Introduction

Neurodegenerative disorders of the aging population, such as Alzheimer’s disease (AD), Parkinson’s disease (PD) and Frontotemporal dementia (FTD), are characterized by the progressive accumulation of misfolded protein aggregates that initially trigger synaptic damage and network dysfunction, and that eventually lead to loss of selected neuronal populations [1, 2]. In AD, the proteins amyloid-β (Aβ) and tau accumulate in the neocortex, limbic system, and basal forebrain in the form of plaques and neurofibrillary tangles [3]. In PD and related disorders such as PD dementia, dementia with Lewy bodies (DLB), and multiple system atrophy (MSA), the protein α-synuclein (α-syn) accumulates in neuronal and non-neuronal cells in cortical and subcortical nuclei as Lewy bodies, neuronal cytoplasmic inclusions, or glial cytoplasmic inclusions [4, 5]. Furthermore, in FTD (amyotrophic lateral sclerosis spectrum disorder) aggregates of either tau, superoxide dismutase 1, TAR DNA-binding protein 43 (TDP-43), or fused in sarcoma are found [6, 7]. In addition, recent studies have shown that α-syn can accumulate in selected brain regions in AD [8], and that TDP-43 aggregates are found in the limbic system in AD and DLB [9]. These findings reinforce the idea that abnormal protein accumulation is key in most neurodegenerative disorders. Under native conditions, most of these proteins can be found as poorly structured monomers or as dimers or tetramers associated with the plasma membrane [1012]. However, under pathological conditions such as those associated with AD, PD, and FTD, various molecular weight aggregates of these protein are detected, ranging from small oligomers to protofibrils and fibrils [1317].

Most recent evidence suggests that oligomers and probably also protofibrils are toxic to neurons by disrupting synaptic function, membrane permeability, calcium homeostasis, gene transcription, mitochondrial activity, autophagy, and/or endosomal transport [1821]. Moreover, recent studies have shown that propagation and seeding of Aβ, tau, and α-syn in a prion-like manner might also contribute to neurodegeneration [2228]. Remarkably, there is also evidence that these various protein aggregates can interact with each other [29]. For example, Aβ promotes the aggregation of α-syn and tau in AD and DLB [30, 31], α-syn and tau interact in the brain of patients with PD and DLB [32, 33], α-syn and Aβ can form hetero-oligomers [34, 35], and α-syn can modulate the fibrillization state of Aβ [36].

Progressive misfolding and accumulation of neurotoxic Aβ, tau, and α-syn have been associated with an imbalance in the levels of their synthesis, aggregation, and clearance (Fig. 1). Mechanisms of clearance include proteolysis, autophagy, and proteasomal degradation [37, 38]. In this context, it has been suggested that Aβ, tau, and α-syn toxic aggregates might be major therapeutic targets for these neurodegenerative disorders (Fig. 1). Thus, therapeutic strategies for AD, PD, and FTD might require reducing the synthesis, preventing the aggregation and/or enhancing the clearance of Aβ, tau, or α-syn. Numerous strategies directed at reducing the accumulation of these proteins have been developed, including the use of small interfering RNA, antisense RNA [3943], degrading enzymes (e.g., cathepsin D, neurosin, neprilysin) [4446], chaperone-like molecules that modulate aggregation state (e.g., Hsp70, β-syn) [4750], anti-aggregation compounds (e.g., polyphenols) [5153], and immunotherapy (passive, active, and T-cell-based) [54]. Moreover, the recent discovery that toxic oligomeric forms of α-syn and tau accumulate in the plasma membrane and are secreted to the extracellular environment has provided further rationale for the development of immunotherapeutic approaches for PD, DLB, MSA, FTD, and other neurodegenerative disorders characterized by the abnormal accumulation of these proteins [24, 26, 5558].

