ESCRT-mediated uptake and degradation of brain-targeted α-synuclein single chain antibody attenuates neuronal degeneration in vivo.

Parkinson's disease and dementia with Lewy bodies are neurodegenerative disorders characterized by accumulation of α-synuclein (α-syn). Recently, single-chain fragment variables (scFVs) have been developed against individual conformational species of α-syn. Unlike more traditional monoclonal antibodies, these scFVs will not activate or be endocytosed by Fc receptors. For this study, we investigated an scFV directed against oligomeric α-syn fused to the LDL receptor-binding domain from apolipoprotein B (apoB). The modified scFV showed enhanced brain penetration and was imported into neuronal cells through the endosomal sorting complex required for transport (ESCRT) pathway, leading to lysosomal degradation of α-syn aggregates. Further analysis showed that the scFV was effective at ameliorating neurodegenerative pathology and behavioral deficits observed in the mouse model of dementia with Lewy bodies/Parkinson's disease. Thus, the apoB modification had the effect of both increasing accumulation of the scFV in the brain and directing scFV/α-syn complexes for degradation through the ESCRT pathway, leading to improved therapeutic potential of immunotherapy.

Introduction

α-Synuclein (α-syn) is a 140-amino-acid synaptic protein[1,2] that plays a role as a scaffolding protein in vesicle trafficking, dopamine release, and synaptic remodeling.[3] Misfolded, aggregated α-syn has been implicated in neurological disorders with parkinsonism, including Parkinson's disease (PD), dementia with Lewy bodies (DLB), and multiple systems atrophy (MSA) (reviewed in ref. 4). Oligomeric aggregates of α-syn are located in axons and presynaptic terminals where they might damage synapses and dendrites.[5,6] In addition, recent studies have shown that these α-syn oligomers can be released from neurons and propagate from cell to cell in a prion-like fashion exacerbating neurodegeneration.[7,8,9,10]

Given the potential toxicity of various α-syn aggregates, therapeutic approaches for synucleinopathies might involve decreasing the levels of intracellular α-syn or reducing the cell-to-cell propagation of extracellular α-syn. We have recently shown that active[11] and passive[12] immunotherapy might represent a useful approach to increase the clearance of α-syn by activating autophagy[13] or by facilitating the microglia-mediated degradation of extracellular α-syn.[14]

Passive immunization has been performed with monoclonal IgG antibodies that do not distinguish monomer and oligomeric α-syn species. Given the central role of oligomeric species of α-syn in the pathogenesis of PD, targeting specific conformational species of α-syn (i.e., dimers, oligomers, versus monomers) is important. Single-chain fragment variables (scFVs) are easier to utilize for this purpose due to the use of large phage display libraries for identifying specific binding characteristics (reviewed in ref. 15). Single-chain antibodies consist of the antigen-binding domain but lack the Fc region. Another advantage of the scFV is the lack of binding to the Fc receptor that can lead to phagocytosis, cell-mediated cytotoxicity, and release of inflammatory mediators (reviewed in ref. 16). However, the lack of an Fc region prevents the endocytosis/degradation by the Fc receptor (FcR) that is useful for the elimination of the antibody/antigen complex.

Recently, the generation of a series of scFVs directed against α-syn has been described[17,18]}. These scFVs have been developed to specifically bind oligomeric α-syn and contain the antigen-binding domain of the antibody while lacking the Fc portion. Thus, a specific scFV designed against the pathogenic oligomeric species of α-syn may be able to reduce the oligomer α-syn species without subsequent reduction in the monomer α-syn and without inducing inflammation through the Fc region. Delivery of the scFV to the CNS could be enhanced by the addition of a blood–brain transport peptide.[19]

We recently showed that the addition of the low-density lipoprotein (LDL) receptor-binding domain of apolipoprotein B (apoB) was sufficient to transport proteins from the blood to the CNS.[19,20] Under physiological conditions, the low-density lipoprotein receptor (LDLR) binding to lipoproteins can be endocytosed and directed to autophagy for degradation via the endosomal sorting complex required for transport (ESCRT) system.[21] The ESCRT pathway facilitates the trafficking of proteins from the endosome to the lysosome via multivesicular bodies (MVBs).[21] Thus, endocytosis/degradation via the ESCRT pathway may be a method for removing α-syn in the absence of an Fc receptor-binding domain.

In this report, we tested the efficacy of a modified scFV directed against the oligomeric forms of α-syn, by addition of the LDL receptor-binding domain of apoB, which enhances brain penetration and cellular endocytosis/degradation. We found that this modified scFV reduced α-syn accumulation while also ameliorating the neuropathology and improving behavioral deficits. The brain-targeting peptide was also responsible for the import and lysosomal degradation of the scFV/α-syn oligomer through the LDL-receptor-mediated ESCRT import pathway. Thus, the development of scFV directed specifically against the pathogenic species of α-syn may be a useful immunotherapeutic option for PD/DLB.

Results

The single-chain variable fragment antibody containing the apoB motif reduces accumulation of α-syn in a neuronal cell system

The single-chain variable fragment (scFV) antibody clone D5 was generated by phage display library against oligomeric α-syn as previously described.[17] We, and others, have shown that antibodies against neuronal proteins such as α-syn and Aβ delivered peripherally are capable of crossing the blood–brain barrier and affect levels of protein accumulation in the CNS.[11,22] Typical delivery of monoclonal antibody (mAb) by intravenous injection results in ~0.1% of injected protein transported to the CNS.[23] Targeting proteins or antibodies to receptors on the blood–brain barrier for transport to the CNS can improve the efficiency of transport to 2–3% of injected dose or 20–30 times more than the antibody alone.[24,25] This approach would be more desirable as delivery of the therapeutic protein would be less intrusive and delivery by the vascular system would provide for widespread CNS delivery of the scFV protein.

