Reactions to Multiple Ascending Doses of the Microtubule Stabilizer TPI-287 in Patients With Alzheimer Disease, Progressive Supranuclear Palsy, and Corticobasal Syndrome: A Randomized Clinical Trial

Key Points

Question  Do patients with different tauopathies react differently to the microtubule stabilizer TPI-287?

Findings  Two parallel, double-blind, placebo-controlled randomized clinical trials of near-identical design enrolled patients with Alzheimer disease, progressive supranuclear palsy, and corticobasal syndrome. Severe hypersensitivity reactions were observed only in patients with Alzheimer disease who received TPI-287, whereas clinical worsening and biomarker changes were observed in those with progressive supranuclear palsy and corticobasal syndrome.

Meaning  Although the present study does not support continued development of TPI-287 for tauopathies, the results highlight the value of measuring treatment effects in multiple clinical tauopathy syndromes at early stages of development.

Importance  Basket-design clinical trials that allow investigation of treatment effects on different clinical syndromes that share the same molecular pathophysiology have not previously been attempted in neurodegenerative disease.

Objective  To assess the safety, tolerability, and pharmacodynamics of the microtubule stabilizer TPI-287 (abeotaxane) in Alzheimer disease (AD) or the 4-repeat tauopathies (4RT) progressive supranuclear palsy (PSP) and corticobasal syndrome (CBS).

Design, Setting, and Participants  Two parallel-design, double-blind, placebo-controlled phase 1 randomized clinical trials in AD and 4RT were conducted from December 20, 2013, through May 4, 2017, at the University of California, San Francisco, and University of Alabama at Birmingham. A total of 94 patients with clinically diagnosed AD (n = 39) and 4RT (n = 55) were screened; of these, 3 refused to participate, and 10 with AD and 11 with 4RT did not meet inclusion criteria. A total of 29 patients with AD, 14 with PSP, and 30 with β-amyloid–negative CBS (determined on positron emission tomography findings) were enrolled. Data were analyzed from December 20, 2013, through May 4, 2017, based on modified intention to treat.

Interventions  Randomization was 8:3 drug to placebo in 3 sequential dose cohorts receiving 2.0, 6.3, or 20.0 mg/m² of intravenous TPI-287 once every 3 weeks for 9 weeks, with an optional 6-week open-label extension.

Main Outcomes and Measures  Primary end points were safety and tolerability (maximal tolerated dose) of TPI-287. Secondary and exploratory end points included TPI-287 levels in cerebrospinal fluid (CSF) and changes on biomarker, clinical, and neuropsychology measures.

Results  A total of 68 participants (38 men [56%]; median age, 65 [range, 50-85] years) were included in the modified intention-to-treat analysis, of whom 26 had AD (14 women [54%]; median age, 63 [range, 50-76] years), and 42 had 4RT (16 women [38%]; median age, 69 [range, 54-83] years). Three severe anaphylactoid reactions occurred in TPI-287–treated patients with AD, whereas none were seen in patients with 4RT, leading to a maximal tolerated dose of 6.3 mg/m² for AD and 20.0 mg/m² for 4RT. More falls (3 in the placebo group vs 11 in the TPI-287 group) and a dose-related worsening of dementia symptoms (mean [SD] in the CDR plus NACC FTLD-SB [Clinical Dementia Rating scale sum of boxes with frontotemporal dementia measures], 0.5 [1.8] in the placebo group vs 0.7 [1.6] in the TPI-287 group; median difference, 1.5 [95% CI, 0-2.5]; P = .03) were seen in patients with 4RT. Despite undetectable TPI-287 levels in CSF, CSF biomarkers demonstrated decreased chitinase-3–like protein-1 (YKL-40) levels in the 4RT treatment arm (mean [SD], −8.4 [26.0] ng/mL) compared with placebo (mean [SD], 10.4 [42.3] ng/mL; median difference, −14.6 [95% CI, −30.0 to 0.2] ng/mL; P = .048, Mann-Whitney test).

