The immunological potency and therapeutic potential of a prototype dual vaccine against influenza and Alzheimer's disease

Alzheimer's disease (AD) is the most common form of dementia in the elderly which is clinically characterized by progressive loss of memory and general cognitive decline. The neuropathological features of AD include neurofibrillary tangles (NFT), deposition of soluble (monomeric, oligomeric) and insoluble fibrillar Aβ (senile plaques) forms, and neuronal loss in affected brain regions [1]. Pre-clinical and clinical trials have revealed that anti-Aβ antibodies are beneficial in clearing Aβ deposits [2‐13]. The first clinical trial of active immunization against Aβ was of the vaccine AN 1792, which comprised of fibrillar Aβ₄₂ formulated in a strong Th1-type biasing adjuvant, QS21. Patients treated with this vaccine were suffering mild-to-moderate AD. The trial was halted due to development of meningoencephalitis in some of the patients, which was believed to be associated with anti-Aβ specific T cell immune responses [8, 9, 14‐16]. One possible way to avoid these side effects is the replacement of the self-T helper epitope(s) present in the Aβ₄₂ peptide by a foreign epitope(s) while leaving self-B cell epitope(s) of Aβ₄₂ intact. Another important, but overlooked, result from the AN-1792 clinical trial was that the majority of AD patients generated only low titers of anti-Aβ antibodies, and approximately 50% of the patients failed to produce a measurable antibody response [12, 17]. The cause of the low anti-Aβ antibody titers and non-responsiveness observed in AN-1792 trial could be due to immune tolerance induced by self-Aβ₄₂ antigen. The mammalian immune system normally fails to generate antibodies specific to self-molecules; however, B cell tolerance is not rigorous, while T cell tolerance is more stringent [18, 19]. Previously we suggested that replacement of the Th cell epitope of Aβ₄₂ by a foreign Th epitope will help to overcome not only T cell tolerance induced by self antigen, but also side effects caused by autoreactive T cells. In our previous work we generated peptide- and DNA-based epitope vaccines based on amyloid-specific B-cell epitopes Aβ_1-15 or Aβ_1-11 attached to the promiscuous foreign Th epitope pan HLA DR-binding peptide (PADRE) and demonstrated the feasibility of this strategy in wild-type [20‐22] and APP/Tg mice [23‐25]. In this study we hypothesized that for therapeutic purposes AD epitope vaccines could be delivered to patients by a conventional viral vaccine [26]. Specifically, chimeric influenza viruses expressing the B cell epitope of Aβ may not only induce anti-viral immunity, but also generate higher titers of anti-Aβ antibodies in adult individuals with pre-existing influenza virus-specific memory Th cells. Accordingly, we generated and tested for the first time the immunogenicity and protective efficacy of chimeric inactivated flu virus vaccines expressing 1-7 or 1-10 aa of Aβ₄₂ (flu-Aβ_1-7 and flu-Aβ_1-10) in mice and demonstrated that these dual vaccines induced therapeutically potent anti-Aβ and anti-influenza antibodies.

Different approaches that aimed to prevent Aβ over-production or accelerate its degradation are currently being developed for treatment of AD. However all available treatments have only relatively small symptomatic benefits and could not delay or halt the progression of the disease. As a result, there is no cure from AD today. A potentially powerful strategy is immunotherapy with anti-Aβ antibody that can facilitate the reduction of pathological forms of Aβ in the brain [42‐52] via several pathways, including catalytic dissolution of amyloid deposits by antibodies; Fc mediated macrophage phagocytosis of amyloid; non-Fc mediate macrophage amyloid clearance; a peripheral sink, whereby Aβ is drawn out of the brain into the peripheral circulation [53, 54].

The results of the first AD clinical trial using the AN-1792 vaccine confirmed that anti-Aβ antibodies are beneficial for AD patients and may at least slow the progression of a disease. However this trial raised concerns about the safety and the efficacy of the active immunization strategy with Aβ₄₂ self-peptide. Although the results from the Phase I trial showed good tolerability, in the phase IIa portion of the AN-1792 immunotherapy a subset of individuals developed adverse events in the central nervous system [8‐11, 14‐17]. Further examinations demonstrated that these adverse effects were presumably due to the infiltration of autoreactive T cells, rather than anti-Aβ antibody. In addition, the relatively low antibody titers generated even after multiple immunizations and non-responsiveness in ~80% of patients indicating that the Aβ self-antigen vaccine was not a strong immunogen, suggest that alternative immunotherapeutic strategies should be pursued.

