Abstract
Background: Aβ42-immunization reduces plaque burden and improves cognition in transgenic mouse models of Alzheimer disease (AD). This phase 1 study evaluated the safety, tolerability, and immunogenicity of AN1792 (human aggregated Aβ42) in patients with mild to moderate AD.
Methods: Twenty patients were enrolled into each of four dose groups and randomly assigned to receive IM AN1792 (50 or 225 μg) with QS-21 adjuvant (50 or 100 μg) or QS-21 only (control) in a 4:1 active:control ratio on day 0 and at weeks 4, 12, and 24. Patients could receive up to four additional injections of a polysorbate 80 modified formulation at weeks 36, 48, 60, and 72. Safety, tolerability, immunogenicity, and exploratory evidence of efficacy were evaluated.
Results: Treatment-related adverse events were reported in 19 (23.8%) patients, but no relationship was observed between AN1792 dose and incidence. One patient developed meningoencephalitis that was diagnosed after death (not directly related to study treatment) and 219 days after discontinuing from the study. Five deaths occurred during the study follow-up, but none was considered to be directly related to study treatment. During the period of the first four injections, 23.4% of AN1792-treated patients had a positive anti-AN1792 antibody titer (an anti-AN1792 antibody titer of ≥1:1,000). This increased to 58.8% after additional injections with the modified formulation. Disability Assessment for Dementia scores showed less decline among active compared with control patients at week 84 (p = 0.002). No treatment differences were observed in three other efficacy measures.
Conclusions: AN1792 + QS-21 elicited a positive antibody response to Aβ42 in more than half of this elderly study population.
Therapeutic options for treatment of Alzheimer disease (AD) have so far focused on modifying neurotransmitter systems, in particular the cholinergic system, to maximize the remaining activity in affected neuronal circuits.1,2⇓ However, these approaches are likely to treat only the symptoms of AD without altering the underlying disease process or slowing disease progression.
AD is characterized pathologically by the presence of cerebral β-amyloid (Aβ) plaques (with Aβ42 as the major peptide constituent), neurofibrillary tangles, and neuronal loss.3 Furthermore, genetic linkage indicates that mutations affecting Aβ production are causative for some cases of AD.4–6⇓⇓ The notion of Aβ as an AD-causing factor has also been supported by experiments showing fibrillary tangle formation after intracerebral microinjections of Aβ into P301L mutant tau transgenic mice,7 and an influence of APP or Aβ on tangle formation in double mutant tau/APP mice.8 Accordingly, several strategies for removing or reducing the pathologic amyloid burden in AD are under investigation.3
Active immunization with human aggregated Aβ42 (AN1792) effectively reduces cerebral Aβ plaque burden in mice that express mutations implicated in familial AD.9–11⇓⇓ In young PDAPP mice, immunization generated anti-Aβ antibodies that almost entirely prevented the development of Aβ plaques, neuritic dystrophy, and gliosis.9 Furthermore, immunizing older PDAPP mice (with an existing extensive plaque burden) markedly reduced the extent and progression of AD-like neuropathology.10 Immunization also resulted in improvements in cognitive performance, with12,13⇓ or without14 a reduction of cerebral amyloid deposits. These studies strongly suggest Aβ immunization as a potential disease-modifying therapeutic strategy for patients with AD.
Immunization of Aβ is the first immunotherapeutic approach to AD to enter clinical studies. The aim of the current study was to determine the safety, tolerability, and immunogenicity of four dose combinations of AN1792 + QS-21 (adjuvant) vs QS-21 alone as a control, in patients with mild to moderate AD, and to determine the optimal dosing regimen for a larger phase 2a study. Clinical assessment measures were also included to explore the potential efficacy of AN1792 + QS-21, although the study was not powered to detect differences in efficacy between treatment groups.
Patients and methods.
Patients.
