AD - BIIB etc
Further notes from Sherman after discussion with key opinion leaders
>>>>>>>>>
We recently hosted a Key Opinion Leader (KOL) dinner at AAIC 2018 focused on the current understanding of Alzheimer’s disease pathogenesis and updates in the clinical landscape. Our note below describes some of the most interesting parts of the discussion from the event. Our KOLs, Dr. George Bloom (University of Virginia) and Dr. John Hardy (University College London), are renowned research scientists in the Alzheimer’s field. Dr. Bloom’s laboratory focuses on deciphering the metabolic links that connect ß-amyloid and tau to damage neurons. Dr. Hardy has focused his research on molecular genetic analysis of these neurological diseases and discovered APP mutations as the first known cause of Alzheimer's disease in 1990. The following companies are mentioned in our note: Biogen (NasdaqGS: BIIB), AbbVie (NYSE: ABBV), Roche (SIX: ROG), Eisai (TYO: 4523), AC Immune (NasdaqGM: ACIU), and MorphoSys (NasdaqGS: MOR).
Analysts
David Sherman, Ph.D. (AC)
Dinner Summary
The key theme from the dinner was the uncertainty that remains in the field of Alzheimer’s disease (AD). Even things that may seem obvious, such as the functions of the amyloid beta and tau proteins, remain surrounded by controversy. The lack of clarity regarding the pathogenesis of AD has affected development of therapeutics for the disease. For example, even if an agent were to remove amyloid beta with a high success rate (as some agents do appear to effectively do), there may not be a cognitive or functional improvement for AD patients. This suggests that our fundamental understanding of AD, based on the amyloid hypothesis, remains up for debate. The two KOLs agreed that while amyloid is a key feature of the disease early on, by the time the plaques have built up, there are many downstream pathogenic mechanisms which are already in motion and cannot be stopped by reducing amyloid levels. Further research regarding disease pathogenesis, development of drugs with alternative targets for treating AD, studies evaluating combination therapy, and improved biomarkers will all be required in order to advance the field.
Discussion of BAN2401. The most anticipated news at this year’s AAIC was data from Eisai (TYO: 4523) and Biogen’s (NasdaqGS: BIIB) Phase II trial evaluating BAN2401. The highest dose achieved statistical significance on a number of endpoints (plaque reduction, ADCOMS, ADAS-cog) but failed to show significant difference over placebo on the CDR-SB. In addition, imbalances in the proportion of ApoE4 carriers across the different treatment groups (70% of total patients in the placebo group compared to ~30% in the 10 mg/kg biweekly group) and lack of a clear dose-response on the ADCOMS raises concerns. Despite concerns regarding the novel ADCOMS endpoint, one KOL did note that we still “don’t even know if we’re measuring dementia correctly” with currently used endpoints. The KOL stated that the work put into developing suggest that it is a “reasonable endpoint” for measuring dementia.
“If it just works a little bit, that’s a problem.” – As for many of the anti-Aß antibodies, BAN2401 has shown potential signs of efficacy. Had BAN2401 met all endpoints with statistical significance, the different treatment groups had similar baseline characteristics/demographics, and a clear dose-response been established, the path forward for BAN2401 would be clear. Similarly, had the drug failed to meet any of the endpoints, the path forward would also be clear (the path ends). However, the mixed results raise questions regarding whether BAN2401 will continue to be studied in a Phase III trial. Subgroup analyses may answer some of these questions, but it seems unlikely that BAN2401 will be the clear answer some were hoping for regarding the validity of the amyloid hypothesis for Alzheimer’s disease.
Importance of the CDR-SB for Read Through to Aducanumab – Going into the data readout, many considered the drug’s effect on the Clinical Dementia Rating – Sum of Boxes (DR-SB) to be the most critical metric. CDR-SB is a measure of both cognition and function in AD and is generally believed to be an appropriate endpoint in trials evaluating early AD patients. Notably, CDR-SB is the primary endpoint in Biogen’s Phase III trial evaluating aducanumab. Therefore, the effect of BAN2401 on CDR-SB was thought to have direct read through to aducanumab’s potential effect on this endpoint. However, this is the one endpoint where BAN2401 did not show an effect. While BAN2401 and aducanumab differ in notable ways (targeting different epitopes and different Aß species), this raises some concerns regarding aducanumab’s likelihood of success in the Phase III trial. Data from this trial is expected in 2020.
