychiatry. 2019;Epub ahead of print. 24. Fitzpatrick AWP, Falcon B, He S, et al. Cryo-EM structures of tau filaments from Alzheimer’s disease. Nature. 2017;547:185–190. 25. Falcon B, Zhang W, Murzin AG, et al. Structures of filaments from Pick’s disease reveal a novel tau protein fold. Nature. 2018;561:137–140. 26. Maji SK, Perrin MH, Sawaya MR, et al. Functional amyloids as natural storage of peptide hormones in pituitary secretory granules. Science. 2009;325:328–332. 27. Krishnaswamy S, Lin Y, Rajamohamedsait WJ, Rajamohamedsait HB, Krishnamurthy P, Sigurdsson EM. Antibody-derived in vivo imaging of tau pathology. J Neurosci. 2014;34:16835–16850.
Alzheimer’s therapy development: A few points to consider Einar M. Sigurdsson* Departments of Neuroscience and Physiology, and Psychiatry, Neuroscience Institute, New York University School of Medicine, New York, NY, United States *Corresponding author: e-mail address: [email protected]
Contents 1. Introduction 2. Targeting tau intracellularly or extracellularly 3. Drug-screening models: Preventing seeding/spread vs neurotoxicity 4. Strains of Aβ and tau: Influence on therapeutic development 5. Targets other than Aβ and tau 6. Concluding remarks Acknowledgments References
206 207 210 212 214 215 215 216
Abstract Development of therapies for Alzheimer’s disease has only resulted in a few approved drugs that provide some temporary symptomatic relief in certain patients. None of these compounds in clinical use halts or slows the progression of the disease. To date, several drugs targeting the amyloid-β peptide, and some against the tau protein, have failed in clinical trials. While there are various reasons for these failures, considering the following points may aid in improving the outcome of future trials. First, the tau protein should ideally be targeted intracellularly because most of tau pathology is within cells, neurons in particular. Second, an overriding emphasis in recent years has been on implementing drug-screening models that focus on prevention of seeding/spread of aggregates. Much less attention has been paid to identify compounds that inhibit neurotoxicity of these aggregates, which may not necessarily relate to their seeding/spread propensity. Ideally, all these markers should be readouts in the same assay or model. Third, diversity in conformers/strains of aggregates complicates drug development of small molecule aggregation inhibitors but is likely to be less of an issue for antibody-based treatments. Lastly, other more general targets associated with neurodegeneration should continue to be pursued but are in many ways more difficult to address than clearing amyloid-β and tau, the defining hallmarks of AD.
Progress in Molecular Biology and Translational Science, Volume 168 ISSN 1877-1173 https://doi.org/10.1016/bs.pmbts.2019.06.001
#
2019 Elsevier Inc. All rights reserved.
205
206
Einar M. Sigurdsson
1. Introduction Research on Alzheimer’s disease (AD) has expanded greatly in the last few decades and many important advances have been made in clarifying the etiology and pathogenesis of the disease, in the development of diagnostic probes, and in identifying therapeutic targets. However, this progress has been laced with several disappointing outcomes in clinical trials targeting the key hallmarks of the disease, the amyloid-β peptide (Aβ) and the tau protein, as well as in trials for other targets thought to be associated with the disease. Most of the failures have been in trials targeting Aβ since it received earlier and more extensive attention as a target for therapy than the tau protein. The reasons for more drug development effort having been spent on Aβ are logical and relate largely to the discovery of mutations in the amyloid precursor protein (APP) around its enzymatic cleavage sites that generate Aβ. These APP mutations increase Aβ levels or its more pathogenic Aβ42 form, and are associated with a small number of familial cases of AD. In addition, Aβ is mostly found extracellularly in AD and thereby is a more accessible target than tau, which is primarily an intracellular protein. The smaller size of Aβ (40–42 amino acids) compared to tau (352–441 amino acid) also render it easier to study and target with therapies. In hindsight, the failures of the Aβ targeting therapies are not particularly surprising, since when clinical symptoms are evident, Aβ levels have or are about to plateau and Aβ pathology may no longer be closely involved in the progression of the disease. APP is primarily located in synapses and increased generation, or decreased clearance, of Aβ (or other APP fragments) leads to synaptic damage, influx of calcium, activation of protein kinases, hyperphosphorylation of tau and subsequent cytoskeletal instability, which leads to further degeneration of synapses and accelerates their loss. Hence, the source of Aβ diminishes as the disease progresses, and it may be too late to target Aβ when memory impairments have become obvious. Sporadic AD cases have a relatively unknown etiology and no reliable presymptomatic biomarkers, and thus cannot be treated prophylactically before clinical symptoms have manifested. However, ongoing Aβ targeted immunotherapy trials on non-symptomatic individuals with familial forms of AD will hopefully be more positive. A comparable approach on control subjects with high Aβ load detected on PET scans may also provide valuable insight, although it is unclear at this point what percentage of these subjects would develop AD.
