|Home | About | Journals | Submit | Contact Us | Français|
One of the challenges of our society is to find a treatment or cure for Alzheimer's disease (AD). AD is the leading form of age-related dementia and with the increase of life expectancy worldwide, the social and economic burden from this disease will increase dramatically. It is a progressive and, in regard to clinical scores, a highly variable disease. AD pathogenesis has been associated with the accumulation, aggregation, and deposition of amyloid beta (Abeta) peptides in the brain. Hallmarks of AD are the amyloid plaques consisting of fibrillar Abeta and neurofibrillary tangles which are intracellular fibrils of hyperphosphorylated tau protein that develop later in this disease. The amyloid cascade hypothesis postulates that Abeta deposition is an initial event in the multifactorial pathogenesis and Abeta deposition may precede AD symptoms in some patients by at least 20 years. Amyloid beta therapy with active and passive immunizations against Abeta has a high possibility to be effective in removing Abeta from brain and might thus prevent the downstream pathology. Since 2000 a number of clinical trials for AD immunotherapy have started, have failed, and are continuing to be pursued. This article will review these clinical trials and ongoing research in this regard.
Alzheimer's disease (AD) is a devastating disease with no current cure or treatment. AD worsens over time and affects many layers of mental function: memory, thinking, and behavior. AD is the sixth leading cause of death in the United States and people 65 and older with AD survive an average of four to eight years (www.alz.org). Two of the pathophysiological hallmarks are the amyloid beta (Abeta) plaques which develop in the very early stages for this neurological disease and neurofibrillary tangles which begin to form later on (Hardy, 1992; Selkoe, 1996; Hardy and Selkoe, 2002; Bateman et al., 2012). Both of these aberrant features consist of normal self proteins: amyloid plaques develop when excess Abeta1–42 (Abeta42), a small proteolytic fragment from the amyloid precursor protein (APP), is present. Abeta42 has a tendency to build insoluble fibrils, and the insoluble neurofibrillary tangles are formed by hyperphosphorylation of the microtubule associated protein tau, which is also a normal structure in the cytoskeleton of the cell. Particularly neurotoxic forms of Abeta are the oligomeric forms of Abeta42 and it has been shown in rodents that a dodecameric form of Abeta42, which was isolated from brains of APP transgenic mice, and Abeta dimer, which had been isolated from human AD brains, impair directly synaptic plasticity and memory (Lesne et al., 2006; Shankar et al., 2008). Furthermore, Abeta dimers isolated from human AD brain induce tau hyperphosphorylation in rat primary hippocampal neuronal cultures (Jin et al., 2010), providing another link between Abeta accumulation and tau phosphorylation. Two recent papers described new methods with measurement of Abeta oligomers to classify more precisely patients with clinical dementia of the Alzheimer type in comparison to cognitively normal older patients with Abeta deposition in brain (McDonald et al., 2010; Esparza et al., 2013; Handoko et al., 2013).
It is a common belief that treatment is too late when these two features, amyloid plaques and neurofibrillary tangles, are already present. Intervention has to start early and from the current treatment options immunotherapy has the highest possibility to be effective. The medical term immunotherapy addresses the manipulation of the immune system by inducing, enhancing, or suppressing immune responses in vivo. Immunotherapy in AD covers two types of vaccination: active vaccination against Abeta42 in which patients receive injections of the antigen itself or passive vaccination in which patients receive injections of preformed antibodies against Abeta42. These antibodies are predicted to help with Abeta clearance via different pathways: antibody binding to Abeta in plasma might cause a gradient effect leading to Abeta removal from brain (peripheral sink effect) or the antibody binding might label Abeta in brain for the recognition by professional phagocytic cells in brain (microglia) to remove the Abeta deposits. Preliminary data from our laboratory showed that antibodies from Abeta42 peptide immunizations can lead to high titers of Abeta42 antibodies in APP/PS1 double transgenic mice and that these antibodies directly bind to the plaques in brain and help to remove excess Abeta42 from brain (Figure 1, unpublished data).
Immunotherapy targeting the pathological forms of tau protein are in preclinical testing, analyzing this treatment in the respective mouse models in which they show positive results as well (Boutajangout et al., 2010; Chai et al., 2011; Troquier et al., 2012).
