The formation of small aggregates contributes to the neurotoxic effects of tau_45–230

Abstract

Intracellular deposits of hyperphosphorylated tau are commonly detected in tauopathies. Furthermore, these aggregates seem to play an important role in the pathobiology of these diseases. In the present study, we determined whether the recently identified neurotoxic tau_45–230 fragment also formed aggregates in neurodegenerative disorders. The presence of such aggregates was examined in brain samples obtained from Alzheimer’s disease (AD) subjects by means of Western blot analysis performed under non-denaturing conditions. Our results showed that a mixture of tau_45–230 oligomers of different sizes was easily detectable in brain samples obtained from AD subjects. Our data also suggested that tau_45–230 oligomers could be internalized by cultured hippocampal neurons, mainly through a clathrin-mediated mechanism, triggering their degeneration. In addition, in vitro aggregation studies showed that tau_45–230 modulated full-length tau aggregation thereby inducing the formation of smaller, and potentially more toxic, aggregates of this microtubule-associated protein. Together, these data identified alternative mechanisms underlying the toxic effects of tau_45–230.

1-. INTRODUCTION

The microtubule-associated protein (MAP) tau plays an important role in several neurodegenerative diseases collectively known as tauopathies (reviewed by Iqbal et al., 2005; Ding and Johnson; 2008; Hernandez and Avila, 2008; Mandelkow and Mandelkow, 2012; Iqbal et al., 2016; Guo et al., 2017). In these diseases, hyperphosphorylated tau is abundant in affected brain areas either in monomeric or aggregated forms (Grudke-Iqbal et al., 1986; Kosik et al., 1986; Wood et al., 1986; Takashima et al., 1993; Ferreira et al., 1997; Gamblin et al., 2003; Rapoport et al., 2002; Huang et al., 2016; Quinn et al., 2018). The tau_45–230 fragment is also easily detectable in tauopathies (Ferreira and Bigio, 2011). This fragment is the result of the proteolytic activity of calpain, a Ca²⁺-dependent protease which activity is enhanced in tauopathies (Park and Ferreira, 2005; Kelly and Ferreira, 2006 & 2007; Kelly et al., 2005; Park et al., 2007; Sinjoanu et al., 2008; Nicholson and Ferreira, 2009; Nicholson et al., 2011). Furthermore, tau_45–230 accumulation in neurons that develop either in situ or in culture resulted in progressive degeneration of axons and dendrites, synapse loss, and eventually cell death (Afreen et al., 2017; Afreen and Ferreira, 2019; Lang et al., 2014; Park and Ferreira, 2005). These morphological changes were accompanied by behavioral changes in mice expressing this tau fragment as early as six months after birth (Lang et al., 2014). In contrast, mutating the calpain cleavage sites responsible for the generation of tau_45–230 was sufficient to preclude its formation and to abrogate tau toxicity both in hippocampal neurons and in a Drosophila tauopathy model (Park and Ferreira, 2005; Reinecke et al., 2011). Several studies have identified potential mechanisms by which tau_45–230 could induce this phenotype. Thus, we have shown that a pool of this tau fragment was associated with microtubules, affecting their polymerization and stabilization (Afreen et al., 2017; Afreen and Ferreira, 2019). In turn, these microtubular changes could alter neurite elongation and/or regeneration (Afreen and Ferreira, 2019). Tau_45–230 was also associated with membrane-bound organelles, partially impairing their transport both anterogradely and retrogradely along axonal processes; a defect that could contribute to neurodegeneration (Afreen et al., 2017).

We have also shown that tau_45–230 formed dimers when transfected into cultured hippocampal neurons (Afreen et al., 2017). On the other hand, no information is available regarding the presence of tau_45–230 oligomers in Alzheimer’s disease (AD) and other tauopathy subjects. Neither is it known whether the presence of this tau fragment affects full-length tau aggregation.

In the present study, we have used brain samples obtained from AD subjects as well as in vitro and culture model systems to assess the potential role of tau_45–230 aggregates as mediators of tau toxicity. Our results indicated that a heterogenous mixture of tau_45–230 aggregates was present in brain areas affected by the disease process in AD. In addition, the data presented herein indicated that tau_45–230 oligomers could be internalized, inducing neuronal degeneration in hippocampal neurons. Moreover, they suggested that the presence of tau_45–230 altered full-length tau aggregation, leading to the formation of smaller, and potentially more toxic, full-length tau aggregates. Together, these data identified novel mechanisms by which tau_45–230 could contribute to the degenerative process in tauopathies.

2-. MATERIALS AND METHODS

2–1-. Preparation of human cortical samples.

All human brain tissue (autopsy tissue) used to generate the data included in this manuscript was obtained from the Mesulam Center for Cognitive Neurology and Alzheimer’s Disease (MCCNAD) Brain Bank, at Northwestern University. All samples were harvested and pathologically classified according to the National Institute on Aging-Alzheimer’s Association guidelines (Montine et al., 2012; King et al., 2021) by the MCCNAD personnel. They also de-identified of all information, with the exception of age, gender, and postmortem interval. Therefore, this research did not meet the definition of research in human subjects.

