FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1585083630> ?p ?o ?g. }

• W1585083630 endingPage "205" @default.
• W1585083630 startingPage "185" @default.
• W1585083630 abstract "AD is responsible for approximately 70% of all dementias.5 Currently, a confirmed diagnosis of AD requires the presence of plaques (deposited amyloid β [Aβ] peptide) and tangles (intracellular, aggregated, hyperphosphorylated, tau protein) found via postmortem neuropathological examination of the brain. Although there are many abnormalities within an AD brain, neuronal death, particularly within the hippocampus, entorhinal cortex, and frontal cortical regions, contribute to cognitive impairment. The amount and regional distribution of plaques in AD brains does not correlate well with the extent of neuronal loss or with the clinical severity of dementia.6 There have been studies suggesting a better correlation with soluble Aβ in AD brain.7 However, deposited Aβ comprises approximately 95% of total Aβ (soluble plus deposited).8 The role of Aβ in AD has been long debated; does it trigger the disease process, is there some threshold amount that is required to sustain the disease, or does deposited Aβ drive the disease forward in a continuous fashion?9 The presence of tau pathology, in the form of insoluble paired helical filaments (PHFs), correlates much better both with the areas of the brain that suffer from neurodegeneration and also with the extent of cognitive impairment.10, 11 However, the numbers of PHFs do not account for all the neuronal loss.6 Finally, brain volume remains the best pathological correlate of dementia in AD.12 There is extensive literature demonstrating that proteoglycans bind to Aβ peptide and can accelerate the transition of soluble Aβ to a β-sheet structure that is required for the formation of plaques.13, 14 Tramiprosate (3-amino-1-propanesulfonic acid) is a glycosaminoglycan mimetic that was discovered in a screen that measured the heparin-stimulated conversion of soluble Aβ40 from a random coil to the β-sheet structure that is characteristic of aggregated Aβ.15 Tramiprosate was tested for its ability to bind to soluble Aβ and thereby prevent its aggregation. Mechanistically, this would prevent the accumulation of aggregated Aβ and increase the levels of soluble Aβ in AD brain. The published data on tramiprosate are not as comprehensive as might be expected for a clinical candidate. A 20-fold molar excess of tramiprosate prevented the conversion of Aβ40 from random coil/alpha helix to β-sheet, as assessed by circular dichroism spectral analysis. No data were available for Aβ42.16 Experiments to determine the interaction between tramiprosate and Aβ were performed using electrospray mass spectrometry analysis, which provided evidence that tramiprosate was able to bind to both Aβ40 and Aβ42 with a 10-fold molar excess of drug required to give 50% binding, a finding that was replicated by others.17 However, there are no data on binding affinity or dose–response relationships, and the concentration of Aβ and tramiprosate used were very high for these experiments: 20μM of Aβ42 and 200μM of tramiprosate. This contrasts with a concentration of soluble Aβ42 in human cerebrospinal fluid (CSF) of ∼45pM.18 Furthermore, the relevance of this method to an aqueous phase system is not known. A more relevant approach was taken where Aβ42 was coated onto microtiter plates and test compounds together with fluorescently labeled Aβ42 were added to assess the potency of tramiprosate to prevent Aβ aggregation.19 Key elements of this assay were validated using fresh-frozen brain slices taken from AD brains. In this assay, inhibitory concentration of 50% (IC50) values of test compounds required to block aggregation of 0.22pM of Aβ were calculated. Tramiprosate was shown to be inactive at the highest concentration tested (718.6nM); that is, at a 3.2 × 106 molar excess over Aβ. Using similar concentrations of Aβ and tramiprosate as had been used in the mass spectrometry analysis, but conducting the experiment in the aqueous phase, also failed to demonstrate any activity.20 At a 20-fold molar excess, tramiprosate was able to inhibit the cell death caused by 5μM Aβ42 applied to primary rat neurons.16 These data are difficult to interpret; there were no dose–response data, and the protective mechanism was not explored, so it is not possible