S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

BIOCHEMISTRY Copyright © 2024 The

Authors, some rights
D-­peptide-­magnetic nanoparticles fragment tau fibrils reserved; exclusive
licensee American
and rescue behavioral deficits in a mouse model of Association for the
Advancement of
Alzheimer’s disease Science. No claim to
original U.S.
Government Works.
Ke Hou1,2,3,4,5, Hope Pan1,2,3,4,5, Hedieh Shahpasand-­Kroner6,7,8, Carolyn Hu1,2,3,4,5, Distributed under a
Romany Abskharon1,2,3,4,5, Paul Seidler1,2,3,4,5,9, Marisa Mekkittikul6,7,8, Melinda Balbirnie1,2,3,4,5, Creative Commons
Carter Lantz1, Michael R. Sawaya1,2,3,4, Joshua L. Dolinsky1,2,3,4,5, Mychica Jones6,7,8, Attribution
Xiaohong Zuo6,7,8, Joseph A. Loo1,2,3,4, Sally Frautschy6,7,8, NonCommercial
Greg Cole6,7,8, David S. Eisenberg1,2,3,4,5* License 4.0 (CC BY-­NC).

Amyloid fibrils of tau are increasingly accepted as a cause of neuronal death and brain atrophy in Alzheimer’s disease
(AD). Diminishing tau aggregation is a promising strategy in the search for efficacious AD therapeutics. Previously,
our laboratory designed a six-­residue, nonnatural amino acid inhibitor D-­TLKIVW peptide (6-­DP), which can pre-
vent tau aggregation in vitro. However, it cannot block cell-­to-­cell transmission of tau aggregation. Here, we find
D-­TLKIVWC (7-­DP), a d-­cysteine extension of 6-­DP, not only prevents tau aggregation but also fragments tau fibrils
extracted from AD brains to neutralize their seeding ability and protect neuronal cells from tau-­induced toxicity.
To facilitate the transport of 7-­DP across the blood-­brain barrier, we conjugated it to magnetic nanoparticles

Downloaded from https://www.science.org on May 06, 2024

(MNPs). The MNPs-­DP complex retains the inhibition and fragmentation properties of 7-­DP alone. Ten weeks of
MNPs-­DP treatment appear to reverse neurological deficits in the PS19 mouse model of AD. This work offers a direc-
tion for development of therapies to target tau fibrils.

INTRODUCTION agents and drug carriers in AD (9). MNPs are benign, easily func-
Alzheimer’s disease (AD) is a progressive and fatal neurodegenera- tionalized, and able to cross the BBB.
tive disorder characterized by two major neuropathologic hallmarks: To diminish tau aggregation, our laboratory has developed
extracellular senile plaques containing fibrils of amyloid beta (Aβ) structure-­based design of compounds aimed to inhibit tau aggrega-
and intracellular neurofibrillary tangles of tau (1). Increasing evi- tion or disassemble tau fibrils (10–12). For example, Sievers et al.
dence indicates that cognitive symptoms and severity in AD are more (10) used the atomic structure of a segment from the tau fibril core
strongly correlated with tau pathology than Aβ plaques (2). The fail- as a template to design a six-­residue, nonnatural amino acid inhibi-
ure of multiple phase 3 clinical trials with Aβ-­targeting drugs (3) has tor D-­TLKIVW peptide (6-­DP) that binds the ends of tau fibrils to
redirected AD therapeutics to focus on targeting tau pathology. Re- cap their growth; 6-­DP was demonstrated to prevent tau aggrega-
searchers have pursued diverse approaches to diminish tau aggrega- tion in vitro. Compared to l-­enantiomeric peptides, D-­enantiomeric
tion in AD, including reducing expression of the MAPT gene which peptides are known to be less immunogenic, less protease-­sensitive
encodes tau, modulating posttranslational modifications of tau, pre- in vitro, and more resistant to degradation in vivo (13). While D-­
venting tau aggregation, passive antibody immune neutralization, or TLKIVW inhibits fibril growth, it does not prevent AD-­extracted
clearance of different tau species, and stabilizing the microtubules to tau seed spreading in biosensor cells (14) and is theoretically un-
which tau binds (4). So far, tau antibodies have failed in the clinic (5), able to cross the BBB, which limits its clinical potential.
possibly due to their inability to penetrate neurons. Tau aggregation Here, we designed a seven-­residue D-­TLKIVWC peptide (7-­DP),
inhibitors such as curcumin and methylene blue/LMTX have pro- which is a C-­terminal d-­cysteine extension of D-­TLKIVW. We
gressed to phase 2 and phase 3 clinical trials, respectively, but have yet found 7-­DP to be a better inhibitor of tau fibril growth than 6-­DP
to show efficacy in treating disease (4). These molecules fell short in vitro and in addition to be a tau fragmentor. To facilitate its trans-
partly due to their limited biocompatibility (6) and low blood-­brain port across the BBB and improve its stability, we conjugated 7-­DP
barrier (BBB) permeability (7). Recently, nanomaterials applied as with MNPs. We found the treatment of 7-­DP–conjugated MNPs
drug candidates or drug carriers have been reported to overcome (MNPs-­DP) reversed the behavioral deficits of PS19 mice, a model
these limitations (8). Among them, magnetic nanoparticles (MNPs) of AD tauopathy. 7-­DP not only prevents tau aggregation but also
have attracted attention as magnetic resonance imaging contrast dissolves existing tau aggregates.

1
Department of Chemistry and Biochemistry, UCLA, Los Angeles, CA, USA. 2Depart- RESULTS
ment of Biological Chemistry, UCLA, Los Angeles, CA, USA. 3UCLA-­DOE Institute, D-­TLKIVWC is a better inhibitor of tau aggregation
Los Angeles, CA, USA. 4Molecular Biology Institute, UCLA, Los Angeles, CA, USA.
5
Howard Hughes Medical Institute, Los Angeles, CA, USA. 6Department of Neurology, than D-­TLKIVW
UCLA, Los Angeles, CA, USA. 7Veterans Administration Greater Los Angeles Health- We first evaluated D-­TLKIVWC (7-­DP) as an inhibitor of tau ag-
care System, Geriatric Research and Clinical Core, Los Angeles, CA, USA. 8Depart- gregation by comparing its effects on fibril formation of tau K18+ to
ment of Medicine, UCLA, Los Angeles, CA, USA. 9Department of Pharmacology and
Pharmaceutical Sciences, University of Southern California, Los Angeles, CA, USA. that of the previously published inhibitor D-­TLKIVW (6-­DP) (10).
*Corresponding author. Email: david@​mbi.​ucla.​edu K18+ is a recombinant construct of tau, spanning residues 244 to

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 1 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

380, which forms the ordered core visible in the cryo–electron or 6-­DP (fig. S2), but 7-­DP reduced the abundance of fibrils more
microscopy structure of tau fibrils extracted from postmortem than 6-­DP, especially at 100 μM concentration, where no tau K18+
brains of patients with AD (15). The aggregation kinetics of tau fibrils (red arrows) were observed; only tau oligomers (black ar-
K18+ monomer incubated alone or with increasing concentrations rows) and fibrils formed from 7-­DP itself (blue arrow) were seen.
of 7-­DP or 6-­DP were monitored by thioflavin T (ThT) assays in Together, these data show that D-­TLKIVWC is a better tau fibril-­
parallel. Tau K18+ fibrillation exhibited typical sigmoidal curves binding inhibitor than D-­TLKIVW.
(black line in Fig. 1A and fig. S1A); both 7-­DP (Fig. 1A) and 6-­DP
(fig. S1A) showed dose-­dependent inhibition of tau K18+ aggrega- D-­TLKIVWC fragments recombinant tau K18+ fibrils
tion by prolonging the lag phase and decreasing the ThT fluores- To characterize its fragmentation ability, we added 7-­DP to pre-
cence signal. However, 7-­DP is a stronger inhibitor than 6-­DP formed tau K18+ fibrils and observed an immediate decrease in ThT
because its estimated median inhibitory concentration (IC50) is fluorescence intensity, whereas adding H2O or 6-­DP had no notable
12.1 ± 0.3 μM (fig. S1B) compared to that of 19.6 ± 1.4 μM (fig. S1C) effect on the ThT signal (Fig. 1B). TEM characterization revealed
for 6-­DP. Consistent with these results, in parallel circular dichroism that incubation of preformed tau K18+ fibrils with 7-­DP led to the
(CD) spectroscopy further revealed that 7-­DP can better prevent tau disappearance of tau fibrils and the appearance of globular species,
K18+ monomers from forming β sheet structures (probed by the whereas 6-­DP could not change the morphology of tau K18+ fibrils
negative peak around 220 nm) and keep tau in a random coil con- (Fig. 1C). After incubation with 7-­DP, we observed a slight shift
formation with a negative peak near 200 nm (fig. S1, D and E). in the broad peak of tau fibrils (shown by the blue dashed line in
Transmission electron microscopy (TEM) showed the counts and fig. S3) toward lower mass/charge ratio in their native electrospray
length of tau K18+ fibrils decreased with the increased dose of 7-­DP ionization mass spectra, suggesting that the fibrils were fragmented