Human cortical tissue from the superior temporal gyrus (Brodmann’s area 22) obtained from control (63–90 years of age; cognitively intact, with maximum Braak stages I and II according to the criteria described by Braak and Braak, 1991) and AD cases (64–89 years of age; clinically demented and pathologically severe AD, Braak stages V and VI) were used for the preparation of whole cell extracts (Table 1). The post mortem interval (PMI) for all subjects ranged between 3 and 71 hours, with a median value of 19 hours (Table 1). Control and AD brain samples were homogenized on ice in Tris-Glycerol buffer (125 mM Tris, 20% glycerol, 0.01 NP-40), centrifuged at 16,000 × g for 30 min at 4° C. The supernatant was incubated with DNAse and MgCl₂ at 37° C for 30 min, centrifuged at 16,000 × g for 30 min at 4° C, and the protein concentration was determined by the method of Lowry et al. (1951) as modified by Bensadoun and Weinstein (1976). Samples were diluted 1:1 in sample buffer or treated with 1% sarkosyl for 90 min at room temperature. The detergent-treated supernatant was then centrifuged at 16,000 × g for 30 min at 4° C and the pellet resuspended in sample buffer. Equal amounts of total protein of these samples were separated using 10% Native PAGE and analyzed by means of quantitative Western blot as described below.

Table 1:

Human brain tissue analyzed in this study.

Pathological Diagnosis	Age Onset	Age Death	Gender	PMI (hr)	Clinical Assessment	Braak Stage
Control	NA	90	M	3	MCI	II
Control	NA	89	F	4	MCI	I
Control	NA	88	M	12	NCI	II
Control	NA	88	M	11	NCI	II
Control	NA	71	F	40	NCI	II
Control	NA	89	M	18	NCI	II
AD	68	74	M	20	MCI	V
AD	68	77	F	24	SCI	VI
AD	51	64	M	12	SCI	VI
AD	54	71	M	6	SCI	VI
AD	77	89	F	71	SCI	VI
AD	78	87	F	16	SCI	VI

AD: Alzheimer’s disease: Braak stages V and VI. PMI: post mortem interval; NCI: No Cognitive Impairment; MCI: Mild Cognitive Impairment; SCI: Severe Cognitive Impairment; NA: not applicable.

2.2-. Hippocampal culture preparation.

Embryonic day E18 Sprague Dawley rat embryos were used for the preparation of hippocampal cultures as described previously (Banker and Goslin, 1998; Ferreira and Loomis, 1998; Rapoport et al., 2002). In brief, hippocampi were dissected, stripped of meninges, and trypsinized (0.25%) for 15 min at 37°C. Neurons were dissociated by pipetting gently through a fire-polished Pasteur pipette and plated (~800,000 cells/60 mm dish) in Minimum Essential Medium (MEM) containing 10% horse serum (MEM10) on poly-L-lysine coated dishes. After 4 hr, the medium was replaced with glia-conditioned MEM containing N2 supplements, ovalbumin 0.1%, and 0.1 mM sodium pyruvate (N2 medium; Bottenstein and Sato, 1979). For immunocytochemical analysis, neurons were plated (150,000 cells/60 mm dish) on poly-L-lysine-coated coverslips in MEM10. After 4 hr, the coverslips were transferred to dishes containing an astroglial monolayer and maintained in N2 medium (Banker and Goslin, 1998; Ferreira et al., 2011). All animal experiments were performed according to protocols approved by the Institutional Animal Care Committee of Northwestern University according to the National Institutes of Health guide for the care and use of laboratory animals (NIH Publications No. 85–23).

2.3-. Preparation of astrocyte monolayer cultures.

Astrocyte cultures were prepared from the cerebral cortex of E18 Sprague Dawley rat embryos as previously described (Ferreira and Loomis, 1998; Ferreira et al., 2011). Briefly, embryos were removed and their cerebral cortex dissected and freed of meninges. The cells were dissociated by trypsinization (0.25% for 35 min at 37° C) and then centrifuged in MEM10 medium at 500 × g for 10 min. The cells were resuspended in fresh MEM10 medium, triturated with a fire-polished pipette, and plated at high density (1,600,000 cells/60 mm dish) on non-coated culture dishes. After the cells reached ~70% confluency, the medium was replaced with N2 medium to produce N2-glial-conditioned medium or to co-culture with coverslips containing dissociated hippocampal neurons.

2.4-. Electrophoresis and immunoblotting.

