Our findings indicate that transgenic expression of anti-tau single domain antibodies (sdAbs) have therapeutic potential in the well-established Drosophila model of tauopathy. Among the four different sdAbs examined, one was clearly superior, two had moderate efficacy and one was ineffective. These differences were attributed at least in part to their different binding epitope on the tau protein. The effective sdAbs to a different degree extended the lifespan of the tauopathy flies, cleared pathological tau, attenuated or blocked tau induced: 1) eye abnormalities; 2) loss of neurons important for memory; 3) axonal transport defects, and; 4) locomotor dysfunction.
The transgenic anti-tau sdAbs exhibit no toxicity when expressed independently (Figure 1). This is promising for their potential application as gene therapy to treat tauopathies. When co-expressed neuronally with mutant tauR406W, three of the four sdAbs significantly improved the survival of flies expressing the familial tauR406W mutation (Figure 1). This tauopathy model has previously been shown to have neurodegeneration and shortened lifespan (16). The efficacy differences may at least in part be attributed to their varying binding epitopes and affinity for tau. Although all four sdAbs bind within the microtubule-binding repeat region of 2N4R tau protein, we found that they differ in their binding affinity to tau epitopes within this region both in solid and solution phases (Figure 2). This region of tau is well established to be involved in tau aggregation (41-43), and is a target of some of the tau antibodies in clinical trials (17). The most efficacious sdAb, 2B8, binds strongly to tau 271-295 and moderately to tau 307-331 in solid phase as well as strongly to positive control PHF-enriched tau from human tauopathy brain. In solution phase, it binds with high affinity to these two tau regions. The less effective 1D9 sdAb has a similar binding profile to 2B8 in these experiments with a bit less binding to 307-331 in solid phase and to 307-331 in solution phase. In contrast, the other less effective sdAb, 1F12 binds strongly to tau 307-331 in solid phase but less to tau 271-295 or PHF tau in this phase. In solution phase, it binds to tau 271-295 but with substantially lower affinity than 2B8 or 1D9 and does not recognize tau 307-331 in this phase. On the other hand, 2F12 binds strongly to tau 271-295 in solid phase but less to tau 307-331 and PHF in this phase. In solution phase, it does not recognize tau 271-295 but binds with high affinity to tau 307-331 in this phase. Taken together, solid phase binding to tau epitopes does not appear to relate well to efficacy. Instead, strong to moderate binding to tau 271-295 in solution appears to be required for efficacy.
The results of the survival assay are supported by western blot analysis of tau levels in the fly heads (Figure 3). Of the four sdAbs, neuronal expression of the 2B8 sdAb, which extended the lifespan of the tauR406W flies close to that of normal flies, led to the lowest tau and phospho-tau levels in the tauopathy flies. The three effective sdAbs, 2B8, 1D9 and 1F12, all reduced these levels in 5 day old flies whereas only 2B8 reduced phospho-tau levels in 30 day old flies. In contrast, the ineffective 2F12 sdAb did not reduce tau or phospho-tau levels at either age.
The Drosophila eye has been used by several groups to identify genetic modifiers of tau toxicity. It is a powerful model for understanding pathways that direct cell fate (44). Developmental expression of human mutant tau has been shown to result in neurodegeneration of the adult Drosophila eye (16). Of the four sdAbs, only 2B8 rescued degeneration of the photoreceptors when expressed pan-neuronally (Figure 4). These results align with 2B8 being more effective in prolonging the lifespan of flies and clearing pathological tau in their brains. Thus, eye morphology could be used as a quick screening tool to assess the therapeutic potential of sdAbs. For example, we have identified over 50 sdAbs with unique complementarity determining regions (CDRs) from a phage display library derived from a llama that we immunized with tau immunogens (22). We split these into families based on the binding profile of their cell culture supernatants to various tau immunogens but have only examined a few of those in detail. By transgenically expressing them individually in tauopathy flies, we could relatively quickly identify the most promising ones by examining their eyes.