Downloaded from https://www.science.org on May 06, 2024

Fig. 1. 7-­DP inhibits tau aggregation and fragments tau fibrils in vitro and in cells. (A) ThT assays of 20 μM tau K18+ monomer incubated alone or in the presence of
5, 10, 20, 50, and 100 μM 7-­DP. (B) ThT fragmentation assay of 20 μM tau K18+ fibrils with H2O, 50 μM 6-­DP, or 50 μM 7-­DP added at 16 hours. (C) TEM image of 20 μM
recombinant tau K18+ fibrils before and after 24 hours of incubation with 50 μM 7-­DP or 50 μM 6-­DP at 37°C. Scale bars, 200 nm. (D) CD spectra of tau K18+ monomer, tau
K18+ fibrils before and after incubation with various concentrations of 7-­DP for 48 hours at 37°C. (E) Representative TEM images of AD-­extracted tau fibrils (indicated by
red arrows) before and after 48 hours of incubation with 500 μM 7-­DP or 500 μM 6-­DP at 37°C. (F) Quantification of AD tau fibrils in TEM images (n = 10 images). (G) West-
ern blotting analysis of cell lysate from HEK293 cells expressing 1N4R tau treated with 20 μM or 50 μM 7-­DP, using Dako tau antibody. β-­Actin was used as a loading control.
a.u., arbitrary units.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 2 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

by 7-­DP. To further verify that 7-­DP can fragment tau K18+ fibrils, puncta visible under fluorescence microscopy (indicated by white ar-
a dose-­dependent experiment was performed. As expected, 7-­DP rows in fig. S5E). The introduction of 7-­DP later significantly de-
treatment caused a dose-­dependent decrease of (i) β strand content creased the number of puncta compared with phosphate-­buffered
of tau K18+ fibrils as measured by CD spectroscopy (Fig. 1D), (ii) saline (PBS)–treated group (fig. S5, E and F). These experiments
ThT fluorescence intensity (fig. S4A), (iii) binding of the anti-­ indicated that 7-­DP can fragment tau aggregates in cells.
amyloid OC antibody (fig. S4B), and (iv) number of tau K18+ fibrils
under TEM (fig. S4, C to I). It was noteworthy that the globular D-­TLKIVWC blocks AD-­extracted tau seeding in tau
species in Fig. 1C and fig. S4I was significantly larger than tau oligo- biosensor cells
mers observed in fig. S2A (fig. S4J), indicating that the fragmented It has been suggested that tau pathology propagates through the
products of tau fibrils by 7-­DP were probably the tau K18+ and 7-­DP brain following a prion-­like mechanism, in which aggregates in a
complex. Collectively, these data establish that 7-­DP can fragment diseased cell can travel to adjacent cells and seed further protein
recombinant tau K18+ fibrils, whereas 6-­DP cannot. The fragmented aggregation (19, 20). Blocking the prion-­like spread of tau may halt
products were probably composed of tau K18+ and 7-­DP. the progression of AD. We used a HEK293T cell line stably express-
ing yellow fluorescent protein (YFP)–fused tau (tau biosensor cell)
D-­TLKIVWC fragments AD-­extracted tau fibrils to investigate whether 7-­DP could block the spread of tau pathology
Previous studies have revealed that heparin-­induced recombinant (21). We transfected AD tau fibril seeds to induce the aggregation of
tau K18+ fibrils adopt different structures than pathological tau the endogenous fluorescent tau, which formed bright puncta visible
fibrils in AD (16). Here, we investigated whether 7-­DP can fragment by fluorescence microscopy (indicated by white arrows in Fig. 2A).
tau fibrils extracted from autopsied AD brains (AD tau fibrils) (17). Automated image analysis of visible puncta provided an objective
TEM characterization showed that most of the AD tau fibrils disap- quantification of seeded aggregation (Fig. 2B). When AD tau fibrils
peared after incubation with 500 μM 7-­DP for 2 days, while abun- were incubated with 7-­DP overnight before transfection, tau puncta

Downloaded from https://www.science.org on May 06, 2024

dant AD tau fibrils remained in the 6-­DP–treated sample (Fig. 1E). were reduced significantly with an IC50 of 7-­DP around 33.5 μM
Quantification confirmed that 7-­DP significantly reduced the number (Fig. 2, A and B). We also tested whether 7-­DP could block AD-­
of AD tau fibrils, but 6-­DP did not (Fig. 1F). To further verify that extracted tau seeding when 7-­DP and AD tau seeds were directly
7-­DP can fragment AD tau fibrils, we incubated AD tau fibrils with transfected to biosensor cells without preincubation overnight. 7-­DP
various concentrations of 7-­DP. After 48 hours, a dot blot experi- showed comparable performance in blocking tau spreading [fig. S6
ment was performed with the monoclonal antibody GT38, which (A and B) versus Fig. 2 (A and B)]. This reduction suggested the in-
specifically recognizes AD tau fibrils. The dose-­dependent decrease hibition and fragmentation capacity of 7-­DP, which can change or
of chemiluminescence intensity in fig. S5 (A and B) indicated that destroy the structure of AD tau fibrils and render them no longer
7-­DP reduced the amount of AD Tau fibrils in a dose-­dependent seeding competent. Consistent with our previous report (14), 6-­DP
manner. Western blotting of the supernatant and pellet of AD tau had no measurable inhibition of AD-­extracted tau seeding with
fibrils before and after 48 hours of incubation with 7-­DP indicated concentrations up to 100 μM (fig. S6, C to E).
that the fragmented products were not soluble tau monomers; in-
stead, they were a series of insoluble species, containing dimers and D-­TLKIVWC protects neuronal cells from tau-­induced toxicity
other degraded fragments (fig. S5C). Next, we investigated the ability of 7-­DP to protect neurons from
tau-­induced toxicity. The 3-­(4,5-­dimethylthiazol-­2-­yl)-­2,5-­diphenyl
D-­TLKIVWC fragments tau aggregates in cells tetrazolium bromide (MTT) metabolic assay showed that Neuro2a
As 7-­DP is able to fragment recombinant and AD-­extracted tau (N2a) cells exposed to tau K18+ aggregates, formed by tau K18+
fibrils, we tested whether 7-­DP could fragment intracellular tau ag- monomer incubated alone, had a significantly reduced cell viability
gregates. We expressed 1N4R tau in human embryonic kidney of ~35% (Fig. 2C). In contrast, the aggregates of tau K18+ and 7-­DP
(HEK) 293 cells and seeded them with AD-­extracted tau fibrils for significantly increased cell viability of N2a cells with median effec-
24 hours to accumulate tau aggregates. The successful formation of tive concentration (EC50) of 7-­DP around 21.1 μM (Fig. 2C). This
tau aggregates in the cell was confirmed by Western blot in fig. S5D, indicated that 7-­DP can protect neuronal cells from tau-­induced
where AD tau-­seeded group showed more tau contents in the pellet toxicity by inhibiting tau aggregation. 6-­DP also showed a slight
compared with nonseeded group. The cells were then transfected improvement in cell viability, but it was not as effective as 7-­DP
with 7-­DP followed by a 24-­hour incubation, before lysis and ultra- (fig. S7A) and an estimate of EC50 was unobtainable.
centrifugation. Tau content in the pellet and supernatant was ana- We performed another MTT assay by treating N2a cells with pre-
lyzed by Western blot using antitau rabbit polyclonal antibody (Dako, formed tau K18+ fibrils. Incubation of tau K18+ fibrils with 7-­DP
A0024). The result showed that the 7-­DP reduced the tau content in gave rise to higher cell viability compared with tau K18+ fibrils alone
both the pellet and supernatant in a 7-­DP dose-­dependent manner (Fig. 2D), indicating that unlike putative toxic tau oligomers, the
(Fig. 1G). Cotreatment with proteasome inhibitor MG132 reduced fragmented products of tau K18+ fibrils by 7-­DP were nontoxic or
the effect of 7-­DP, indicating that the tau was possibly degraded by less toxic than tau fibrils. As expected, 6-­DP did not reduce the
the ubiquitin-­proteasome pathway (fig. S5D) (18). Whether tau is cytotoxicity of tau fibrils (fig. S7B), presumably because it cannot
degraded alone or in a 7-­DP–tau complex remains to be determined. disassemble tau fibrils. Both 7-­DP (Fig. 2E) and 6-­DP (fig. S7C)
In addition, we used another HEK293T cell line stably expressing alone were nontoxic to N2a cells under all tested concentrations.
enhanced yellow fluorescent protein (EYFP)–fused tau 1N4R P301S
to visualize whether 7-­DP could fragment tau aggregates in cells. Conjugation of MNPs-­DP complex
Similarly, we first transfected AD tau fibril seeds to induce the ag- Encouraged by the inhibition and fragmentation properties of 7-­DP,
gregation of the endogenous fluorescent tau, which formed bright we aimed to evaluate its clinical potential in vivo. However, 7-­DP is