In 1999, Schenk and colleagues published results on active Abeta1–42 peptide immunization in APP/PS1 double transgenic mice, an AD mouse model, which led to reduction in Abeta levels in the brains of the treated mice as well as improvements in memory tasks tested in these mice (Games et al., 1995; Schenk et al., 1999; Janus et al., 2000; Morgan et al., 2000). A subsequent clinical trial, AN1792, in which early AD patients received very similar Abeta1–42 peptide injections in combination with an adjuvant to improve the immune response, was suddenly stopped in 2002 when 6% of the trial participants receiving the active vaccine developed encephalitis (Orgogozo et al., 2003; Fox et al., 2005; Gilman et al., 2005). Despite this negative side effect which was caused by an inflammatory T cell response directed against brain Abeta, clinical evaluations of the patients showed that the immunizations with Abeta42 peptide to initiate production of the respective antibodies had worked and had led to reduction of Abeta levels and lesser plaque counts in these patients. The progression of the AD dementia and cognitive decline was unchanged and did not correlate to the reduction in brain amyloid. On the other hand, a clear relation between level of amyloid removal and the level of Abeta42 antibodies in plasma of the treated patients was observed (Holmes et al., 2008).
Since then much work has been done to improve the safety of immunotherapy to treat AD and was continued with alternative approaches. In the view of potentially dangerous autoimmune responses caused by newly activated Abeta42 specific inflammatory T cells, new peptide vaccines were designed in which the parts responsible for T cell activation were deleted and only the parts needed for generation of Abeta specific antibodies remained. Three of these peptide vaccines for active immunizations, CAD106, ACC001, and Affitope, are currently in phase 2 clinical trials (Table 1). These peptides contain only the B cell epitope, Abeta1–6, which is coupled to another carrier protein to optimize the immune response or as in the Affitope vaccine, the peptide resembles only the B cell epitope but has no sequence similarities, thus minimizing the risk of an inflammatory T cell response against the natural Abeta42 in brain (Ryan and Grundman, 2009; Schneeberger et al., 2009; Winblad et al., 2012). Phase 1 trials showed for all three approaches positive antibody responses with no signs for adverse autoimmune inflammation and these active immunization therapies are continued in clinical trials.
Another approach using similar players, anti-amyloid beta antibodies, in which already preformed and manufactured humanized antibodies are injected as passive immunization therapy are also ongoing. Three of the antibodies, solanezumab, gantenerumab, and crenezumab (Table 1), are in phase 2 and 3 clinical studies and showed benefits, at least in mild forms of Alzheimer's disease (Adolfsson et al., 2012; Blennow et al., 2012; Bohrmann et al., 2012; Farlow et al., 2012; Garber, 2012; Ostrowitzki et al., 2012).
Bapineuzemab was the first monoclonal antibody to be used as passive immunotherapy in AD patients. This humanized monoclonal antibody derived from a mouse antibody directed against Abeta1–5 and was reported to bind fibrillar Abeta42 as well as amyloid plaques. The occurrence of vasogenic edema in antibody treated patients particularly in APOE ε4 carriers was a negative side effect in a phase 2 clinical trial (Salloway et al., 2009). APOE ε4 is one of the alleles of the polymorphic apolipoprotein E which is involved in cholesterol metabolism and this particular allele increases the risk of AD. After the evaluation of a completed phase 3 clinical trial, use of this antibody was discontinued when the outcomes did not meet the predicted results. While there was a slight reduction in cerebrospinal fluid (CSF) tau, no significant changes were found for CSF Abeta in patients who had received bapineuzemab passive immunotherapy in comparison to the placebo treated control group (Blennow et al., 2012).
Solanezumab is a humanized monoclonal antibody which binds to the mid region of the Abeta42 peptide (Abeta13–28) and showed preferential binding to soluble Abeta but not to fibrillar Abeta. The antibody promotes the clearance of excess Abeta from brain and phase 2 studies showed a good safety profile for this antibody. It was given in variable doses and high doses with 400 mg antibody weekly were tolerated very well. In patients who had received 400 mg weekly an increase of Abeta42 in CSF was noted. Therefore, even though the patients which received a 12-week high dose antibody treatment did not show improved cognition on the Alzheimer's Disease Assessment Scale, this antibody might indeed mobilize Abeta from brain amyloid depositions as indicated by the increased Abeta42 levels in CSF and studies are ongoing (Farlow et al., 2012). In a different report, patients with mild AD showed slowing of the cognitive decline by one-third (Rachelle S. Doody, report from the Alzheimer's Disease Cooperative Study, oral presentation at the Annual Meeting of American Neurological Association, Boston, Oct. 2012). Also the analysis of pooled data from several clinical studies indicated a significant reduction in cognitive decline (Neurology Today, Nov. 1, 2012).
