Doxycycline interferes with tau amyloid aggregation abolishing its associated neuronal toxicity

Tauopathies are neurodegenerative disorders with increasing incidence and still without cure. The extensive time required for development and approval of novel therapeutics highlights the need for testing and repurposing known safe molecules. Since doxycycline impacts α-synuclein aggregation and toxicity, herein we tested its effect on tau. We found that doxycycline reduces amyloid aggregation of the different isoforms of tau protein in a dose-dependent manner, remodeling the resultant species. Furthermore, doxycycline interacts with tau microtubule-binding domain preventing its aggregation. In a cell free system doxycycline also prevents tau seeding and in cell culture reduces toxicity of tau aggregates. Overall, our results expand the spectrum of action of doxycycline against aggregation-prone proteins, opening novel perspectives for its repurposing as a disease-modifying drug for tauopathies.

α -synuclein, preventing the buildup of pathogenic species and ultimately redirecting the process towards non-toxic off-pathway oligomers [23].
In the present work, we studied the effect of doxycycline on tau amyloid aggregation pathway and its associated toxicity. By using heparin-induced 2N4R tau fibrillization as well as the 4R truncated species self-aggregation, we assessed the capacity of doxycycline to hinder the tau amyloid pathway. Our results suggest the relevance of tau microtubule binding domain in doxycycline:tau interaction. Additionally, doxycycline also halted the ability of tau seeds to recruit monomers, which is essential for the progression of the pathology. By using cell culture, we also demonstrate the ability of this tetracycline to abate the toxicity related with tau-aggregated species. Taking together these results, with the well-known brain bioavailability, safety, antiinflammatory and antioxidant abilities of doxycycline, endorse this old drug as an ideal compound to be repurposed for tauopathies.

Preparation of heterologous 2N4R tau
Expression and purification of recombinant human tau was performed as

Thioflavin T assay
Formation of cross-β structure during tau aggregation was followed by addition of Thioflavin T (ThT) fluorescent probe on aliquots withdrawn from the incubation mixture at different times, according to LeVine [26,27]. Changes in the emission fluorescence spectra with the excitation wavelength set at 450 nm were monitored using an ISS (Champaign, IL) PC1 spectrofluorometer.
Doxycycline dose-response assay on tau aggregation was fitted to the equation: where IF is the normalized fluorescence intensity, [Ligand] is doxycycline concentration, and IC 50 is the concentration at which aggregation is inhibited at a 50%.

Bis-ANS fluorescent assay
Tau was aggregated as described above in the absence or presence of doxycycline. Aliquots were taken at 0 and 24 h incubation, and bis (1-Anilinonaphthalene-8-Sulfonic Acid) also known as bis-ANS was added to a final concentration of 5 μ M. Bis-ANS was excited at 395 nm and fluorescence emission was measured from 410 nm to 610 nm.

Infrared Spectroscopy
Samples of tau (100 µM), tau:heparin (100 µM:0.8 mg/ml) and was expressed as a percentage relative to the untreated cell control.

Statistical analyses
All data were obtained from at least three independent experiments and expressed as mean ± SD. Multiple-group comparisons were performed with one-way ANOVA and t-test. Differences were considered as statistically significant at p < 0.05. Statistical analyses were carried out with GraphPad Prism 5 (San Diego, California, USA).

Doxycycline hinders tau amyloid fibril formation yielding novel species
The ability of doxycycline to interfere with 2N4R tau aggregation was studied in the presence of heparin, a classical model system which induces the formation of tau fibrillary elements similar to NFT [10,13]. In agreement with previous reports, we found that heparin efficiently triggered full-length tau amyloid aggregation as monitored by Thioflavin T (ThT) fluorescence emission.
Doxycycline inhibited tau aggregation in a dose-dependent manner (Fig. 1A), showing an IC 50 of 29 µM (Fig. 1B). Since doxycycline exerted optimal inhibition of tau amyloid aggregation at 100 µM, all experiments henceforth were performed at this concentration. Kinetics of heparin-induced tau aggregation showed a classical sigmoidal behavior (Fig. 1C), characteristic of the nucleation-polymerization process as previously described [31][32][33]. The short lag phase was followed by an exponential increase that finally reached a plateau after 24 h of incubation. In the presence of doxycycline the system followed a hyperbolic curve, suggesting that early species formed in the presence of the tetracycline poisoned the cooperative characteristic of this aggregation process. Moreover, the amount of cross-β-containing species finally formed was significantly reduced in the presence of doxycycline, with a fluorescence steady state remaining stable up to 168 h (Fig. 1C). The different kinetics observed in the presence or absence of doxycycline suggests that distinct tau species were formed. To evaluate this hypothesis, we compared the hydrophobic patches exposed to solvent of these species. For this, we studied the interaction of bis-ANS with species prepared with or without doxycycline ( Fig. 1D). Basal bis-ANS emission spectrum of tau increased when heparin was added to trigger amyloid aggregation, reflecting a gain of hydrophobic surfaces exposure during the process. On the contrary, in the presence of doxycycline, the fluorescence intensity diminished significantly (61%) (Fig. 1D). These results indicate that these species were structurally different and the presence of doxycycline led to less hydrophobic residues exposed.