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 3 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Downloaded from https://www.science.org on May 06, 2024

Fig. 2. 7-­DP blocks tau seeding and reduces tau-­induced cell toxicity. (A) Representative fluorescent images of HEK293 cells expressing YFP-­labeled tau K18 trans-
fected with AD tau fibril seeds in various concentrations of 7-­DP. Fluorescent puncta were marked by white arrows. Scale bars, 100 μm. (B) Quantification of puncta formed
in tau biosensor cells seeded with AD tau fibrils with various concentrations of 7-­DP. (C) Cell viability of N2a cells treated with tau K18+ monomer incubated alone or in
various concentrations of 7-­DP, measured by MTT assay. 7-­DP treatment of tau K18+ monomer rescues cell viability. (D) Cell viability of N2a cells treated with tau K18+
fibrils in various concentrations of 7-­DP, measured by MTT assay. 7-­DP treatment of tau K18+ fibrils rescues cell viability. (E) N2a cells treated with a range of 7-­DP concen-
trations shows no toxicity. Results shown as mean + SD of triplicate wells. Statistical significance was analyzed by one-­way analysis of variance (ANOVA) (*P < 0.05;
**P < 0.01; ***P < 0.001; NS, not significant).

likely to have poor BBB permeability and structural stability due to aggregation, indicating that MNPs-­DP retained the specificity of
its large mass (>800 Da) and the active sulfhydryl group in its side 7-­DP in inhibiting tau aggregation.
chain. To address these challenges, we conjugated 7-­DP to MNPs,
which have been approved for clinical use (22) and are reported to MNPs-­DP fragment recombinant tau K18+ fibrils and
be stable, benign, and able to cross the BBB (23). The conjugation AD-­extracted tau fibrils
involves two steps (fig. S8). Briefly, the carboxylic acid-­stabilized In addition to inhibiting tau K18+ aggregation, MNPs-­DP also frag-
iron oxide MNPs (CD Bioparticles, WHM-­G062) were functional- mented preformed tau K18+ fibrils as evidenced by the decrease of
ized with a maleimide group (MNPs-­MAL) and then covalently ThT signal (Fig. 3C), as well as by the reduction of tau K18+ fibrils
bonded to the sulfhydryl group of 7-­DP through the thiol-­maleimide and the appearance of globular species under TEM (fig. S14, A and
Michael addition click reaction (24). Successful conjugation be- B). However, MNPs-­DP (83 μM DP, red line in Fig. 3C) did not
tween MNPs and 7-­DP was confirmed by morphological changes work as efficiently as 7-­DP (50 μM, blue line in Fig. 3C) to fragment
(from dispersed to aggregated and then dispersed) observed with tau K18+ fibrils (Fig. 3C and fig. S14, B and C). This result is reason-
TEM (fig. S9, A to C), a notable increase of hydrodynamic size and able, considering that conjugation of 7-­DP onto the surface of MNPs
reduction of zeta potential measured by dynamic light scattering may limit the geometry of interaction between 7-­DP and tau K18+
(fig. S9D). The conjugation efficiency of 7-­DP to MNPs was ~94% fibrils. Fragmentation of tau K18+ fibrils by MNPs-­DP was verified
estimated by the ultraviolet absorbance spectrum of 7-­DP (fig. S10). to be dose dependent via (i) reduction of ThT fluorescence intensity
This implied that the concentration of 7-­DP on the MNP surface (Fig. 3D), (ii) decrease of binding to anti-­amyloid OC antibody
was ~4.3 mM/mg of MNPs. (fig. S15A), and (iii) reduction of the abundance of tau K18+ fibrils
and a concurrent increase in the number of globular species ob-
MNPs-­DP show a dose-­dependent inhibition of tau served by TEM (Fig. 3E and fig. S15, B and C). MNPs-­DP fragment-
K18+ aggregation ed AD-­extracted tau fibrils after a 48-­hour incubation (Fig. 3, F and
ThT assay (Fig. 3A) and TEM characterization (Fig. 3B and fig. S11) G). In contrast, the naked MNPs alone could not disassemble tau
demonstrated that like 7-­DP, MNPs-­DP inhibited tau K18+ aggrega- K18+ fibrils (Fig. 3C and fig. S14D) or AD tau fibrils at any dose
tion and reduced the length of tau fibrils in a dose-­dependent (Fig. 3, F and G).
manner. To rule out the possibility that the inhibition of MNPs-­DP
originated from the MNPs or the conjugation process, naked MNPs MNPs-­DP block AD-­extracted tau seeding in tau
and a scrambled peptide D-­LKTWIVC–conjugated MNPs (MNPs-­ biosensor cells
DPsc) were tested following the same aggregation protocols. Neither After incubating AD tau fibril seeds with MNPs-­DP overnight and
naked MNPs (fig. S12) nor MNPs-­DPsc (fig. S13) prevented tau then transfecting them into tau biosensor cells, we observed a

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 4 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Downloaded from https://www.science.org on May 06, 2024

Fig. 3. MNPs-­DP retain the properties of 7-­DP to inhibit tau aggregation and fragment tau fibrils in vitro. (A) ThT assays of 20 μM tau K18+ monomer incubated
alone or in the presence of MNPs-­DP (0.005, 0.01, 0.02, and 0.04 mg/ml). The concentration of 7-­DP on the surface of MNPs is 21.5, 43, 86, 172 μM, respectively. (B) TEM
images of tau K18+ incubated alone or in the presence of MNPs-­DP (0.005 and 0.04 mg/mL) for 48 hours. Scale bars, 200 nm. (C) ThT fragmentation assay of 20 μM tau K18+
fibrils with H2O, MNPs (0.02 mg/ml), MNPs-­DP (0.02 mg/ml; 86 μM 7-­DP), or 50 μM 7-­DP added at 10 hours. (D) ThT fragmentation assay of 20 μM tau K18+ fibrils with
MNPs-­DP (0.005, 0.01, and 0.02 mg/ml) added at 16 hours. (E) TEM images of tau K18+ fibrils before and after 24 hours of incubation with MNPs-­DP (0.02 mg/ml). Scale
bars, 100 nm. (F) Representative TEM images of AD-­extracted tau fibrils (indicated by red arrows) before and after 48 hours of incubation with MNPs (0.02 mg/ml) or MNPs-­
DP (0.02 mg/ml; 86 μM 7-­DP) at 37°C. (G) Quantification of AD tau fibrils in TEM images (n = 10 images).