Alzheimer’s therapy development
207
Within the tau field, more than half of the trials are on immunotherapies and are still in the Phase 1–2 stages, which focus on safety and pharmacokinetics rather than efficacy.1,2 Of the 10 trials that have started, only 1 has failed. It occurred early in Phase 1 and it may have been because of a short half-life of the antibody since there were apparently no safety issues. It was being conducted in healthy subjects that likely did not have much if any of the phospho-tau epitope being targeted (P-Ser422). Hence, target engagement could not have been properly analyzed, and lack thereof could then not have been the reason for halting the trial. Beside tau targeting immunotherapies, various tau therapeutic approaches have entered clinical trials and some have failed. Those that have been discontinued include a microtubule stabilizer to counteract presumed loss of function of hyperphosphorylated tau, an aggregation inhibitor to inhibit tau assembly or promote its disassembly, and a GSK-3β enzyme inhibitor to block phosphorylation of tau on certain sites. Similar or other approaches are in Phase 1–2 and have been outlined elsewhere.2 In this overview, I would like to focus on the following three issues relevant to AD therapy and/or pathogenesis: (1) Is it likely to be sufficient to target tau extracellularly or is an intracellular interaction necessary for clinical benefits? (2) In drug-screening models, is prevention of seeding/spread of the protein aggregates the key outcome measure or should the focus be on or include prevention of neurotoxicity? (3) How may diversity in conformers/strains of Aβ and tau influence therapeutic development?
2. Targeting tau intracellularly or extracellularly Tau is primarily an intracellular protein, both in its normal physiological form and in its pathological form as aggregates of various sizes that can form neurofibrillary tangles. Low levels of tau have been measured in brain interstitial fluid and cerebrospinal fluid (CSF) in experimental animals and in CSF in humans. Antibodies have been shown to target tau intra- and extracellularly. The primary form of tau that is being targeted is likely to be soluble, although it presumably consists of small aggregates.1,2 Evidently, secreted tau is soluble and extracellular ghost tangles are relatively few and not a very useful target. Intracellularly, antibodies associate mainly with tau in the endosomal-lysosomal system and the small tau aggregates in these vesicles can be considered rather soluble. Hence, soluble tau is likely the
208
Einar M. Sigurdsson
primary target of the antibodies and thus it is helpful to consider the size of the intra- vs extracellular pools of soluble tau. A recent study assessed these two pools of tau in a group of Alzheimer’s patients and subjects with mild cognitive impairment, compared to age-matched controls.3 Their findings indicate that CSF tau is about 0.001–0.0001% of total brain soluble tau. It is likely that tau released from cells has been largely degraded when it reaches the CSF and the extent of this is not clear in human subjects. However, mouse studies indicate that CSF tau is approximately 10% of tau levels in brain interstitial fluid, although these levels do not correlate in individual animals.4 Assuming a similar relationship in humans, it can be estimated that extracellular tau available to interact with antibodies is then about 0.01–0.001% of intracellular soluble tau. Measurements in a more artificial culture system echo these differences with extracellular tau measuring about 0.1% of intracellular tau.5 Now, in this context, one has to consider that most of soluble tau is in the cytosol and not readily accessible to the antibodies. However, at least some tau antibodies find their way into the cytosol following endocytic uptake, presumably because of leakage or rupture of endosome and lysosomes.6 Once there, the antibodies have access to various forms of tau. These include soluble physiological forms of tau to which antibodies that recognize normal tau epitopes can bind. This would likely initiate the clearance of this complex, possibly via TRIM 21 binding or other proteasomal degradation pathways.7 Such clearance of normal tau could be considered to have detrimental consequences. However, it is unlikely to be extensive and other microtubule-associated proteins can take over tau functions as evident by the relatively limited deficits seen in tau knockout mice. Reducing normal tau levels with an antisense oligonucleotide is also being pursued in clinical trials.8 In diseased neurons, the cytosol will also contain tau aggregates of various sizes that are