Control and AD brain lysates were separated by Native Discontinuous Polyacrylamide (10%) Gel Electrophoresis (PAGE) as previously described (Guttmann et al., 1995; Gallagher, 2001; Vintilescu et al., 2016; Feinstein et al., 2016). Transfer of protein to Immobilon membranes (Millipore, Burlington, MA, USA) and immunodetection were performed as previously described (Towbin et al., 1979). Membranes were processed using a tau_45–230 specific antibody recently generated and characterized in our laboratory (1:1000; Lang et al., 2014). A secondary antibody conjugated to horseradish peroxidase (1:1000; Promega, Madison, WI, USA) followed by enhanced chemiluminescence reagents were used for the detection of proteins (Yakunin and Hallenbeck, 1998). Immunoreactive bands were imaged using a ChemiDoc XRS apparatus (Bio-Rad, Hercules, CA, USA). Density of these bands was quantified using Quantity One software (Bio-Rad). To determine the molecular weight of tau_45–230 aggregates, standard proteins were separated using gels prepared with different polyacrylamide concentrations (5%, 7.5%, 10%, and 12%). The relative mobility of each marker was estimated by measuring the migration distance to the dye front and plotted against the gel concentration. The slope of the curve was determined and compared to the one obtained for the tau_45–230 aggregates using the Ferguson plot analysis as previously described (Guttmann et al., 1995; Gallagher, 2001; Vintilescu et al., 2016). Quantitative Western blot analysis was also performed using samples obtained from cultured neurons. Hippocampal neuron cultures were washed once in warmed (37° C) phosphate-buffered saline (PBS), scraped in Laemmli buffer containing a protease inhibitor cocktail [Sigma, 4-(2-Aminoethyl)-benzolsulfonylfluorid-hydrochloride (AEBSF, 1mM), aprotinin (800 nM) bestatin (40 μM), E-64 (14 μM) leupeptin (20 μM), and pepstatin A (15 μM)] and homogenized by boiling in a water bath for 10 min (Laemmli, 1970). These samples were analyzed by means of SDS-PAGE followed by immunoblotting (Towbin et al., 1979). The following primary antibodies were used for immunodetection of specific proteins: α-tubulin (1:1,000; clone DM1A, Sigma, St Louis, MO, USA), both full-length tau and tau_45–230 (1:1,000; clone tau5, BioSource International, Camarillo, CA, USA) and full-length tau (1:6,000; clone T46, Invitrogen, Carlsbad, CA, US). Incubation with secondary antibodies conjugated with horseradish peroxidase and enhanced chemiluminescence reagent were used to detect protein expression (Yakunin and Hallenbeck, 1998). Immunoblotting was performed sequentially using different primary antibodies. For these experiments, analyzed membranes were stripped in 60 mM Tris, 2% SDS and 0.8% β-mercaptoethanol at 55 °C for 30 min and then reprobed with a different primary antibody. Imaging and densitometric quantification of proteins were performed using a ChemiDoc XRS system (Bio Rad). Scanning of the membranes demonstrated the curve to be linear in the range used for each antibody.

2.5-. Expression and purification of recombinant tau_45–230 and full-length tau.

Recombinant histidine-tagged human tau_45–230 and full-length tau (tau_1–441) were prepared by Genescript (Piscataway, NJ, USA) using a customized protein production in a bacterial expression system. Recombinant proteins were purified by metal affinity chromatography by two-step purification on a nickel-based HisTrap Fast Flow column. Target fractions were identified by Coomassie staining of SDS-PAGE gels, and each crude tau recombinant protein was further purified via size-exclusion chromatography using the Superdex 75 gel filtration column (GE Healthcare Life Sciences, Marlborough, MA, USA). Recombinant proteins were concentrated using Amicon-ultra centrifugal filters. Target proteins were dialyzed and sterilized with a 0.22 um filter. Protein concentration was determined as previously described (Lowry et al., 1951; Bensadoun and Weinstein, 1976). Recombinant proteins were characterized by Western blot analysis using the anti-histidine antibody (Genescript) and liquid chromatography-mass spectrometry (LC-MS) for accurate intact molecular mass confirmation. These studies confirmed our previous results indicating that recombinant tau_45–230 run at ~32 kDa molecular weight; a molecular weight higher than the expected one (Ferreira and Bigio, 2011).

2.6-. Tau_45–230 and full-length tau aggregation.

Pre-cleared tau_45–230 and full-length tau (tau_1–441) proteins (Genescript) were aggregated alone or in combination with freshly prepared 150 μM arachidonic acid (Cayman Chemicals, Ann Arbor, MI, USA) in 250 μl of polymerization buffer (10 mM HEPES pH 7.4, 200 mM NaCl, 5 mM DTT) for 24 hr at room temperature. Aggregates were diluted in an equal volume of 2X Laemmli buffer (Laemmli, 1970) and the levels of tau_45–230 aggregates were analyzed by means of quantitative Western blot analysis as described above. For these experiments, a total of 9 independent aggregated protein preparations were used.

2.7-. Inhibition of endocytosis.

Seven days after plating, E18 cultured hippocampal neurons were pre-incubated with the clathrin inhibitors chloropromazine (CPZ; Signa) and monodansylcadaverine (MDC; Sigma) at a final concentration of 10 μM and 50 μM, respectively, in 1 ml of N2 medium for 1 hr at 37° C as described (Falcon et al., 2018). Tau_45–230 aggregates were then added directly to the culture medium (100 nM of monomeric tau_45–230 incubated with AA as described above). After 48 hr incubation, cultures were washed three times to remove surface-bound aggregates as previously described (Wu et al., 2013). Neurons were scraped in 2X Laemmli and levels of tau_45–230 were quantified by means of quantitative Western blot analysis as described above.

To inhibit dynamin-1-mediated endocytosis, hippocampal neurons kept in culture for 7 days were incubated with the membrane-permeable dynamin 1 inhibitory peptide (P4, QVPSRPNRAP, Tocris, Ellisville, MO, USA) at a final concentration of 100 μM for 30 min before the addition of tau_45–230 aggregates as described above (Kelly and Ferreira, 2007). P4 prevents the recruitment of dynamin 1 to the endocytic complex. Biochemical assays showed that this peptide is highly specific and binds with high affinity to a unique site in the dynamin-proline rich domain (Grabs et al. 1997; Kittler et al. 2000). Treated neurons were then processed for quantitative Western blot analysis as described above.