to determine whether this effect was associated with inhibition of Aβ aggregation. In 8-week-old TgCRND8 mice that carry the human APP K670N/M671L and V717F mutations, tramiprosate was administered subcutaneously (s.c.) daily at 30 or 100mg/kg for 8 weeks.16 The levels of compound in the brain at the end of this dosing regimen were not assayed. In a separate experiment, continuous infusion of 14C-tramiprosate for 10 days was used to estimate brain and plasma levels of drug in rats. At 2 doses, 1 and 10mg/kg/h, tramiprosate demonstrated brain drug levels of about 1μg/ml (70nM) and 10μg/ml (700nM), respectively. The drug half-life was between 2 and 4 hours in plasma and ≥16 hours in brain. However, the concentrations of total rather than free drug were assayed, and it is not known how these data might compare with the s.c. bolus administration that was used to determine efficacy. The efficacy experiment demonstrated a significant effect on the percentage of the cortex occupied by plaques at 100mg/kg but not at 30mg/kg, and the drug had no effect on the number of thioflavin S–positive plaques at either dose. A more complete analysis would have required a wider range of doses and using one mouse brain hemisphere for histology and the other for quantitative biochemical analysis of Aβ species. Surprisingly, the levels of soluble plasma Aβ40 and Aβ42 were both reduced in a dose-related manner by tramiprosate. A reduction of circulating levels of Aβ is consistent with some type of facilitated clearance mechanism, although this was not explored further. A different cohort of TgCRND8 mice were bred that for unknown reasons showed a 4- to 5-fold increase in cerebral Aβ levels. In these mice, 9-week administration of 500mg/kg/day tramiprosate (a much larger dose) resulted in significant reductions in brain of both soluble and insoluble Aβ40 and Aβ42 peptides, data that are difficult to reconcile for an antiaggregation mechanism. The preclinical data provided some support for an effect of tramiprosate on Aβ levels in brain, but the data were incomplete. An experimental design that incorporated a range of drug doses, mice analyzed at different times, histological and biochemical analysis performed simultaneously, and analysis of free and total drug levels would have provided a clearer picture of the therapeutic potential of the drug. Target engagement was not assayed; there was no detection of Aβ/tramiprosate complexes. The use of different doses in mice for histology and biochemistry, and the use of different mouse Aβ phenotypes for the 2 experiments, do not assist interpretation. These data would have confirmed or refuted the mechanistic hypothesis that at the administered doses tramiprosate binds to Aβ and prevents aggregation. In the phase 2 program, tramiprosate was administered at 50, 100, and 150mg twice daily (b.i.d.) for 3 months to mild–moderate AD patients with a Mini-Mental State Examination (MMSE) scores between 13 and 25.21 It is not possible to determine how these doses were computed from the preclinical studies. Tramiprosate exposure did not increase in a dose-proportional manner between the 100 and 150mg b.i.d. dosing regimens. At 5 hours postdose, CSF samples were taken, and tramiprosate could be detected in ∼60% of patients at between 18 and 50nM. The concentration of total Aβ in CSF is approximately 1nM,22 and thus it is likely that tramiprosate achieved the 20-fold molar excess demonstrated to be required to bind to Aβ in some of the in vitro studies. However, it was not demonstrated whether tramiprosate/Aβ complexes were found in the CSF. Furthermore, some data suggest that Aβ concentrations in the extracellular space in the brain parenchyma might be as much as 100-fold greater than that found in CSF,9 which would mean that efficacious levels of tramiprosate may not have been achieved. Nonetheless, there was a striking dose-dependent reduction in CSF Aβ42 levels of up to 70% after 3 months of treatment, with greater reductions seen in the mild AD population. If this reduction were seen in a therapeutic approach that was designed to inhibit Aβ production, it would have been an encouraging sign of efficacy and proof of mechanism. In AD, a reduction in CSF Aβ42 is interpreted as heralding an increase in Aβ42 deposition.23, 24 Thus, an agent designed to prevent aggregation should elevate Aβ42 CSF levels to the normal