To determine if the D5 scFV fused to the apoB LDL-R binding domain might be a viable therapeutic approach for the reduction of oligomeric α-syn, we cloned the myc-epitope tagged D5 scFV cDNA into the third-generation lentivirus vector along with the CD5 secretory signal[26] under the control of the human CMV promoter to generate LV-sD5. The 38-amino-acid LDL-R binding domain of apolipoprotein B (apoB)[19] was fused to the C-terminus of the D5 scFV and cloned into the third-generation lentivector to generate LV-sD5-apoB (). Initial examination of the vector in vitro was carried out to determine if addition of the apoB binding/transport domain affected the function of the D5 scFV.

B103 rat neuroblastoma cells were infected with this vector and examined for production and secretion of the D5 scFV protein. The B103 neuronal cell line was derived from a rat neuroblastoma and shares many typical neuronal properties with other commonly used neuronal cell lines, including outgrowth of neurites upon differentiation, synthesis of neurotransmitters, possession of neurotransmitter receptors, and electrical excitability of surface membranes.[27] Two days later, cells were analyzed for expression of the D5 scFV. With the anti-myc antibody, the apoB D5scFV was detected by western blot in the cell lysate and in the cell culture supernatant () as well as by immunohistochemistry in the neuronal cells (). To verify the binding specificity of apoB D5scFV to α-syn, blots containing whole-brain extracts from α-syn knockout, nontransgenic control, and α-syn transgenic (tg) (line D and line 61)[22,28] mice were probed with conditioned media from cells infected with LV-sD5-apoB. With the anti-myc antibody, bands ranging in size from 28 to 80 kDa corresponding to the oligomeric forms of α-syn were detected only in homogenates from the α-syn tg mice but not in the α-syn knockout or nontransgenic mice (). Some cross-reacting bands were observed in the α-syn knockout and nontransgenic mouse lanes.

To test if the apoB D5scFV produced from the lentiviral vector was capable of clearing α-syn aggregates, we coinfected neuronal cells with a vector overexpressing human α-syn (LV-α-syn) and the vector expressing apoB D5scFV (LV-sD5 apoB) (,). Coinfection of neuronal cells with the α-syn overexpressing lentivector and LV-sD5-apoB expressing the scFV resulted in significant reduction in α-syn (–) and a shift of the cytoplasmic staining to more perinuclear, punctate staining (). Western blot analysis confirmed that apoB D5scFV reduced only the α-syn aggregates (). Finally, to examine the binding of D5scFv to α-syn, cultured supernatant containing apoB D5scFv in the presence of oligomerized α-syn was immunoprecipitated with the myc antibody (scFV) and then analyzed by western blot with an antibody against the α-syn protein. Cultured supernatant from sD5-apoB-expressing cells pulled down the oligomeric α-syn, indicating specific binding to the α-syn protein (). The results with LV-sD5-apoB are consistent with those observed for LV-sD5 alone (data not shown) and support the contention that sD5-scFV with apoB has anti-α-syn activity.

Delivery of the brain-targeted sD5 scFV reduces α-syn accumulation and associated neurodegenerative pathology in α-syn transgenic mice

In order to determine if the brain-targeted scFV would be efficacious in vivo, we delivered the LV-sD5-apoB scFV vector or the controls LV-sD5 or LV-control vectors to PDGFβ-α-syn mice (line D) by a single intraperitoneal injection (). Intraperitoneal delivery of the lentivector primarily transduces cells of the liver and the spleen where the transgene can be expressed and secreted into the bloodstream;[19] thus the liver and spleen become the depot organs expressing the recombinant protein. One month after intraperitoneal injection of the LV-sD5 or LV-sD5-apoB vectors, immunohistochemical analysis of the brain showed that sD5-apoB accumulated in the neocortex and hippocampus (,). Addition of the apoB transport peptide significantly increased the accumulation of sD5 scFV observed in the CNS with a 2–2.5-fold increase in immunostaining with the anti-myc antibody (,). Equivalent amounts of recombinant sD5 and sD5-apoB were detected in the blood (), and both sD5 and sD5-apoB were detected in the brain by immunoblot analysis; however, sD5-apoB accumulated to a significantly greater extent than sD5 scFV (,).

Double labeling studies were performed to determine the colocalization of sD5-scFV with the neuronal marker NeuN. Delivery by intraperitoneal injection resulted in colocalization of sD5 scFV (detected with an anti-myc antibody) with neurons in the neocortex (,) and hippocampus (not shown). Addition of the apoB binding/ transport peptide to sD5 scFV resulted in significantly greater colocalization with neurons (threefold) ().