dose-­dependent reduction in the number of tau puncta (Fig. 4, A (fig. S17A). Our control MNPs were found to be harmless, as previ-
and B). The IC50 was calculated to be 0.0067 mg/ml MNPs-­DP ously reported (fig. S17B) (23) but had no effect on reducing tau-­
(28.9 μM 7-­DP), which is slightly better than that of 7-­DP alone induced cell toxicity (fig. S17, C and D).
(33.5 μM; Fig. 2, A and B). This improvement may be attributed to
the enhanced stability of 7-­DP on the MNP surface, which protects MNPs-­DP cross the BBB in C57BL/6J mice
it from intracellular proteases and the complex redox environment To assess the ability of MNPs-­DP to penetrate the BBB, we adminis-
within cells. Our results suggest that, like free 7-­DP, MNPs-­DP can tered MNPs-­DP (10 mg/kg body weight) to 2-­month-­old wild-­type
inhibit the prion-­like propagation of tau pathology by capping and (WT) C57BL/6J mice via tail vein injection, following intranasal
fragmenting AD tau fibrils. In contrast, the naked MNPs had no injection of 20 μl of 2% mannitol. After 2 hours, the mice were
effect on preventing tau seeding in tau biosensor cells (fig. S16). euthanized, and their brains were perfused to wash out residual
MNPs-­DP from the blood vessels. We analyzed the Fe levels in the
MNPs-­DP reduce tau-­induced cell toxicity brain by inductively coupled plasma mass spectrometry (ICP-­MS)
We used MTT assays to show that MNPs-­DP can protect N2a neu- after digesting the brain tissues. We observed a 1.5-­fold increase in
ronal cells from tau-­induced toxicity by inhibiting tau K18+ aggre- estimated Fe concentration in the brains of mice that were injected
gation (Fig. 4C) and fragmenting tau K18+ fibrils (Fig. 4D). MNPs-­DP with MNPs-­DP compared to PBS-­injected mice (n = 3 per group;
alone were nontoxic to N2a cells under all tested concentrations fig. S18), suggesting that MNPs-­DP can cross the BBB.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 5 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Fig. 4. MNPs-­DP block tau seeding and reduce tau-­induced cell toxicity in cells. (A) Representative fluorescence images of HEK293 cells expressing YFP-­labeled tau

Downloaded from https://www.science.org on May 06, 2024

K18 transfected with AD tau fibril seeds in various concentrations of MNPs-­DP. Fluorescent puncta were marked by white arrows. Scale bars, 100 μm. (B) Quantification of
puncta formed in tau biosensor cells seeded with AD tau fibrils in various concentrations of MNPs-­DP. (C) Cell viability of tau K18+ monomer incubated alone or in various
concentrations of MNPs-­DP, assessed by MTT assay. (D) Cell viability of N2a cells treated with tau K18+ fibrils, in various concentrations of MNPs-­DP measured by MTT
assay. Results shown as mean + SD of triplicate wells. Statistical significance was analyzed by one-­way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001).

MNPs-­DP rescue memory deficits of the PS19 mouse Also, PS19 mice treated with MNPs-­DP displayed a comparable per-
model of AD formance to WT mice and a significant improvement in memory
We used the PS19 transgenic mice that overexpress the P301S mutant function compared with PBS-­treated PS19 mice, while MNPs did not
form of human microtubule-­associated protein tau (MAPT) (25) as a show obvious improvement, indicating the specific activity of 7-­DP
tauopathy model to investigate the therapeutic efficiency of MNPs-­ (Fig. 5, C to E). These results suggest that 10 weeks of treatment with
DP against AD-­related tauopathy in vivo, although the structures of MNPs-­DP tend to rescue spatial learning and memory deficits in the
tau filaments from PS19 mice differ from those of WT tau filaments PS19 model mice of tauopathy.
from human brains (26). Two-­month-­old male PS19 mice were ste- In addition to the Barnes maze test, we used the novel object rec-
reotaxically injected with tau K18+ fibril seeds to induce similar and ognition test to evaluate the cognitive ability of mice (29). Briefly, the
consistent levels of tau pathology in cohorts of PS19 mice (27). Simi- mice were first exposed to two identical objects for 10 min. After a
larly, 6-­month-­old female PS19 mice were also seeded and adopted as 90-­min interval, the mice were allowed to explore one of the familiar
a second cohort of PS19 mice. Two weeks after the stereotaxic sur- objects and a novel one. Cognitive ability can be evaluated by the dif-
gery, MNPs-­DP, MNPs, and vehicle (PBS) were tail vein–injected ference between the exploration time of novel and familiar objects.
into PS19 transgenic mice weekly for 10 weeks at a dose of 10 mg/kg We observed that WT mice and MNPs-­DP–treated PS19 mice ex-
of body weight (n = 13 per group, eight male and five female mice for plored the novel object significantly longer than the familiar one,
each group). WT, age-­matched control C57BL/6J mice also received while unliganded MNP-­treated PS19 mice did not show a difference
identical stereotaxic surgery but were injected with vehicle (PBS) as in exploring the two objects and vehicle-­treated PS19 mice spent
controls for surgery, injection, and anesthesia. The body weights of more time exploring the familiar object (Fig. 5F). MNPs-­DP treat-
mice were recorded weekly, and MNPs-­DP significantly prevented ment appeared to significantly improve the discriminatory index (DI;
weight loss of PS19 mice associated with late-­stage tau pathology that Fig. 5G) and discriminatory ratio (DR) (fig. S19) of the PS19 mice
afflicted vehicle-­or naked MNP-­treated PS19 mice after 10 weeks of compared to vehicle-­treated groups, while MNPs did not, indicating
treatment (Fig. 5A). At the end of the 10-­week treatment, we assessed the specific activity of 7-­DP.
the spatial cognition and memory of these mice using the Barnes
maze test (28). During the learning period, the mice searched for the MNPs-­DP do not cause obvious toxicity in PS19 mice
escape hole, and the latency time was recorded daily for 6 days. WT After the behavior tests, the mice were euthanized, and their blood,
mice and MNPs-­DP–treated PS19 mice learned a new task quicker brains, and main organs were collected. It has been reported that MNPs
with a shorter latency time to find the escape hole compared to the are mainly cleared through the liver and spleen (23). Here, the aspartate
vehicle-­or naked MNP-­treated PS19 mice (Fig. 5B). After training, aminotransferase (AST) blood test was used to assess the potential tox-
we examined their long-­term memory by conducting a probe trial icity of MNPs-­DP on liver function. The AST levels of MNPs-­DP–treated
with the escape hole removed. Compared to the WT mice, the PBS-­ mice were comparable to the vehicle groups, demonstrating that the
and MNP-­treated PS19 model mice exhibited obvious memory defi- long-­term treatment of MNPs-­DP under current dose did not lead to
cits, including longer latency times (Fig. 5C), fewer target hole entries obvious liver damage (Fig. 5H). Histological analysis of the liver, spleen,
(Fig. 5D), and more random motion-­like wandering paths (Fig. 5E). heart, lung, and kidney (fig. S20) revealed no evidence of toxic changes

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 6 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Downloaded from https://www.science.org on May 06, 2024

Fig. 5. MNPs-­DP rescues memory deficits in PS19 mice without causing obvious toxicity. (A) Body weight of WT and PS19 mice treated with PBS, MNPs and MNPs-­DP, respec-
tively (n = 13 per group, eight male and five female mice for each group). (B) Total latency time to find the escape hole in the learning period for each group of mice (n = 13 per group,
eight male and five female mice for each group). Statistical significance was analyzed by two-­way ANOVA (P < 0.05; *MNPs-­DP versus PBS; #MNPs-­DP versus MNPs). (C) Primary
latency for finding escape hole in the probe trial for each group of mice. (D) Entry number to target hole in probe test for each group of mice. (E) Representative track of probe test
for each group of mice. (F) Time spent during the exploration of the familiar and novel objects in novel object recognition test (n = 13 per group, eight male and five female mice for
each group). (G) Discriminatory index obtained in novel object recognition test. (H) Serum aspartate aminotransferase (AST) levels for each group of mice (n = 13 per group, eight
male and five female mice for each group). Results shown as mean + SD (n = 13). Statistical significance was analyzed by two-­way ANOVA (*P < 0.05; **P < 0.01).