Doxycycline affects tau seeding ability and toxicity
Considering that hydrophobicity is one of the primary driving forces behind protein self-assembly processes [34,35], and that brain-derived tau oligomeric species that can spread the pathology have affinity for bis-ANS [36], we assessed the ability of doxycycline at halting the pro-aggregating properties of heparin-induced tau aggregates on monomeric species. For this, tau seeds were produced by incubating monomeric species with heparin at 37°C under orbital agitation for 24 h. Then, aliquots of this solution were added to fresh tau samples (deprived of heparin) with and without 100 µM of doxycycline. Only in the absence of the tetracycline, the seeding effect of pre-aggregated tau on the monomeric protein could be observed ( Fig. 2A). In the absence of seeds, monomeric tau did not evolve into amyloid species, as indicated by reduced ThT signal. As an internal control, heparin at the residual concentration present in seed aliquots, was added to monomeric tau and the mix incubated in the same experimental conditions. Upon incubation, no amyloid aggregation was observed according to ThT signal (Data not shown).
As our results showed that doxycycline remodeled tau aggregates affecting the formation of hydrophobic patches, we evaluated the ability of these species to bind and disrupt cellular membranes. For this, we analyzed cellular viability in a  (Fig. 2B). A solution of heparin was tested without significant differences from control (Data no shown).
In summary, Fig. 2 shows that the presence of doxycycline interfered with the seeding-ability of tau and rendered tau aggregates that were less toxic for cultured cells.

β -structuration of tau aggregates
The impact of doxycycline on the structure of heparin-induced tau aggregates was analyzed by using Fourier-Transform infrared spectroscopy (FTIR).
Comparative analysis of the conformationally sensitive band Amide I' (1,700-  1C), a band located at around 1,614 cm -1 assignable to cross-β structure emerges, together with a significant increase in β -sheet structure. This β -structuration occurred at the expense of the random coil (Fig. 3B, D and Table   I). The PPII-helix contribution was undetectable ( Fig. 3F) even using high derivative or deconvolution factors.
The presence of doxycycline significantly diminished the 1,614 cm -1 band contribution, suggesting that this tetracycline hindered the heparin-induced gain of cross-β structures associated with amyloid aggregation of tau ( Fig. 3C and G). Moreover, the overall β -structuration of the protein was decreased (curves shaded in yellow and light blue) (Fig. 3F), while non-structured regions (curve shaded in pink) remained more conserved (Fig. 3D).

Doxycycline interferes with heparin-induced tau PHF-like formation
To . The removal of the fuzzy coat from the PHF left the fibril core exposed which was retained in the stacking gel as previously reported [47]. Digestion products were resolved on a 12% Tris-Glycine SDS-PAGE gel (Fig. 4D), and the corresponding densitometric analysis are shown in Fig. 4E. The data revealed that monomeric tau was digested into peptides of less than 11 kDa, while heparin-induced tau fibrils exhibited fragments resistant to both proteases, excised from the fibril core. In presence of doxycycline the SDS-PAGE and the densitogram showed a significant change in the digestion profile for both proteases, whose digestion products were enriched in low molecular weight elements. This data suggests that the doxycycline interacts with the microtubule-binding region, inducing novel conformational arrangements and exposing sites previously inaccessible to protease cleavage.

Doxycycline interacts with tau microtubule binding domains inhibiting its self-aggregation
To confirm the hypothesis that doxycycline interacts with the microtubule- Our results showed that doxycycline inhibited in a dose dependent manner K18 peptide amyloid aggregation, by blocking monomer incorporation to the growing fibrils. We suggest that doxycycline binds to the microtubule-binding region, considered the core of tau fibrils [50], disrupting the intermolecular interactions of tau at the level of these key repetitions in the process of amyloid aggregation.

DISCUSSION
Recent results from our group and others [17,23,51-53] have shed light regarding the ability of tetracyclines in interfering with the toxicity and seeding of aggregate-prone proteins. In the present work, we extend these studies to tau.
Our results demonstrate that doxycycline can disrupt tau aggregation in both heparin-induced and K18 self-assembling heparin-free systems (Fig. 1, 5 B and C). It is important to note that tau amyloid formation in the presence of doxycycline produces, in addition to oligomers, short fibrillary species (Fig. 4C).  (Fig 2A), which could explain why tau species produced in the presence of this tetracycline are incapable of seeding ( Fig. 2A) and ultimately trigger decreased toxicity in cell cultures (Fig. 2B).
Structural studies revealed that tau species produced in the presence of doxycycline have less β-structuration (Fig. 3C, F  Moreover, these mechanisms observed in vitro could have impact in neuroprotection, since the presence of smaller aggregated species with higher susceptibility to proteases would also favor their "clearance" by the intracellular degradation systems. Likewise, taking into account that amyloid fibrils trigger microglia activation and chronic inflammatory response, which in turn lead to neurodegeneration [62][63][64], the doxycycline-increased susceptibility to proteases could be beneficial for neuron survival.
From a clinical perspective, doxycycline has been used for decades in human health and has proven to be a safe and well tolerated drug [65,66].
Moreover, due to its anti-inflammatory actions, doxycycline therapy is not restricted to the treatment of microbial infections, but has also been demonstrated to be useful for the management of periodontal and skin inflammatory pathologies [67,68]. When used as an antimicrobial compound, doxycycline (200 mg twice a day) achieves a concentration of about 3 µM in the cerebral spinal fluid (CSF) [69]. Therefore, considering its brain penetration, the In summary, our results extend the spectrum of the anti-aggregation action of doxycycline over tau, a key player during the development and progression of tauopathies such as AD. Altogether, the anti-aggregant ability on tau reported herein, in addition to its long safety record, as well as its brain penetration and low concentration required, prompt us to propose doxycycline as a valuable candidate for the development of a therapy against Alzheimer's disease and related tauopathies.  I  n  s  t  i  t  u  t  d  u  C  e  r  v  e  a  u  e  t  d  e  l  a  M  o  e  l  l  e  E  p  i  n  i  è  r

Funding sources and disclosure of conflicts of interest
The authors declare no conflict of interest.      Monomer incorporation rate per RT-QuiC cycle. n = 5 ns: not significant. * p < 0.05. ** p < 0.005. *** p < 0.0005. Error bars represent SD.