derived from ruptured lysosomes. Neuronal accumulation of lysosomes filled with tau aggregates is a well-known phenomenon in AD and related transgenic tauopathy mouse models.9 Presumably, with advancing tau pathology, the lysosomal clearance system exceeds its capacity, resulting in accumulation of lysosomes with tau aggregates that eventually leak into the cytosol. There, these small rather soluble aggregates then seed the formation of larger aggregates, which eventually results in the formation of neurofibrillary tangles. The tangles may be rather inert but eventually these neurons die, perhaps in part because the accumulating fibrils affect the function of various cellular organs. Antibody interaction with these larger aggregates may in some instances, depending on the epitope, prevent further aggregation and/or lead to the disassembly of the aggregates. Overall,
209
Alzheimer’s therapy development
this may be beneficial or detrimental, depending on if these aggregates cause dysfunction or have been assembled to prevent toxicity caused by the smaller assemblies. The stage of the disease may also matter in this context. Regardless of the exact mechanism of action of antibody-mediated clearance of tau pathology, it is very clear that intracellular tau levels exceed extracellular tau levels by several orders of magnitude (Fig. 1A).
A
B
Extracellular Tau
Neurotoxicity
Intracellular
Tau
seeding/ sequestration
propagation
Conformers / Strains
C
Small molecules
Antibodies
Fig. 1 (A) The tau protein is primarily found intracellularly so the most efficacious therapies targeting tau need to be able to enter the cell to access its largest pool. (B) The seeding/sequestration and propagation pathway may be less neurotoxic (broken arrows) than other misfolding/aggregation pathway (solid arrows). (C) Small molecules may be less likely to recognize different conformers/strains than antibodies. To make this point, the panel depicts three different small molecules that are specific for each of the three conformers whereas one of the two antibodies shown recognizes two out of three conformers equally well.
210
Einar M. Sigurdsson
Considering these large differences in intra- vs extracellular pools of tau, an antibody that gets into neurons is likely to be more efficacious than an antibody that can only work extracellularly. Our findings indicate that antibody charge has a great influence on antibody uptake and that antibody chimerization/humanization can greatly affect its charge and thereby uptake.10 This needs to be carefully examined for clinical candidate antibodies. However, the extent of animal studies that can be conducted on humanized antibodies is rather limited because of species related immunogenicity issues. It is conceivable to use mice with human-like immune system for this type of studies but it is unclear how such a drastic alteration affects tau pathology in transgenic models. However, the efficacy of these humanized antibodies can be easily examined in human-derived culture systems. We have shown in such a model that a partial humanization of an antitau mouse monoclonal antibody, with well-established therapeutic potential in various models, prevents its intracellular efficacy in blocking tau toxicity and promoting tau clearance.6,10–13 The reason for this loss of efficacy is that the chimeric antibody is not taken up into neurons because of substantial change in antibody charge. It is not clear how well the current humanized antibodies have been re-examined for efficacy prior to clinical trials and it raises concerns that some of these may fail in clinical trials for reasons unrelated to their direct interaction with the target. With regard to binding itself, we have reported that at least for one prominent tau epitope, the phospho-serine 396,404 region, stronger binding does not convey improved efficacy.10,12,13 This important finding should also be kept in mind as the humanization may also influence binding, even though the binding regions remain the same, as we have reported.10 Lastly, tau levels in the CSF are not altered in non-AD tauopathies, suggesting that tau lesions in these subjects may spread in a cell autonomous manner, not involving extracellular tau. Therefore, in those subjects, targeting tau intracellularly is even more important than in AD.
3. Drug-screening models: Preventing seeding/spread vs neurotoxicity It is now well-established that various peptide/protein aggregates can be seeded in vitro and in vivo, and that these seeds can spread through anatomically connected pathways.14–16 Incorporating this phenomenon into antibody efficacy studies has been popular and is in many ways appropriate.