Next, we determined if delivery of LV-sD5 and LV-sD5-apoB reduces the accumulation of α-syn in the tg mice. By immunocytochemistry, compared with the non-tg controls, in the α-syn tg mice, α-syn accumulates in the neuronal cell bodies and neuropil in the neocortex and hippocampus (). Delivery of the scFV antibody resulted in decreased accumulation of α-syn in the neocortex and hippocampus (,). Addition of the apoB transport peptide significantly decreased the accumulation of α-syn in tg mice compared with treatment with LV-control or LV-sD5 (,) with a twofold reduction in accumulated α-syn compared with tg mice treated with LV-sD5 and a fivefold reduction compared with mice treated with LV-control (). To verify the effects of LV-sD5 and LV-sD5-apoB by an independent method, western blot analysis was performed with antibodies against α-syn. Consistent with the immunohistochemistry, immunoblot analysis showed that total α-syn was reduced in animals that received either LV-sD5 or LV-sD5-apoB but was reduced to a greater extent when the apoB binding/transport peptide was added (, and Supplementary Figure S1). Oligomeric α-syn was reduced in animals that received either LV-sD5 or LV-sD5-apoB, but was reduced in the membrane fraction only in animals that received the sD5-apoB vector. Similarly, there was a significant reduction in phosphorylated α-syn (Ser129) (,).

Real-time PCR analysis of α-syn mRNA levels in the CNS showed no significant change in human () or mouse () expression levels when animals were treated with any of the viruses, suggesting that the change in α-syn levels was at the protein level only.

Finally, we analyzed the effects of the systemic administration of LV-sD5 and LV-sD5-apoB on neurodegeneration. Compared with non-tg mice treated with LV-control, α-syn tg mice treated with LV-control displayed significant reductions on NeuN cell counts (,) and MAP2 and synaptophysin immunoreactivity in the neuropil (,) accompanied by astrogliosis (GFAP) (,) in the neocortex. Compared with the LV-control, intraperitoneal injection of LV-sD5-apoB significantly ameliorated the alterations in NeuN (,), synaptophysin (,), and MAP2 (,) as well as reducing levels of GFAP in the α-syn tg mice (,), resulting in values comparable to the non-tg controls. Delivery of LV-sD5 resulted in improvements in markers of neurodegeneration but to a lesser extent when compared with LV-sD5-apoB ().

Brain-targeted scFV ameliorates the behavioral deficits associated with accumulated α-syn

Although motor deficits are a prominent clinical feature of PD/DLB, many patients also experience memory and cognitive deficits.[29] The D-line of α-syn tg mice also experiences memory deficits as measured by the Morris water maze.[30] To determine if delivery of D5 scFV could ameliorate these deficits, α-syn-tg mice were tested 1 month after intraperitoneal delivery of the LV-control, LV-sD5, or LV-sD5-apoB vectors. During the training period of the test, all groups of mice performed similarly at locating the visible platform after 3 days of testing. In the subsequent days of testing (days 4–7) with the platform submerged, non-tg mice located the platform in the water pool, with progressively shorter path distances over time (). At day 7 of testing with the hidden platform, non-tg mice and α-syn tg mice treated with LV-sD5-apoB discovered the platform in a significantly shorter time than the LV-control or LV-sD5-treated mice. In fact, LV-sD5 did not appear to have any effect on the memory deficits in α-syn tg mice as measured by the time to platform or the number of entrances to the target quadrant (). These results are consistent with neuropathological studies which suggest that D5 scFV containing the apoB transport peptide (sD5-apoB) has the capability of ameliorating the functional deficits associated with α-syn accumulation in the limbic system in transgenic mice.

Addition of the LDL-R apoB sequence to the scFV facilitates cellular uptake/degradation of α-syn by the ESCRT/autophagy pathway

Previous reports have shown that among other pathways,[31] α-syn can be degraded via autophagy.[13,32,33] At first, we determined if α-syn co-localized with scFV in mice treated with LV-sD5 and LV-sD5-apoB. By double labeling and confocal microscopy, α-syn and scFV as detected by an anti-myc antibody were found to be co-localized in granular structures in the neuronal cytoplasm (). The % of neurons displaying double labeled punctae was greater in tg mice treated with LV-sD5-apoB than with sD5 alone or control (). To determine if sD5 scFV-directed α-syn degradation occurs by the autophagy pathway, we double immunolabeled brain sections for α-syn and the autophagy vesicular marker LC3. As expected, in the LV-control-treated non-tg mice, there were LC3 positive granules in the cytoplasm of neurons but no co-localization with α-syn (). In the LV-control-treated α-syn tg mice, LC3 and α-syn colocalized in ~10% of the neurons (,). Mice treated with LV-sD5 displayed a trend toward increased colocalization between α-syn and LC3; in contrast, mice treated with LV-sD5-apoB displayed a significant twofold increase in colocalization compared with LV-control (,). To investigate intracellular trafficking of the sD5 scFV and α-syn complex, sections were double labeled with an anti-myc antibody that recognizes the sD5 and sD5-apoB scFV's and then for the early endosomal marker, Rab5. Animals that received the LV-sD5 lacking the apoB peptide showed little colocalization of scFV with Rab5. In contrast, addition of the apoB LDL-R binding domain to sD5 scFV significantly increased the uptake and colocalization of scFV with the Rab5 endosomal marker (,).

Rab5 is an endosomal protein associated with the ESCRT pathway of internalization.[34] To determine if internalization and lysosomal sorting of scFV-α-syn occurred through the ESCRT pathway, we double immunolabeled brain sections for the ESCRT protein CHMP2B and the scFV via the myc epitope tag. The CHMP2B protein and scFV co-localized in neurons of α-syn-tg mice that had received LV-sD5-apoB but not in mice that received LV-sD5 (,). Further ultrastructural analysis of these neurons showed that sD5-apoB scFV was localized to multivesicular bodies whereas sD5 scFV was not (). This evidence led us to hypothesize the model () consisting of uptake of the antibody-α-syn complex by the LDL-receptor through the ESCRT-mediated pathway followed by degradation in the autophagosome.