Conclusions and Relevance  In this randomized clinical trial, TPI-287 was less tolerated in patients with AD than in those with 4RT owing to the presence of anaphylactoid reactions. The ability to reveal different tau therapeutic effects in various tauopathy syndromes suggests that basket trials are a valuable approach to tau therapeutic early clinical development.

Trial Registration  ClinicalTrials.gov identifiers: NCT019666666 and NCT02133846

Alzheimer disease (AD), progressive supranuclear palsy (PSP), and corticobasal syndrome (CBS) are distinct neurodegenerative syndromes that are associated with pathologic accumulation of tau protein in the brain. The clinical hallmarks of each disorder are unique: memory loss in AD, postural instability with gaze palsy in PSP, and asymmetric parkinsonism in CBS.¹ Neuropathologically, AD tau aggregates consist of neurofibrillary tangles of 3- and 4-microtubule binding repeat (3R/4R) tau, whereas in PSP and corticobasal degeneration, the most common neuropathology associated with CBS, different tau aggregates consisting primarily of 4R tau accumulate in neurons and glia.¹ Although neuropathologic tau accumulation correlates with disease severity and clinical phenomenology in all 3 diseases, tau-targeted therapeutic trials mostly have included participants with a single clinical diagnosis of either AD or PSP.² Basket studies are clinical trials that include participants defined by a specific underlying molecular cause that may be associated with treatment response independently of clinical phenotype. In oncology, basket clinical trials with treatments targeting a common oncogenic mutation found in different cancers have provided new insights into molecular pathophysiology and evidence of treatment response.³ To date, similar approaches have not been attempted for neurodegenerative tauopathies.

Tau is a microtubule-associated protein that promotes the assembly of tubulin, stabilizes microtubules, and regulates other cellular processes.⁴ In tauopathies, tau disengages from microtubules and may contribute to instability and axonal transport deficiencies owing to a hypothesized loss of tau function.⁵ Microtubule-stabilizing agents have been proposed to compensate for this loss of function when used at low doses. Tau transgenic mouse models treated with the microtubule-stabilizing agents paclitaxel⁶ or epothilone D⁷ had improved axonal transport and motor function and reduced tau pathology.

TPI-287 (abeotaxane) is a blood-brain barrier–penetrable microtubule stabilizer. In animal models, TPI-287 accumulates at higher concentrations in the brain than blood and reduces brain metastatic colonization of breast cancer cells. In melanoma and neuroblastoma clinical trials, the maximal tolerated dose (MTD) of TPI-287 was 125 mg/m².^8-10 In PS19 tau-transgenic mice, TPI-287 reduced hyperphosphorylated tau levels in brain and improved performance on the Morris Water Maze.¹¹ Because it is not clear which sporadic human tauopathy best corresponds to the PS19 model, we investigated the safety and tolerability of TPI-287 intravenous infusions in patients with AD, PSP, and CBS in a randomized clinical trial with a parallel-group, double-blind, placebo-controlled basket design. For funding reasons, 2 clinical trials were conducted in parallel; one in patients with AD and the other in patients with PSP or CBS (referred to as the 4-Repeat Tauopathy [4RT] trial).

Participants were evaluated at the University of California, San Francisco (UCSF), for the AD and 4RT trials and the University of Alabama at Birmingham for the 4RT trial only. Ethics approval was obtained at UCSF and University of Alabama. All participants gave written informed consent at screening. The study was overseen by an independent data and safety monitoring board and followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. The complete AD and 4RT trial protocols are found in Supplement 1.

Patients in the AD cohort met the National Institute on Aging–Alzheimer Association workgroups criteria for probable AD dementia and had a Mini-Mental State Examination (MMSE) score ranging from 14 to 26 (inclusive) at screening (scores range from 14 to 24, with higher scores indicating better cognitive function).¹² Patients in the PSP cohort met the National Institute of Neurological Disorders and Stroke–Society for Progressive Supranuclear Palsy probable or possible PSP criteria, as modified for the Neuroprotection and Natural History in Parkinson Plus Syndromes clinical trial.¹³ Participants in the CBS cohort met the 2013 criteria for possible or probable corticobasal degeneration, CBS subtype.¹⁴

Exclusion criteria included history of significant peripheral neuropathy. Because as much as 30% of CBS can be caused by underlying AD pathologic features,^1,15,16 participants with CBS had an additional exclusion criterion of elevated cortical β-amyloid levels on visual read of florbetapir fluorine 18–labeled positron emission tomography (PET) scans (complete exclusion criteria list is found in the eMethods in Supplement 2).