Alzheimer’s therapy development
211
However, we have observed that antibodies’ ability to prevent tau seeding and block tau toxicity does not always go hand in hand.10 For example, one antibody that we developed against the P-[Thr212, Ser214] tau region, prevented tau seeding but not tau toxicity.10 This can be explained by the tau species to which it binds, which may promote seeding without influencing toxicity. Antibodies against some other regions can prevent both seeding and toxicity.10 With toxicity, we mean loss of neurons as detected by LDH ELISA assay or retraction of neuronal processes as detected by reduced NeuN signal on Western blots. At least in culture, the toxicity pathway appears to be more acute whereas the seeding pathway is more gradual. It can be envisioned in vivo that certain forms of tau assemblies may have an acute toxic effect, whereas other tau forms may be more prone to initiate tau seeding. Such seeding is eventually likely to have detrimental effects but may to some extent be an effort of the cell to sequester toxic forms of tau (Fig. 1B). Hence, prevention of seeding may unleash some toxic forms of tau that are better kept sequestered in aggregates. Reports from the prion field supports this view, showing that prion replication can be separated from toxicity and that some forms of prion aggregates may actually delay the onset of clinical symptoms.17,18 Specifically, levels of the infectious scrapie form of the prion protein, PrPSc, have been shown to plateau before clinical onset,17 analogous to Aβ levels in AD. While this may also relate to neuronal reserve that can be differentially affected by PrPSc and other perhaps more toxic forms of PrP in different models, it is likely that different species of the amyloidogenic peptide/protein in amyloid diseases can have varying effects on seeding/propagation and toxicity. However, this is not always the case, at least in vitro. A culture study on tau strains reported that seeding efficiency correlated with inhibition of growth.19 How this plays out in vivo is less clear as in that particular study, tau strain inoculations led to robust tau pathology without overt neuronal loss.19 As pointed out above, using antibodies as tools to neutralize certain tau species provides an insight into this matter, showing some overlap between seeding and toxicity but also that these can proceed via separate pathways. To reiterate, how antibodies affect the different pathways of acute toxicity vs chronic spreading is likely to be epitope-dependent, as we have observed for some of our antibodies.10 This and their efficacy in general may also vary depending on subtle binding differences within the same epitope region.12 It may also depend on the exact tauopathy under study. To choose the best antibodies to advance to clinical trials, it is important to include various assays that can distinguish between these two pathways,
212
Einar M. Sigurdsson
the acute toxicity vs chronic seeding/toxicity. Ideally, a clinical candidate antibody should perform well in both types of assays, assuming that preventing the latter does not exacerbate the former. Hence, each model should ideally allow examining both pathways at the same time as we have done in culture models.10,12 Subsequent in vivo studies should likewise incorporate both pathways into data analysis. Based on publications in this field, it appears that much more attention is being paid to interfering with the seeding/aggregation/propagation pathway (protein-protein interaction) as reflected by reports on the development of and testing of drug candidates in high throughput in vitro assays. Prevention of toxicity is typically a secondary measure conducted on key hits from the protein-protein interaction assay. Hence, these assays miss compounds that may be preferable for clinical benefits. This is also evident in in vivo assays that focus on the ability of drug candidates to prevent seeding and spread in animal models. Neurotoxicity/ neuronal loss may not be very evident in such injection models and is more time consuming to analyze. The neurotoxicity pathway may be more relevant for identifying therapies that more quickly provide cognitive improvements, whereas the functional benefits of targeting the more chronic seeding pathway may take longer to manifest themselves.