Other Targets for Treating Alzheimer’s Disease. While the amyloid hypothesis for the disease has dominated in the last few decades, the failures of anti-Aß antibodies has led researchers to evaluate other strategies for treating the disease. As one KOL noted, “if you had a drug that was 100% successful against amyloid, would dementia still exist? I think it would.” It is likely that once the “amyloid problem” is solved, the brain may be exposed to other pathogenic processes that result in dementia. Because of this, it seems that there is a need to study treatments for AD with alternative targets. Recently, there has been interest in targeting tau protein, motivated by the strong temporal and spatial correlation between tau protein and AD neuropathology over the course of the disease. Other potential interesting novel targets include glial cells and the mTOR pathway.
Growing Excitement over Anti-Tau Antibodies, Despite Much Remaining Unknown – Recently, companies have been developing anti-tau antibodies, including Biogen’s BIIB092, AbbVie’s (NYSE: ABBV) ABBV-8E12, and Genentech (division of Roche, SWX: ROG) and AC Immune’s RO 7105705. These candidates are earlier in development compared to anti-Aß antibodies, with the most advanced candidates currently in Phase II development. The KOLs, however, discussed the difficulties in targeting tau. Tau is a relatively large protein, consisting of roughly 400 amino acids, which is roughly 10 times the size of amyloid beta. The large size of tau makes it difficult to target. Additionally, it remains unclear exactly what area of tau would serve as the best target for antibodies. It is also important to consider that tau exists primarily intracellularly, which means that an anti-tau antibody would have to cross the cell’s plasma membrane in addition to the blood brain barrier. The anti-tau antibodies currently in development only target extracellular tau, which may be beneficial in preventing the prion-like spreading of pathological forms of tau after it is “dumped” into the extracellular space. Additionally, because the full range of functions for tau remains unknown, it is possible that knocking out tau may have deleterious effects. Several strains of knockout mice exhibit normal lifespans but show neurological deficits when the mice are older. Therefore, when targeting tau, it may be important to avoid knocking down tau past a certain threshold (Lei et al., 2014).
Renewed Attention Towards the Immune System with Modulation of Microglia – Because cells other than neurons are involved in AD pathogenesis, targeting other types of cells may be an interesting strategy for treating the disease. In recent years, the role of microglia has received attention in studying AD pathogenesis (Wes et al., 2016). Research has suggested that, in AD, microglia may enter a “primed” state following chronic activation. Once in this state, microglia may respond with greater sensitivity to any further insults. Therefore, the aim of microglia-targeted therapy for AD may be to change the phenotype of microglia back to that which is seen in cognitively normal individuals. However, the difficulty with this approach is that increased activation of microglia may be beneficial in certain brain regions while lower activation may be beneficial in other areas in a certain AD patient. Techniques for accomplishing this have not yet been developed. Interestingly, microglia may affect antibody treatments for AD because microglia are believed to be responsible for Fc receptor-mediated antibody effector activity, suggesting that combination therapy with an antibody and microglia-targeting agent may represent an attractive combination strategy.
Regulation of mTOR – The mammalian target of rapamycin (mTOR) pathway is involved in a number of pathways in the brain and may be specifically involved in mediating synaptotoxicity. One KOL stated that mTOR may be a “key player” in AD pathogenesis and mediate the toxicity caused by amyloid and tau. Research has shown that activation of mTOR increases production and deposition of Aß through changing the metabolism of APP and upregulating secretases involved in the APP processing pathway. Additionally, mTOR, as an inhibitor of autophagy, decreases the clearance of Aß (Wang et al., 2014). In studies looking at the relationship between tau and mTOR, an increase in mTOR signaling was shown to facilitate tau pathology (Caccamo et al., 2013). Rapamycin, an mTOR inhibitor, has been studied in AD animal models and was able to bring about improvement in AD symptoms. Other mTOR inhibitors with stronger safety profiles may represent an interesting therapeutic approach.
Developments in AD Biomarkers. With many recent failures in therapeutics for AD and concerns regarding which stage of patients may respond best in clinical trials, researchers have made advances in the development of AD biomarkers. Historically, CSF biomarkers (Aß 42, Aß 42/Aß 40, tau and pTau181) have been used primarily to assess AD (Lewczuk et al., 2018). While lumbar puncture (LP) is a routine procedure for clinicians, it is invasive and inconvenient for patients. Because of this, there has been an unmet need for blood-based biomarkers to allow for non-invasive screening. One of the KOLs noted that plasma neurofilament light may be one of the most promising biomarkers currently being developed.