Eligible patients were aged ≤85 years, lived with their caregivers, and met the criteria for a diagnosis of probable AD as defined by the criteria of the National Institute of Neurologic and Communicative Disorders and Stroke-AD and Related Disorders Association.15 Additional criteria included a score of 14 to 26 on the Mini-Mental State Examination (MMSE),16 a Rosen-Modified Hachinski Ischemic score ≤4, and written, informed consent at time of entry into the study. The local independent ethics committee for each study site approved the protocol, amendments, and informed consent forms before study initiation. The study was conducted in accordance with the International Conference on Harmonisation Tripartite Guideline on Good Clinical Practice and in compliance with the Declaration of Helsinki and subsequent revisions.17 An independent Safety Monitoring Committee assessed safety data throughout the study.
Study design.
This was a multicenter, phase 1, randomized, multiple-dose, dose escalation, double-blind study conducted at four study sites in the United Kingdom and involved 80 patients with mild to moderate AD. Twenty patients were enrolled in each of four dose groups and randomly assigned in a double-blind manner to receive treatment in a 4:1, active:control ratio (within each dose group patients were randomly assigned in blocks of five using statistical random number tables generated through the Clinical Trials Materials Group at Elan). During the conduct of the study, the code was not to be shared with the sponsor, the investigators, or study site personnel (other than the study pharmacist who was unblinded to treatment), unless specifically required for safety reasons. In each dose group, treatment was administered as a 1 mL intradeltoid injection of QS-21 (immunogenic adjuvant, surface-active saponin, Aquila Biopharmaceuticals, Inc.) alone or in combination with AN1792 at weeks 0, 4, 12, and 24. The doses for each group were as follows: Group A, AN1792 50 μg + QS-21 50 μg or QS-21 50 μg alone; Group B, AN1792 50 μg + QS-21 100 μg or QS-21 100 μg alone; Group C, AN1792 225 μg + QS-21 50 μg or QS-21 50 μg alone; Group D, AN1792 225 μg + QS-21 100 μg or QS-21 100 μg alone. Study assessments were performed at weeks 1, 2, 4, 6, 8, 12, 14, 16, 20, 24, 26, 28, 32, 39, 52, and 65 after the first immunization.
After an interim data review to assess tolerability and immunogenicity, the study protocol was amended to add an optional extension phase (up to a total of 84 weeks after the first injection) that permitted patients to receive additional study injections at weeks 36, 48, 60, and 72. In the extension phase, the formulation of active and control immunizations was modified by the addition of 0.4% polysorbate 80 (PS-80) to improve product stability. Patients from Groups B, C, and D could receive up to four additional injections, and patients from Group A could receive up to three additional injections (because of the timing of the amended protocol). The study schedule was amended so that assessment visits after week 32 took place at weeks 36, 38, 40, 44, 48, 50, 52, 56, 60, 62, 64, 68, 72, 74, 76, 80, and 84. In some patients, week 64 visits actually occurred during week 65.
Outcome measures.
The primary outcome measures were safety and tolerability, which were monitored throughout the trial by adverse event (AE) reporting, physical and neurologic examinations, vital signs, and routine laboratory evaluations (biochemical and hematologic tests and urinalyses). The secondary outcome measure was the immunogenicity of the four dose combinations as assessed by ELISA analysis of serum samples. The ELISA was able to detect immunoglobulins (Ig) of the IgG, IgM, and IgA classes against AN1792 and had a lower limit of detection of a titer of 1:100 ELISA units. Based on preclinical studies, a positive response was defined as an anti-AN1792 antibody titer ≥1:1,000. Anti-AN1792 antibody responders were defined as patients with an antibody titer ≥1:1,000 at 4 weeks after an injection or a titer ≥1:5,000 at any time point after an injection.
Exploratory evaluations of efficacy were assessed using four scales: two cognitive scales, the AD Assessment Scale-Cognitive Subscale (ADAS-Cog; score range 0 to 70, higher score indicates greater impairment),18 and the MMSE (score range 0 to 30, lower score indicates greater impairment)16; a global rating scale, the AD Cooperative Study-Clinical Global Impression of Change (ADCS-CGIC; 7-point scale rating an independent investigator’s assessment of change from baseline, higher score indicates greater impairment)19; and a functional disability scale, the Disability Assessment for Dementia (DAD).20 The total DAD score is the sum of 28 items expressed as a percentage of all items answered; it contains items related to basic self-care and instrumental activities of daily living and can range from 0 to 100%, with lower scores indicating greater impairment.