range, unless the therapeutic agent acts both to prevent aggregation and to increase clearance or degradation. Furthermore, there was no effect on CSF Aβ40 levels, yet the preclinical in vitro data had shown no difference in the binding potential between Aβ40 and Aβ42. Tramiprosate had no effects on cognitive and clinical assessments, which is unsurprising given the short duration of the trial. The biomarker effects on CSF Aβ42 were considered sufficiently interesting to promote the Alphase phase 3 trial. Alphase was a double-blind, placebo-controlled multicenter study that enrolled 1,052 patients in North America and Canada.25 Tramiprosate was administered at 100mg b.i.d. and 150mg b.i.d. for 78 weeks. The primary endpoint measures were the Alzheimer's Disease Assessment Scale–Cognitive Subscale (ADAS-cog) and Clinical Dementia Rating–Sum of Boxes (CDR-SB). The study was powered to detect a 25% reduction in clinical deterioration. Hippocampal volume changes were assessed by magnetic resonance imaging (MRI) and used as a measure of disease modification. Unfortunately, this trial failed its primary and secondary endpoints. For unknown reasons, there was a significant variance introduced at different clinical trial sites that confounded the prespecified statistical analysis. Post hoc analysis showed some evidence of reduced hippocampal volume loss. Given that a surprising feature of the phase 2 data was a reduction in Aβ42 in the CSF, it is regrettable that these data are not available from the Alphase study. Tramiprosate is currently marketed as an over-the-counter supplement, Vivimind, for memory improvement. Epidemiological data suggest that the use of nonsteroidal anti-inflammatory drugs (NSAIDs) may offer some protection against the onset of AD,26 especially longer-term use,27, 28 although this has not been seen by others.29-31 Interventional studies have been negative.32 However, anti-inflammatory agents were tested for their ability to affect Aβ production,33 and remarkably several commonly prescribed NSAIDs reduced Aβ42. Sulindac, indomethacin, and ibuprofen reduced the production of Aβ42, and this suppression was compensated for by an increase in the shorter Aβ metabolites, especially Aβ38. This work opened a new field of pharmacological intervention: the γ-secretase modulators. These agents are not inhibitors of γ-secretase, but shift the cleavage sites in favor of the production of shorter forms of Aβ. Most importantly, they do not affect the processing of an important substrate of γ-secretase, Notch.34 The effects on Aβ42 production were not mediated via inhibition of the NSAIDs' primary pharmacological target, the cyclooxygenase (COX) enzymes COX1 and COX2. It was shown that flurbiprofen racemate and the S- and R- enantiomers were equipotent,35 allowing use of R-flurbiprofen enantiomer (a less active COX inhibitor), thus reducing unwanted side effects, especially gastrointestinal toxicity. The hypothesis being tested was that R-flurbiprofen, subsequently named tarenflurbil, would provide a disease-modifying therapeutic agent for AD by reducing the production of Aβ42 in the brains of AD patients. The original published work on tarenflurbil35 did not establish full in vitro dose responses for inhibition of Aβ42 production. Other workers have used photoaffinity ligands attached to tarenflurbil and demonstrated that these were able to bind to an APP-derived substrate, but not to components of γ-secretase itself.36 However, subsequent studies have shown that tarenflurbil most likely binds allosterically to the γ-secretase complex37-39 that mediates a change in spectrum of Aβ metabolites in favor of shorter species39-41 with a median effective concentration (EC50) of Aβ42 inhibition of ∼250μM. Although the rationale of using tarenflurbil to reduce the potential side effect liability of COX inhibition has been widely accepted, it remains a more potent COX1 inhibitor (IC50 = 44μM) than an inhibitor of Aβ42 production.42 Thus, doses of tarenflurbil that suppressed Aβ42 production would always have COX1 suppression as a potential liability, or as an additional mechanism of efficacy, depending on the context. The first publication of in vivo pharmacology was not comprehensive; 3 doses (10, 25, and 50mg/kg) of tarenflurbil were administered for 3 days to Tg2576 mice with levels of brain Aβ42 and Aβ40 measured.35 All 3 doses