Schenk DB, Seubert P, Lieberburg I, Wallace J. β-Peptide ImmunizationA Possible New Treatment for Alzheimer Disease. Arch Neurol. 2000;57(7):934-936. doi:10.1001/archneur.57.7.934
Research on the pathophysiological characteristics of Alzheimer disease (AD) over the past decade has been directed toward the ultimate goal of developing a disease-modifying treatment to control or prevent the disease. A report recently published in Nature1 describes for the first time a treatment that may both prevent and treat the progression of Alzheimer disease. The treatment ironically involves immunization with the pathological β-amyloid peptide itself to generate an immune response targeted against the amyloid plaques in the brain. These findings have raised a number of scientific and clinical questions that are discussed below.
Pathologically, AD is characterized by atrophy of the cerebral cortex with loss of neurons and neuropil, the widespread appearance of senile amyloid plaques, and neurofibrillary tangles in the brain parenchyma of patients.2,3 The relationship of these lesions to the cognitive dysfunction and neuronal loss characteristics of AD has been the basis for a number of hypotheses, including the amyloid hypothesis, which states that β-amyloid deposits in the brains of patients are the cause of the disease. Substantial evidence supports this hypothesis: (1) The presence of such β-amyloid deposits are absolutely required to confirm a diagnosis of AD at autopsy.4(2) All known autosomal dominant mutations that cosegregate with early onset familial AD (including multiple mutations found in the amyloid precursor protein and presenilin 1 and presenilin 2 genes) mechanistically result in overproduction of the 42-amino-acid form of the peptide of β-peptide.5- 9 (3) The presence of β-amyloid plaques precedes clinical symptoms of the disease.10 And (4) clearance of Aβ42 in the brains of patients is reduced, as evidenced by reduced cerebrospinal fluid (CSF) levels of Aβ42 in patients with sporadic AD.11
Although these findings, on balance, collectively present a compelling case in favor of the amyloid hypothesis, some questions remain. In particular, the actual relevance of any disease-related hypothesis is defined by how well it predicts treatment outcomes or assists in management in the clinical setting. To this end, many approaches have been tried to test the amyloid hypothesis in the clinical setting. Currently, one of the most advanced amyloid-based approaches includes efforts to inhibit γ-secretase, the enzyme responsible for the carboxy-terminal cleavage of β-amyloid peptide.12,13 This approach is designed to block the formation of Aβ assuming that this will lead to the reduction of plaque burden and the consequent neuronal dysfunction and nerve cell death. A number of pharmaceutical companies have γ-secretase inhibitor compounds in preclinical development that are expected to advance to human trials soon. Many other approaches, such as treatments with β-secretase inhibitors14(the enzyme responsible for the amino-terminal cleavage event of β-peptide), Aβ aggregation inhibitors, and neurotoxicity and inflammation inhibitors such as cyclooxygenase II inhibitors, are also being pursued. All of these approaches are intended to either block the peptide from being formed or to lessen its impact once it is present in the brain.
An important tool for AD research has been the widespread development and use of mice transgenic for amyloid precursor protein (APP) containing 1 or more familial AD mutations.15- 18 These transgenic animals overproduce APP and hence β-amyloid, which leads to widespread β-amyloid deposition, neuritic dystrophy, and synaptic loss in brain regions similar to that in AD. These animals are therefore extremely useful for testing therapeutic hypotheses related to AD because one can investigate the effect of a given therapeutic treatment directly on the underlying disease in the animal. This cannot currently be done in patients with the disease because current imaging technologies do not permit such an analysis in life.
These transgenic animals play a critical role in the confirmation of a therapeutic vaccine approach. This approach involves the immunization with β-amyloid peptide as the immunogen together with an appropriate adjuvant,1 and results in both the prevention and reduction of progressive plaque buildup in the brain tissue of an animal model of AD, namely, the PDAPP transgenic mice.16 Specifically, when the mice were immunized at a young age, they developed very little, if any, β-amyloid plaque deposition with advancing age. Moreover, the progression of both neuritic dystrophy (a marker of neuronal dysfunction) and astogliosis (a marker of inflammation) were significantly reduced in the treated animals, suggesting that the immunization had benefits beyond simply reducing amyloid deposition. In a different type of experiment, presumably more predictive of the effect in treating existing AD, immunization with β-amyloid peptide was not begun until the animals were aged 11 to 12 months, an age when they exhibit substantial AD-like neuropathological characteristics.1 These animals were then observed until approximately age 18 months to assess whether vaccination reduced progression of existing disease. Remarkably, further β-amyloid formation was blocked and somewhat reversed; both neuritic dystrophy and astogliosis were statistically reduced. In addition, remaining amyloid plaque deposits often were actively metabolized by microglial cells, suggesting that the treatment can, in addition to preventing new amyloid deposition, lead to clearance of existing deposits. From the studies presented in the Nature report,1 there is evidence that Fc-mediated phagocytosis accounts, in part, for the reduction in β-peptide deposition. Evidence in favor of this possibility includes the observation that of the few plaques remaining in the immunized animals, a percentage of them were labeled with IgG. In addition, microglia were colocalized with β-peptide within plaques using confocal microscopic techniques. Importantly, microglia/monocytes, as detected by the presence of major histocompatibility complex II–positive staining, were found almost entirely near the few remaining plaques.1 Further, levels of APP were not altered by the immunization treatment, nor was there any evidence of toxic effects such as nephritis or brain disease in the treated animals. These results show that the effects in the PDAPP mice are a result of altering clearance of the β-peptide, which contrasts with most of the other attempts to treat AD currently under investigation.