Fig. 1
figure 1

Mechanisms of action of immunotherapy for neurodegenerative disorders. The misfolding and accumulation of amyloid-β, α-synuclein, and tau has been associated with an imbalance in the levels of their synthesis, aggregation, and clearance. The toxicity of these proteins is correlated with their ability to adopt specific conformations (oligomers, protofibrils) and to propagate from cell to cell, leading to neurodegeneration. Disease-modifying therapeutic strategies may require reducing the synthesis, preventing the aggregation and/or enhancing the clearance of amyloid-β, α-synuclein, and tau. Specifically, immunotherapeutic approaches are able to target specific conformational species and inhibit cell-to-cell propagation of these proteins

Among these strategies, the development of immunotherapeutic approaches targeting Aβ, tau, and α-syn has received considerable attention in recent years. In this sense, both humoral (active and passive) and T-cell-based approaches have been explored. While active immunization stimulates the immune system to produce antibodies against target proteins, passive immunization consists in directly administering antibodies that confer temporary protection against the disease. A third type of immunotherapy involving the activation of regulatory T-cells has also been explored for the potential treatment of AD and PD [59, 60]. The advantage of humoral immunotherapy over other approaches is that it allows for the generation of antibodies targeting specific conformations of Aβ, tau, or α-syn (monomers, oligomers, and/or fibrils) (Fig. 1). Moreover, antibodies against these proteins can regulate inflammation and facilitate the clearance of target proteins via autophagy or microglia (Fig. 2) [6166]. Neurodegenerative diseases are associated with signs of chronic neuroinflammation and elevated levels of several proinflammatory cytokines released from microglia, such as interleukin-1β, interleukin-6, and tumor necrosis factor-α [67]. Microgliosis, astrogliosis, and peripheral immune infiltration contribute to the cognitive and motor deficits, and lead to a toxic increase in the levels of reactive oxygen species [68], and to secondary neurodegeneration, characteristic of late disease stages [69]. Immunotherapy has been shown to induce a physiological microglial response (M2 type) and reduce the production of proinflammatory cytokines [65, 66], thus exerting an anti-inflammatory effect in neurodegenerative disorders [70, 71]. However, among the disadvantages of immunotherapy are the potential for autoimmune responses, nonspecific inflammatory reactions such as perivascular edema, need for repetitive administration, lack of response due to senescence of the innate immune system, and limited penetration of antibodies into the central nervous system.

Fig. 2
figure 2

Antibodies promote microglia-mediated clearance of aggregated toxic proteins in neurodegenerative disorders. Toxic oligomeric forms of amyloid-β (Aβ), α-synuclein (α-syn), and tau are released by neurons to the extracellular environment in Alzheimer’s disease, Parkinson’s disease, and frontotemporal dementia brains, where they propagate to neuronal or glial cells leading to neuroinflammation and neurodegeneration. Antibodies bind extracellular or membrane-bound toxic oligomeric conformations of Aβ, α-syn, and tau, and might accelerate their clearance by microglia-mediated mechanisms, probably involving interaction with Fcγ receptors (FcγR) and autophagy degradation (bottom panel). MVB = multivesicular body

Active and passive immunization strategies are being explored in several clinical trials for AD and PD (Table 1). The most advanced studies are those on Aβ immunotherapy, for which 2 active immunization and 5 passive immunization programs are currently underway. For α-syn, there are 2 active and 2 passive immunization programs in phase I. For tauopathies, 2 vaccines and 3 antibodies are currently in phase I studies. Here we will review the most recent contributions and advances on humoral immunotherapy approaches for AD, PD, and FTD targeting Aβ, α-syn, and tau.

Table 1 Current clinical trials on immunotherapies for neurodegenerative diseases

Immunotherapy Targeting Aβ

The Aβ peptide is derived from a larger amyloid precursor protein (APP) by proteolytic cleavage at the β- and γ-secretase sites, resulting in the formation of Aβ1-38, Aβ1-40, Aβ1-42, and Aβ1-43. Patients suffering from AD often produce the longer forms of Aβ (Aβ1-42, Aβ1-43), which are more prone to aggregation and exhibit higher toxicity, while healthy subjects produce more of the shorter Aβ varieties [89, 90]. Therefore, developing antibodies against the longer forms of Aβ allows for the selective targeting of the more fibrillogenic and toxic species in therapy.