showed a reduction, but there was no dose response and the group sizes were low, ranging from 4 to 7 mice per group. The brain and plasma levels of tarenflurbil also did not increase in a dose-proportional manner. The measured brain levels of tarenflurbil were between 1.5 and 2.6μM, some 100-fold lower than the in vitro EC50 concentration. This discrepancy makes the suppression of Aβ42 levels difficult to interpret. A follow-up study in TG2576 mice looked at a longer-term dosing preventative paradigm (Fig 1), where 2 parallel groups were dosed at 10mg/kg tarenflurbil for 4 months between the ages of 8–9 and 11.5–12 months.43 Another group was dosed for 2 weeks from 17.5–18 to 18–19 months. The brains of the mice were analyzed using enzyme-linked immunosorbent assays (ELISAs) for Aβ40 and Aβ42 levels after formic acid and detergent extraction. Plaque burden was measured using immunohistochemistry. None of the treatment paradigms reduced Aβ42 or Aβ40. Surprisingly, plaque burden was reduced in the therapeutic paradigm but not the preventative paradigm. The discrepancy between a lack of effect on quantitative measurements of Aβ42 and Aβ40, and a significant lowering of plaque burden in the group that received just 2 weeks of tarenflurbil versus 4 months administration in the preventative group, makes these experiments difficult to interpret. Brain analysis of tarenflurbil and S-flurbiprofen showed evidence of significant enantiomeric biotransformation, and the total flurbiprofen concentration was 1μM, about 250-fold lower than the in vitro EC50. Other workers delivered tarenflurbil in food to 4- to 5-month-old Tg2576 mice at an average dose of 32mg/kg/day for 9 days.41 The study duration was cut short due to toxicity, which reduced the group size to N = 5. The tarenflurbil brain concentration was 1.3μM. There was a reduction in brain Aβ40, but not Aβ42, which increased compared to the control group. Given the very low N, and evidence of toxicity, interpretation of this study is challenging. Another study examined Aβ42 and Aβ40 levels in the cortex and hippocampus of 7- to 8-month-old Tg2576 mice following extraction with guanidinium HCl.44 The mice received 25, 50, and 100mg/kg flurbiprofen for 3 days. There were no effects on Aβ42 and Aβ40. A second experiment at lower doses of 10 and 25mg/kg showed a significant reduction in Aβ40 in the cortex but not in the hippocampus; Aβ42 was unaffected in both brain regions at both doses. Finally, another group administered 25mg/kg/day to 7- to 8-month-old Tg2576 mice for 7 days. There was no inhibition of brain Aβ42 or Aβ40 levels extracted using guanidinium HCl.42 (A) The time course of the development of amyloid plaque in a typical APP transgenic mouse model. (B) Preventative paradigm. A potential amyloidocentric therapeutic agent is administered with dosing starting prior to the onset of amyloidosis. The therapeutic acts to delay the initial amyloid seeding events in a concentration-dependent manner but does not affect the rate of amyloid deposition. (C) Therapeutic paradigm. A potential amyloidocentric therapeutic agent is administered with dosing starting after the onset of amyloidosis. The therapeutic agent acts to slow the rate of amyloid deposition in a concentration-dependent manner. The dose responses of the therapeutic agent are very similar in B and C, but are potentially mediated via different mechanisms, and their construct validity in regard to the clinical situation has to be carefully considered. In summary, the preclinical science identified a new pharmacological approach to the suppression of Aβ42 production. The in vitro data provided evidence for suppression of Aβ42 from cells with an EC50 of ∼250μM, although a clear modulator effect—a suppression of Aβ42 coupled to an increase in shorter forms of Aβ—was not always demonstrated. The in vitro EC50 concentration of tarenflurbil was never approached in the brain in the in vivo experiments due to the lack of brain penetration (about 1.5% of plasma levels42) and dose-limiting toxicity. A dose response of Aβ42 suppression, coupled to brain drug levels that were consistent with the EC50, was not demonstrated. The clinical development of tarenflurbil appears to have been based on the few preclinical experiments that showed a reduction in brain Aβ42 levels. In a phase 1 study, 3 cohorts of 16 healthy aged subjects