Statistical analysis.
As this trial was designed to explore the safety of AN1792 + QS-21, 16 control and 64 treated patients were deemed adequate to demonstrate initial safety. No formal power calculations were made with respect to the ability to detect significant treatment differences on any efficacy measures. Patients in Groups A, B, C, and D were combined into a single active group and a single control group for safety and exploratory efficacy analyses.
All efficacy variables were analyzed at day 0, week 32, week 64 (data for patients with week 65 visits were combined with week 64 data for analyses), and week 84. Analyses were also conducted at a final visit time point, which consists of data from each patient’s last observation carried forward (LOCF; only post-baseline values were carried forward). Exploratory efficacy was evaluated in the intent-to-treat (ITT) population (all patients who had at least one baseline efficacy evaluation and at least one postinjection efficacy evaluation; patients who had no post-baseline efficacy evaluations were excluded from the ITT population).
ADAS-Cog, MMSE, and DAD data were analyzed using analysis of covariance (ANCOVA) with baseline as the covariate and treatment and center as main effects. ADCS-CGIC data were analyzed using analysis of variance (ANOVA) with treatment and center as main effects. In addition, the Cochran-Mantel-Haenszel test using equally spaced scores for the ordered levels of the response variable was used to evaluate the effect of treatment on ADCS-CGIC scores.
Results.
Patients.
Eighty patients were enrolled in the study and received study treatment: 64 received AN1792 + QS-21, 16 received QS-21 alone (figure 1). A total of 64 patients entered the protocol extension phase (51 patients from the active treatment group and 13 from the control group), during which up to a further four injections of the PS-80 modified formulation were administered. Study enrollment began in April 2000 and the final observations were made in June 2002. Patient demographics, baseline characteristics, and concomitant medications are summarized in table 1.
Figure 1. Flow chart of study patients. Early withdrawals from the study and the reasons for withdrawal are listed. †A maximum of three injections was allowed in the extension phase because of the timing of the protocol amendment; ††patients did not receive injection 8 due to the sponsor’s decision to halt further dosing of AN1792. ‡One patient died from carcinoma of the lung after completing treatment according to the original protocol. *One patient died from pulmonary embolism after withdrawing from the study; **both patients died (cerebrovascular accident, chest infection) after withdrawing from the study.
Table 1 Patient demographics, baseline characteristics, and concomitant medications
Exposure and concomitant medications.
All 80 patients received at least two injections of study treatment and 91.3% received all four injections scheduled in the original protocol. Of the 64 patients who entered the protocol extension phase, 89.1% received seven injections and 51.6% received all eight injections. For these injections, 0.4% PS-80 was added to the AN1792 formulation to improve stability. The 13 patients in Group A who entered the protocol extension phase could only receive up to seven injections (at weeks 0, 4, 12, 24, 48, 60, and 72) due to the timing of the protocol amendment. Patients in Group D could also only receive up to seven injections because of the sponsor’s decision to suspend further treatment with AN1792 following the reporting of encephalitis in several patients in an ongoing phase 2a study of AN1792. Details of these cases have been published.21
The most frequently used concomitant medications included donepezil hydrochloride (75% of patients in the active group and 75.1% in the control group), influenza vaccine (50% and 31.5%), aspirin (39.1% and 43.8%), and paracetamol (acetaminophen) (23.4% and 50%).
Safety and tolerability results.
Adverse events.
AEs were reported in 97.5% of patients during the study period (table 2 and table E-1 on the Neurology Web site). The most frequently reported AEs were accidental injury (38.8%), infection (31.3%), and confusion (31.3%); these events were reported at a similar frequency in the active and control groups (see table 2). In the majority of patients, the maximum intensity of AEs was either mild (active: 21.9% and control: 25%) or moderate (54.7% and 56.3%).
Table 2 Summary of treatment-emergent adverse events reported by >10% patients overall during treatment with AN1792 + QS-21 or QS-21 alone
AEs that were considered treatment related by the investigators were reported in 28.1% of patients who received active treatment, compared with 6.3% of patients in the control group (table E-2 on the Neurology Web site); the majority were mild in intensity. The most frequently reported treatment-related AE in patients receiving active treatment was injection site pain (15.6%; see table E-2). No relationship was observed between the AN1792 dose and the incidence of treatment-related AEs.