For these experiments, E18 embryos from a total of 4 pregnant rats were used for the preparation of hippocampal cultures.

2.8-. Immunocytochemistry.

Hippocampal neurons cultured in the presence or absence of fluorescein-labeled tau_45–230 aggregates for 24 hr and then washed as described above were fixed in 4% paraformaldehyde in PBS containing 0.12 mM sucrose for 15 min and permeabilized in 0.3% Triton X-100 in PBS for 4 min. Coverslips were then incubated with 10% bovine serum albumin (BSA) in PBS at room temperature for 1 hr before overnight incubation with a tubulin primary antibody (clone DM1A; 1:1,000, Sigma). Anti-mouse AlexaFluor secondary antibody (1:200; Molecular Probes, Eugene, OR, USA) was used for tubulin detection. Cultures prepared from E18 embryos from 4 pregnant rats were used for these experiments.

2.9-. Statistical analysis.

Quantitative analysis and statistical comparisons were performed using Prism 5.0 (GraphPad Software, Inc. San Diego, CA, USA). We tested for and found that all sampled distributions satisfied the normality criteria. Data were analyzed using Student’s T test. Values of p<0.05 were considered significant. The numbers in the graphs represent the mean ± S.E.M. and the statistical significance is indicated.

3-. RESULTS

3.1-. Tau_45–230 aggregates were present in AD brain samples.

We have previously shown that monomeric tau_45–230 was easily detectable in brain samples obtained from AD subjects (Ferreira and Bigio, 2011). On the other hand, no information was available regarding the presence of aggregated forms of this tau fragment in human brains in the context of these neurodegenerative disorders. In the present study, we first assessed whether tau_45–230 aggregates were present in affected brain areas in AD subjects. For these experiments, we prepared extracts from AD and age-matched control brain samples under non-denaturing conditions. Native-PAGE was performed using equal amounts of total protein and under conditions favorable for the analysis of tau_45–230 and that precluded the incorporation of full-length tau into the gel (Vintilescu et al., 2016; Feinstein et al., 2016). Immunoblotting was performed using a tau antibody that only recognizes tau_45–230 and not full-length tau, recently generated and characterized in our laboratory (Lang et al., 2014). All AD samples analyzed showed easily detectable tau_45–230 immunoreactive bands at ~68 kDa molecular weight. In addition, a strong ~168 kDa tau_45–230 immunoreactive band was detected in at least half of the AD brain extracts analyzed (Fig. 1A). In contrast, faint tau_45–230 immunoreactive bands were detected at both ~68 kDa and ~168 kDa molecular weights in age-matched control samples (Fig. 1A). The intensity and area of tau_45–230 immunoreactive bands were quantified as described in the Materials and Methods section. This quantitative analysis showed a significant increase in tau_45–230 aggregates of ~68 kDa molecular weight in AD and samples when compared to age-matched controls (Fig. 1B). On the other hand, no significant differences were detected when the intensity of the 168 kDa immunoreactive bands were quantified using pooled disease samples and compared to controls due to the great variability among AD samples (Fig. 1B).

Figure 1: Tau_45–230 aggregates were present in AD brain samples.

(A) Selected representative lanes of Native/non-denaturing PAGE followed by Western blot analysis of the presence of tau_45–230 aggregates in brain samples of Alzheimer’s disease (AD) and aged-matched controls (Ctl) subjects using a tau_45–230 antibody. Tau_45–230 immunoreactive bands of ~68 kDa and ~168 kDa molecular weights were detected in AD samples analyzed. (C) Immunoreactive bands were quantified and referred to age-matched controls. (B) Determination of detergent-stability of tau_45–230 aggregates present in AD brain samples by means of Native/non-denaturing PAGE followed by Western blot analysis. Homogenates prepared in the presence or absence of sarkosyl (1%) from post-mortem cortical samples. The content of sarkosyl-resistant ~68 kDa tau_45–230 aggregates were significantly decreased in AD samples. A slight decrease in sarkosyl-resistant ~168 kDa tau_45–230 aggregates was also detected in AD samples when compared to age-matched controls. (D) Quantification of immunoreactive bands shown in B. Numbers represent the mean ± S.E.M. control subjects (n = 10) and subjects with AD (n = 7). *Differs from control values, p<0.05. ** Differs from control values, p<0.01.

We next assessed the stability of these tau_45–230 aggregates. For these experiments, brain extracts were incubated in the presence of sarkosyl, centrifuged, and the pellet was resuspended in sample buffer. Samples were analyzed by means of Native-PAGE followed by immunoblotting. Tau_45–230 immunoreactive bands at ~68 kDa and ~168 kDa molecular weights were detected both in extracts incubated in the presence or absence of detergent (Fig. 1C). These tau_45–230 immunoreactive bands varied in intensity. Quantitative analysis showed only a slight decrease in the amount of the ~168 kDa tau_45–230 aggregates after detergent-incubation in AD samples when compared to those in untreated AD brain extracts (Fig. 1D). On the other hand, a quantitative analysis of ~68 kDa tau_45–230 oligomers showed that only half of these aggregates present in AD were present in brain samples after their incubation in the presence of sarkosyl (Fig. 1C). These results suggested that larger tau_45–230 aggregates were more detergent-resistant than smaller ones in AD brains.

3.2-. Tau_45–230 formed aggregates in vitro and modulated full-length tau aggregation.