received either 400, 800, or 1600mg/day (n = 12, administered in 2 doses) or placebo (n = 4) for 21 days.45 The drug was well tolerated at all doses, and there was a dose-proportional increase in tarenflurbil concentrations in the CSF. At the highest dose, 800mg b.i.d., the mean tarenflurbil concentration in the CSF was ∼1.2μM, some 200-fold below its EC50 concentration for the inhibition of Aβ42 production in cell culture. There was no lowering of Aβ42 levels in the CSF at any dose. The phase 2 clinical trial studied 210 mild–moderate AD patients with MMSE scores between 15 and 26.46 Patients received either tarenflurbil 400mg b.i.d. (n = 69), 800mg b.i.d. (n = 70), or placebo (n = 71) for 12 months in a multicenter, placebo-controlled double-blind study. The primary outcome measures were ADAS-cog and 1 functional assessment, either the Alzheimer's Disease Cooperative Study–Activities of Daily Living Inventory (ADCS-ADL) or the CDR-SB. An analysis showed an apparent interaction between the baseline cognitive and functional scores and treatment effect, so that efficacy analyses were performed separately for the mild (MMSE > 20) and moderate AD patients. When the analyses were performed in this way, 800mg b.i.d. tarenflurbil-treated patients showed a significantly slower rate of decline of ADCS-ADL; ADAS-cog and CDR-SB showed similar effect sizes but were not statistically significant. In the moderate AD group (MMSE ≤ 19), the placebo group demonstrated a significantly lower rate of decline in all 3 outcome measures than did the 800mg b.i.d. tarenflurbil group. These somewhat paradoxical findings are difficult to interpret. However, the change in ADCS-ADL over the 12-month period was higher in the mild AD placebo group than the moderate AD placebo group, whereas for the 2 other primary outcome measures, ADAS-cog and CDR-SB, the moderate AD group showed greater deterioration, as might be expected. Thus, the effect seen at 800mg b.i.d. tarenflurbil in mild AD patients might have been due to an unusually large placebo group deterioration rather than a bona fide treatment effect. Importantly, proof of mechanism—a change in the spectrum of Aβ metabolites in the CSF in favor of shorter forms—was not assessed. The phase 3 study enrolled 1,646 mild AD patients in a multisite, randomized, double-blind placebo-controlled trial comparing 800mg b.i.d. tarenflurbil versus placebo for 18 months.47 The primary outcome measures were change at 18 months from baseline on ADAS-cog and ADCS-ADL. There was no difference between the drug-treated and placebo-treated groups on the primary outcome measures, and CSF analyses of Aβ metabolite spectrum were not performed. γ-Secretase activity is required to release the Aβ peptide,48, 49 hence inhibitors of γ-secretase should reduce Aβ production. In the simplest interpretation of the amyloid hypothesis, which posits that continued deposition of Aβ drives pathological processes resulting in neuronal dysfunction and death, a γ-secretase inhibitor (GSI) that reduced Aβ production would slow the progression of AD. Semagacestat is a classical GSI,50 acting as a noncompetitive enzyme inhibitor with an allosteric binding site. γ-Secretase is responsible for the final cleavage of the APP C-terminal domain following cleavage by either α- or β-secretase and also cleaves a wide range of substrates, including Notch.51, 52 The Notch signaling pathway is critical for cell fate determination in many dividing cells and is therefore a significant potential safety liability for a GSI. Several drug discovery programs have sought compounds that were selective for Aβ versus Notch inhibition so as to provide a margin of safety.53-55 However, the in vitro assays (cell-free and cell-based) employed, although they do allow compounds to be compared with each other, are of unknown predictive validity for the in vivo situation. Semagacestat inhibited Aβ production with EC50 = 14.9nM in HEK293 cells stably transfected with hAPPSwe cDNA.56 In HEK293 cells stably transfected with the Notch δE cDNA construct, semagacestat inhibited the production of the Notch intracellular domain with EC50 = 46nM (P. C. May, personal communication). This indicated that semagacestat has a cell-based Aβ inhibition/Notch inhibition ratio of ∼3. The dose-related inhibition of Aβ production in cell-based assays has been widely replicated but with slightly