Gantenerumab is the only fully human monoclonal antibody used in clinical studies and it binds to the fibrillar forms of Abeta (Bohrmann et al., 2012). The antibody action is very likely the induction of phagocytosis of Abeta fibrils by brain microglia which had been shown in vitro. Antibody treated patients showed a decrease of brain amyloid of up to 30% in an antibody dose dependent manner using positron-emission tomography (PET) scans with the fibrillar Abeta specific Pittsburgh B compound (Ostrowitzki et al., 2012).
Crenezumab is derived from a mouse antibody binding to Abeta12–23 and is further modified to carry the human IgG isotype IgG4, which is a so-called Th2 antibody which has low binding affinities to antibody binding receptors on leucocytes (Fc receptors) and low probabilities of causing inflammatory immune responses. In a phase 1 clinical study patients treated with crenezumab had a reduced risk of brain microhemorrhage and vasogenic edema. Treatment of AD patients with this antibody showed favorable results as well (Adolfsson et al., 2012; Garber, 2012).
Of note, all the monoclonal antibodies used in clinical trials are different in regard to the Abeta binding capacities described.
A slightly different passive immunotherapy approach in which AD patients received injection of concentrated immunoglobulins (IVIg) from healthy persons showed benefits in the studied patient population as well. This study was based on the fact that IVIg contains naturally occurring autoantibodies that specifically recognize and block the toxic effects of Abeta (nAbs-Abeta). This study showed positive results: Anti-Abeta antibodies in the serum from AD patients increased in proportion to the IVIg dose administered while the CSF Abeta decreased significantly at 6 months, returned to baseline after washout, and decreased again after IVIg treatment was continued. This showed clearly that the nAbs-Abeta mobilized Abeta from brain. In this patient group even the mini-mental state (MMS) scores increased an average of 2.5 points after 6 months of treatment and remained stable with continuous IVIg treatment (Relkin et al., 2009; 2012; Shayan et al., 2012; Dodel et al., 2013).
It is well established that some of the biomarkers, such as Abeta deposition in brain, precede the clinical symptoms of AD by about twenty years and therapy to prevent or treat AD must be started before symptoms become apparent (Rosenberg, 2011; Bateman et al., 2012; Miller, 2012). This is also a strong argument for the lack of more definitive positive results from the ongoing trials: “Treatment was started too late and too little.” All the patients enrolled in the trials had already a positive AD diagnosis with mild to moderate dementia. A longitudinal study showed that the Abeta42 concentrations in CSF declined already 25 years before the onset of clinical symptoms as an indicator for Abeta accumulation in brain; and using PET scans with fibrillar Abeta specific Pittsburgh compound B, the authors showed that Abeta deposition in brain is visible already 15 years before clinical symptoms occur (Bateman et al., 2012). It has been shown in a small patient cohort that, in the common late-onset form of AD, the Abeta clearance rates from brain are significantly reduced in comparison to cognitively normal controls while the production rates are very similar (Bateman et al., 2006; Mawuenyega et al., 2010), supporting the idea that immunotherapy and the involvement of antibodies as well as cellular components of the immune system will greatly facilitate these clearance processes.
Three new trials will investigate when exactly AD therapy has to be started. These are the DIAN (Dominantly Inherited Alzheimer Network) study, the Alzheimer Prevention Initiative (API), and the study for Treatment of Asymptomatic Alzheimer (A4) (Table 2). These studies will focus on therapy in patient cohorts before the onset of clinical symptoms of AD. It appears critical for these preventive studies that the set-up (study design, enrollment of patients, drug use) and the data analysis (use of biomarkers and cognitive testing, statistical tools) are highly coordinated in national and international trials. Other factors such as detailed information about the treatment, its possible health improvements for the respective participants, and trust of patients and the treating physicians into these treatments highly influence participation and adherence to these trials which are of major importance for the final evaluations (Vellas et al., 2011).
The DIAN study will analyze patients who are carriers of genetic mutations which make it highly likely that these individuals develop AD at early age, a form of AD which is called Familial Alzheimer's Disease (FAD). In this study three different treatment methods will be used: two of them are monoclonal anti-Abeta antibodies, solanezumab and gantenerumab, and the third treatment will be a beta secretase inhibitor (LY2886721), which will also lead to a reduction in Abeta deposition because this inhibitor blocks the enzyme involved in the pathway of amyloid precursor protein turnover (Bateman et al., 2012; Morris et al., 2012).
The API study will be undertaken in a large group of families in Columbia who are FAD carriers and develop AD early. Three hundred symptom-free people will be treated from which two-thirds carry the familial mutation that makes them certain to develop end stage dementia around age 50, which is decades earlier than the typical sporadic AD case. These patients will receive passive immunizations with the monoclonal antibody crenezumab (Reiman et al., 2011; Garber, 2012).