To further analyze the ability of this tau fragment to aggregate, we incubated recombinant tau_45–230 in the presence of arachidonic acid (AA) for 24 hrs. The use of AA to induce full-length tau at physiological protein concentrations has been extensively characterized (reviewed by Gamblin et al., 2003). Samples were then analyzed by SDS-PAGE followed by Western blot analysis. Under these experimental conditions, a tau immunoreactive band was detected at ~32 kDa molecular weight corresponding to monomeric recombinant tau_45–230 (17 kDa tau_45–230 + his-tag) when the fragment was incubated in the absence of AA (Fig. 2A). On the other hand, several immunoreactive bands were detected when tau_45–230 was incubated in the presence of AA. The strongest band was detected at ~70 kDa molecular weight. Bands were also detected at ~100 to 150 kDa and at ~225 kDa molecular weights (Fig. 2A). To determine the stability of such aggregates, samples were incubated in the presence or absence of sarkosyl as described in the Material and Methods section. Immunoblot analysis of detergent-resistant aggregates showed that most of the tau_45–230 aggregates were present under both experimental conditions, albeit at different concentrations (Fig. 2B). Thus, monomeric tau_45–230 was significantly higher in sarkosyl-treated aggregates as compared to non-treated ones (162 ± 26% vs. 100% respectively, p< 0.05). On the other hand, a significant decrease in ~70 kDa, ~100–150 kDa, and ~225 kDa bands was detected in sarkosyl-treated aggregates when compared to untreated controls (Fig. 2C) (75 ± 11%*, 74 ± 7 %*, and 31 ± 9 %**, respectively, vs. untreated aggregates considered 100%. *p< 0.05; **p< 0.01).

Figure 2: Recombinant tau_45–230 aggregation.

(A) Detection of recombinant tau_45–230 aggregates induced by arachidonic acid (AA) by Western blot analysis using a tau antibody (clone tau5). (B) Stability of tau_45–230 aggregates incubated in the presence or absence of sarkosyl (1%) was determined by quantitative Western blot analysis using a tau antibody (clone tau5). (C) Quantitative analysis of immunoreactive bands. Values represent mean ± S.E.M. from 3 independent experiments per condition. *Differs from control values, p<0.05. ** Differs from control values, p<0.01.

We determined next whether tau_45–230 aggregates could be internalized by hippocampal neurons affecting neuronal viability. For these experiments, 7 days in culture hippocampal neurons were incubated for 24 hr in the presence or absence of fluorescein-labeled tau_45–230 aggregates added directly to the culture medium. In the absence of tau_45–230 aggregates, hippocampal neurons displayed an elaborated network of axons and dendrites with no signs of degeneration (Fig. 3A). The entirety of this neuritic network was uniformly stained using a tubulin antibody. As expected, no fluorescein immunoreactivity was detected in these neurons (Fig. 3B). On the other hand, intense diffused fluorescein immunoreactivity was detected in cell bodies of hippocampal neurons incubated with labeled tau_45–230 aggregates (Fig. 3D). In addition, numerous fluorescent spots along neurites could be identified in almost all these neurons (Fig. 3D). These fluorescent spots corresponded to fragmented neuritic processes also labeled using a tubulin antibody (Fig. 3C).

Figure 3: Internalization of tau_45–230 aggregates induced neurodegeneration in cultured hippocampal neurons.

(A-D) Seven days in culture hippocampal neurons incubated for 24 hr in the absence (A & B) or the presence (C & D) of fluorescein-labeled recombinant tau_45–230 aggregates and counterstained with a tubulin antibody (A & C). Untreated hippocampal neurons did not show any sign of degeneration (A) and were not fluorescein-positive (B). On the other hand, hippocampal neurons incubated with fluorescent tau_45–230- aggregates (C & D) were intensely fluorescein-positive and showed numerous signs of degeneration (D). Scale bar = 20 μm.

We assessed next whether the presence of tau_45–230 could affect the polymerization of full-length tau. For these in vitro experiments, monomeric tau_45–230 was incubated with monomeric full-length tau in the presence of AA. After 24 hr aggregation, samples were analyzed by immunoblotting using antibodies that recognize only full-length tau (clone T46) or both tau_45–230 and full-length tau (clone tau5). As previously described, AA induced the polymerization of full-length tau into aggregates of ~150 kDa and several aggregates of molecular weight ranging from ~200 to 250 kDa as recognized by T46 antibody (Fig. 4A). In the presence of tau_45–230, all of these bands were easily detectable (Fig. 4A). However, the band corresponding to monomeric full-length tau was significantly stronger than in the absence of tau_45–230 (466 ± 125 %* vs. 100%, *p<0.05. Fig. 4B). Similar pattern was detected when ~150 kDa molecular weight immunoreactive bands were analyzed (135 ± 11%*, vs. 100%, *p<0.05. Fig. 4B). On the other hand, a significant decrease was observed in the higher molecular weight T46 immunoreactive bands (~225 kDa) when full-length tau was aggregated in the presence of tau_45–230 as compared to its aggregation without this tau fragment (70 ± 10*, vs. 100%, *p<0.05. Fig. 4B). Together, these results suggest that the presence of tau_45–230 modulated full-length tau polymerization, decreasing the formation of bigger tau aggregates and favoring the formation of smaller full-length tau aggregates.

Figure 4: Effects of tau_45–230 on full-length tau aggregation.