The mushroom bodies (MBs) play a crucial role in olfactory learning and memory. However, tau-induced degeneration of these neurons compromises this important function (13, 45). Here, we only examined two of the tau sdAbs, 2B8 and 1D9 (Figure 5). Like in the other assays, 2B8 significantly suppressed tauR406W-induced damage to the MBs, whereas 1D9 and a control sdAb were ineffective. Specifically, it was partially effective when expressed pan-neuronally but completely blocked MB degeneration when expressed via an MB-specific GAL4. Again, these findings bode well for the therapeutic potential of 2B8, in particular as a gene therapy where its expression could at least in theory be directed to the most vulnerable brain regions affected by tau pathology.
The role of axonal microtubules in the anterograde transport of synaptic vesicles is crucial for proper neuronal function (46), and deficits in axonal transport are well established in tauopathies (47). To examine if 2B8 and 1D9 could rescue these tau-induced deficits, we expressed tau with or without the sdAbs in neurons that express PDF, a neuropeptide that regulates the output of circadian behavioral rhythms. Notably, circadian rhythm is dysregulated in tauopathies (39). Again, 2B8 significantly improved axonal transport as evident by analyzing the PDF neuronal projections in the flies (Figure 6), whereas 1D9 and control sdAb were ineffective.
Finally, we conducted a climbing assay as a behavioral measure of neuronal dysfunction and compared the effectiveness of the four sdAbs to attenuate climbing defects in tauR406W flies. While all four sdAbs suppressed the climbing defects during the first two weeks, only 1D9 and 2B8 were effective on Day 28, with 2B8 bringing their locomotor ability closest to the normal control group (Figure 7).
We previously reported that neuronally expressed anti-tau single chain variable fragment (scFv) prevents tauopathy-induced phenotypes in Drosophila models (18). It is notable that in that prior report, we showed extensive neurotoxicity in the tauopathy flies as measured by loss of Elav levels on western blots, whereas here Elav levels did not differ between control and tauopathy flies, although the survival profile of the two groups were comparable in both studies. A likely explanation is that in the prior report, protein levels were not normalized before loading the crushed head samples onto the gels, whereas here they were. The heads of the tauopathy flies appear smaller than in control flies resulting in lower protein levels for the same volume of homogenization solution. Therefore, by normalizing the protein levels before loading the samples onto the gel, we are masking the loss of neurons in the tauopathy flies. Regardless, the antibody-mediated clearance of tau in the two studies can be compared, since in the prior study tau levels were normalized to Elav levels. In both studies, the extended lifespan was associated with tau clearance in young flies examined at ages prior to tau-induced death.
Only a few anti-tau sdAbs have been reported previously (21, 22, 48, 49). One was shown by two-photon imaging via a skull opening to target tau aggregates in two tauopathy mice (49). Another one from a synthetic phage display library showed some efficacy after its direct brain injection in a lentivirus (48). In addition, we demonstrated that 2B8 and 1D9, cleared pathological tau and prevented tau toxicity in culture and reduced tau levels in brain interstitial fluid in tauopathy mice as assessed by microdialysis (21). Subsequently, we reported on the diagnostic imaging potential of 2B8, showing that it allowed non-invasive and specific imaging of tau pathology in mice following its intravenous injection, with the brain signal correlating strongly with lesion burden (22). Notably, here we show for the first time the therapeutic efficacy of these sdAbs in vivo.
In summary, our findings demonstrate in vivo molecular, cellular, and behavioral effects of anti-tau sdAbs in minimizing or preventing the development of various tau-induced phenotypes in Drosophila tauopathy models. Of the four sdAbs that we examined, 2B8 was the most effective, increasing survival rate, clearing tau and phospho-tau, and attenuating or preventing tau-induced axonal transport defects, neurodegeneration and functional deficits. The use of Drosophila as a model allows for further investigation into the underlying mechanisms of this therapeutic approach and assessment of its efficacy in other cell types. In particular, transgenic expression of sdAbs in Drosophila will guide sdAb-focused gene therapies in mammalian models and hopefully eventually in humans.