Five (6.3%) patients (mean age 74 years) died during the study period; these deaths were due to arterial anomaly, pulmonary embolus, cerebrovascular accident, infection (all in the active treatment group), and carcinoma of the lung (control group). No deaths were considered to be directly related to study treatment. Non-fatal serious AEs (SAEs) were reported in 22 (27.5%) patients (26 events in the active group and 9 in control patients) during the study (see table E-1). Four patients had non-fatal SAEs that were considered related to study treatment; these events were rash, confusion and syncope, encephalitis, and worsening of dementia. The encephalitis case was diagnosed histologically after the patient’s family consented to an autopsy; the results have been previously reported.22 This patient had experienced dizziness and disorientation, and showed signs of functional deterioration on day 288, approximately 36 days after study injection 5 (the first dose of the altered formulation of AN1792 50 μg + QS-21 100 μg with 0.4% PS-80). These symptoms were initially attributed to a primary CNS neoplasm, and the patient was treated with dexamethasone. As a consequence of these symptoms, she was withdrawn from the study on day 380. The patient subsequently died from a pulmonary embolism secondary to deep vein thrombosis 347 days after administration of the last study injection (injection 5) and 219 days after her withdrawal from the study. At the postmortem examination, there were no signs of a CNS neoplasm, but pathologic changes consistent with T-lymphocyte meningoencephalitis were present.22
A total of 10 patients (9 active [14.1%] and 1 control [6.3%]) prematurely withdrew from the study due to AEs (see table E-1). The AEs resulting in discontinuation were hostility (4 patients), depression (2 patients), hallucinations (2 patients), accidental injury (2 patients), and infection (2 patients); some patients reported more than one AE. Treatment-related AEs (hostility, n = 3; hallucinations, n = 1) led to discontinuation in four patients (3 active [4.7%] and 1 control [6.3%]).
No apparent dose-related trends were observed in the clinical laboratory measures studied. A total of 20 (25%) patients developed one or more clinically important abnormal laboratory values (table 3); these were reported more frequently in AN1792-treated patients than in control patients. However, only one of these abnormalities (a moderate event of increased alkaline phosphatase that was reported in a patient receiving QS-21 100 μg alone) was considered to be treatment related by the investigator.
Table 3 Summary of anti-AN1792 antibody responders among the population of patients who entered the protocol extension
Mean decreases in blood pressure, especially systolic pressure, were observed immediately following each injection in the active treatment group. The greatest decreases were typically observed 60 minutes after an injection, with mean systolic blood pressure changes after an injection of −1.2 mm Hg at week 0, −6.8 mm Hg at week 24, and −8.8 mm Hg at week 72 among patients receiving active treatment. Changes in blood pressure after control injections were +2.9 mm Hg at week 0, −2.4 mm Hg at week 24, and −3.0 mm Hg at week 72.
Immunologic findings.
Control patients had anti-AN1792 antibody titers that were below the lower limit of detection (<1:100) at nearly all time points. No control patients developed a positive anti-AN1792 antibody titer (>1:1,000) at any time during the study. Among study patients who received active treatment (n = 64), no positive anti-AN1792 antibody titers were reported within 2 weeks after injection 1. After injection 2, one positive antibody titer (in the AN1792 225 μg + QS21 50 μg dose group) was reported. More positive anti-AN1792 antibody titers became evident after injections 3 and 4 (see table 3, supplementary figure E-1). A notable increase in the proportion of patients with positive anti-AN1792 antibody titers was observed after injection 5 for all dose groups except AN1792 50 μg + QS-21 50 μg. Patients in this latter dose group did not receive injection 5 because of the timing of the amended protocol. Of all patients who received AN1792 (n = 64), 53.1% had a positive anti-AN1792 antibody titer at some point during the study. Of those who entered the protocol extension (n = 51), 58.8% of patients had a positive antibody response and 56.9% were considered anti-AN1792 antibody responders at some point during the study (see table 3). Of the 13 AN1792-treated patients who did not enter the protocol extension, 15.4% were considered anti-AN1792 antibody responders.