To examine the ESCRT-mediated uptake more closely, we developed an in vitro model of neuronal cultures exposed to the antibody-α-syn complex while manipulating the expression of the ESCRT component CHMP2B. For this assay, B103 neuronal cells were infected with lentivectors overexpressing CHMP2B or expressing an shRNA to knock down CHMP2B. Expression of CHMP2B was confirmed in these cultures by immunoblot () and immunohistochemistry (). LV-CHMP2B, LV-shCHMP2B, or LV-control infected neuronal cells were treated with conditioned media (CM) containing sD5-apoB and oligomerized α-syn (,). Cells were incubated with the antibody-α-syn complex for 6 hours to allow uptake and subcellular distribution and then were analyzed by immunohistochemistry for α-syn and the lysosomal protein cathepsin D (,). In neuronal cell cultures treated with the control CM and oligomerized α-syn, the α-syn was taken up by neuronal cells as previously described;[7] however, the α-syn appeared to remain in the cell and was probably not rapidly cleared (,). In contrast, overexpression of CHMP2B appeared to reduce the α-syn immunoreactivity. Addition of the sD5-apoB scFV to the cultures significantly reduced the accumulation of α-syn immunoreactivity in the neuronal cells (,). In contrast, blocking the ESCRT pathway through downregulation of CHMP2B in cells treated with CM containing sD5-apoB resulted in accumulation of α-syn around the cells and intracellularly (,), thus supporting the possibility that reducing the ESCRT pathway blocks the scFV ability to clear α-syn.

Next, to further determine if in fact the α-syn/scFv complex trafficking via the ESCRT pathway is targeted for degradation via autophagy, similar experiments as in were performed but this time the coverslips were double labeled with antibodies against α-syn and the autophagy marker LC3. As expected, neuronal cells exposed to α-syn oligomers in the presence of control CM displayed intraneuronal accumulation of α-syn in granular structures (arrowheads) that were LC3 positive (Supplementary Figure S3a,b). Addition of sD5-apoB containing CM resulted in increased colocalization with LC3 (Supplementary Figure S3a,b) and reduced accumulation of α-syn around the cells (Supplementary Figure S3c). This effect was observed with or without the LV-CHMP2B. In contrast, blocking the ESCRT pathway with the LV-shCHMP2B resulted in reduced co-localization of α-syn with LC3 (Supplementary Figure S3a,b) and increased accumulation of α-syn in the pericellular region (Supplementary Figure S3a,c).

Taken together, these results suggest that the D5 scFV containing the apoB binding/ transport peptide (sD5-apoB) is efficiently transported to the CNS of α-syn tg mice and has the capability of ameliorating the neurodegenerative pathology while reducing the accumulation of α-syn. Furthermore, the addition of the apoB LDLR binding domain to the scFV facilitates uptake and degradation by the novel ESCRT-mediated import/degradation pathway in neurons.

Discussion

Although still in the early stages, immunotherapy against α-syn may be a potentially useful treatment for PD and other synucleinopathies.[2] In support of the feasibility of such an approach, we have developed a single-chain antibody (scFV) specifically targeted to the oligomeric species of α-syn. This scFV was effective at reducing the accumulation of α-syn and related deficits in a DLB-like tg mouse model. Addition of the 38-amino-acid LDLR binding domain of apolipoprotein B functioned as a brain penetrating peptide.[19,20] This modified scFV (sD5-apoB) accumulated in the CNS following systemic delivery and reduced accumulated α-syn in the transgenic mouse. The addition of the apoB LDLR domain also facilitated binding and uptake through the LDL receptor and ESCRT pathway, thus bypassing the Fc receptor. Additionally, the brain-targeted sD5-apoB was able to reverse neurodegenerative pathology and memory deficits observed in the DLB-like tg mouse model.

Previous studies in vitro have shown that intracellular antibodies (intrabodies) can inhibit α-syn aggregation[18,35,36], and in vivo immunotherapy with copolymer-1 reduces neurodegeneration in the MPTP model of PD.[37] Moreover, we found that in actively vaccinated α-syn tg mice that produced high-affinity antibodies against α-syn, there was decreased accumulation of aggregated human α-syn in neuronal cell bodies and synapses that was associated with reduced neurodegeneration.[11] More recently, we have shown that passive immunization with antibodies that recognize the C-terminus of α-syn is effective at promoting clearance of α-syn aggregates and reducing the associated behavioral deficits via activation of autophagy in neurons.[12] Antibodies against the C-terminus of α-syn are also effective at reducing the neuron to astroglial cell propagation of α-syn.[14] In this study, it was found that the α-syn-antibody complex was cleared by microglial cells via the Fc-receptor. In such a model, the microglia-assisted clearance of α-syn resulted in reduced neuro-inflammation and functional deficits in transgenic mice.[14]

Therefore, the present immunotherapy study is different from others in that we used scFV, that scFV was selected to recognize α-syn oligomers, and that scFV was coupled to a brain penetrating peptide (apoB) that allowed higher trafficking into the CNS and enhanced efficacy. The scFV used in this study, D5, was shown to specifically target oligomers in the range of dimers to pentamers.[17] In this report, we showed that the D5 scFV reacted to a wider range of α-syn species in whole-brain homogenates from transgenic mouse brain including higher order oligomers and some binding to monomers. However, the predominant binding still remained to the 56 kDa pentamer and the 28 kDa dimer bands previously described for the D5 scFV. The discrepancy may be due to the original characterization of the D5 scFV having been performed with purified α-syn while our characterization was performed with whole-brain homogenates. A true measure of specificity of binding could better be measured in solution such as by an inhibition ELISA assay. This would provide a better understanding of the scFV-α-syn interaction that would occur in the brain extracellular milieu that cannot be replicated on a solid-surface immunoblot assay.