Immunotherapy for neurodegenerative disorders was first established for AD by targeting the Aβ peptide. The group of Schenk et al. [91] pioneered active immunization strategies by vaccinating APP transgenic mice with Aβ1-42 (AN-1792) and adjuvant, while Solomon et al. [92] developed a passive immunization approach using monoclonal antibodies against Aβ and showing that they reduce fibrillization in vitro. More recent active immunization strategies have included CAD106 [93], vanutide cridificar [94], and AD02, a synthetic peptide that mimics the N-terminus structure of the Aβ peptide (AFFITOPE; AFFiRiS AG, Vienna, Austria) [95]. Unfortunately, none of these approaches has resulted in significant clinical improvements. However, one of the placebo formulations for AD02 (renamed AD04) had a greater benefit on the primary outcome than the other placebo formulation or any of the 3 AD02 treatment groups in a phase II study, suggesting that it could be further developed. Although the composition of this placebo formulation has not been disclosed, it is possible that includes the adjuvant alum, which has been shown to boost adaptive immunity and induce uric acid production [96], a natural peroxynitrite scavenger [97]. Currently, 2 active immunization trials against Aβ are undergoing (Table 1). Lu AF20513 consists of 3 repetitions of a modified Aβ1-12 sequence, in which the natural T-helper cell epitopes are engineered to reduce the possibility of inducing harmful autoreactive T-cell responses and to improve the ability to mount an effective immune response [74]. ACI-24 (Europe, phase I/II) is a liposome vaccine designed to elicit an antibody response against aggregated Aβ peptides without concomitant proinflammatory T-cell activation. An array of Aβ1-15 sequences are anchored to the surface of liposomes adopting an aggregated β-sheet structure that acts as conformational epitope. In preclinical studies, repeated subcutaneous injection of ACI-24 into AD transgenic mice generated high titers of anti-Aβ antibodies, decreasing the concentration of insoluble Aβ1-40 and Aβ1-42, and of soluble Aβ1-42 [72, 73]. ACI-24 also improved novel object recognition without triggering proinflammatory responses [73]. Finally, DNA vaccines against Aβ1-42 [98, 99], alone or in combination with protein antigens [100, 101], have shown promising results at the preclinical stage. DNA-based vaccination utilizes direct injection of DNA-encoding genes for protein or peptide antigens, and it does not require the use of adjuvants. Co-immunization with a mixture of Aβ1-42 DNA and protein is capable of inducing Th2-type Aβ-specific antibodies while simultaneously suppressing unwanted inflammatory reactions and avoiding T-cell-mediated autoimmune responses [100].

Initial active immunization using AN-1792 (full-length Aβ1-42) highlighted the risk of autoimmune responses when using this type of therapeutic strategy, as 6 % of the patients in that study suffered with meningoencephalitis associated with T-cell infiltration [102]. Since then, numerous efforts have been devoted to reduce antigen-induced inflammatory T-cell activation, including the use of multiple small fragments of Aβ instead of the full sequence (e.g., CAD106, vanutide cridificar), and the use of small synthetic peptides that mimic the original epitope without carrying its sequence (AD02). However, the marginally positive effects observed with active immunization against Aβ suggest that other type of approaches might be more beneficial for patients with AD. In this sense, the use of antibodies directed against specific epitopes or conformations of Aβ has yielded promising results. Passive immunization approaches using monoclonal antibodies against Aβ1-40 [103], Aβ1-42 [104], pyroglutamate Aβ [105], oligomers [106], or protofibrils [107109] have been developed. Currently, clinical trials with the antibodies BAN2401 (recognizing protofibrils) [75], crenezumab (aggregated species) [76], gantenerumab (fibrils) [7880], and solanezumab (Aβ mid-domain) [81, 82] are ongoing (Table 1). However, other programs using antibodies such as bapinezumab (N-terminus) [110] and ponezumab (C-terminus) [111] have been discontinued as they did not meet expected goals. Finally, owing to the lack of significant disease modification in phase II and III trials, the use of nonspecific strategies such as intravenous immunoglobulin [Gammagard (Baxter Healthcare Corp., Deerfield, IL, USA), Octagam (Octapharma, Hoboken, NJ, USA), Flebogamma (Instituto Grifois SA, Barcelona, Spain) [77]] is relatively losing momentum for the treatment of AD [112]. These findings suggest that passive immunization against Aβ holds promise for disease modification in AD; however, more research is needed to improve the outcome of the immunotherapeutic treatments. This might include using immunotherapy as a preventive approach prior to the onset of symptoms, co-immunizing with both anti-Aβ and anti-tau antibodies in clinical trials, developing antibodies with better specificity to the toxic forms of Aβ, and boosting antibody penetration into the brain [113].