Measures of efficacy—exploratory results.
ADAS-Cog scores showed similar increases in the active and control groups during the study (figure 2A, table E-3 on the Neurology Web site), indicating a decline in cognition. No significant treatment effects were observed at any visit. Similarly, no significant treatment effects on MMSE were observed at any visit, as MMSE scores showed declines from baseline in both active and control groups during the study period (figure 2B, see table E-3).
Figure 2. Summary of changes in efficacy measures for patients who received AN1792 + QS-21 or QS-21 alone. (A) Alzheimer Disease Assessment Scale-Cognitive Subscale (ADAS-Cog); (B) Mini-Mental State Examination (MMSE); (C) Disability Assessment for Dementia (DAD); (D) frequency distribution of AD Cooperative Study-Clinical Global Impression of Change (ADCS-CGIC) scores at week 84. Data presented are adjusted mean changes (± standard error of the adjusted mean change) from baseline; ADAS-Cog, MMSE, and DAD scores were evaluated by analysis of covariance, ADCS-CGIC scores were evaluated by analysis of variance. For patients with week 65 visits, data were combined with those of patients with week 64 visits for ADAS-Cog, MMSE, and DAD. **p = 0.002 and ***p < 0.001 vs QS-21 alone; all other comparisons were not significant at the 0.05 level.
Reductions in DAD scores (worsening functional ability) were observed throughout the study (figure 2C, see table E-3). At week 64, patients in the active treatment group showed marginally less decline than control patients (adjusted mean values −11.68 vs −22.23, p = 0.077). The difference between treatment groups was significant at week 84 (−14.15 vs −36.42, p = 0.002) and at the final visit (−12.64 vs −34.13, p < 0.001).
Figure 2D displays the frequency distribution of ADCS-CGIC scores by treatment at week 84; observed mean ratings (SD) for active and control groups were 5.36 (1.19) in the active group and 5.31 (1.32) in the control group (see table E-3). No significant treatment differences on the ADCS-CGIC were observed at any visit.
Discussion.
No safety concerns arose during the course of this study to preclude the initiation of a larger phase 2a trial; however, one patient died approximately 1 year after her last injection; although the death was not directly attributable to study treatment, the patient was found to have evidence of encephalitis at autopsy.
Positive anti-AN1792 antibody titers were reported in patients treated with AN1792 + QS-21, but not with QS-21 alone. There appeared to be a correlation between dose and positive anti-AN1792 antibody titer, as a greater proportion of patients who were treated with the higher dose of AN1792 (225 μg) had a positive anti-AN1792 antibody titer at some point during the study than patients treated with AN1792 50 μg (77% vs 40%). The increase in the number of anti-AN1792 antibody responders after injection 5 may reflect the increased cumulative exposure to the antigen or enhanced immunogenicity associated with the addition of PS-80 to the formulation. As AN1792 225 μg + QS-21 50 μg resulted in the highest percentage of anti-AN1792 antibody responders of the four doses tested, this dose was selected for the subsequent phase 2a study (manuscript in preparation).
Although active treatment induced an immunogenic response in approximately one-half of patients, it is apparent that a considerable proportion of elderly individuals did not mount an immune response to the current formulation. Accordingly, it may be necessary to modify the adjuvant or other aspects of the formulation to achieve a higher response rate and thus a more useful therapeutic entity. Given the acknowledged difficulties of stimulating an immune response in elderly individuals, the use of passive immunization may be an attractive alternative approach to AD immunotherapy.
The overall incidence of AEs occurring during the study was similar between active and control treatment groups. However, the incidence of treatment-related AEs was higher in patients who received active treatment, although most AEs were mild or moderate in intensity. The incidence of severe treatment-related AEs and treatment-related SAEs was similar between active and control groups.