4. Strains of Aβ and tau: Influence on therapeutic development In recent years, it has become popular to refer to strains of Aβ and tau, analogous to the well-established concept of prion strains. Within the prion field, a definition of a strain is that it retains its properties following multiple passages, ideally between animals. There is some indication that this may occur for certain Aβ and tau preparations but the fidelity of this propagation is not yet entirely clear, at least not in animal models. Changes in tau profiles with continued passages can indicate mixtures of strains and the presence of substrains.20 It can also be interpreted to reveal certain conformational variability that does not clearly support the presence of a strain. A key reason why the presence of Aβ and tau strains is not as clear as for PrP prions is that unlike PrPSc, Aβ and tau aggregates are much less transmissible, presumably because they are more easily degraded. Hence, because of subtle differences in degradation profiles, some diversity of Aβ or tau profiles can be expected. Regarding how possible strain/conformer presence may affect therapeutic development, this is likely to differ for small molecules vs antibodies. Generally speaking, development of small molecule aggregation inhibitors
Alzheimer’s therapy development
213
targeting Aβ or tau has been disappointing. Several classes of small molecules have been identified that recognize the characteristic β-sheet conformation, which defines amyloids in general. These have been effective in various in vitro assays and some in vivo models but very few have advanced to clinical trials, presumably because of low blood-brain barrier permeability and/or toxicity. Those that have advanced to late stage clinical trials have either failed to show efficacy or are still under study, in what appears to be limited capacity as defined by small trial size.21 Some insight into how this may play out can be derived from related positron emission tomography (PET) studies on compounds targeting Aβ or tau. Approved PET ligands targeting Aβ are comparable in many ways, which may relate to the fact that Aβ deposits are large and extracellular. Even if the different Aβ ligands may have different affinities for certain β-sheet subtypes of Aβ, each deposit likely comprises a diversity of β-sheet structures because it is formed by accretion of many different types of large Aβ assemblies. However, at least some of the Aβ PET ligands, like PiB, do not bind well to plaques in most mouse models, which can be explained by differences in binding sites within Aβ in humans vs mice.22 Hence, potential therapeutic β-sheet binders that are effective in the mouse may not work in humans, and vice versa. A similar scenario can be envisioned in the tau field, but it is not as clearcut. Tau aggregates are primarily found intracellularly and the various small molecule tau β-sheet ligands show some diversity in their binding profile.23 Many of the first tau ligands that were examined only recognized tau in Alzheimer’s patients but not in other tauopathies. Those early findings suggest differences in tau conformations in the different tauopathies, as recently shown more clearly in cryo-electron microscopy studies.24,25 In addition, some of the small molecule tau PET ligands turned out to bind strongly to unrelated proteins.23 Even if all the tau deposits from different tauopathies contain β-sheets and all the ligands are defined as β-sheet binders, subtle conformational differences likely explain why the ligands bind differently to tau from different tauopathies. However, to date, at least some of the ligands have been reported to give a positive PET signal in tauopathies other than AD.23 Future studies will likely reveal the molecular explanation for these differences in binding profile. In summary, for these small molecule β-sheet binders to work as a therapy, it may be necessary to develop strain/conformer specific inhibitors, at least for the tau protein (Fig. 1C). Because of its size, disordered structure, different isoforms and diversity of posttranslational modifications, its
214
Einar M. Sigurdsson
secondary structure is much more variable than for Aβ. Beside these issues, there are some concerns that chronic targeting of β-sheets is likely to result in adverse reactions. Various proteins have at least some segments of β-sheets, and β-sheet assemblies are used, for example, to store hormones for later use.26 In addition, chronic targeting of a subgroup of tau strains/conformers, may lead to an expansion of other tau subgroups, analogous to changes in bacterial flora following antibiotic treatment, which can have unforeseen detrimental consequences. For antibody therapies, this is likely to be less of an issue as many antibodies against a particular epitope will recognize different strains/conformers (Fig. 1C). For example, most of tau antibodies will recognize tau deposits in all tauopathies. The exception may be certain antibodies against discontinuous epitopes (conformational) and antibodies against regions that are buried in some strains/conformers. For example, for tau, an antibody directed at P-Ser262 binds well to tau aggregates in AD but not in Pick’s disease, in which this epitope is buried.25 Typically, the N- and C-terminal regions of full-length tau are accessible to antibodies, as well as large segments of the middle region. However, tau exists in various fragments in the brain and it is unclear if some of these may constitute certain strains/conformers of tau, which would then be differentially affected by antibodies against different epitopes. Overall, it is unlikely that subtle structural differences between strains/ subtypes of tau will greatly affect therapeutic efficacy of most antibodies. This ability of most antibodies to recognize all the different tauopathies bodes also well for their diagnostic potential. Whole antibodies are likely too large to be used as PET ligands for brain imaging but small antibody derivatives have great potential as imaging agents because of better brain penetration.27 Compared to β-sheet small molecule dye based ligands, the antibody derivatives should be more specific and can provide important information on the epitope profile of tau aggregates. That information can then be used to personalize antibody treatment to target these aggregates.