(A) Quantitative Western blot analysis of aggregates formed by tau_45–230, full-length tau (FL-tau), and both tau forms combined when incubated in the presence of arachidonic acid (AA). Immunoblots were reacted with a tau antibody that recognizes only full-length tau (clone T46) and a tau antibody that recognizes both tau forms (clone tau5). Note the significant increase in the levels of both monomeric FL-tau and ~150 kDa FL-tau aggregates and a significant decrease in the levels of ~225 kDa FL-tau aggregates in the presence of tau_45–230. (B & C) Quantitative analysis of immunoreactive bands when tau_45–230 and full-length tau were aggregated alone or combined and detected with T46 (B) and tau5 (C) antibodies. Values represent mean ± S.E.M. from 9 independent experiments per condition. *Differs from FL-tau alone values, p<0.05. ** Differs from FL-tau alone values, p<0.01.

3.3-. Tau_45–230 aggregates internalized via a clathrin-mediated mechanism and induced neurite degeneration.

A growing body of evidence suggests that oligomeric protein aggregates are released from degenerating neurons and could be internalized by surrounding neurons further propagating the disease process (Danzer et al., 2007, 2009, & 2011; Kfoury et al., 2012; Saman et al., 2012; Wu et al., 2013). This seems to be the case of full-length tau aggregates which are internalized by a dynamin-dependent mechanism. Internalized full-length tau aggregates, in turn, are involved in the transmission of tau pathology (Wu et al., 2013; Falcon et al., 2018). To determine whether tau_45–230 aggregates could be taken up also by dynamin-mediated endocytosis, neurons were preincubated with the dynamin inhibitor peptide for 1 hr before the addition of tau_45–230 aggregates as previously described (Kelly and Ferreira, 2007). Twenty-four hr later, control and treated hippocampal neurons were scraped in Laemmli buffer and the content of internalized tau_45–230 aggregates were assessed by quantitative Western blot analysis. Strong immunoreactive bands corresponding to tau_45–230 aggregates were detected both in DIP-treated and untreated hippocampal neurons incubated in the presence of these aggregates (data not shown). Quantitative analysis of these immunoreactive bands showed no significant differences when DIP-treated cultures were compared to untreated controls (98% ± 2% vs. 100%). In contrast, experiments that were performed in parallel with full-length tau aggregates as an internal control showed a very significant decrease in full-length tau aggregate uptake in DIP-treated neurons as compared to untreated controls (50% ± 5%** vs. 100 %; **p<0.01) as previously described (Wu et al., 2013; Falcon et al., 2018). These results suggested that the internalization of tau_45–230 aggregates was not dynamin-dependent.

We determined next whether tau_45–230 oligomers were internalized by a clathrin-mediated mechanism instead. For these experiments, cultured hippocampal neurons incubated in the presence or absence of clathrin-mediated inhibitors (CPZ or MDC) were cultured for 24 hr in the presence or absence of aggregated tau_45–230 as described in the Material and Methods section. Immunoblots of whole cell extracts obtained from untreated control neurons showed only one intensely reactive full-length tau for both tau5 and T46 antibodies (Fig. 5A). On the other hand, neurons incubated with tau_45–230 aggregates in the absence of clathrin inhibitors and reacted with a tau antibody, that recognized both full-length tau and tau_45–230, showed intense immunoreactive bands at ~52, ~68 and ~150 kDa molecular weights (Fig. 5A). Only one immunoreactive band at 52 kDa molecular weight was recognized when these membranes were reprobed using a tau antibody that only recognized endogenous full-length tau (Fig. 5A). These results suggested that dimers, trimers and oligomeric forms of tau_45–230 were internalized by cultured hippocampal neurons. Similar immunoreactive bands were present when cultured neurons were incubated with clathrin inhibitors before the addition of tau_45–230 aggregates. However, bands at ~68 and ~150 kDa appeared weaker than those in controls suggesting that inhibitors of clathrin significantly diminished the internalization of tau_45–230 aggregates as compared to untreated controls (Fig. 5A). Quantification of these bands supported these results. Thus, tau_45–230 aggregates of ~150 kDa decreased to 54% ± 5 %* in the presence of CPZ and to 42% ± 11%* in the presence of MDC when compared to untreated controls (100%, *p<0.05) (Fig. 5B). Similar results were detected when tau_45–230 aggregates of ~68 kDa molecular weight were quantified under these experimental conditions (CPZ-treated neurons: 69% ± 8%**; and MDC-treated neurons: 69% ± 2%** vs. untreated controls: 100%.**p <0.01) (Fig. 5C).

Figure 5: Tau_45–230 aggregates were internalized by a clathrin-mediated mechanism in hippocampal neurons.

(A-B) Quantitative Western blot analysis of untreated control hippocampal neurons and hippocampal cultures incubated with aggregated tau_45–230 in the presence or absence of chloropromazine (CPZ) and monodansylcadaverine (MDC) (A) Immunoblots were reacted with tau antibodies (clone tau5 and clone T46), and normalized to α-tubulin (clone DM1A) as protein controls. (B & C) Note the decrease in the levels of aggregated tau_45–230 in hippocampal neurons cultured in the presence of CPZ or MDC when compared to control neurons incubated with aggregated tau_45–230 in the absence of CPZ and MDC. Values represent the mean ± S.E.M. from 4 independent experiments per condition. The levels of aggregated tau_45–230 in untreated neurons were considered 100%. **Differs from neurons incubated with aggregated tau_45–230 in the absence of CPZ and MDC controls, p<0.01

4-. DISCUSSION

In the present study, we identified tau_45–230 oligomers in affected brain areas in AD subjects. In addition, our results suggested that these aggregates could induce degeneration when internalized by otherwise healthy central neurons. Furthermore, the data presented herein suggested that the presence of tau_45–230 could affect full-length tau aggregation, modulating tau-mediated toxicity.