All participants and study personnel were masked to treatment assignment. Participants were randomly assigned (by an unblinded pharmacist) in an 8:3 ratio of TPI-287 to placebo into sequential dose panels (11 per panel) of 2.0, 6.3, and 20.0 mg/m². In the 4RT trial, separate doses of 2.0 mg/m² for PSP and CBS and of 6.3 and 20.0 mg/m² for combined PSP and CBS panels were enrolled. Doses were selected based on extrapolation from previous microtubule stabilizer preclinical experiments that were confirmed with TPI-287 in PS19 tau-transgenic mice.^9,11 Dose escalation rules specified that if no more than 1 of the 8 participants receiving active treatment experienced dose-limiting toxic effects during the first 4 weeks, the dose of TPI-287 would be escalated to the next higher dose. If 2 of 8 participants in a dose panel experienced a dose-limiting toxic effect, this dose was determined to be the MTD, and no further escalation occurred. A dose-limiting toxic effect was defined as (1) any grade 3 or higher adverse event per National Cancer Institute Common Terminology Criteria for Adverse Events, (2) any grade 2 or higher adverse event in the nervous system considered clinically significant for which there is reasonable possibility that TPI-287 caused the event, or (3) any infusion-related toxic effect (eg, allergic reaction or hypersensitivity) that did not resolve with supportive care. Anaphylactoid hypersensitivity reactions were a known risk of TPI-287 and were considered adverse events of special interest.

Study drug was provided by Cortice Biosciences, Inc, as a concentrate (10 mg/mL) and diluted in normal saline solution. Placebo infusion consisted of normal saline solution. TPI-287 was given as a 1-hour intravenous infusion once every 3 weeks during 9 weeks for a total of 4 infusions for the placebo-controlled period. An additional 6 weeks of open-label extension for a total of 3 infusions was offered to all participants. Lumbar puncture for cerebrospinal fluid (CSF) collection was performed at screening and 1 week after the last double-blind infusion. Magnetic resonance imaging (MRI) scans were performed at screening and 2 weeks after the last double-blind dose.

Study visits occurred from December 20, 2013, through May 4, 2017. After 2 patients experienced a rash, the protocol was amended in May 2014 with addition of 25 mg of intravenous diphenhydramine hydrochloride pretreatment. After an anaphylactoid hypersensitivity reaction in October 2014, the protocol was further amended with addition of 10 mg of intravenous dexamethasone and 20 mg of intravenous famotidine as pretreatment for all participants. At the time of the second amendment, 10 patients had been enrolled in the AD and 6 in the 4RT trial cohorts.

The primary outcome was the safety and tolerability (MTD) of TPI-287 intravenous infusions. Safety assessments included adverse effects, physical and neurological examination, weight, vital signs, electrocardiography, blood tests, CSF cell count with differential, glucose levels, total protein levels, and qualitative reads of the brain MRI. Secondary outcomes included the plasma pharmacokinetics (PK) profile of TPI-287 after a single infusion and steady state TPI-287 CSF level 1 week after completion of the fourth infusion (eMethods in Supplement 2).

Exploratory end points included the following CSF biomarkers: neurofilament light chain (NfL) (NF-light assay; Uman Diagnostics), total tau, phosphorylated tau (p-tau), and β-amyloid 1-42 (Aβ42) (INNO-BIA AlzBio3; Fujirebio). Chitinase-3–like protein 1 (YKL-40) (Quidel Corporation) was added as a post hoc biomarker based on a published hypothesis that taxanes could alter central nervous system inflammation.^7,17 The MRI methods are described in the eMethods in Supplement 2.