Many investigators have shown that larger species of α-syn are toxic, including proto-fibrils; however, there is significant evidence suggesting that small oligomers cause toxicity to neurons.[38,39,40,41,42] In addition, there is growing evidence showing that small oligomers are involved in the cell-to-cell propagation of synuclein in PD/DLB.[7,8,9,10,43] Thus we believe the targeting of dimers and pentamer α-syn species is a viable target for D5 scFV immunotherapy.

The mechanisms as to how the antibodies against α-syn specifically identify affected neurons are not completely clear. However, antibodies might be able to recognize abnormal α-syn accumulating in the neuronal plasma membrane;[11,39] furthermore, recent studies have shown that α-syn is secreted in vitro[7,44] and that α-syn is present in the cerebrospinal fluid (CSF).[45] Given that circulating antibodies might be able to recognize membrane-bound or secreted α-syn, there are several possibilities as to how they might promote the clearance of aggregates (see review paper).[2]

mAbs against α-syn are taken up by nonneuronal cells such as microglia through the Fc receptor;[14] however, other mechanisms might be at play. Since the scFV protein does not contain the Fc portion of the antibody, it is unclear how the scFV would be taken up by neurons in its native state. The addition of the apoB LDL-R binding peptide (sD5-apoB) appears to significantly increase the uptake of the protein by neurons and colocalizes with the early endosomal marker, Rab5. This may be a result of uptake by neurons expressing the LDL receptor.[19,46] The LDL receptor family members (LDLR, Megalin)[46] bind to apolipoproteins such as apoB and internalize the cargo for delivery through the endosomal pathway for transport to the lysosome. In the low pH of the lysosome, the receptor and apolipoprotein separate and the receptor is recycled to the surface (reviewed in [46]). Addition of the apoB LDL-R binding peptide to the scFV may facilitate uptake by neurons in this manner and would account for co-localization with the endosomal protein, Rab5, as well as the increased degradation of α-syn by the autophagy pathway as measured by co-localization with LC3.

In fact, recent evidence suggests the LDL-R can be internalized by a clathrin-independent, ESCRT-dependent pathway that leads to the formation of multivesicular bodies.[21] The CHMPs have received considerable attention in the last few years for their potential role in autophagy and neurodegenerative disorders.[47] Mammalian ESCRT-III proteins, also known as charged MVB proteins (CHMPs), consist of 11 isoforms[48] that play a central role in endosomal transport to the autophagosome. This possibility is consistent with the finding that the α-syn/sD5-apoB complex co-localizes with CHMP2B and is present in the MBV. Thus, the method of internalization and intracellular degradation of the α-syn/scFV-apoB may involve the following steps: (1) binding to the LDL receptor through the apoB LDLR binding domain, (2) internalization through the Rab5/ESCRTIII-mediated endosomal pathway, (3) formation of multivesicular bodies, and (4) fusion with lysosomes to form autophagosomes resulting in degradation of α-syn (). It remains to be seen if this route of LDLR internalization results in recycling of the LDLR receptor to the cell surface.

Previous studies with mAbs for the treatment of Alzheimer's disease have shown that ~0.1% of injected antibodies cross the blood–brain barrier.[23] This low rate of transport may at least partially account for the clinical failures of several monoclonal anti-Aβ antibodies. In contrast, we have shown that apoB fusion proteins cross the blood–brain barrier at a rate of 2–3% of injected dose or 20–30 times more than antibodies alone.[24,25] This increased blood–brain barrier transport may be sufficient to move some of the current antibodies from clinically ineffective to effective. In this study, we used the lentiviral vector to express the scFV-apoB from the liver for constant expression into the blood. This protein is circulated to the CNS and is then transported to the CNS. Future work with this scFV-apoB could include recombinant protein directly infused into the bloodstream to generate a more clinically feasible route of administration while also removing the use of viral vectors.

Recombinant protein administration would also allow us to more accurately examine the pharmacokinetics of scFV delivery to the CNS as well as immunogenicity of the scFV-apoB fusion protein. With regard to immunogenicity, the use of scFV was partially chosen due to the low immunogenicity compared to whole mAbs.[15] Additionally, previous work of ours has shown no immunogenicity towards the 38-amino-acid sequence of the apoB LDLR binding domain used in this fusion construct.[24]

The addition of the brain transport peptide (i.e., the 38-amino-acid LDL-R binding domain of apolipoprotein B) effectively increased the uptake of systemically delivered anti-α-syn scFV to the brain and facilitated internalization/degradation of the α-syn protein. This resulted in a significant reduction in α-syn aggregates, neuronal pathology, and behavioral improvements in the α-syn tg mouse. The increased uptake following the addition of the brain transport peptide has also been effective when applied to other proteins and antibodies.[19,20] For example, we have recently shown that addition of the apoB peptide to the Aβ degrading enzyme neprilysin increases the trafficking into the CNS and results in reduced accumulation on Aß and amelioration of the neurodegenerative and behavioral deficits in an amyloid precursor protein-tg model of Alzheimer's disease.[20,24]

Another significant advantage of the scFV over the traditional mAb approach is the cost of production. Because scFVs do not express the full Fab or the Fc portion of the antibody, production in eukaryotic cells is not necessary and production can occur in bacterial cell lines, thus significantly decreasing cost.[15] Thus, the development of brain-targeted scFV directed specifically against the pathogenic species of α-syn may be a useful immunotherapeutic option for PD/DLB.