or cell death in cellular structures between the MNPs-­, MNPs-­DP–, AT8-­positive staining in the CA1 or CA3 regions, whereas a signifi-
and vehicle-­treated mice (30). In addition, the consistent increase in cant amount of AT8 positive ptau neurons with labeling resembling
body weights of MNPs-­DP–treated PS19 mice (blue line in Fig. 5A) tau neurofibrillary tangles were found in the hippocampus of vehicle-­
also supported the biosafety of MNPs-­DP with current dosing. Together, treated PS19 mice (n = 6 per group; Fig. 6A and fig. S21). Impres-
10 weeks of treatment of MNPs-­DP significantly rescue the spatial sively, MNPs-­DP treatment significantly limited tau pathology in both
learning and memory impairments in a mouse model of AD without the CA1 and CA3 regions, relative to the weak effect of unliganded
apparent toxic side effects in other body systems. MNPs (Fig. 6, A to C, and fig. S21). In addition, mouse brain tissues
were analyzed for radioimmunoprecipitation assay (RIPA)–soluble
MNPs-­DP clear tau pathology in PS19 mice and RIPA-­insoluble tau by Western blot, with ptau levels normalized
To assess tau pathology of the mice, we immunostained brain sections by β-­actin levels (Fig. 6, D and E, and fig. S22). MNPs-­DP significant-
with AT8, an antibody specific for phosphorylated tau (ptau) pSer202/ ly decreased both the insoluble and soluble ptau levels in the brain of
Thr205 that is widely used to identify neurofibrillary tangles in PS19 mice (n = 8 per group; Fig. 6, F and G), consistent with the im-
postmortem human and PS19 mouse brains. WT mice did not show munohistochemistry results.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 7 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

DISCUSSION in vitro antitau amyloid activities have been identified; however, none
The development of tau-­targeted therapeutics intended to slow the on- have shown substantial success in placebo-­controlled clinical trials.
set and progression of AD and other tauopathies remains an unfulfilled Their lack of in vivo efficacy could be due to factors such as limited bio-
goal. Although numerous approaches have shown promise in animal availability, poor metabolic stability, lack of specificity, and limited BBB
models and preclinical studies, antitau therapeutics have yet to progress permeability. New approaches are needed, and here and in other recent
successfully beyond the early phases of clinical trials. Compounds with work (11, 12), we suggest exploration of fragmentors of amyloid fibrils.

Downloaded from https://www.science.org on May 06, 2024

Fig. 6. MNPs-­DP reduce phosphorylated tau in the hippocampus of PS19 mice. (A) Representative immunohistochemical images of the hippocampus (CA1 and CA3
regions) of WT and PS19 mice that treated with PBS, MNPs, and MNPs-­DP, respectively, at ×10 magnification, stained with AT8 antibody for phosphorylated tau at residues
Ser202/Thr205. The neurofibrillary tangles were indicated by red arrows. (B) Quantification of AT8-­positive staining in the hippocampus CA1. (C) Quantification of AT8-­positive
staining in the hippocampus CA3. Representative Western blot for AT8 of (D) soluble and (E) insoluble hippocampus homogenates from WT and PS19 mice treated with PBS,
MNPs, and MNPs-­DP, respectively. β-­Actin was used as a loading control. Quantification of (F) soluble phosphorylated tau (ptau) and (G) insoluble ptau normalized to β-­actin
(n = 8 per group, four male and four female PS19 mice for each group). Statistical significance was analyzed by two-­way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001).

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 8 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Here, we describe a nonnatural amino acid fragmentor of tau fibrils, for effective uptake remains unknown and requires further testing. If
D-­TLKIVWC (7-­DP), by modifying our previous structure-­based in- it is found to be necessary, then coadministration of MNPs-­DP and
hibitor of tau fibrils, 6-­DP (10). 7-­DP not only inhibits the formation of mannitol could be considered in a phase 1 trial through intravenous
tau aggregation but also fragments recombinant tau fibrils and tau fibrils injection, similar to the US Food and Drug Administration–approved
isolated from the autopsied brains of patients with AD. Apparently, 7-­ iron oxide nanoparticle drug Feraheme injection (22) (AMAG Phar-
DP can inhibit the prion-­like spread of tau in biosensor cells and reduce maceuticals) that contains mannitol. We believe that brain-­deliverable
tau-­induced cellular toxicity in N2a cells by fragmentation. A potential peptides that fragment tau fibrils are promising precursors to possible
concern about fragmenting amyloid fibrils is that the products of frag- future therapeutics of AD, but successful development will necessitate
mentation could act as toxic oligomers, killing rather than protecting future studies of toxicity, pharmacokinetics, and distribution.
cells or acting as seeds that nucleate additional fibril formation (31). This In summary, we provide preclinical data demonstrating the in vi-
is not the case for 7-­DP as shown by the experiments described in Fig. 1. tro tau-­fragmenting activity and the in vivo therapeutic potential of
Given the substantial stability of pathogenic amyloid, particularly 7-­DP and one intravenously bioavailable formulation on MNPs,
of tau fibrils from AD postmortem brain (32), it is unexpected that a MNPs-­DP. Our work supports future phase 1 evaluation of 7-­DP for-
seven-­residue peptide can fragment these fibrils readily at 37°C with mulations including MNPs-­DP in humans with AD and other tauop-
no energy source other than of ambient thermal agitation and of mild athies. If safe and acutely effective against tauopathy biomarkers,
shaking. While we have not yet fully determined the molecular mech- then this new class of highly specific structurally designed tau frag-
anism of fragmentation, we have established some properties of 7-­DP mentors could be optimized and further developed for clinical use.
essential to its fragmenting ability. It must be at least seven residues in
length; its six-­residue analog D-­TLKIVW does not fragment. Also,
the sulfhydryl group of cysteine is not essential; it can be linked to MATERIALS AND METHODS
MNP through a S─C bond without impairing fragmenting activity. Recombinant tau protein expression and purification

Downloaded from https://www.science.org on May 06, 2024

The ability of 7-­DP to form amyloid-­like fibrils itself may be key, and Recombinant tau K18+ (residues Gln244 to Glu380 of 4R tau) was ex-
we hypothesized a fragmentation mechanism as illustrated in Fig. 7. pressed in a pNG2 vector in BL21-­Gold Escherichia coli cells grown
We conjugated 7-­DP to MNPs to facilitate its transport across the in LB to an A600 = 0.8. Cells were induced with 0.5 mM isopropyl
BBB. In addition, we intranasally injected 20 μl of 2% mannitol into 1-­thio-­β-­​d-­galactopyranoside for 3 hours at 37°C and lysed by soni-
mice to increase BBB permeability, but whether mannitol is necessary cation in 20 mM MES buffer (pH 6.8) with 1 mM EDTA, 1 mM

Fig. 7. Hypothesized schematic mechanism of fragmentation of tau fibrils. (A) Scheme illustrates the interaction of tau fibrils with 7-­DP. (B) A metastable interface
forms between 7-­DP and tau fibrils because of self-­assembly of 7-­DP along tau fibrils. (C) A gap opens between tau molecules as the 7-­DP assembly changes polymorph
to attain a lower energy state and tau strains to maintain its interface with the new polymorph of 7-­DP assembly. (D) Water invades the newly formed gap to fragment
tau fragments.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 9 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