5. Targets other than Aβ and tau It makes sense to focus on Aβ and tau aggregates for AD therapy as they are the defining hallmarks of AD. However, there has always been interest in other more general targets related to neurodegeneration, such as promoting synaptic health, manipulating inflammation/microglial
Alzheimer’s therapy development
215
activation by decreasing or enhancing it, influencing lipid metabolism via ApoE or by other means, and enhancing intracellular protein degradation via proteasomal or lysosomal degradation. While these are certainly valid targets, they are fraught with their own difficulties that are more formidable than targeting Aβ and tau. Cognitive enhancers designed to improve synaptic health have so far not worked in the clinic. The benefits of inhibiting or promoting inflammation/microglial activation may depend on the severity of the pathology, which varies between brain regions as the disease progresses. Therefore, what is beneficial for the entorhinal cortex at some stage of the disease may be detrimental for the neocortex at the same time. Lipid metabolism is complex and not well understood. ApoE targeted approaches have not yet made it into the clinic over 25 years after its association with AD was discovered. Promoting intracellular protein degradation by activating the pathways involved is a double-edged sword that can be very toxic and is difficult to fine-tune, which explains why some of the experimental compounds that have been used to study these pathways have not made it into the clinic.
6. Concluding remarks While there are many explanations for the failures of drugs to attenuate progression of AD, there are reasons to be optimistic about the future of therapeutic development in this field. In particular, targeting the tau protein is likely to provide therapeutic benefits later in the disease process than Aβ since tau pathology correlates better with the degree of dementia than Aβ lesions. However, as outlined in this brief overview, several uncertainties remain about the ideal preclinical features of a drug that is likely to succeed in clinical trials. I hope that the points raised here will improve the selection process and thereby accelerate the development of effective therapies. Other more general targets associated with neurodegeneration should continue to be pursued but are in many ways more difficult to address than targeting the defining hallmarks of AD.
Acknowledgments E.M.S. is currently funded by NIH grants R01 AG032611, R01 NS077239, R21 AG058282, and R21 AG059391. He is an inventor on several patents that are assigned to New York University. Some of the patents that focus on tau immunotherapy and related diagnostics are licensed to and are being co-developed with H. Lundbeck A/S.
216
Einar M. Sigurdsson
References 1. Sigurdsson EM. Tau immunotherapies for Alzheimer’s disease and related tauopathies: progress and potential pitfalls. J Alzheimers Dis. 2018;64:S555–S565. 2. Congdon EE, Sigurdsson EM. Tau-targeting therapies for Alzheimer disease. Nat Rev Neurol. 2018;14:399–415. 3. Han P, Serrano G, Beach TG, et al. A quantitative analysis of brain soluble tau and the tau secretion factor. J Neuropathol Exp Neurol. 2017;76:44–51. 4. Yamada K, Cirrito JR, Stewart FR, et al. In vivo microdialysis reveals age-dependent decrease of brain interstitial fluid tau levels in P301S human tau transgenic mice. J Neurosci. 2011;31:13110–13117. 5. Chai X, Dage JL, Citron M. Constitutive secretion of tau protein by an unconventional mechanism. Neurobiol Dis. 2012;48:356–366. 6. Gu J, Congdon EE, Sigurdsson EM. Two novel tau antibodies targeting the 396/404 region are primarily taken up by neurons and reduce tau protein pathology. J Biol Chem. 2013;288:33081–33095. 