Experimental Procedures

Animal care.

All experiments described were carried out in strict accordance with good animal practice according to NIH recommendations, and all procedures for animal use were approved by the Institutional Animal Care and Use Committee at the University of California at San Diego (UCSD) under protocol #S07221.

Construction of lentivirus vectors.

The anti-oligomeric (D5) scFV[17] cDNA was PCR amplified and cloned into the third-generation self-inactivating lentivirus vector plasmid[49] with the CMV promoter driving expression and the secretory signal from the human CD5 gene[26] producing the vectors LV-sD5, LV-sD5-apoB, and LV-control.

The Flag tagged mouse CHMP2B was a generous gift from Fen-Biao Gao.[47] The CHMP2B cDNA was PCR amplified and cloned into the lentivector to generate LV-CHMP2B. A short-hairpin RNA (shRNA) designed against the mouse CHMP2B (575–596) was cloned into the pSuper-GFP vector as previously described[47] to generate LV-shCHMP2B. The control shRNA lentivector (LV-shLuc) contains an shRNA directed against firefly luciferase.

The lentivirus vector expressing the human wild-type α-syn has been previously described.[13] Lentiviruses were prepared by transient transfection in 293T cells.[49]

Establishment of a neuronal cell line expressing α-syn and scFV.

For these experiments we used the rat neuroblastoma cell line B103. This model was selected because overexpression of α-syn in these cells results in mitochondrial alterations, reduced cell viability, defective neurite outgrowth, and abnormal accumulation of oligomeric α-syn.[13] For all experiments, cells were infected with LV expressing wt α-syn at a multiplicity of infection (MOI) of 20. Cells were co-infected with LV-sD5, LV-sD5-apoB, or empty vector (LV-control). After infection, cells were incubated in a humidified, 5% CO2 atmosphere at 37 °C. All experiments were conducted in triplicate to ensure reproducibility.

Cultured supernatant containing sD5-apoB or control was prepared by BBS transfection of 293T HEK cells with the plasmids pLV-sD5-apoB or pLV-control. At 72 hours after transfection, supernatant was collected and filtered with a 0.22 µmol/l cellular acetate filter (Corning, Corning, New York) to remove cellular debris. Protease inhibitor cocktail was added to the supernatant (Roche, Pleasanton, CA) to prevent protease degradation of scFVs during 4 °C storage.

B103 cells were infected with LV-shCHMP2B, LV-CHMP2B, or LV-control at MOI=20 in normal growth media. After 48 hours, cells were replated onto poly-l-lysine-coated coverslips in a 12-well dish at 5 × 104 cells/well. After 24 hours, normal media was removed and replaced with cultured supernatant containing sD5-apoB or control cultured supernatant. For cells treated with scFV + oligomerized α-syn,[49] 100 nmol/l oligomerized α-syn was added to the media at the time of treatment of cells. Briefly, oligomerized α-syn was prepared by incubation in water for 16 hours at 37 °C followed by 6 hours at 56 °C (Supplementary Figure S2). Cells were incubated with cultured supernatant for 6 hours to allow uptake of scFV and then were washed twice with ice-cold PBS followed by fixation with 4% PFA.

Immunoblot analysis.

Cells were infected with lentivirus vectors for 72 hours and then lysed in TNE buffer (50 mmol/l Tris–HCl, pH 7.4, 150 mmol/l NaCl, 1 mmol/l EDTA; all from Sigma-Aldrich, St. Louis, MO) containing 1% Nonidet P-40 (Calbiochem, Billerica, MA) with protease and phosphatase inhibitor cocktails (Roche, Pleasanton, CA). Total cell extracts were centrifuged at 6,000 × g for 15 minutes, and the protein concentration of supernatants was assayed with a BCA protein assay kit (Pierce Biotechnology, Rockford, IL). For western blot analysis, 20 µg of lysate per lane was loaded into 4–12% Bis–Tris SDS-PAGE gels and blotted onto polyvinylidene fluoride (PVDF) membranes. Blots were incubated with antibodies against α-syn (Chemicon, Temecula, CA), myc epitope tag (Sigma, St. Louis, MO), and actin (Chemicon, Temecula, CA) followed by secondary antibodies tagged with horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA), visualized by enhanced chemiluminescence and analyzed with a Versadoc XL imaging apparatus (BioRad). Analysis of actin levels was used as a loading control.

Immunoprecipitation.

Cultured supernatant containing sD5-apoB (myc tagged) was prepared as described above. Supernatant containing sD5-apoB or control cultured media in the presence or absence of oligomerized α-syn (100nM) were incubated with 50 μl of anti-myc or control IgG-coated magnetic beads (Fisher, Hampton, NH) and rotated at 4 °C for 1 hour before spinning down and saving supernatant. The beads were spun down and washed 3× with PBS before removing all liquid. The beads were resuspended in 10 μl reducing agent and 20 μl loading buffer and then heated at 92 °C for 5 minutes prior to loading on a 4–12% Bis–Tris gel with MOPS buffer. Gels were transferred to PVDF and probed with anti-α-syn overnight.

Real-time PCR analysis.