MgCl2, 1 mM dithiothreitol, and HALT protease inhibitor before the One (Thermo Fisher Scientific). The conjugation efficacy (%) of
addition of NaCl (500 mM final concentration). Lysate was boiled 7-­DP on MNPs can be calculated following (Mi − MS)/MS, where
for 20 min and then clarified by centrifugation at 15,000 rpm for Mi is the initial amount of 7-­DP used, and MS is the amount of 7-­DP
15 min and dialyzed to 20 mM MES buffer (pH 6.8) with 50 mM in the supernatant. The morphology of the nanoparticles was ob-
NaCl and 5 mM dithiothreitol. Dialyzed lysate was purified on a served by FEI Tecnai T12 Quick room temperature TEM at 120 kV. The
5-­ml HighTrap SP ion exchange column and eluted over a gradient zeta potential and hydrodynamic diameters of the nanoparticles
of NaCl from 50 to 550 mM. Protein was further purified on a were measured using a ZetaSizer Nano ZS90 (Malvern Instruments).
HiLoad 16/600 Superdex 75 pg column (GE Healthcare) in 10 mM
tris (pH 7.6) with 100 mM NaCl and 1 mM dithiothreitol and ThT inhibition assay
concentrated to 20 to 60 mg/ml by ultrafiltration using a 3-­kDa cut- Frozen aliquot of the purified recombinant tau K18+ was thawed
off (Millipore-­Sigma, Burlington, MA). on ice and diluted into PBS (pH 7.4) to a final concentration of
20 μM. In vitro aggregation was initiated by incubating 150 μl, 20 μM
Preparation of crude and purified AD brain tau fibrils tau K18+ monomer in the presence of 40 μM ThT, 15 μM heparin
Tissue sections of 0.2 to 0.3 g were excised on a block of dry ice (H3393, Sigma-­Aldrich), and 1 mM dithiothreitol in a black Nunc 96-­
and manually homogenized in a 15-­ml disposable tube in 1.5 ml of well optical bottom plate (Thermo Fisher Scientific). To test the effect
sucrose buffer with 1 mM EGTA and 5 mM EDTA. Samples were of D-­TLKIVWC, D-­TLKIVW, MNPs-­DP, MNPs, and MNPs-­DPsc on
then aliquoted to polymerase chain reaction tubes and sonicated inhibiting tau K18+ aggregation, tau K18+ monomer was incubated in
in a cuphorn bath for 120 min under 30% power at 4°C in a recir- the presence of different concentrations of D-­TLKIVW/D-­TLKIVWC
culating ice water bath to obtain the crude AD brain tau fibrils. (5, 10, 20, 50, and 100 μM), MNPs-­DP/MNPs/MNPs-­DPsc (0.005,
For purification of tau fibrils from AD brain tissue, the crude AD 0.01, 0.02, and 0.04 mg/ml) in parallel. Kinetic fluorescence data were
brain tau samples were further heated at 90°C for 20 min in a ther- collected in a microplate reader (FLUOstar Omega, BMG Labtech)

Downloaded from https://www.science.org on May 06, 2024

mal cycler and centrifuged at 20,100g for 30 min. The supernatant at 37°C with double orbital shaking at 700 rpm. Fluorescence mea-
was collected and then spun at 90 K for 30 min on a Airfuge surements were recorded every 10 min with excitation and emission
(Beckman-­Coulter). Last, the pellet was the purified AD brain-­ wavelengths of 440 and 480 nm. All samples were added in triplicate,
derived tau fibrils and resuspended in 20 μl of 20 mM tris buffer and experiments were repeated at least three runs.
with 100 mM NaCl and stored at 4°C.
ThT fragmentation assay
Synthesis of D-­TLKIVWC liganded iron oxide Tau K18+ monomer (20 μM) was grown in the presence of 40 μM
nanoparticles MNPs-­DP ThT, 15 μM heparin, and 1 mM dithiothreitol overnight to form
Peptides were purchased from GenScript (Piscataway, NJ) and puri- the fibrils. The tested D-­TLKIVWC, D-­TLKIVW, MNPs-­DP, or
fied to ≥98%, as determined by mass spectrometry and high-­ MNPs were added separately in parallel to designated wells, and
performance liquid chromatography. The synthesis of D-­TLKIVWC the ThT assay was further continued. Kinetic fluorescence data
liganded iron oxide nanoparticles involves two steps. First, the car- were collected in a microplate reader (FLUOstar Omega, BMG
boxylic acid iron oxide MNPs (CD Bioparticles, WHM-­G062) were Labtech) at 37°C with double orbital shaking at 700 rpm. Fluores-
reacted with 2-­maleimidoethylamine hydrochloride to gain the cence measurements were recorded every 10 min with excitation
maleimide-­functionalized iron oxide nanoparticles, which were and emission wavelengths of 440 and 480 nm. All samples were
further conjugated with D-­TLKIVWC to obtain the MNPs-­D-­ added in triplicate, and experiments were repeated at least twice.
TLKIVWC through the thiol-­maleimide Michael addition click re- After 24 hours of incubation, the samples were collected for TEM
action. Briefly, 1 mg of N-­hydroxysulfosuccinimide (Thermo Fisher and dot blot studies.
Scientific) and 1.5 mg of N-­ethyl-­​N′-­(3-­(dimethylamino)propyl)
carbodiimide (Thermo Fisher Scientific) were dissolved in 50 μl Transmission electron microscopy
activation buffer [0.1 M MES and 0.5 M NaCl (pH 6.0)] separately Six microliters of sample was applied to a glow discharged carbon
before they were added to the 200 μl MNPs (5 mg/ml) with 100 μl coated electron microscopy grid (CF150-­Cu, Electron Microscopy
of activation buffer and kept shaken at room temperature for Sciences) for 5 min. Then, grids were stained with 2% uranyl acetate
15 min. Next, 2 mg of 2-­maleimidoethylamine hydrochloride that for 2 min. Samples were visualized using a FEI Tecnai T12 Quick
dissolved in 500 μl of 20× PBS was added to the above solution and room temperature transmission electron microscope using an ac-
kept shaken at room temperature for another 2 hours. After that, the celeration voltage of 120 kV equipped with a Gatan 2048 × 2048
solution was kept in a magnetic separator in 4°C. After three times charge-­coupled device camera.
of magnetic separation and washing, the supernatant was removed,
and the nanoparticles were resuspended in 1 ml of 1× tris buffer Circular dichroism
(pH 7.04) by sonication for several hours. In the second step, 4 mg The measurements were conducted on Chirascan Q100 CD spec-
of D-­TLKIVWC was dissolved in 1× tris buffer (pH 7.04) and then trometer (Applied Photophysics Limited, UK). Samples were mea-
added to the above MNP solution. The mixed solution was kept sured in a quartz cell (0.5 mm), and the wavelength was set in the
rotated at room temperature for 1 hour before left in a magnetic range between 190 and 250 nm. Accumulation was conducted three
separator. All samples were washed with deionized water for three times continuously for each sample.
times to remove the surface-­adsorbed 7-­DP. The supernatant in all
washing process was combined with the supernatant in the mag- Native electrospray–ionization mass spectrometry
netic separator process. Then, the absorbance intensity of the com- Samples were buffer-­exchanged into 20 mM ammonium acetate
bined supernatant at 280 nm was measured using the NanoDrop with 0.5 ml of 10 kDa Amicon filters (Millipore-­Sigma, Burlington,

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 10 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