Immunotherapy Targeting α-Syn

α-Syn is a synaptic protein involved in synaptic transmission and vesicle release that is specifically upregulated in a discrete population of presynaptic terminals during acquisition-related synaptic rearrangement [114, 115]. α-Syn was initially identified in AD brains associated with plaque formation and neurodegeneration [116, 117]. The abnormal aggregation of α-syn is correlated with the neuropathological changes observed in PD and other synucleinopathies [13, 118], and therefore inhibiting α-syn aggregation would be a key mechanism for preventing its toxicity.

Initial immunotherapeutic studies were performed using vaccination with the full human α-syn protein [119]. Active immunization of α-syn transgenic mouse models of LBD decreased accumulation of aggregated α-syn and reduced neurodegeneration [119]. Furthermore, antibodies produced by immunized mice promoted the degradation of α-syn aggregates, probably via lysosomal pathways [119]. These results suggested that α-syn vaccination is effective in reducing neuronal accumulation of α-syn aggregates and that further development of this approach might have a potential role in the treatment of synucleinopathies.

Other active immunization approaches using AFFITOPEs (AFFiRiS AG) that mimic abnormal conformations of α-syn have been studied in animals model of PD and MSA [65, 66]. AFFITOPEs that mimic the C-terminus region of α-syn are able to elicit an immune response specific to α-syn oligomers [66]. Vaccination with one of these AFFITOPEs (AFF 1) resulted in high antibody titers against α-syn aggregates, decreased accumulation of α-syn oligomers, reduced degeneration of tyrosine hydroxylase fibers in the caudoputamen nucleus, and improved motor and memory deficits in 2 α-syn transgenic models [66]. Moreover, when administered to a transgenic model of MSA, AFF 1 also induced a reduction in neurodegeneration and demyelination in neocortex, striatum, and corpus callosum [65]. The clearance of α-syn induced by AFF 1 involved activation of microglia, increased anti-inflammatory cytokine production, and reduced spreading of α-syn to astroglial cells [65, 66]. These studies suggested that vaccination with AFFITOPEs could help ameliorate the neurodegenerative pathology in synucleinopathies. In this sense, phase I clinical trials with the AFFITOPEs PD01A and PD03A for PD and MSA, respectively, are currently ongoing (Table 1).

Passive immunization approaches using antibodies against α-syn are also being actively pursued. Different groups have investigated which region of the α-syn protein is the best target for the development of disease-modifying monoclonal antibodies. We and others have observed that antibodies that recognize an epitope in the C-terminus of α-syn are more effective at ameliorating the pathology in transgenic mouse models of PD, as they clear intracellular aggregates, inhibit α-syn propagation, and prevent C-terminus cleavage of the protein, which may lead to increased aggregation [54, 84, 120, 121]. However, other groups have reported that antibodies against the N-terminus are also effective at clearing α-syn aggregates, reducing their propagation, and diminishing motor dysfunctions [122, 123]. Together, these reports support the value of immunotherapy with antibodies directed against α-syn for PD, and in this sense the C-terminus antibody PRX002 (AFFiRiS AG) and the antibody BIIB054 (Biogen, Cambridge, MA, USA) are currently being tested in phase I clinical trials (Table 1).