Based on data that the immunodominant B cell epitope of Aβ₄₂ has been mapped to the N-terminus of this peptide (aa spanning residues 1-5, 1-7, 1-8, 1-11, 1-15, 1-16, or 4-10) [34, 35, 37, 39, 55] and that this Aβ_1-11 peptide does not contain a T cell epitope in mice [35] or in humans [56], we proposed to use a prototype epitope vaccine that contains the small immunodominant self-B cell epitope of Aβ in tandem with promiscuous foreign T helper cell epitope/s, in order to reduce the risk of an adverse T cell-mediated immune response to Aβ-immunotherapy [20]. The efficacy and immunogenicity of our peptide and DNA-based epitope vaccines have been previously tested in the pre-clinical trials [23‐25]. Other groups of scientists and different pharmaceutical companies are working on development of epitope-based AD vaccines composed of self-Aβ B cell epitope attached to the carrier protein rather than small foreign Th epitope [57]. Another category of epitope vaccines are those based on viral-like particles (VLP) [58‐61]. Incorporation of the Aβ B cell epitope into a viral capsid protein or scaffold proteins allows the expression of this epitope on the surface of VLP in a repetitive and ordered array. Such organization of the epitope may induce T cell-independent B cell activation and production of anti-Aβ antibodies of IgM isotype. On the other hand, T cell epitopes from the viral proteins may help B cells to induce T cell-dependent humoral responses and produce antibodies of other isotypes. In fact, high titers of persisting long-term anti-Aβ antibodies were induced by recombinant protein based on pyruvate dehydrogenase complex of B. stearothermophilus fused with Aβ_1-11 B cell epitope. This protein self assembles in vitro into a high molecular mass scaffold with icosahedral symmetry exposing Aβ B cell epitope on a surface [62]. Therapeutically potent anti-Aβ antibodies (up to 1:10000 titer) were generated in APP/Tg mice using VLP based on papillomavirus [58, 61], retrovirus [59], Qβ bacteriophage [58, 60]. Qβ-based vaccine comprising the Aβ_1-6 epitope (CAD106) covalently linked to VLPs [63] is currently in Phase II clinical trials conducted by Novartis. Report from Phase I trial on safety, tolerability and Aβ-specific antibody responses in a group of patients with mild to moderate AD following three subcutaneous injections of 50 μg (cohort I) and 150 μg (cohort II) CAD106 was encouraging and showed that adverse events were predominantly mild. Although CAD106 induced low titers of specific antibody with a 2-fold increase in cohorts II vs I, 16/24 and 18/22 of subjects in cohort I and cohort II, respectively, responded to the vaccine [64, 65].

Our chimeric vaccine strategy described in this paper is different from VLP-based vaccines. First of all it is based on whole chimeric virus instead of non-replicative particles and therefore it could be used as either killed or live attenuated virus based vaccine. The use of chimeric influenza viruses whose backbone is widely used as a human influenza vaccine has the advantages of having quite well known antigenic properties in humans, of its immunogenicity being helped in humans by memory T cell responses against the backbone virus. More importantly, our strategy aimed to generate dual vaccine and test the feasibility of this approach.

Accordingly, we decided to take advantage from our previously developed plasmid-based reverse genetic technique [26] and generate a dual vaccine expressing the short B-cell epitope of amyloid within the HA of influenza virus. The HA and NA glycoproteins of influenza A viruses contain the major antigenic determinants of the virus responsible for the induction of neutralizing (protective) immune response. The appropriate mutations or insertions that may attenuate virus without compromising the immunogenicity of the vaccine allowed generating chimeric viruses (vectors) that can express heterologous polypeptides [66]. Because influenza viruses are potent inducers of antigen-specific B and T cell immune responses [66] they can also be attractive candidates as delivery vectors for amyloid-β B-cell epitope. In fact, previously it was shown that appropriate chimeric influenza viruses delivered heterologous small antigen (usually about 10-12 aa) into the host [67] and induced potent antibody [68] or cellular [69] immune responses specific to grafted peptide.

Here we generated and studied dual vaccines based on chimeric viruses, expressing Aβ_1-10 or Aβ_1-7 epitopes of Aβ_42. These B-cell epitopes of amyloid-β were inserted between amino acids 171 and 172 of HA, while the other four antigenic sites of HA remained intact (Figure 1A). The WB analysis demonstrated that chimeric, but not WNT-WT virus expressed HA of correct size containing Aβ_1-10 (Figure 1C) or Aβ_1-7 (data not shown) peptides. Importantly, the insertion of Aβ into HA did not change the capability of virus to infect host MDCK cells (Figure 2) or the conformation of the HA molecule (Figure 2 and 3).