7. McEwan WA, Falcon B, Vaysburd M, et al. Cytosolic Fc receptor TRIM21 inhibits seeded tau aggregation. Proc Natl Acad Sci U S A. 2017;114:574–579. 8. ClinicalTrials gov. Safety, Tolerability, Pharmacokinetics, and Pharmacodynamics of IONISMAPTRx in Patients With Mild Alzheimer’s Disease. https://www.clinicaltrials.gov/ct2/ show/NCT03186989 US National Library of Medicine, ClinicalTrials gov; 2019. 9. Lin WL, Lewis J, Yen SH, Hutton M, Dickson DW. Ultrastructural neuronal pathology in transgenic mice expressing mutant (P301L) human tau. J Neurocytol. 2003;32:1091–1105. 10. Congdon EE, Chukwu JE, Shamir DB, et al. Tau antibody chimerization alters its charge and binding, thereby reducing its cellular uptake and efficacy. EBioMedicine. 2019;42:157–173. 11. Congdon EE, Gu J, Sait HB, Sigurdsson EM. Antibody uptake into neurons occurs primarily via clathrin-dependent fcgamma receptor endocytosis and is a prerequisite for acute tau protein clearance. J Biol Chem. 2013;288:35452–35465. 12. Congdon EE, Lin Y, Rajamohamedsait HB, et al. Affinity of tau antibodies for solubilized pathological tau species but not their immunogen or insoluble tau aggregates predicts in vivo and ex vivo efficacy. Mol Neurodegener. 2016;11:62–85. 13. Wu Q, Lin Y, Gu J, Sigurdsson EM. Dynamic assessment of tau immunotherapies in the brains of live animals by two-photon imaging. EBioMedicine. 2018;35:270–278. 14. Sigurdsson EM, Wisniewski T, Frangione B. Infectivity of amyloid diseases. Trends Mol Med. 2002;8:411–413. 15. Goedert M, Falcon B, Clavaguera F, Tolnay M. Prion-like mechanisms in the pathogenesis of tauopathies and synucleinopathies. Curr Neurol Neurosci Rep. 2014;14:495. 16. Walker LC, Diamond MI, Duff KE, Hyman BT. Mechanisms of protein seeding in neurodegenerative diseases. JAMA Neurol. 2013;70:304–310. 17. Sandberg MK, Al-Doujaily H, Sharps B, Clarke AR, Collinge J. Prion propagation and toxicity in vivo occur in two distinct mechanistic phases. Nature. 2011;470:540–542. 18. Diaz-Espinoza R, Morales R, Concha-Marambio L, Moreno-Gonzalez I, Moda F, Soto C. Treatment with a non-toxic, self-replicating anti-prion delays or prevents prion disease in vivo. Mol Psychiatry. 2018;23:777–788. 19. Kaufman SK, Sanders DW, Thomas TL, et al. Tau prion strains dictate patterns of cell pathology, progression rate, and regional vulnerability in vivo. Neuron. 2016;92:796–812. 20. Sharma AM, Thomas TL, Woodard DR, Kashmer OM, Diamond MI. Tau monomer encodes strains. Elife. 2018;7:e37813.
Alzheimer’s therapy development
217
21. Alzforum Therapeutics Database, https://www.alzforum.org/therapeutics/search?fda_ statuses¼&target_types%5B%5D¼170&therapy_types¼&conditions¼&keywords-entry¼ &keywords¼#results. Accessed 05/08/19 2019. 22. Klunk WE, Lopresti BJ, Ikonomovic MD, et al. Binding of the positron emission tomography tracer Pittsburgh compound-B reflects the amount of amyloid-beta in Alzheimer’s disease brain but not in transgenic mouse brain. J Neurosci. 2005;25: 10598–10606. 23. Leuzy A, Chiotis K, Lemoine L, et al. Tau PET imaging in neurodegenerative tauopathies-still a challenge. Mol Psychiatry. 2019;Epub ahead of print. 24. Fitzpatrick AWP, Falcon B, He S, et al. Cryo-EM structures of tau filaments from Alzheimer’s disease. Nature. 2017;547:185–190. 25. Falcon B, Zhang W, Murzin AG, et al. Structures of filaments from Pick’s disease reveal a novel tau protein fold. Nature. 2018;561:137–140. 26. Maji SK, Perrin MH, Sawaya MR, et al. Functional amyloids as natural storage of peptide hormones in pituitary secretory granules. Science. 2009;325:328–332. 27. Krishnaswamy S, Lin Y, Rajamohamedsait WJ, Rajamohamedsait HB, Krishnamurthy P, Sigurdsson EM. Antibody-derived in vivo imaging of tau pathology. J Neurosci. 2014;34:16835–16850.