different potencies and consequently different Aβ/Notch inhibition ratios: for example, 1.3,57 0.8,55 and 20.5.58 The most informative study was performed using a cell-free, quantitative γ-secretase in vitro assay where Notch and APP substrate concentrations were accurately controlled.59 This demonstrated an Aβ/Notch ratio of 0.1. These data suggest that for semagacestat, the separation of inhibition of Aβ production over Notch inhibition was marginal. Preclinical in vitro and in vivo studies revealed that the pharmacology of semagacestat and of GSIs in general was complex. This led to a biphasic stimulation/inhibition of Aβ production determined by both substrate availability and compound concentration.58, 60, 61 The mechanistic explanation for this effect remains obscure. In vivo experiments demonstrated a similar stimulation/inhibition effect of semagacestat on plasma Aβ levels, but this was not demonstrated in mouse brain,62 guinea pig brain,61 or rat brain.58 Semagacestat was also orally administered at 2mg/kg acutely to beagle dogs to assess the pharmacokinetic and pharmacodynamic profile in plasma and in CSF.63 This study showed that Aβ40 and Aβ42 peptides were lowered in the CSF by up to 60% and that suppression of Aβ production was sustained for longer in the CSF than in the plasma compartment. With lower doses of semagacestat, or at longer time-points at which point compound concentrations are declining, there was an elevation of Aβ in plasma that was not seen in the CSF.64 These data can be rationalized as follows. At low GSI and substrate concentrations, γ-secretase is stimulated. APP expression in peripheral tissues is lower than in the brain, hence peripherally derived Aβ is initially suppressed following an oral dose (when compound levels are high), but then stimulated as compound levels diminish. In the brain, where APP expression is higher, the stimulation of Aβ is less apparent. As Aβ is trafficked out of the brain rapidly,65 it might also be technically challenging to detect GSI-induced increases in Aβ levels. In PDAPP transgenic mice, which overexpress the hAPP717 mutation,66 dose-related inhibition of brain Aβ production was demonstrated after acute and 7-day dosing.67 In a chronic study, semagacestat was administered daily to 5-month-old PDAPP mice for 5 months at 3, 10, and 30mg/kg.68 This resulted in dose-related reduction in insoluble brain Aβ that was significantly different from control groups at the highest dose for both Aβ40 and Aβ42. There was no significant reduction in plaque as measured immunohistochemically. Interestingly, semagacestat was a more potent inhibitor of Aβ40 than Aβ42 production, an effect seen by others using semagacestat61 and other GSIs of this class.69 Importantly, the dosing of semagacestat was initiated prior to the onset of Aβ plaque deposition in the PDAPP mice, and thus reflects a preventative rather than a therapeutic dosing paradigm. This is an important concept from 2 perspectives: first, in how it relates to its proposed clinical use; and second, because a therapeutic agent can inhibit Aβ deposition via fundamentally different mechanisms (see Fig 1). Several studies have investigated this issue and demonstrated that GSIs prevent the formation of new Aβ plaques, but even with significant suppression of Aβ production, do not mediate the clearance of existing plaques.69-72 In healthy volunteers, semagacestat has a time to reach maximum concentration in plasma of 1 to 1.5 hours and a plasma half-life of 2.5 hours when administered daily for 14 days at doses ranging from 5 to 50mg/person. There was a dose-related reduction in plasma Aβ, followed by a stimulation of up to 500% over baseline for the lowest dose of semagacestat.73 In this study, no reduction in CSF Aβ could be detected when sampled 6 hours after compound dosing. In a phase 2 study, semagacestat was given at 30mg every day (q.d.) for 1 week followed by 40 mg q.d. for 5 weeks to 33 mild–moderate AD patients.74 At the end of the study, there was evidence of Notch-related effects on lymphocytes, but on the whole the drug was well tolerated. There was a 38% suppression of plasma Aβ40 but no effect on CSF Aβ40/42. In the preclinical studies in the PDAPP mouse, 30mg/kg given once daily for 5 months reduced deposited Aβ and suppressed plasma Aβ by approximately 60% at maximal drug concentration. Thus, this level