MA). The samples were then measured using a Synapt G2-­Si Fragmentation of tau aggregates in tau biosensor cell line
(Waters Corporation Milford, MA). All protein solutions were loaded HEK293 cell lines stably expressing EYFP-­fused tau 1N4R P301S
into in-­house pulled capillaries coated with gold and electrosprayed were engineered by M. Diamond’s laboratory at the University of
by applying a capillary voltage between 1 and 2 kV on the nanoelec- Texas Southwestern Medical Center and used without further char-
trospray ionization source. The native electrospray–ionization mass acterization or authentication. Crude AD brain tau fibrils were soni-
spectrometry experiment was repeated twice to demonstrate repro- cated for 1 min before they were transfected to cells. After 24 hours,
ducibility. 7-­DP or PBS was transfected into cells using Lipofectamine 2000
(Life Technologies, catalog no. 11668019). After another 24-­hour in-
Fragmentation of purified AD brain-­derived tau fibrils cubation, the number of seeded aggregates was determined by imag-
Purified AD brain tau fibrils were incubated with D-­TLKIVWC (25, ing the entire well of a 96-­well plate in triplicate using a Celigo image
50, 100, 250, and 500 μM) and D-­TLKIVW (500 μM) at 37°C for cytometer (Nexcelom) in the YFP channel. Aggregates were counted
48 hours, respectively. The samples were collected for TEM charac- using ImageJ by subtracting the background fluorescence from un-
terization, dot blot, and Western blot studies. The TEM protocol was seeded cells and then counting the number of peaks with fluores-
the same as above. The number of AD tau fibrils in each image was cence above background using the built-­in particle analyzer. The
counted (n = 10 images). For dot blot study, 2.5 μl of samples number of aggregates was normalized to the confluence of each well,
was added on nitrocellulose membrane (0.2 μm, Bio-­Rad). The and dose-­response plots were generated by calculating the average
membrane was blocked by 5% (w/v) nonfat dry milk in TBS-­T and SD values from triplicate measurements. For high-­quality im-
[T = 0.1% (v/v) Tween-­20] at room temperature for 1 hour. After ages, cells were photographed on a ZEISS Axio Observer D1 fluores-
blocking, the membrane was incubated with GT38 antibody obtained cence microscope using the YFP fluorescence channel.
from Virginia Lee’s lab (1:1000) in 5% (w/v) milk in TBS-­T at 4°C
overnight. Then, the membrane was washed in TBS-­T for three times Crude AD brain tau fibril seeding in tau biosensor cell line

Downloaded from https://www.science.org on May 06, 2024

and incubated with goat anti-­mouse immunoglobulin G horseradish HEK293 cell lines stably expressing tau-­K18-­YFP were engineered
peroxidase (HRP; 1:5000; catalog no. AB205719, Abcam) in TBS-­T by M. Diamond’s laboratory at the University of Texas Southwestern
for 1 hour at room temperature. The membrane was washed three Medical Center and used without further characterization or au-
more times, and the signal was developed with Pierce ECL Western thentication. Cells were maintained in Dulbecco’s modified Eagle’s
blotting substrate (170-­5061, Bio-­Rad). The samples were centrifuged medium supplemented with 10% (v/v) fetal bovine serum, 1%
at 21,000g for 30 min at 4°C (Eppendorf Centrifuge 5424R). Western antibiotic-­antimycotic, and 1% GlutaMAX at 37°C, 5% CO2 in a
blot was performed with antitau rabbit polyclonal antibody (1:1000; humidified incubator. Crude AD brain tau fibrils were incubated
Dako, A0024) and anti-­rabbit HRP-­conjugated secondary antibody with D-­TLKIVWC (5, 10, 20, 50, 75, and 100 μM), D-­TLKIVW
(1:5000; Thermo Fisher Scientific). (5, 10, 20, 50, 75, and 100 μM), MNPs-­DP (0.005, 0.01, 0.02, and
0.04 mg/ml), or MNPs (0.005, 0.01, 0.02, and 0.04 mg/ml) overnight
Fragmentation of tau aggregates in HEK293 cells and sonicated in a cup horn water bath for 3 min. Then, these in-
HEK293 cell lines were purchased from the American Type Culture hibitor/fragmentor-­treated tau seeds were mixed with 1 volume of
Collection (CRL-­1573). Cells were maintained in Dulbecco’s modi- Lipofectamine 3000 prepared by diluting 1 μl of Lipofectamine in
fied Eagle’s medium (Life Technologies, catalog no. 11965092) supple- 19 μl of OptiMEM. After 20 min, 10 μl of fibrils was added to 90 μl
mented with 10% (v/v) fetal bovine serum (Life Technologies, catalog of tau biosensor cells. In some cases, the D-­TLKIVWC and AD tau
no. A3160401), 1% penicillin/streptomycin (Life Technologies, cata- seeds were directly transfected to the cells without incubation over-
log no. 15140122), and 1% GlutaMAX (Life Technologies, catalog no. night. After 24 hours of incubation, the number of seeded aggregates
35050061) at 37°C, 5% CO2 in a humidified incubator. pcDNA3.1 for was determined by imaging the entire well of a 96-­well plate in trip-
1N4R tau with a P301S mutation was a gift from M. Goedert at the licate using a Celigo image cytometer (Nexcelom) in the YFP chan-
Medical Research Council Laboratory of Molecular Biology in Cam- nel. The data analysis was described before. For high-­quality images,
bridge, UK. Crude AD brain tau fibrils were sonicated for 1 min be- cells were photographed on a ZEISS Axio Observer D1 fluorescence
fore they were transfected to cells together with tau pcDNA3.1. After microscope using the YFP fluorescence channel.
24 hours. 7-­DP at various concentrations with or without 10 μM
MG132 (Thermo Fisher Scientific) were transfected into cells using Cell viability of tau K18+ aggregates
Lipofectamine 2000 (Life Technologies, catalog no. 11668019). After Tau K18+ aggregates were produced by incubation of tau K18+
another 24-­hour incubation, cells were collected using 100 μl of lysing (20 μM) in the absence or presence of D-­TLKIVWC (5, 10, 20, 50,
buffer [10 mM tris (pH 7.4), 200 mM NaCl, 1 mM EDTA, 0.5% de- and 100 μM), D-­TLKIVW (5, 10, 20, 50, and 100 μM), MNPs-­DP
oxycholate, and 0.5% Triton X-­100] and sonicated in a cuphorn water (0.005, 0.01, 0.02, and 0.04 mg/ml), or MNPs (0.005, 0.01, 0.02,
bath for 10 min. The lysed cells were ultracentrifuged at 90,000 rpm in and 0.04 mg/ml) for 48 hours in PBS, respectively. PBS without
a Beckman Coulter Airfuge Air-­driven Ultracentrifuge. The superna- any protein was used as controls. N2a cells were cultured in min-
tant was collected and labeled “supernatant,” and the pellet was resus- imum essential medium supplemented with 10% (v/v) fetal bo-
pended in 100 μl of PBS and labeled “pellet.” An equalized amount of vine serum, 1% antibiotic-­antimycotic, and 1% GlutaMAX in a
each sample was run on a NuPAGE 4 to 12%, bis-­tris, 1.0 mm, Mini 5% CO2 humidified environment at 37°C. Cells were plated at a
Protein Gel (15-­well). Western blot was performed with antitau rabbit density of 10,000 cells per well on 96-­well plates in 90 μl of fresh
polyclonal antibody (1:1000; Dako, A0024) and anti-­rabbit HRP-­ medium. After 24 hours, 10 μl of the above tau K18+ aggregates
conjugated secondary antibody (1:5000; Thermo Fisher Scientific). was added, and the cells were incubated for another 24 hours at
β-­Actin was used as a loading control (1:2000; sc-­47778, Santa Cruz 37°C. Cytotoxicity was measured using the MTT assay. Ten micro-
Biotechnology). liters of the stock MTT reagent (5 mg/ml) was added to each well.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 11 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

After 3 hours of conversion into the formazan product, dimethyl 10 weeks. The weight of mice was recorded weekly. Treatment with
sulfoxide (DMSO; Thermo Fisher Scientific) was added to dis- unliganded MNPs or MNPs-­DP was always in combination with
solve the purple crystals left in the dark for 10 min before the intranasal administration of mannitol. WT, age-­matched C57BL/6
measurement of absorbance at 570 nm with 700 nm as a sub- mice were stereotaxically injected with vehicle to serve as controls.
tracted reference wavelength using a microplate reader (Spectra-
Max M5, Molecular Devices, USA). Cells treated with 100% Behavioral tests
DMSO were designated as 0% viable, and those treated only with All behavioral experiments were conducted in the Behavioral Test-
vehicle were designated as 100% viable. To further investigate the ing Core (BTC) at UCLA. Mice were transferred to the BTC and
cytotoxicity of compounds alone, various concentrations of D-­ allowed to acclimate for 1 week. All the animals were handled for at
TLKIVWC (final concentration in wells is 2, 5, 10, 20, 50, 100, or least three to four consecutive days before testing. They were tested
200 μM), D-­TLKIVW (final concentration in wells is 2, 5, 10, 20, in random order, and the experimenter conducting the tests was un-
50, 100, or 200 μM), MNPs (final concentration in wells is 0.0025, aware of the genotype/group.
0.005, 0.01, 0.02, or 0.04 mg/ml), or MNPs-­DP (final concentra-
tion in wells is 0.0025, 0.005, 0.01, 0.02, or 0.04 mg/ml) were dis- Barnes maze
pensed into the N2a cells and further incubated for 24 hours at The maze consisted of a circular platform with 20 holes around the
37°C and in 5% CO2 incubator. Cytotoxicity was measured using periphery with an escape box attached to the bottom of one of the
a MTT assay. holes and shallow boxes attached to the bottom of the other holes.
Bright light and white noise were used to motivate mice to find and
Cell toxicity of tau K18+ fibril fragmentation species enter the escape box. Visual extra-­maze cues were present on three
Tau K18+ fibril fragmentation species were produced by incubating walls of the room. For all trials, mice were placed individually in a
tau K18+ fibrils with D-­TLKIVWC (5, 10, 20, 50, or 100 μM), D-­ cylindrical start chamber in the center of the maze for 30 s, which