In the AD trial, exploratory clinical end points included the MMSE, Alzheimer Disease Assessment Scale–Cognitive,¹⁸ and the Alzheimer Disease Cooperative Study Activities of Daily Living scale¹⁹ scores. In the 4RT trial, exploratory outcomes included the PSP Rating Scale score,²⁰ Schwab and England Activities of Daily Living scale score, the sum of boxes score from the Clinical Dementia Rating (CDR) Dementia Staging Instrument plus the NACC FTLD Module (FTLD-SB [CDR scale sum of boxes with frontotemporal dementia measures]; scores range from 0 to 10.5, with higher scores indicating worse dementia),²¹ MMSE, and phonemic fluency²² scores. Mood was assessed by the Geriatric Depression Screen score.²³

Data were analyzed from December 20, 2013, through May 4, 2017. Sample size was based on a standard phase 1, PSP, multiple ascending-dose escalation scheme²⁴ with 3 additional participants per panel to increase power for exploratory analyses. Data from all participants who received at least 1 infusion were included in the safety analysis.

A modified intention-to-treat analysis of all participants who received 2 or more infusions and had end-of-study efficacy assessments was used for the secondary and exploratory outcomes. A previous study²⁵ showed no differences between PSP and CBS in decline on the PSP Rating Scale or Schwab and England Activities of Daily Living score during 6 and 12 months, and in our study, CBS and PSP were combined for exploratory analyses. Baseline demographic variables and baseline to end-of-study outcome differences between treatment and placebo arms were analyzed with a Shapiro-Wilk normality test, and nonparametric tests were used as appropriate. For outcome differences between treatment and placebo arms, median difference and 95% CI were assessed by Mann-Whitney test and Hodges-Lehmann method. Differences between treatment dose panels and placebo were compared with a Kruskal-Wallis test with post hoc Dunn test and Benjamini-Hochberg correction. The significance level was set to 2-sided P ≤ .05. Statistical analyses were conducted using Stata, version 13.1 (StataCorp LLC)

A total of 94 patients (39 in the AD trial and 55 in the 4RT trial) were screened from local and national referrals; 29 with AD and 44 with 4RT were randomized to TPI-287 or placebo, and all were included in the safety analysis (Figure 1). In the AD trial, only 4 patients were randomized to the high-dose panel before the trial was terminated for safety reasons. A total of 68 patients (38 men [56%] and 30 women [44%]; median age, 65 [range, 50-85] years), including 26 patients with AD (14 women [54%]; median age, 63 [range, 50-76] years) and 42 with 4RT (16 women [38%]; median age, 69 [range, 54-83] years), had sufficient data to be included in the modified intention-to-treat analysis. In the modified intention-to-treat population, baseline characteristics did not differ between the placebo and treatment arms in the 2 trials except for lower CDR plus NACC FTLD-SB scores in the CBS low-dose panel (median, 0 [range, 0-4.0]) compared with the medium- (median, 4.0 [range, 0-10.5]) and high-dose (median, 4.0 [range, 2.5-7.0]) panels (Table 1). Post hoc analyses demonstrated expected differences in CSF β-amyloid, p-tau, and NfL levels among the 3 diagnostic cohorts. The AD cohort had the lowest Aβ42 level (mean [SD], 283 [72] pg/mL) compared with the PSP and CBS cohorts (mean [SD], 462 [135] and 519 [132] pg/mL, respectively). Levels of p-tau were highest in the AD cohort (mean [SD], 64 [25] pg/mL) compared with the PSP and CBS cohorts (mean [SD], 26 [7] and 28 [7] pg/mL, respectively). Levels of NfL were highest in the CBS cohort (mean [SD], 6097 [3552] pg/mL) compared with the AD and PSP cohorts (mean [SD], 2643 [1102] and 3482 [1717] pg/mL, respectively) (Figure 2A and the eFigure in Supplement 2). Post hoc analysis of baseline volumetric MRI measurements in each diagnostic cohort demonstrated atrophy in expected regions, including the temporal parietal lobes for AD, frontal white matter for CBS, and midbrain pontine white matter for PSP (Figure 2B).