As an ad hoc evaluation, we also summarized AEs according to anti-AN1792 antibody titers. Several AEs occurred in patients who had a positive anti-AN1792 antibody titer with twice the incidence rate than in patients with a negative anti-AN1792 antibody titer. These events (including rash, asthenia, tremor, abnormal gait, flu syndrome, infection, injection site pain and inflammation, pain, and peripheral edema) primarily affected the nervous system and the body as a whole. Conversely, the following AEs occurred in patients who had a negative anti-AN1792 antibody titer with twice the incidence than in patients with a positive anti-AN1792 antibody titer: headache, dyspepsia, fecal incontinence, ecchymosis, hostility, insomnia, prostatic disorder, breast neoplasm, and hematuria. Some of these events may have been associated with the progression of AD; others might be expected when following an elderly population for >18 months.
No relationship was observed between AN1792 dose and the number of patients who reported treatment-related AEs. However, several SAEs, including confusion, convulsion, and encephalitis (diagnosed after death), were considered to be possibly related to treatment with AN1792. Some of these AEs led to premature discontinuation from the study: one patient receiving active treatment withdrew because of confusion and hallucinations and three patients (two in the active treatment group and one in the control group) withdrew because of hostility. In addition, three patients who received active treatment had convulsions during the study; two of these cases were severe. Five patients (mean age 74 years) died during the study follow-up, but no deaths were considered to be directly related to study treatment. One of these cases was the patient who died on day 599 after a pulmonary embolism, and was found at autopsy to have evidence of a meningoencephalitic reaction.22 The patient died 347 days after administration of the last (fifth) injection of study treatment and 219 days after leaving the study. The death occurred a few weeks after cases of encephalitis21 resulted in the sponsor halting dose administration in a phase 2a study of AN1792. In this latter study, all patients who had encephalitis had received the AN1792 + QS-21 formulation containing PS-80. Studies are presently underway to understand the etiology of encephalitis in relation to AN1792 + QS-21 and to develop alternative immunotherapeutic approaches for AD.
The autopsy report of the patient who died after participating in the current study indicated that there were extensive areas of neocortex with very few Aβ plaques in relation to the density of tangles and neuropil threads, and in some of these areas, Aβ-immunoreactivity was associated with microglia.22 These findings resemble the changes seen after Aβ immunotherapy in mouse models of AD10–13,23⇓⇓⇓⇓ and suggest that the immune response generated against AN1792 may have elicited the clearance of Aβ plaques in this patient.
This phase 1b study was not powered to demonstrate differences between groups on any of the clinical outcome measures. Indeed, three of the exploratory measures of efficacy showed no significant differences among treatment groups during the study period, but interestingly, the decline from baseline in DAD scores (caregiver assessment) appeared slower among patients receiving AN1792 + QS-21 compared with those in the control group. These differences emerged approximately 18 months after the study began; however, data from only a small number (n = 12) of control patients were available for analysis at this timepoint. Further studies in larger cohorts of patients are needed to determine the efficacy of AN1792 + QS-21.
Acknowledgments
The study was sponsored by Elan Pharmaceuticals, Inc.
The authors thank the patients, families, and carers of those who participated. They also thank Helen Cartwright, Carol Hall, Dr. Venkatesh, Dr. Jill Mann, Catriona Rainsford, Jo Waring, and Viv Hopkins for their help and Dr. Edgar Fenzl (Elan Pharma Limited, Stevenage, UK) for assistance with the study and review of the manuscript. Complete Medical Communications Ltd. (Macclesfield, UK) assisted in the development of the manuscript: Mary Gaskarth made substantial contributions to the analysis and interpretation of data and the drafting of the article; Delyth Clemett made substantial contributions to the analysis and interpretation of data, revision of the article for important intellectual content, and assisted with the final approval of the version to be published.
Footnotes
-
Additional material related to this article can be found on the Neurology Web site. Go to www.neurology.org and scroll down the Table of Contents for the January 11 issue to find the title link for this article.
-
A.J. Bayer, R. Bullock, and Drs. Jones, Wilkinson, and Paterson have received grants or honoraria from Elan Pharmaceuticals, Inc. S.B. Millais and Dr. Donoghue are full-time employees of Elan Pharma Limited, in which Dr. Donoghue also holds equity. Dr. Jenkins is a full-time employee of, holds equity in, and has received honoraria from Wyeth Research.
- Received December 3, 2003.
- Accepted July 20, 2004.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.