Next we decided to analyze the immunogenic potency of the chimeric virus and compare it with that of wild-type influenza virus. Purified WSN-Aβ_1-10, WSN-Aβ_1-7, or WSN-WT viruses (Figure 1B and data not shown) has been used for preparation of inactivated vaccines that have been formulated into Th1 type adjuvant prior to immunization of experimental and control mice. We demonstrated that WSN-Aβ_1-10 was more immunogenic than WSN-Aβ_1-7 (Figure 4) and it induced the highest titers of anti-amyloid and anti-viral antibodies at 50 μg/mouse dose (Figure 5). WSN-Aβ_1-10 induced as good anti-viral humoral immune responses as WSN-WT after 3-4 immunizations (Figure 5, 6). These results support our hypothesis that chimeric influenza virus could be an excellent delivery platform for Aβ epitope, and at the same time provide T helper cell help to Aβ specific B cells. Of note, using peptide, recombinant protein and DNA based epitope vaccines we showed that Aβ_1-11 region did not possess epitopes for H2-b and H-2d mice [20, 23, 25]. More importantly, it was shown that Th epitope of Aβ₄₂ mapped to C-terminal region of this peptide [56]. Based on these data currently several companies are conducting Phase I/IIa studies with carriers fused with N-terminal regions of amyloid [70, 71].

The data represented above implied that a dual vaccine strategy is feasible since vaccinations of mice induced strong anti-viral and anti-amyloid humoral immune responses. At the same time these results did not demonstrate the therapeutic potency of anti-influenza and anti-Aβ antibodies. To test that, we performed in vitro assessment using HI [29] and neurotoxicity [24, 32] assays routinely used in our laboratories. These analyses showed that chimeric virus maintained the ability to induce the production of (i) virus neutralizing antibodies that inhibited the hemagglutination of red cells by the both chimeric and wild-type viruses (Figure 9, 10); and (ii) anti-Aβ antibodies that are binding to various Aβ₄₂ forms (Figure 8A) and inhibiting Aβ₄₂ fibrils- and oligomer-mediated toxicity of human neuroblastoma SH-SY5Y cells (Figure 8B). Data presented above suggest that anti-viral antibody could block viral infection while anti-Aβ antibody could be an effective modulator of Aβ₄₂ aggregate formation regardless of the nature of the aggregated species. Indeed, anti-Aβ antibody bind not only Aβ₄₂ fibrils and oligomers in vitro, but also Aβ plaques present in brain sections of cortical AD tissue (Figure 7).

To our knowledge this is the first attempt for generation dual vaccine based on conventional seasonal Flu vaccine and therefore designed to protect the elderly from both AD and seasonal Flu infection. Annual administration of seasonal Flu vaccine is currently proposed, therefore it is important to study the persistence of anti-Aβ antibodies and optimized schedule for vaccination with dual vaccine. However, in mice that are leaving in average 2.2-3.2 years it is not accurate testing annual vaccination strategy used for vaccination of elderly people. Thus, we are currently planning to study the doses, type of vaccine (killed or live attenuated), as well as schedule for vaccination in non-human primates, including aged animals with immunosenescence. The major complication connected with vaccination of elderly people is the poor response to the vaccines due to the immunosenescence. One possible strategy to counteract the immunosenescence is to recruit previously generated memory T cells produced during prior vaccinations and/or exposure to human pathogens. The majority of people already possess memory T cells specific for influenza due to yearly vaccinations and/or infection by virus. Thus, immunization of elderly people with our dual vaccine may in theory recruit memory T helper cells specific to influenza epitopes and induce rapid and potent anti-Aβ antibody production, while continuing to boost anti-viral cellular and humoral responses. This hypothesis is the subject of studies in progress in our laboratories.

Another important aspect of a dual vaccine is related to the safety issues. Since the majority of people including children and elderly are vaccinated with influenza vaccine yearly and the safety of this vaccine is observed for a long period of time, the chance that the dual vaccine is safe is very high. Finally, we think that the availability of a safe dual vaccine will allow the treatment of pre-symptomatic people rather than AD patients. Based on both preclinical studies and the results from the AN1792 clinical trials [70, 71] we may assume that early intervention in the disease process, pre-symptomatic if possible, is likely to be significantly more beneficial than attempting to intervene in the disease process after clinical diagnosis of the disease. In addition, early intervention is likely to significantly reduce the probability of adverse events in response to active immunization [14]. We believe that the recent breakthroughs in the development of biomarkers for AD provide a hope that patients can be accurately identified while they are still in the preclinical stages of AD [72‐77], which should facilitate the usage of dual vaccines before extensive neuronal damage and cerebral amyloid angiopathy has occurred in the brain in the general population. At the same time it should be mentioned that many groups including us have not observed infiltration of autoreactive T cells (presumed Th1 response that likely occurred in AN1792 vaccinated patients) in the brains after immunizations of APP/Tg or wild-type mice with the original Schenk et al. protocol [2] or with other Aβ vaccines (unless pertussis toxin widely used to induce brain T cell penetration in experimental autoimmune encephalomyelitis have been co-administered [78]). Thus, obviously only clinical trials may help us to conclude that any epitope vaccine including our chimeric flu vaccine is safe and do not induce harmful proinflammatory T cell responses in vaccinated AD patients.