Adverse events are listed in Table 2. No safety signals were reported on laboratory assessments, weight, vital signs, electrocardiography, physical and/or neurological examination findings, or MRI for both trials. In the AD trial, all 3 serious adverse events (15%) were anaphylactoid reactions (narrative descriptions are given in the eResults in Supplement 2) that occurred in the treatment arm. Because 2 anaphylactoid reactions occurred early in the high-dose panel, the AD trial was stopped for safety considerations. In the 4RT trial, no serious adverse events or anaphylactoid reactions were reported, although more participants in the treatment arm experienced falls (11 of 32 [34%]) compared with the placebo arm (2 of 12 [17%]; χ² = 1.3; P = .16). Patients in the treatment arms for both trials reported a higher incidence of headaches, dizziness, constipation, diarrhea, and nausea. Five participants in the AD trial discontinued early, 3 owing to anaphylactoid reactions that resulted in serious adverse events and 1 owing to mild infusion-related rash and 1 owing to withdrawal of consent. In the 4RT trial, 1 participant withdrew owing to a mild infusion-related rash and another withdrew owing to disease burden.

Pharmacokinetics data were available on 14 patients in the AD cohort and 38 in the 4RT trial cohort (eTable in Supplement 2). Plasma TPI-287 concentrations were highest from 30 to 60 minutes after the infusion, with dose-dependent increases seen across dose panels at 5 and 60 minutes after infusion (mean [SD] peak 60-minute postinfusion plasma concentration after 20-mg/m² dose: 75.4 [17.3] ng/mL in the AD cohort and 59.5 [15.5] ng/mL in the 4RT trial cohort). No steady-state CSF concentrations of TPI-287 were detectable 1 week after the last infusion, although plasma TPI-287 concentrations were measurable in highest-dose panel in the 4RT trial cohort at this time (mean [SD], 1.1 [2.1] ng/mL).

For the exploratory CSF biomarker end points, no differences were found in rate of change between treatment and placebo arms in the AD or 4RT trial cohort. Post hoc exploratory analysis demonstrated a reduction in YKL-40 levels in the 4RT cohort treatment arm (median [range], −15 [49 to 76] ng/mL) compared with the placebo arm (median [range], 4 [−30 to 134] ng/mL; median difference, −14.6 [95% CI, −30.0 to 0.2] ng/mL; P = .048) (Table 3). No treatment-related differences were noted in the exploratory imaging end points.

For exploratory clinical end points in the AD trial, less decline in MMSE scores was observed in the treatment (median [range], 0 [−4 to 4]) compared with the placebo arms (median [range], −3 [−4 to 1]; median difference, 2.0 [95% CI, 0-4.0]; P = .04) (median and range are given in Table 3). In the 4RT trial, a more severe worsening in CDR plus NACC FTLD-SB score was observed in the 4RT treatment arm (median [range], 0.5 [−3 to 5]) compared with the placebo arm (median [range], −0.75 [−3 to 3]; median difference, 1.5 [95% CI, 0-2.5]; P = .03), with a dose-dependent trend for greater worsening with higher dose. A slight worsening of Geriatric Depression Screen in the 4RT treatment arm (median [range], 1 [−8 to 6]) compared with placebo (median [range], −1 [−4 to 1]; median difference, 2.0 [95% CI, 0-3.0]; P = .03) was also observed.