Interestingly, and as mentioned before, antibodies against α-syn may not only reduce α-syn levels, but also reduce its oligomerization and fibrillization in living cells, thus reducing the pathology in mouse models of PD [66, 124, 125]. Furthermore, antibodies may also prevent cell-to-cell propagation of α-syn and facilitate the clearance of extracellular α-syn [84, 121, 122] (Fig. 2) . Importantly, both aggregation and cell-to-cell propagation are intimately related to α-syn toxicity and PD pathology, suggesting that these processes are promising therapeutic targets for immunotherapy.

Immunotherapy Targeting Tau

In AD and other tauopathies such as FTD, hyperphosphorylated tau accumulates within neurons in the form of neurofibrillary tangles [126131]. Importantly, as cognitive impairments closely correlate with the extension of tau pathology [132, 133], removing neurofibrillary tangles has become one of the main therapeutic goals for the treatment of AD and FTD [134, 135]. In this regard, it has been shown that both active and passive immunization against tau reduce its accumulation and slow or prevent behavioral deficits in transgenic mouse models of tauopathy [135142].

Active immunotherapy using phosphorylated tau epitopes has shown promising results in animal models [136, 142], and 2 tau vaccines are currently in phase I trial for AD, AADvac-1, and ACI-35 (Table 1). AADvac-1 consists of a synthetic peptide derived from amino acids 294–305 of the tau sequence, although the precise molecular nature of the antigen has not been disclosed. ACI-35 is a liposome-based vaccine that elicits an immune response against pathological conformers of phosphorylated tau without mounting autoimmune B- or T-cell responses against physiological tau conformations. The vaccine contains 16 copies of a synthetic tau fragment phosphorylated at S396 and S404, and anchored into a lipid bilayer. In the tau P301L transgenic mice, ACI-35 injection rapidly generates high titers of polyclonal antibodies specifically directed against phosphorylated tau. The resulting antibodies bind neurofibrillary tangles in mouse brain tissue sections and are able to reduce soluble tau, as well as insoluble, aggregated tau in brain extracts [86].

Antibodies against phospho-tau and tau oligomers have also been developed and tested at preclinical levels. Mechanistically, these antibodies seem to act either by promoting microglial clearance (Fig. 2), or by blocking neuronal uptake of the protein [143]. Passive immunization with anti-phospho-tau antibodies reduce tau pathology and functional deficits [137, 144146], and antibodies targeting tau oligomers have also shown promise in transgenic models [136, 147], including a concomitant upstream reduction in Aβ pathology [148]. In this sense, there are 3 anti-tau antibodies currently being studied in phase I trials (Table 1). BMS-986168 targets extracellular, N-terminally fragmented forms of tau (eTau) that can induce an increase in Aβ production and contribute to the spreading of the pathology [87]. This antibody reportedly neutralizes eTau toxicity in mouse models of FTD. RG7345 is a humanized monoclonal antibody targeting phospho-tau (pS422). Phosphorylation of tau at S422 has been linked to the relocalization of tau away from microtubules and toward the somatodendritic compartment of the neuron [149]. It has been shown that targeting the pS422 tau epitope with active vaccination decreases levels of insoluble phosphorylated tau and improves behavioral performance in transgenic mouse model of tauopathy [142]. Moreover, in a transgenic mouse model of AD, anti-tau pS422 antibodies are able to reduce accumulation of tau and induce its clearance via lysosomal pathways [88]. Anti-tau antibodies are internalized by neurons with tau aggregates via interaction with Fcγ receptors, and this internalization leads to the clearance of tau pathology in primary neurons [150], a mechanism that is probably shared with antibodies against α-syn (Fig. 2) [121]. Finally, the recombinant humanized anti-tau antibody C2N-8E12 has recently begun a phase I clinical study in patients with progressive supranuclear palsy.