of plasma Aβ reduction was sought in human studies as a translational biomarker. Accordingly, a phase 1 study investigated the Aβ pharmacodynamic effect of 3 doses of semagacestat: 60, 100, or 140mg in normal humans.75 Blood samples were taken at regular intervals up to 24 hours postdose for analysis of compound, and Aβ concentration and CSF samples were collected 4 hours after dosing. The maximum percentage decrease in plasma Aβ from baseline values was 50% for the 60mg group and 73% for the 140mg dose, and occurred between 4 and 6 hours postdose, returning to baseline values between 8 and 13 hours later, depending on the dose. There were slight reductions in CSF Aβ that were significant for Aβ40 at the 140mg dose. As seen previously, there was a large increase in plasma Aβ that followed the initial suppression phase. Although the plasma biomarker response confirmed that γ-secretase was being inhibited in a dose-related manner, there was no evidence that the production of brain Aβ was being affected. Given the excellent brain penetrant properties of semagacestat, it was unlikely that brain γ-secretase was unaffected by the compound, and the most likely explanation for the lack of a measurable Aβ response lay in the technical challenge of measuring CSF Aβ. The inhibition of brain Aβ production by semagacestat was measured using the stable isotope kinetic effect assay.65 Humans were given a continuous intravenous (i.v.) infusion of 13C-leucine for 9 hours to isotope-label proteins. CSF was collected via a spinal tap every hour for up to 36 hours, and Aβ species were immunoprecipitated using a mid-domain antibody before mass spectrometry analysis. The fractional incorporation of 13C-leucine was used to analyze the rate of production and clearance of Aβ. Semagacestat was administered in a single oral dose of 100, 140, and 280mg, and the effects on brain Aβ synthesis and clearance were measured.76 This crucial study proved that semagacestat was able to inhibit brain Aβ production by 47%, 52%, and 84% at 100, 140, and 280mg doses, respectively, over a 12-hour period. A phase 2 safety study77 investigated the tolerability of 100 and 140mg once daily dosing over a 12-week period in mild–moderate AD patients. Although the drug was well tolerated overall, there was an increased incidence of skin rashes and hair color changes, which were indicative of inhibition of Notch signaling. In retrospect, it is noteworthy that both doses numerically worsened ADAS-cog scores. Plasma Aβ levels were inhibited by 65% at the 140mg dose. It is apparent that semagacestat was cautiously developed, and that given the side effect profile of Notch inhibition, it was not possible to increase the dose above 140mg q.d. to garner increased efficacy. Two phase 3 trials (Identity 1 and Identity 2, ClinicalTrials.gov identifiers NCT00594568 and NTC0076241122600) planned to enroll 2,600 mild–moderate AD patients who were randomized to placebo, 100mg semagacestat, and 140mg semagacestat once daily for 76 weeks in 2 trials. The ADAS-cog and ADCS-ADL were the coprimary outcome measures. These trials were halted after an interim futility analysis of Identity 1 showed a significantly increased incidence of skin cancer, infections, and white blood cell and other hematologic abnormalities.78 There was no improvement in cognition as measured by the ADAS-cog, and activities of daily living were significantly worsened at the highest dose. The 140mg dose showed a significant worsening of the CDR-SB and the MMSE. CSF levels of Aβ40, Aβ42, and tau were not altered by semagacestat treatment, whereas phospho-tau 181 (ptau) was significantly, but modestly, reduced. There were no drug effects on fluorodeoxyglucose positron emission tomography (PET), 18" @default.
• W1585083630 created "2016-06-24" @default.
• W1585083630 creator A5007171007 @default.
• W1585083630 creator A5020514779 @default.
• W1585083630 date "2014-07-02" @default.
• W1585083630 modified "2023-10-09" @default.
• W1585083630 title "A critique of the drug discovery and phase 3 clinical programs targeting the amyloid hypothesis for Alzheimer disease" @default.
• W1585083630 cites W139300322 @default.
• W1585083630 cites W1520669015 @default.
• W1585083630 cites W1528113046 @default.
• W1585083630 cites W1534001036 @default.
• W1585083630 cites W1546526125 @default.
• W1585083630 cites W1567490604 @default.