In a basket-design randomized clinical trial that recruited patients with the tauopathies AD, PSP, or β-amyloid PET–negative CBS, we examined the safety, tolerability, PK, pharmacodynamic effects, and clinical effects of the microtubule stabilizer TPI-287. Treatment-related immune hypersensitivity reactions that resulted in serious adverse effects were observed in patients with AD but not in those with PSP or CBS, leading to an MTD of 6.3 mg/m² for patients with AD, whereas a higher dose of 20.0 mg/m² was tolerated in patients with PSP and CBS. Of note, hypersensitivity reactions continued to occur in our AD cohort even after the addition of dexamethasone and antihistamine pretreatment, which may reflect stronger genetic links to the innate immune system in AD than in primary tauopathies.²⁶ Instead, a notable increase in falls occurred among patients with PSP in the treatment arm compared with the placebo arm, and a dose-related worsening of global cognitive status on the CDR plus NACC FTLD-SB occurred in the 4RT trial cohort treated with TPI-287. These findings demonstrate that patients with different tauopathies can experience different safety profiles when exposed to the same tau-directed therapy.

The human serum PK profile of TPI-287 has been previously described.⁸ Similar to that study, the highest plasma TPI-287 concentrations were observed 30 to 60 minutes after infusion start.⁸ Animal models have demonstrated excellent TPI-287 blood-brain barrier penetration and persistently elevated brain parenchymal concentrations exceeding plasma levels 96 hours after infusion.⁹ However, CSF TPI-287 levels were not detectable 7 days after the final infusion. Nonetheless, treatment-related decreases in CSF YKL-40 levels in the 4RT trial treatment group suggest that TPI-287 had central nervous system activity. One possible explanation for our inability to detect TPI-287 in CSF might be that it remained in the brain parenchyma owing to its high hydrophobicity.

Although the AD treatment group showed a smaller decline in MMSE scores compared with the placebo group, this outcome was likely driven by the greater-than-expected decline of 2.1 points in the placebo group during 11 weeks and was not an effect of TPI-287.²⁷ Therefore, the treatment differences should not be interpreted as meaningful. Annualized CDR plus NACC FTLD-SB change scores for FTLD syndromes have been estimated to be 3.5.²¹ The 4RT trial demonstrated no clear change in the CDR plus NACC FTLD-SB score in the placebo group but did show worsening scores in a dose-dependent fashion during 11 weeks (higher doses led to greater decline). The dose-dependent worsening in CDR plus NACC FTLD-SB in the 4RT trial combined with more frequent falls as an adverse event, which may be a result of subclinical taxane toxic effects in peripheral nerves, do not support further development of TPI-287 for PSP or CBS.

To the best of our knowledge, this study is the first placebo-controlled randomized clinical trial for β-amyloid PET–negative CBS. The pathologic heterogeneity of CBS—with as many as one-third of cases possibly caused by underlying AD pathology, approximately one-half by underlying 4R tauopathies (roughly 3:1, corticobasal degeneration to PSP), and the remaining by TDP-43 proteinopathy, dementia with Lewy bodies, or other diseases—has complicated clinical trial planning.^1,15,16 We attempted to exclude CBS due to AD pathologic changes using β-amyloid PET. Baseline biomarker analyses suggested that the enrollment criteria were able to accurately identify patients with each of the different tauopathies. The AD cohort had the lowest CSF Aβ42 concentrations and highest p-tau concentrations compared with the PSP and CBS cohorts, similar to previous observations.²⁸ Together, these results suggest that recruitment of patients with different tauopathies in a single trial is feasible and the diagnostic criteria used are reliable. In future studies, the addition of tau PET tracers such as [¹⁸F]flortaucipir may further refine CBS recruitment accuracy.²⁹

The basket design used for this clinical trial revealed differences in adverse events that might reflect differences in the cause of different tauopathy syndromes. Genome-wide association studies have demonstrated genetic overlap between AD and immune-mediated diseases, with many late-onset AD risk genes being highly expressed in microglia and astrocytes.^26,30 Meanwhile, most risk factor polymorphisms identified in corticobasal degeneration and PSP genome-wide association studies do not involve the immune system.^26,31 Our finding of more severe hypersensitivity reactions in the AD cohort suggests clinically meaningful differences between immune system function in AD compared with CBS and PSP.