The A4 study will analyze older patients who are not genetic carriers but whose brains show already early stages of Abeta deposition as measured by PET scan. These patients will be treated with the monoclonal antibody solanezumab to study whether an earlier treatment, before clinical symptoms occur, will provide a better benefit in blocking further Abeta accumulation and will ultimately delay the onset of AD. All three of these studies are designed to ultimately prove whether the amyloid hypothesis is correct and AD can be prevented or at least delayed with an anti-Abeta therapy (Carillo et al., 2013).
Active immunizations provide a number of benefits and considering the large number of possible patients to be treated, active immunization would be effective at a much lower price.
An alternative active immunization approach is the DNA Abeta42 immunization in which not the peptide itself but a DNA encoding Abeta42 is injected. A number of research laboratories, including our own, is currently investigating this approach in preclinical studies (Qu et al, 2004; 2006; 2007; 2010; Kim et al., 2007; Movsesyan et al., 2008; DaSilva et al., 2009; Davtyan et al., 2010). The injected DNA is translated in the immunized individual to produce Abeta peptide which then triggers respective immune responses against Abeta42. We were the first to show that DNA Abeta42 immunization is highly effective to reduce Abeta42 levels in brain by 41% and the Abeta42-containing plaques by 50% in transgenic mouse models (Qu et al., 2006; 2007) which was confirmed later in studies by others (DaSilva et al., 2009).
However, this response differs significantly from peptide immunizations. We and others have previously shown that Abeta42 DNA vaccination via gene gun generated a strongly polarized Th2 cellular immune response (Qu et al., 2004; 2010; Kim et al., 2007; DaSilva 2009; Davtyan et al., 2010). In our studies we determined that in vitro cell proliferation of potentially inflammatory Abeta42 specific T cells was absent in full-length DNA Abeta42 trimer immunized mice when compared to Abeta42 peptide immunized mice, supporting the safety aspect of this approach (Lambracht-Washington et al., 2009; 2011). Different from other Abeta42 DNA vaccine approaches in which only parts of the Abeta peptide were included to avoid a possible harmful Th1 T cell response (Lemere et al., 2007; Maier et al., 2006; Movsesyan et al., 2008; Zou et al., 2008; Davtyan et al., 2010) and which is very similar to active peptide immunizations currently in clinical trials (Table 1), the DNA Abeta trimer vaccine used in our studies is full-length and contains B- and T-cell epitopes.
Our results showed that T cells were clearly present in the immunized mice at earlier immunization time points but were reduced to levels below detection by the time of the cellular recall experiments (Lambracht-Washington et al., 2011). Thus, with the immunization of full-length DNA Abeta42, a potential positive effect of T helper cells in neuroprotection and neuroregeneration, which has been shown in several rodent models (mouse and rat) for neurodegenerative diseases as well as healing responses after a mechanical injury to nerve cells (Hendrix and Nitsch, 2007), is not precluded in this model from the beginning. The need for the inclusion of the analysis of cellular and T cell responses in AD immunotherapy has been reviewed in recent papers (Fulop et al., 2013; Monsonego et al., 2013). In this regard, it is also of interest to point out that the clinical trial for bapineuzumab, an Abeta1–5 monoclonal antibody, was stopped due to failure in reaching the set goals while trials for the other antibodies, detecting Abeta13–28, Abeta12–23, Abeta3–11, and Abeta19–28, are ongoing (Table 1). The full-length DNA Abeta42 vaccine we are pursuing has the advantage that it is open to a wider anti-Abeta response with a broader variety of antibody epitopes.
This is a significant time for Alzheimer's disease research and will provide results whether the amyloid beta hypothesis, postulating that Abeta accumulation is one of the initial events in AD pathology, is correct. Outcomes from ongoing clinical trials together with the planned prevention trials will show whether amyloid beta immunotherapy can indeed prevent or delay the onset of this disease and will demonstrate how translational research can provide an effective therapy for this devastating disease. Considering the high demand for AD immunotherapy on a worldwide scale, when passive immunization trials show positive results, a new focus will concentrate on active vaccinations as preventive treatment for AD.
This study was funded by grants from NIH/NIA Alzheimer's Disease Center (P30AG12300-17), Rudman Partnership, and McCune Foundation.
Disclosure R.N.R. has received clinical trial research grants from Janssen Inc., Novartis, and Pfizer. He holds a U.S. Patent for “Amyloid Beta Gene Vaccines.”