• W1585083630 cites W1572225000 @default.
• W1585083630 cites W1573103890 @default.
• W1585083630 cites W1581267597 @default.
• W1585083630 cites W1605023308 @default.
• W1585083630 cites W1608054934 @default.
• W1585083630 cites W1621971853 @default.
• W1585083630 cites W1655404681 @default.
• W1585083630 cites W1675105744 @default.
• W1585083630 cites W1693283493 @default.
• W1585083630 cites W1817515827 @default.
• W1585083630 cites W1861730929 @default.
• W1585083630 cites W1886532866 @default.
• W1585083630 cites W1928655587 @default.
• W1585083630 cites W1963812982 @default.
• W1585083630 cites W1965275873 @default.
• W1585083630 cites W1965756304 @default.
• W1585083630 cites W1967549235 @default.
• W1585083630 cites W1972163358 @default.
• W1585083630 cites W1972485487 @default.
• W1585083630 cites W1978266853 @default.
• W1585083630 cites W1978924496 @default.
• W1585083630 cites W1979415323 @default.
• W1585083630 cites W1982272424 @default.
• W1585083630 cites W1982816514 @default.
• W1585083630 cites W1985477745 @default.
• W1585083630 cites W1985613394 @default.
• W1585083630 cites W1986688921 @default.
• W1585083630 cites W1986711108 @default.
• W1585083630 cites W1987373495 @default.
• W1585083630 cites W1991993391 @default.
• W1585083630 cites W1998088268 @default.
• W1585083630 cites W2000917982 @default.
• W1585083630 cites W2001436176 @default.
• W1585083630 cites W2001690023 @default.
• W1585083630 cites W2004535959 @default.
• W1585083630 cites W2006219887 @default.
• W1585083630 cites W2007350172 @default.
• W1585083630 cites W2008301592 @default.
• W1585083630 cites W2008645373 @default.
• W1585083630 cites W2011347889 @default.
• W1585083630 cites W2011418880 @default.
• W1585083630 cites W2013078308 @default.
• W1585083630 cites W2013843163 @default.
• W1585083630 cites W2015484447 @default.
• W1585083630 cites W2015675204 @default.
• W1585083630 cites W2016193774 @default.
• W1585083630 cites W2024262813 @default.
• W1585083630 cites W2024917017 @default.
• W1585083630 cites W2027284272 @default.
• W1585083630 cites W2030067782 @default.
• W1585083630 cites W2032061707 @default.
• W1585083630 cites W2032876681 @default.
• W1585083630 cites W2033373102 @default.
• W1585083630 cites W2036288989 @default.
• W1585083630 cites W2039716280 @default.
• W1585083630 cites W2044378879 @default.
• W1585083630 cites W2044975042 @default.
• W1585083630 cites W2048780418 @default.
• W1585083630 cites W2052742260 @default.
• W1585083630 cites W2057838273 @default.
• W1585083630 cites W2060167539 @default.
• W1585083630 cites W2063696444 @default.
• W1585083630 cites W2063808031 @default.
• W1585083630 cites W2065706231 @default.
• W1585083630 cites W2068477037 @default.
• W1585083630 cites W2068660068 @default.
• W1585083630 cites W2069315029 @default.
• W1585083630 cites W2069321794 @default.
• W1585083630 cites W2070880671 @default.
• W1585083630 cites W2076299101 @default.
• W1585083630 cites W2078874988 @default.
• W1585083630 cites W2079749054 @default.
• W1585083630 cites W2083304655 @default.
• W1585083630 cites W2083556882 @default.
• W1585083630 cites W2084453905 @default.
• W1585083630 cites W2085326534 @default.
• W1585083630 cites W2086278983 @default.
• W1585083630 cites W2086572083 @default.
• W1585083630 cites W2087982057 @default.
• W1585083630 cites W2088802253 @default.
• W1585083630 cites W2096576951 @default.
• W1585083630 cites W2096814558 @default.
• W1585083630 cites W2098485296 @default.
• W1585083630 cites W2100610042 @default.
• W1585083630 cites W2103071556 @default.

• next