Our study has important limitations. Although the blood-brain barrier penetration of TPI-287 in animal models has been previously reported,⁹ we were not able to measure steady state concentrations of TPI-287 in CSF 7 days after multiple infusions. The absence of a target engagement biomarker that would show microtubule stabilization or other downstream pharmacodynamic biomarkers related to tau function limited our ability to determine whether TPI-287 engaged its molecular target or exerted influence on tau pathophysiology. Therefore, despite evidence of central nervous system biological activity reflected by CSF YKL-40 changes and adverse cognitive effects, the biological hypothesis that TPI-287 bound to central nervous system microtubules was not tested, and we could not comment on the potential for disease modification. Future clinical trials of microtubule stabilizers will need better pharmacodynamic biomarkers to be successful.

This study demonstrates the potential value of conducting basket clinical trials comparing the effects of tau-directed therapies in AD and the primary tauopathies PSP and β-amyloid PET-negative CBS. These studies may reveal important differences in safety, tolerability, and potential efficacy between neurodegenerative syndromes that share underlying tau pathologic features.

Accepted for Publication: September 13, 2019.

Corresponding Author: Adam L. Boxer, MD, PhD, Memory and Aging Center, Department of Neurology, Sandler Neurosciences Center, University of California, San Francisco, 675 Nelson Rising Ln, Ste 190, San Francisco, CA 94158 (adam.boxer@ucsf.edu).

Published Online: November 11, 2019. doi:10.1001/jamaneurol.2019.3812

Author Contributions: Dr Boxer served as the principal investigator and Dr Koestler as the project manager. Dr Boxer had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Z. Miller, Koestler, Hare, Shamloo, B. L. Miller, Boxer.

Acquisition, analysis, or interpretation of data: Tsai, Z. Miller, Rojas, Ljubenkov, Rosen, Rabinovici, Fagan, Cobigo, Brown, Jung, Hare, Geldmacher, Natelson-Love, McKinley, Luong, Chuu, Powers, Mumford, Wolf, Wang, Roberson, Boxer.

Drafting of the manuscript: Tsai, Koestler, Cobigo, Hare, Powers, Mumford, Wolf, Shamloo, Boxer.

Critical revision of the manuscript for important intellectual content: Tsai, Z. Miller, Koestler, Rojas, Ljubenkov, Rosen, Rabinovici, Fagan, Brown, Jung, Geldmacher, Natelson-Love, McKinley, Luong, Chuu, Wang, B. L. Miller, Roberson, Boxer.

Statistical analysis: Tsai, Ljubenkov, Brown, Jung, Mumford, Wolf, Wang, Boxer.

Obtained funding: B. L. Miller, Boxer.

Administrative, technical, or material support: Z. Miller, Koestler, Rosen, Fagan, Cobigo, Jung, Hare, Geldmacher, McKinley, Luong, Chuu, Powers, Mumford, Shamloo, B. L. Miller, Roberson, Boxer.

Supervision: Z. Miller, Koestler, Shamloo, B. L. Miller, Roberson, Boxer.