Compelling evidence exists regarding the role that protein aggregation plays in neurodegenerative diseases. Thus, numerous studies have shown that the accumulation of beta-amyloid (Aβ), synuclein, and TAR DNA-binding protein 43 aggregates trigger the pathological process in AD and other dementias, Parkinson’s disease, and amyotrophic lateral sclerosis, respectively (reviewed by Soto and Pritzkow, 2018). Full-length tau aggregates also play an important role in AD and related disorders (Maeda et al., 2005; Berger et al., 2007; Hernandez and Avila, 2008; Kolarova et al., 2012; Lasagna-Reeves, et al., 2012; Guo et al., 2017). In these diseases, intracellular deposits of full-length tau organize into insoluble paired helical and straight filaments forming neurofibrillary tangles, one of the hallmark lesions in AD and other tauopathies (Kosik et al., 1986; Grundke-Iqbal et al., 1986; Wood et al., 1986; Iqbal et al., 2005 & 2016; Ding and Johnson, 2008; Mandelkow and Mandelkow, 2012). On the other hand, the presence of aggregates of any fragment of this MAP has yet to be reported in these diseases. Our results showed that, in addition to monomeric tau_45–230, affected areas in brain samples obtained from AD subjects contained aggregates of this neurotoxic tau fragment. The ability of tau_45–230 to aggregate was already observed in tau_45–230-transfected cultured hippocampal neurons. In these neurons, tau_45–230 homodimers were easily detectable as early as 24 hr after transfection (Afreen and Ferreira, 2017). Although this tau fragment does not contain the microtubule-binding repeats, tau_45–230 does contain the sequence that mediates full-length tau dimerization via an electrostatic zipper aligned in an antiparallel manner (Kondo et al., 1988; Kanai et al., 1992; Park and Ferreira, 2005; Rosenberg et al., 2008; Feinstein et al., 2016). Our data indicated that in AD subjects, these dimers were part of a heterogenous mixture in which tau_45–230 oligomers of ~68 kDa and ~168 kDa molecular weights predominated. We observed a similar heterogeneity when recombinant tau_45–230 was aggregated in vitro using inducers of aggregation. These results confirmed and extended a previous atomic force microscopy analysis showing that tau_45–230 dimers, tetramers, and octamers are part of the mixture when this tau fragment is aggregated in vitro (Feinstein et al., 2016). Full-length tau oligomers of marked size heterogeneity have been previously described in AD brains (Maeda et al., 2005; Hernandez and Avila, 2008; Lasagna-Reeves et al., 2012). Our results also suggested that most of the higher molecular weight tau_45–230 aggregates present in brain samples from AD subjects were sarkosyl-resistant. On the other hand, less than half of the smaller aggregates were detergent-resistant. Soluble and insoluble tau_45–230 aggregates could have differential toxic effects, as is the case for full-length tau aggregates. While soluble full-length tau aggregates seemed to be responsible for neuronal loss, insoluble ones were mainly involved in synapse loss (Kimura et al., 2010). The potential biological implications of differences in the size and/or the stability of tau_45–230 aggregates present in tauopathies await further investigation.

Additionally, our results showed that tau_45–230 oligomers present in AD subjects were less susceptible to detergent treatment than those prepared in vitro. Several mechanisms could be responsible for these differences. It is possible that the endogenous enhancers of aggregation are more efficient in inducing tau_45–230 aggregation and/or the stabilization of oligomers than the exogenous agents used for in vitro experiments. These differences could reflect different time course of aggregation as well. While in vitro aggregation was performed for 24 hr, the formation of tau_45–230 oligomers in the context of these diseases could expand throughout the whole neurodegenerative process, enhancing their stability overtime. Alternatively, these brain tau_45–230 aggregates could contain other proteins responsible for increasing their resistance to detergent-induced disassembling. Besides, this tau fragment could be posttranslational modified once aggregated in situ, changing the oligomer stability. One of such posttranslational modifications could be phosphorylation since tau_45–230 contains several tau residues that are highly phosphorylated during the disease process (Kosik et al., 1986; Grundke-Iqbal et al., 1986; Iqbal et al., 2005; Ding and Johnson, 2008; Kolarova et al., 2012; Mandelkow & Mandelkow, 2012; Guo et al., 2017). However, this possibility seems unlikely since cleavage of full-length tau into this fragment preceded the enhanced phosphorylation observed in aggregated Aβ-treated neurons (Ferreira et al., 1997; Alvarez et al., 1999; Ekinci et al., 1999; Park and Ferreira, 2005). Moreover, tau_45–230 was not detected using phosphorylation-dependent tau antibodies when cultured hippocampal neurons were treated with aggregated Aβ (Park and Ferreira, 2005; Ferreira and Bigio, 2011). Further studies will be needed to determine whether tau_45–230 undergoes other posttranslational modifications in the course of the degenerative process.