Total RNA was isolated from mice brains using the RNeasy Lipid mini kit (Qiagen, Germantown, MD) and was reverse transcribed using the RT2 First Strand kit (Qiagen) from 1 μg of total RNA. Real-time PCR analysis was performed using the StepOnePlus real-time PCR system (Applied Biosystems, Carlsbad, CA). PCR reactions were performed in duplicate on four independent sets of templates. Relative quantification of gene expression for the gene targets was achieved by RT2Profiler PCR Array Data Analysis version 3.5 (SABiosciences, Valencia, CA) using β-actin as an internal control. Amplification was performed on a cDNA amount equivalent to 100 ng total RNA. Oligonucleotides were designed using Oligoperfect (Life Technologies, Carlsbad CA, sequences available upon request). Relative quantification of gene expression was calculated by the comparative threshold cycle (Ct) method and expressed as 2-exp (ΔΔCt) using mouse β-actin as an internal control as previously described.[50]

Immunocytochemical analysis and confocal microscopy.

To verify expression levels of α-syn and scFV in cells infected with the different LV vectors, neurons were seeded onto poly-l-lysine-coated glass coverslips, grown to 60% confluence, and fixed in 4% paraformaldehyde (PFA) for 20 minutes. Coverslips were pretreated with 0.1% Triton X-100 in Tris-buffered saline (TBS) for 20 minutes and then incubated overnight at 4 °C with antibodies against human α-syn (Chemicon, Temecula, CA), myc epitope (scFV) (Sigma, St. Louis, MO), cathepsin D (Calbiochem, Billerica, MA), or LC3 (Abcam, San Francisco, CA). The following day, the myc or cathepsin D signals were detected with the FITC-conjugated secondary antibody (Vector Laboratories, Burlingame, CA), and the α-syn signal was detected with the Tyramide Signal Amplification-Direct (Red) system (NEN Life Sciences, Boston, MA). Control samples included empty vector (referred hereafter as LV-control) and immunolabeling in the absence of primary antibodies. Coverslips were mounted with Prolong Gold antifading reagent with DAPI (Invitrogen, Carlsbad, CA). Cells were analyzed with a digital epiflourescent microscope (Olympus BX51, Tokyo, Japan) to estimate the percentage of total cells (DAPI stained) that displayed α-syn or scFV (myc) immunoreactivity.

To verify the coexpression in neuronal cells co-infected with the different LV vectors, coverslips were double labeled with antibodies against α-syn (Chemicon, Temecula, CA) and myc (Sigma, St. Louis, MO) as previously described.[13] Coverslips were air-dried, mounted on slides with antifading media (Vectashield, Vector Laboratories, Burlingame, CA), and imaged with a confocal microscope. An average of 50 cells were imaged per condition and the individual channel images were merged and analyzed with the Image J program to estimate the extent of co-localization between α-syn and myc (scFV).

Transgenic mouse lines and injections of lentiviral vectors.

For this study, mice overexpressing α-syn from the platelet-derived growth factor β (PDGF-β) promoter (line D) were utilized.[22,28] This model was selected because mice from this line develop intraneuronal α-syn aggregates distributed through the neocortex and hippocampus similar to what has been described in LBD

To determine the effects of systemic injections of LV-sD5-apoB briefly as previously described,[20] a total of 36 hα-syn tg mice (n=12 LV-control, n=12 LV-sD5, and n=12 LV-sD5-apoB) from line D (9 months old) and n=6 non-tg mice (LV-control) were injected intraperitoneally with 100 μl of the lentiviral preparations (2.5 × 107 TU). Mice survived for 3 months after the lentiviral injection. Following NIH guidelines for the humane treatment of animals, blood was drawn and mice were anesthetized with chloral hydrate and flush-perfused transcardially with 0.9% saline.

Brains and peripheral tissues were removed and divided sagitally. The right hemibrain was postfixed in phosphate-buffered 4% PFA (pH 7.4) at 4 °C for 48 hours for neuropathological analysis, while the left hemibrain was snap-frozen and stored at −70 °C for subsequent protein analysis.

Immunocytochemical and neuropathological analyses.

Analysis of α-syn accumulation was performed in serially sectioned, free-floating, blind-coded vibratome sections from tg and non-tg mice treated with LV-sD5scFV, LV-sD5apoBscFV, and LV-control vectors.[13] Sections were incubated overnight at 4 °C with an anti-α-syn antibody (affinity purified rabbit polyclonal, Chemicon, Temecula, CA),[22] followed by biotinylated goat antirabbit IgG (Vector Laboratories, Burlingame, CA), avidin D-HRP (ABC Elite, Vector, Burlingame, CA) and detection with the Tyramide Signal Amplification-Direct (Red) system (NEN Life Sciences, Boston, MA) to determine the number of hα-syn immunoreactive inclusions in the neocortex. For each case, three sections were analyzed by the dissector method using the Stereo-Investigator System (MBF Bioscience, Williston, VT) and the results were averaged and expressed as numbers per mm3.

To determine the co-localization between α-syn immunolabeled neurons and scFV, 40-µm-thick vibratome sections were immunolabeled with the rabbit polyclonal antibody against α-syn (Chemicon, affinity purified polyclonal, Temecula, CA)[22] and the mouse mAb against the myc epitope tag of D5 scFV (Sigma, St. Louis, MO). The α-syn immunoreactive structures were detected with the Tyramide Signal Amplification-Direct (Red) system (NEN Life Sciences, Boston, MA) while the D5 scFV was detected with the horse antimouse IgG fluorescein isothiocyanate (FITC) antibody (Vector, Burlingame, CA).