The dramatic results of β-peptide immunization in the PDAPP mice raise many questions. First, how can peripheral stimulation of the immune system result in benefit across the blood-brain barrier in the central nervous system? We are taught that the brain is immunoprivileged and that the blood-brain barrier exists in part to serve this function. A careful review of the relevant literature, however, suggests a slightly different interpretation and viewpoint. For example, it is well established that nearly all serum proteins can be found in the CSF at levels roughly 0.05% to 0.3% of that in plasma.19 Entry of proteins into the CSF and brain occurs at a number of sites throughout the central nervous system, including the circumventricular organs, pial surfaces, and the Virchow Robin spaces.20,21 In addition, it is well established that CSF, to a rough approximation, is a highly diluted form of plasma.22 Consider for example a mouse that has a serum antibody titer of 1:50,000; in this context, one would expect the titer in CSF to be roughly 1:150. Although this represents relatively low levels of antibody, over periods of months such flux may be sufficient to alter the delicate equilibrium of β-peptide deposition. Also, once an antibody has bound to β-amyloid within a plaque, it is unlikely to exit the brain because the high plaque levels of amyloid would tend to recapture dissociating antibody, thus keeping the antibodies within the organ. The 2 most likely ways for its removal are either proteolysis or phagocytosis. In light of these observations, it is not surprising that the high levels of antibody in the periphery generated by β-peptide immunization produced biological effects in the central nervous systems of PDAPP mice. The findings, in fact, open new opportunities for therapeutic approaches. For example, immunization with other amyloidogenic proteins might be of therapeutic value. Also, current attempts with cancer vaccines might be extended to include brain tumors.
Another issue that arises from this work is defining which aspect (or aspects) of the immune response is required to elicit a benefit. Is it cellularly or humorally mediated? The simplest way to test whether a humoral response alone is sufficient is to demonstrate efficacy following passive immunizations with anti–β-peptide antibodies, about which studies are ongoing. These studies have clinical relevance because if the efficacy is entirely due to antibody production against β-peptide, additional strategies such as the use of human monoclonal antibodies as alternative treatments for AD should be considered. Such a passive immunization approach will not be feasible if a cellular response is required. In addition, it is conceivable that a cellular response could lead to safety problems in patients. Therefore, fully understanding the specific T-cell epitopes in the β-peptide might identify ways of circumventing such a response with altered immunogens containing only Aβ B-cell epitopes. Obviously, additional work is required to confirm this mechanism of action, perhaps with the use of Fc receptor–deficient mice. One point not addressed in the study1 was whether soluble forms of β-peptide are altered by the immunization procedure. It is conceivable that antibodies interfere with soluble β-amyloid that might exist as either a dimer or low oligimer in solution, and impede its precipitation into existing plaques. A variety of sophisticated biochemical assays will be required to address this question using the plasma, brain homogenetics, and/or the CSF of immunized animals.
While many interesting ideas relating to the function of the immune system within the brain can be further explored, most importantly, β-peptide immunization provides a testing framework for the amyloid hypothesis. Whether therapeutic benefit to patients with AD will result from a therapy designed to reduce β-amyloid plaque disease remains the key unanswered question. Success through this approach will require not only validity of the hypothesis but also attainment of a sufficient immune response. Several obstacles must be overcome, including breaking tolerance to the β-amyloid peptide in an elderly population. Overcoming tolerance to Aβ has been routinely achieved in other species, including rabbits, which possess an Aβ peptide sequence identical to that in humans. New adjuvants also offer promise as potent immunostimulants; determining immunogenicity in patients with AD will only come through clinical experience. Poor immunogenicity can be overcome in many ways beyond the use of adjuvants, including modification of the immunogen (combining Aβ B-cell and universal T-cell epitopes) and direct administration of a human monoclonal antibody, assuming humoral support of efficacy.
The design of clinical trials to test an amyloid-based disease-modifying therapy is also a novel area of study. Approval of an amyloid-directed therapeutic technique will require slowing of the disease course, as evidenced by well-accepted and widely used metrics such as the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-COG) and the Clinical Interview Based Impression of Change plus (CIBIC plus). Initially, such trials will have to be of sufficient duration to detect meaningful changes in the rate of disease progression.
It is hoped that imaging tests or biomarkers will be validated that may expedite the evaluations of disease-modifying therapies and predict outcomes of individual patients. Assuming β-amydoidosis can be meaningfully altered in humans, the most beneficial application of this therapy may be as a preventive measure. Neuropathological characteristics of AD are thought to precede cognitive symptoms and clinical diagnosis by months to years. Advances in genetic testing, cognitive testing, imaging, and developing biomarkers such as CSF tau and Aβ4211 make it a very real possibility that, in the near future, individuals at high risk for developing AD can be identified. Factors such as age, family history, and apolipoprotein E23 genotype are already known to constitute real risk factors; a well-defined high-risk group therefore exists, which is generally described as having "mild memory impairment," and clinical trials can be conducted in such a group. If successful, such trials would offer an opportunity for earlier intervention in treatment of AD to preserve a high level of cognitive function.
In the past decade substantial progress has been made both in understanding molecular mechanisms of amyloid disease and in the diagnosis of AD. Great strides have been made as well in designing and conducting AD clinical studies. We are now entering a phase where we must choose wisely both the best means of reducing Aβ disease and the most appropriate patient population on whom to try this approach to truly test the amyloid hypothesis in the clinic.
Accepted for publication February 18, 2000.
This work was supported by Elan Pharmaceuticals, San Francisco, Calif.
Reprints: Dale B. Schenk, PhD, Elan Pharmaceuticals, 800 Gateway Blvd, South San Francisco, CA 94080.