Plasma Neurofilament Light as a “Game Changer” in AD Diagnosis – Recently, plasma neurofilament light (NFL) concentrations have been studied as a biomarker for AD. Research has shown that AD patients generally have increased plasma NFL concentrations (Mattsson et al., 2017). Increased NFL concentrations are also associated with cognitive decline, other AD biomarkers, and neurodegeneration. NFL levels may even indicate abnormalities at early stages of the disease, which could allow for its use in clinical trials or even in screening the general population.
Mechanisms of Toxicity in AD. The exact cause of cell death in Alzheimer’s disease remains unknown. There are multiple proposed mechanisms that were discussed by the KOLs and these are described below.
The Unknown Effects of Aß Receptor Binding – Generally, in order to characterize receptor-ligand interactions, one has to demonstrate saturation binding. However, this is not possible for amyloid beta given that the amyloid protein binds to itself as well as receptors. Despite this difficulty, biochemical research has shown that Aß interacts with a number of receptors on the cell surfaces of neurons, such as NMDA receptors, AMPA receptors, prion protein, APP, and others. However, exactly how Aß interacts with these receptors remains unknown. The downstream effects of Aß binding to cell surface receptors, such as activation of protein kinases, may be the most relevant to mechanisms of toxicity in AD.
Tau Protein May be the “Bullet” in AD Pathogenesis – The strong correlation between neurodegeneration and temporal and spatial deposition of tau protein supports the hypothesis that tau is critical for toxicity within AD pathogenesis. Research has shown that protein kinases, often activated by the Aß protein, catalyze site-specific phosphorylation on tau, which then allows for neuronal cell cycle re-entry. Re-entry into the cell cycle is a common apoptotic mechanism for neuronal cells, and this may therefore represent one mechanism for cell death in AD. Also, once hyperphosphorylated, tau dissociates from microtubules, after which it forms aggregates. These aggregates are believed to contribute to synapse dysfunction and mediating a number of downstream effects. As the KOL said, “once tau gets going, it has a life of its own.”
AD is Not a Cell Autonomous Process – The KOLs emphasized that AD is not a cell-autonomous process. In addition to neurons, microglia and astrocytes are also responding to Aß buildup. Once amyloid plaques have built up to a certain threshold, the brain’s immune system has likely initiated a number of inflammatory downstream processes to respond to the stress caused by amyloid buildup and these processes also contribute to the disease progression.
Amyloid Beta in Alzheimer’s Disease. Amyloid beta is a key defining feature of Alzheimer’s disease. It seems clear that some combination between overexpression and failure of clearance brings about decades of amyloid buildup, eventually triggering other pathogenic processes such as hyperphosphorylation of tau and an immune system response. However, despite the clear presence of amyloid buildup in Alzheimer’s disease, much remains unknown regarding the exact nature of amyloid beta’s role in the disease. Researchers are still studying the exact nature amyloid beta species, their mechanisms of toxicity and even what the endogenous function of amyloid beta is in cognitively normal individuals.
No Single Aß Oligomer Type – While we may think of oligomers as a single clear entity, in fact, as the KOLs discussed, amyloid oligomers may take a number of different forms. Monomers may organize into dimers, trimers, hexamers, etc. One could even arrange two identical Aß 42 molecules as dimers and get a number of different types of amyloid dimers, differing in terms of connections, folding, etc. Additionally, once oligomers form, they continue to grow, shorten, and rearrange. The KOLs noted that we currently do not have the tools necessary to properly study these dynamic molecules and all the different possibilities for their structure.