Our results provided insights into the biological relevance of the presence of tau_45–230 aggregates in AD. Tau_45–230 oligomers could be merely a result of neurodegeneration. Alternatively, these aggregates could contribute to tau_45–230 toxicity. Our data showing that the incubation of hippocampal neurons in the presence of tau_45–230 aggregates triggered the formation of neurite varicosities and fragmentation of axons and dendrites favored the latter possibility. In addition, these data identified the mechanisms of tau_45–230 oligomer internalization. Contrary to what has been reported in the case of the majority full-length tau aggregates, the internalization of tau_45–230 oligomers seemed to be mostly independent of dynamin-mediated endocytosis. On the other hand, clathrin inhibitors significantly reduced the amount of tau_45–230 internalized by cultured hippocampal neurons. These differences could be due to the predominantly smaller size of the tau_45–230 aggregates as compared to full-length tau ones. Thus, it has been reported that while bigger filamentous tau was internalized by an active mechanism dependent on dynamin-mediated endocytosis. Conversely, dimers and trimers of full-length tau were endocytosed by a clathrin-mediated mechanism (Wu et al., 2012).

In addition to the intrinsic toxic effects of oligomeric tau_45–230 in hippocampal neurons, our results suggested that the presence of tau_45–230 could also affect the formation of full-length tau aggregates. The cleavage of full-length tau into this fragment is an early event in the degenerative process (Park and Ferreira, 2005). Therefore, the early accumulation of this tau fragment in the cytosol of degenerating neurons could alter the formation of hyperphosphorylated full-length tau aggregates, modulating tau toxicity. Our in vitro experiments showed that tau_45–230 interfered with the formation of full-length tau aggregates, inducing the accumulation of both monomeric and smaller aggregates of this MAP and significantly reducing the formation of larger full-length tau aggregates. These results are consistent with a previous negative-staining electron microscopy study showing the presence of fewer and shorter full-length tau aggregates in the presence of tau_45–230 (Ferreira and Bigio, 2011). Several mechanisms could underlie this effect of tau_45–230 on full-length tau aggregation. One possibility is that by forming heterodimers, tau_45–230 could sequester full-length tau, and in turn, decrease free monomeric full-length tau available to participate in the elongation of aggregates into bigger ones. We have previously shown a similar mechanism in tau_45–230-transfected hippocampal neurons. In these neurons, tau_45–230 reduced the binding of full-length tau to microtubules and induced a pseudo null tau phenotype affecting stabilization of microtubules and neurite elongation (Afreen and Ferreira, 2019). Alternatively, tau_45–230 could bind to oligomeric full-length tau already formed and prevent further incorporation of full-length tau, impairing their conversion into larger aggregates. Our results showing an increase in monomeric full-length tau in addition to a decrease of larger aggregates seemed to support this possibility. Regardless of the mechanisms, our results suggest that tau_45–230 plays a role in the complex mechanisms regulating the formation of full-length tau pathological inclusions favoring the formation of smaller oligomeric full-length tau, the main inducers of neuronal degeneration in the context of neurodegenerative diseases (reviewed by Hernandez and Avila, 2008; Mandelkow and Mandelkow, 2008; Soto and Pritzkow, 2018).

4.1-. Conclusions

The data presented herein indicate that tau_45–230 is not only present as a monomer in brain areas affected in AD subjects, as previously described, but also as oligomers. These abundant aggregates, heterogenous in size and in their stability, could induce progressive neurite degeneration and cell death when internalized by otherwise healthy neurons through a clathrin-mediated mechanism. These findings further support the hypothesis that the aggregation of tau_45–230 has important implications for the pathobiology of AD. First, the intrinsic toxic effects of tau_45–230 aggregates in hippocampal neurons could play a role in the propagation of tau pathology. Second, tau_45–230 oligomers could also modulate the aggregation of full-length tau, inducing the formation of smaller and more toxic forms of such aggregates. Third, considering that tau cleavage leading to the generation of this fragment is an early event in Aβ-induced neurodegeneration, it is tempting to speculate that therapeutic interventions directed to prevent such cleavage and/or tau_45–230 aggregation could be beneficial to ameliorate the consequences of this disease process. Collectively, these results identified additional molecular mechanisms underlying the neurotoxic effects of tau_45–230 and highlighted its role as a mediator of tau pathology in neurodegenerative diseases.

HIGHLIGHTS.

• Tau_45–230 small aggregates are present in affected brain areas in Alzheimer’s disease subjects.

• Hippocampal neurons internalize tau_45–230 aggregates by a clathrin-mediated mechanism

• Internalized tau_45–230 aggregates induce neuronal degeneration

ABBREVIATIONS

AA

Arachidonic Acid

Aβ

Beta Amyloid

AD

Alzheimer’s Disease

CBDG

Corticobasal Degeneration

CPZ

Chloropromazine

DIP

Dynamin Inhibitor Peptide

FL-Tau

Full length tau

MAP

Microtubule-Associated Protein

MDC

Monodansylcadaverine

MEM

Minimum Essential Medium

MEM10

MEM containing 10% horse serum

PAGE

Polyacrylamide Gel Electrophoresis

PiD

Pick Disease

PMI

Post Mortem Interval

PSP

Progressive Supranuclear Palsy

TPSD

Tangle-Predominant Senile Dementia