Conflict of Interest Disclosures: Dr Tsai reported receiving research support from the University of California and grant K23AG055688 from the National Institute of Aging (NIA); consulting for ExpertConnect and Grifols; and current employment at Denali Therapeutics. Dr Z. Miller reported receiving research support from grant K23 AG048291 from the NIA, with additional funds from the Hellman Research Scientist Award, and the Arking Foundation for Frontotemporal Dementia. Dr Rojas reported serving as a site principal investigator for clinical trials sponsored by Eli Lilly and Company and receiving support from the National Institutes of Health (NIH). Dr Ljubenkov reported serving as a site principal investigator for clinical trials sponsored by AbbVie, Inc, and Alector, Inc. Dr Rosen reported receiving research support from Biogen, Inc, and the NIH and consulting agreements with Wave Neuroscience and Ionis Pharmaceuticals. Dr Rabinovici reported receiving research support from Avid Radiopharmaceuticals, Eli Lilly and Company, GE Healthcare, and Life Molecular Imaging; receiving consulting fees from Eisai Co, Ltd, Genentech, Inc, Merck & Co, and Roche Diagnostics; and receiving an honorarium as associate editor for JAMA Neurology. Dr Fagan reported receiving research funding from the NIA of the NIH, Biogen, Inc, Centene Corporation, Fujirebio, Inc, and Roche Diagnostics; serving as a member of the scientific advisory boards for Roche Diagnostics, Genentech, Inc, and AbbVie, Inc; and consulting for Araclon/Grifols, Diadem Pharmacy, and DiamiR. Dr Geldmacher reported receiving grants from UCSF during the conduct of the study and grants from Biogen, grants from Bristol-Myers Squibb, grants from Eisai, grants and personal fees from Genentech, grants from AbbVie, grants from Neurim, grants from Janssen, grants from Lilly, and personal fees from Grifols outside the submitted work. Dr Powers reported receiving grants from Alzheimer's Association, grants from Corticobasal Degeneration Solutions, and grants and nonfinancial support from Cortice Bioscience during the conduct of the study. Dr Shamloo reported receiving grants from the NIH during the conduct of the study. Dr B. L. Miller reported being the medical director for John Douglas French Foundation and scientific director for the Tau Consortium; serving on the Director/Medical Advisory Board of the Larry L. Hillblom Foundation and as a Scientific Advisory Board Member for the National Institute for Health Research Cambridge Biomedical Research Centre and its subunit, the Biomedical Research Unit in Dementia (United Kingdom); and serving as a board member for the American Brain Foundation. Dr Roberson reported receiving research support from grants RF1AG059405, RF1AG059009, and R01NS075487 from the NIH, Bluefield Project to Cure Frontotemporal Dementia, Alzheimer’s Drug Discovery Foundation, BrightFocus Foundation, Weston Brain Institute, the Alzheimer’s Association, AbbVie, Inc, Alector, Inc, Avid Radiopharmaceuticals, Biogen, Inc, Bristol-Myers Squibb, C2N, Eli Lilly and Company, and Roche Diagnostics; consulting for Applied Genetic Technologies Corporation, AVROBIO, Inc, Biogen, Inc, and Novartis International AG; and owning intellectual property related to tau. Dr Boxer reported research support from grants U54NS092089, R01AG038791, U01AG045390, and U24AG057437 from the NIH, Tau Research Consortium, Association for Frontotemporal Degeneration, Bluefield Project to Cure Frontotemporal Dementia, Corticobasal Degeneration Solutions, and the Alzheimer’s Association; serving as a consultant for Aeton, AbbVie, Inc, AGTC, Alector, Inc, Amgen, Inc, Arkuda Therapeutics, Arvinas, Inc, Eisai Co, Ltd, Ionis Pharmaceuticals, iPierian, Inc, Janssen Pharmaceutica, Lundbeck Pharmaceuticals, Merck & Co, Novartis International AG, Pinteon Therapeutics, Inc, Passage Bio, Samumed, LLC, Third Rock Ventures, Toyama Chemical Co, UCB, and Wave Neuroscience; owing stock options in Alector, Inc, and Arvinas, Inc; and receiving research support from Avid Radiopharmaceuticals, Biogen, Inc, Bristol-Myers Squibb, C2N, Cortice Biosciences, Inc, Eli Lilly and Company, Forum Pharmaceuticals, Inc, Genentech, Inc, Janssen Pharmaceutica, Pfizer, Inc, Roche Diagnostics, and TauRx Therapeutics, Ltd. No other disclosures were reported.

Funding/Support: This study was supported by the Alzheimer’s Association (AD trial) and Corticobasal Degeneration Solutions, the Tau Consortium, and UCSF (University of California, San Francisco) Parkinson’s Spectrum Disorders Center (4RT trial). The study drug was donated by Cortice Biosciences, Inc, which provided financial support for preclinical experiments with TPI-287, site monitoring, and some of the pharmacokinetics analyses.

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank the study patients and their caregivers and members of the data safety monitoring board (Carmela Tartaglia, MD, FRCP, Catherine Lomen-Hoerth, MD, PhD, Heidi Kirsch, MD, MS, and Nicholas Butowski, MD) for their contributions to the clinical trials. The members of the board were offered an honorarium.