Developing New Technologies for Immunotherapy

Passive immunization using immunoglobulins, voluminous proteins that do not easily cross the blood–brain barrier (BBB) and recognize a limited variety of epitopes, may yield only modest results. Therefore, efforts have recently been focused on the development of therapeutic single chain antibodies [single chain variable fragment (scFv)], fusion proteins of the variable regions of the heavy and light chains of immunoglobulins connected with a short linker peptide of 10–25 amino acids. scFvs retain antigen-binding properties and can be easily screened for desired affinities using phage display methodology. Using this type of approach, scFvs that detect individual conformational species of α-syn have been identified [151153], and could be potentially used to discriminate among protein conformers for the differential treatment of synucleinopathies or for diagnostic purposes [154]. Moreover, scFvs can be further modified to increase BBB penetrability and facilitate the clearance of α-syn. In this sense, a fusion protein comprising a scFv against α-syn plus the low density lipoprotein domain of apolipoprotein B was recently studied in a transgenic model of DLB. The brain-targeted fusion antibody easily crosses the BBB and gets internalized by neurons using the endosomal sorting complexes required for transport (ESCRT) pathway for enhanced degradation of α-syn aggregates [151], thus attenuating neuronal degeneration in vivo. Similarly, a fusion protein comprising a scFv and a specific protease can further aid in the clearance of aggregation-prone proteins [155]. Finally, the use of gene therapy with intracellular scFv (intrabodies) is also being explored for the detection and clearance of intracellular α-syn aggregates [156158].

Preclinical and clinical studies suggest that immunotherapy against Aβ, α-syn, and tau is a promising approach for the treatment of AD, PD, and FTD. Furthermore, as these proteins may co-aggregate and/or regulate each other [29, 30, 159161], immunization against one of them could reduce the aggregation or toxic modification of the others. For example, immunization against Aβ might be helpful at reducing α-syn and tau if administered at early disease stages [62]; likewise, α-syn antibodies might also be helpful in AD [162], and immunization against tau might be useful for PD [161, 163]. In this sense, identifying and targeting polyvalent antigens or using single-chain polyvalent antibodies targeting simultaneously Aβ, tau, and α-syn could have synergistic effects, as it occurs with polyvalent vaccines for certain cancers and infections [164, 165]. In the case of AD, targeting both Aβ and tau at the same time might improve the outcome of immunotherapeutic clinical trials, as it is likely that both proteins synergistically contribute to the progression of the pathology. Unfortunately, owing to the fact that accumulation of toxic proteins is an early event in neurodegenerative diseases, it is possible that immunization would be more successful as an early or preventive strategy rather than therapeutic one. Therefore, clinical trials using active or passive immunization may yield better results if performed in non-diseased or early-stage patients. Immunotherapy also has anti-inflammatory effects, probably by reducing extracellular levels of proinflammatory antigens, stimulating microglial clearance of toxic protein aggregates, and attenuating microglial inflammatory responses, leading to neuroprotective effects that may be also beneficial in late disease stages. As many of the antibodies described here have been developed against specific pathologic conformers of Aβ, α-syn, or tau, these antibodies could also be used as biomarker tools for diagnosis. In this sense, it has been suggested that autoimmune reactions towards specific proteins involved in the disease pathology can be used as biomarkers of neurodegeneration in both AD and PD, especially in early disease stages [166168]. Antibodies that recognize conformational epitopes specific of amyloid fibrils have been found in sera of healthy and diseased patients [169], suggesting that autoimmune reactivity can play a role as an amyloid clearance mechanism in both health and disease. Many neurodegenerative diseases have similar symptoms as they result from the aggregation of the same protein(s), making diagnosis challenging at times [2]. Using the antibodies developed against specific conformations of these proteins as diagnostic tools would allow the clinician to make accurate decisions about which therapy to prescribe [170172], greatly benefiting the therapeutic regimen and truly opening the door for personalized medicine. Moreover, it will probably be the case in the future that several immunotherapeutic options will be available for each disease, so therapy customization will become crucial. Finally, proteins other than Aβ, α-syn, and tau do accumulate in neurodegenerative disorders and are potential targets for immunotherapy as well. These include β-secretase [173], presenilin-1, leucine-rich repeat kinase 2 (LRRK2), superoxide dismutase-1 [174176], TDP-43, and fused in sarcoma, among others. It is possible that simultaneously targeting these proteins would drastically improve the outcome of the immunotherapeutic approach.