To determine if the scfV gene transfer ameliorated the neurodegenerative alterations associated with the expression of α-syn, briefly as previously described,[13] blind-coded, 40-µm-thick vibratome sections from mouse brains fixed in 4% paraformaldehyde were immunolabeled with the mouse mAbs against synaptophysin (synaptic marker, Chemicon, Temecula, CA), microtubule-associated protein-2 (MAP2, dendritic marker, Chemicon, Temecula, CA), NeuN (neuronal marker, Millipore, Billerica, MA), or glial fibrillary acidic protein (GFAP, astroglial marker, Chemicon, Temecula, CA).[13] After overnight incubation with the primary antibodies, sections were incubated with fluorescein isothiocyanate (FITC)-conjugated horse antimouse IgG secondary antibody (Vector Laboratories, Burlingame, CA), transferred to SuperFrost slides (Fisher Scientific, Hampton, NH), and mounted under glass coverslips with antifading media (Vector Laboratories, Burlingame, CA). Other sets of sections immunostained with antibodies for NeuN and GFAP were incubated with biotinylated secondary antibody and reacted with diaminobenzidine. All sections were processed under the same standardized conditions. The immunolabeled blind-coded sections were serially imaged with the LSCM and analyzed with the Image 1.43 program (NIH, Bethesda, MD), as previously described.[13] For each mouse, a total of three sections from the hippocampus and three from the cortex were analyzed and, for each section, four fields in the frontal cortex and hippocampus were examined. For synaptophysin and MAP2, results were expressed as percent area of the neuropil occupied by immunoreactive terminals and dendrites. All sections were processed simultaneously under the same conditions and experiments were performed twice in order to assess the reproducibility of results. Sections were imaged with a Zeiss 63× (N.A. 1.4) objective on an Axiovert 35 microscope (Zeiss, Cambridge, MA) with an attached MRC1024 LSCM system (BioRad, Hercules, CA).[22]

Double immunolabeling and fluorescence co-labeling.

To determine the co-localization between α-syn, the anti-oligomeric (D5) scFV, ESCRT and autophagy markers, double labeling experiments were performed, as previously described.[12] For this purpose, vibratome sections were immunolabeled with the rabbit polyclonal antibodies against α-syn (Millipore, affinity purified polyclonal, Billerica, MA), LC3 (Abcam, autophagy marker, San Francisco, CA), anti-myc (Santa Cruz), Rab5 (BD Transduction Laboratories, endosomal marker, San Jose, CA), and CHMP2B (Abcam, ESCRT marker, San Francisco, CA). The α-syn, Rab5, and CHMP2B immunoreactive structures were detected with the Tyramide Signal Amplification-Direct (Red) system (NEN Life Sciences, Boston, MA) while LC3 and myc were detected with FITC tagged antibodies (Vector, Burlingame, CA). All sections were processed simultaneously under the same conditions, and experiments were performed twice in order to assess the reproducibility of results. Sections were imaged with a Zeiss 63X (N.A. 1.4) objective on an Axiovert 35 microscope (Zeiss, Cambridge, MA) with an attached MRC1024 LSCM system (BioRad, Hercules, CA).[22]

Electron microscopy and immunogold analysis.

Briefly, vibratome sections were postfixed in 1% glutaraldehyde, then treated with osmium tetraoxide, embedded in epon araldite, and sectioned with the ultramicrotome (Leica, Buffalo Grove, IL). Grids were analyzed with a Zeiss OM 10 electron microscope as previously described.[13] For immunogold labeling, sections were mounted in nickel grids, etched, and incubated with antibodies against myc followed by labeling with 10 nm Aurion ImmunoGold particles (Electron Microscopy Sciences, Hatfield, PA) with silver enhancement. A total of 125 cells were analyzed per condition. Cells were randomly acquired from three grids, and electron micrographs were obtained at a magnification of 25,000×.

Water maze testing.

As previously described,[12] in order to evaluate the functional effects of D5 scFV treatment in mice, groups of α-syn tg and non-tg animals were tested in the water maze. For this purpose, a pool (diameter 180 cm) was filled with opaque water (24 °C) and mice were first trained to locate a visible platform (days 1–3) and then a submerged hidden platform (days 4–7) in three daily trials 2–3 minutes apart. Mice that failed to find the hidden platform within 90 seconds were placed on it for 30 seconds. The same platform location was used for all sessions and all mice. The starting point at which each mouse was placed into the water was changed randomly between two alternative entry points located at a similar distance from the platform. Time to reach the platform (latency) and entrances into the target quadrant were recorded with a Noldus Instruments EthoVision video tracking system (San Diego Instruments, San Diego, CA) set to analyze two samples per second.

Statistical analysis.

All experiments were performed blind coded and in triplicate. Values in the figures are expressed as means ± SEM. To determine the statistical significance, values were compared by using the one-way ANOVA with post hoc Dunnet when comparing the scFV-treated samples with LV-control-treated samples. Additional comparisons were made using Tukey–Kramer or Fisher post hoc tests. The differences were considered to be significant if P values were less than 0.05.

Figure S1. Systemic injection of LV-sD5-apoB reduces the total α-syn in the CNS. Figure S2. Characterization of α-syn oligmomers. Figure S3. CHMP2B is necessary for the transport of the internalized scFV/α-syn complex to the lysosome.