Are Oligomers Truly Toxic? – The KOL noted that while many assume Aß oligomers to be the predominantly toxic species in Alzheimer’s, the evidence for its toxicity has not been shown conclusively. In a review, researchers argued that the lack of correlation between neurodegeneration and location and timing of Aß oligomers deposition (Benilova et al., 2012). Additionally, some researchers posit that the toxicity of Aß is related to their lipophilic nature (Aß is considered a “sticky” molecule) and flexible structure, which allow the molecule to bind to a number of receptors. However, this would maybe mean that Aß’s toxicity is a result of its more generic properties (being lipophilic and having a flexible structure), rather than because of any innate toxic characteristics or pathways associated with the molecule. Further research is required to understand the variety of oligomers that exist in vivo and the respective toxicities of these oligomers. Another hypothesis proposes that, initially, amyloid plaques may in fact have neuroprotective features. Plaques can sequester the supposedly more toxic oligomeric species of amyloid beta. However, it is important to understand that the incorporation of oligomers into plaques is not an irreversible process. In later disease stages, plaques may in fact be sources of the toxic oligomers. This may explain why plaques are later surrounded by dystrophic neurites and microglia. At this stage of the disease, it is likely too late to attempt to target amyloid beta, as downstream toxic effects have already been initiated.
Failure of Aß Clearance – As one KOL mentioned, we are dealing with a “failing motor vehicle” in AD. At older ages, normal mechanisms in the brain start to work less or with lower efficiency. One of these mechanisms may be the normal clearance mechanisms in the brain, such as those related to the lymphatic system. Recent research has implicated the lymphatic system in the normal clearance of Aß. When individuals age, lymphatic clearance systems likely become less efficient, allowing for greater amyloid buildup.
Methods for Targeting Amyloid Beta. Despite the lack of a clear mechanistic role for amyloid in AD, the majority of drugs in development target Aß. Development of anti-Aß antibodies, however, has generally not been successful, with many antibodies failing to show an effect on cognitive/functional endpoints in Phase III clinical trials. Anti-Aß antibodies currently in Phase III of development include Biogen's aducanumab, Roche and MorphoSys' (Nasdaq: MOR) gantenerumab, and Genentech and AC Immune's crenezumab. In recent encouraging news for anti-Aß antibodies, Eisai and Biogen’s BAN2401 showed potential signs of efficacy in a Phase II trial. Despite this recent encouraging news, the validity of targeting Aß in the treatment of AD remains up for debate. One KOL stated that we have learned that “by the time the brain is full of amyloid, amyloid isn’t necessarily the right target.”
More Effective Approaches for Testing Anti-Aß Antibodies – The KOLs emphasized the importance of timing when it comes to effectively using anti-Aß antibodies in treating AD. With improved biomarkers and methods for early identification of high-risk individuals based on genetics, it may be possible to treat patients even before amyloid build-up reaches a significant level. Earlier is considered better in treating AD because, once Aß has built up in the brain past a certain (unknown) threshold, there are a number of downstream effects (activation of protein kinases, hyperphosphorylation of tau, changes to microglia phenotypes) that mediate the majority of the disease’s pathogenesis. Once these downstream effects have been initiated, clearing out amyloid will likely be unable to stop progressive neuronal death. One interesting approach for studying anti-Aß antibodies at a sufficiently early enough stage is targeting amyloid pathology in Down syndrome (DS). Because individuals with DS have a third copy of chromosome 21, which is where the APP gene is located, almost all individuals with DS exhibit Alzheimer’s-like neuropathology by the age of 40 (Head et al., 2012). These individuals have significant deposition of amyloid plaques and neurofibrillary tangles. Therefore, because these individuals almost certainly develop amyloid buildup over a certain age, these individuals may be good candidates for early treatment with anti-Aß antibodies to determine whether the antibodies are effective in delaying AD neuropathology and the associated dementia symptoms. Of note, AC Immune (Nasdaq: ACIU) is currently evaluating an anti-Aß vaccine, ACI-24, as a treatment for delaying cognitive decline in individuals with DS.
Other Methods for Targeting Aß – The KOLs discussed the “obvious attractiveness” of BACE inhibitors and the inherent difficulties associated with developing such drugs. Because BACE has a number of substrates throughout the body aside from APP, there is potential for deleterious off-target effects. Similarly, small molecules targeting amyloid beta, while potentially available at a significantly lower cost compared to antibody treatment, may have significant off-target effects due to their lower target specificity compared to that of antibodies. Following the approval of Biogen’s Spinraza, an antisense oligonucleotide (ASO) therapy, for the treatment of spinal muscular atrophy (SMA), interest has grown surrounding the use of ASO therapy in the treatment of other neurological disorders, such as AD. For ASO in AD, there are a number of potential targets such as APP, BACE1, PSEN1, tau, and ApoER2 (Chakravarthy et al., 2017).
AI