Downloaded from https://www.science.org on May 06, 2024

TLKIVW (5, 10, 20, 50, or 100 μM), MNPs-­DP (0.005, 0.01, 0.02, was then lifted to start the test. During an adaptation period, mice
or 0.04 mg/ml), or MNPs (0.005, 0.01, 0.02, or 0.04 mg/ml) for were guided to the escape tunnel and allowed to stay there for 30 s.
24 hours, respectively. N2a cells were cultured in MEM supplemented During a spatial acquisition period, a total of 10 acquisition trials
with 10% (v/v) fetal bovine serum, 1% antibiotic-­antimycotic, and (two trials per day with an intertrial interval of 15 min) were per-
1% GlutaMAX in a 5% CO2 humidified environment at 37°C. Cells formed; mice were allowed to explore the maze freely for 2 min.
were plated at a density of 10,000 cells per well on 96-­well plates in Each trial ended when the mouse entered the escape tunnel or after
90 μl of fresh medium. After 24 hours, 10 μl of the above tau K18+ 2 min had elapsed. Mice that did not find the tunnel were guided to
fibril fragmentation species was added, and the cells were incubated it. All mice were allowed to remain in the tunnel for 30 s. During the
for another 24 hours at 37°C. Cytotoxicity was measured using probe trial conducted 1 day after the last training trial, the escape
MTT assay as above. tunnel was replaced by a shallow box, and mice were allowed to ex-
plore the maze for 90 s. Animals’ performances were monitored us-
Animals ing Any-­Maze Video Tracking System (Stoelting Co., Wood Dale,
P301S transgenic mice (Prnp-­MAPT*P301S PS19Vle/J, JAX stock IL), which provided data for the acquisition parameter (latency to
#008169, the Jackson Laboratory, Bar Harbor, ME) and C57BL/6J find the platform) and the probe trial parameters (latency to find
mice (the Jackson Laboratory, JAX:000664) were housed in groups the platform, number of entries in the target platform zone of the
of up to four in individually ventilated cages under standard condi- platform).
tions (22°C, 12-­hour light-­dark cycle) receiving food and water
ad libitum. All animal experiments were performed in accordance Novel object recognition
with the National Institutes of Health regulations and approved by Mice were habituated to the open testing arena for 10 min, 24 hours
UCLA Animal Research Committee and performed under over- before the test was performed. During the memory acquisition trial,
sight of the Division of Laboratory Animal Medicine. each mouse was allowed to explore two identical objects for 10 min.
The mice were returned to their home cage for 90 min. For the mem-
Inductively coupled plasma mass spectrometry ory retention phase, animals were exposed for 10 min to the pres-
The iron contents of brain tissues were assessed by ICP-­MS (Nex- ence of one similar object and one novel object (a different shape and
ION 2000, PerkinElmer). The half brain tissues were dried at color). Object exploration time was recorded when the mouse
105°C to constant weight; 1 ml of concentrated nitric acid and touched the object directly with its nose, mouth, or forepaws. Results
0.2 ml of concentrated hydrochloric acid were added to the weight- are expressed as exploration time for each object. The cognitive
ed sample (100 mg) in a tube and heated at 80°C for 2 hours. After function was evaluated by the DR and DI, which is represented by
digestion, the solution was diluted to 10 ml with deionized water TC/(TR + TC) and (TR − TC)/(TR + TC), respectively, where TC is
for ICP-­MS tests. the time spent exploring the novel object, and TR is the time spent
exploring the familiar object.
Treatments
We stereotaxically seeded 24 2-­month-­old, male PS19 mice and 15 Determination of blood AST levels as an index of
6-­month-­old, female PS19 mice with 1.5 μl of recombinant K18+ tau hepatic toxicity
fibril seeds (5 μg/μl) and divided those 39 mice into three groups Blood samples were obtained by cardiac puncture and centrifuged at
(n = 13 per group, eight male and five female mice for each group). 2000g for 20 min to separate the serum for collection. Activities AST
Each group was intravenously administered with vehicle, unliganded was assayed using Mouse AST enzyme-­linked immunosorbent assay
MNPs, or MNPs-­DP (10 mg/kg of body weight) every week for kits (Abcam, ab263882) according to the manufacturer’s protocols.

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 12 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Hematoxylin and eosin staining significant at the *P < 0.05, **P < 0.01, and ***P < 0.001 levels;
Major organs including heart, liver, spleen, lung, and kidney (n = 3) NS, not significant.
were fixed in 4% buffered formalin saline (Sigma-­Aldrich) at 4°C
overnight and then embedded in paraffin blocks. Tissue sections of
5 μm thickness were stained with hematoxylin and eosin (H&E). Supplementary Materials
The morphology of the tissue was observed under a light micro- This PDF file includes:
Figs. S1 to S22
scope (Leica) at ×20 magnification.

Immunohistochemistry REFERENCES AND NOTES

The brain tissues were embedded in paraffin and sectioned at 10-­μm 1. H. Braak, E. Braak, Neuropathological stageing of Alzheimer-­related changes. Acta
Neuropathol. 82, 239–259 (1991).
thickness. Following deparaffinization, rehydration, heat-­induced
2. P. V. Arriagada, J. H. Growdon, E. T. Hedley-­Whyte, B. T. Hyman, Neurofibrillary tangles but
antigen retrieval in citrate buffer (pH 6.0), and blocking of endoge- not senile plaques parallel duration and severity of Alzheimer's disease. Neurology 42,
nous peroxidase activity by 3% hydrogen peroxide, the sections 631–639 (1992).
were incubated with the antitau (1:500; AT8, monoclonal, mouse, 3. S. Asher, R. Priefer, Alzheimer's disease failed clinical trials. Life Sci. 306, 120861
Innogenetics) at 4°C overnight. On the second day, followed by be- (2022).
4. L. VandeVrede, A. L. Boxer, M. Polydoro, Targeting tau: Clinical trials and novel therapeutic
ing washed three times in TBST, 5 min each, the sections were incu- approaches. Neurosci. Lett. 731, 134919 (2020).
bated in biotinylated anti-­mouse secondary antibody (1:1000) in 5. A. Mullard, Anti-­tau antibody failures stack up. Nat. Rev. Drug Discov. 20, 888
1.5% normal serum with 3% bovine serum albumin/tris-­buffered (2021).
saline for 1 hour at 37°C. Immunolabeling was detected using the 6. J. M. Ringman, S. A. Frautschy, E. Teng, A. N. Begum, J. Bardens, M. Beigi, K. H. Gylys,
V. Badmaev, D. D. Heath, L. G. Apostolova, V. Porter, Z. Vanek, G. A. Marshall, G. Hellemann,
avidin-­biotinylated HRP complex method (Elite Kit, Vector, USA)
C. Sugar, D. L. Masterman, T. J. Montine, J. L. Cummings, G. M. Cole, Oral curcumin for
and was visualized with diaminobenzidine (Wako, Japan). Last, the

Downloaded from https://www.science.org on May 06, 2024

Alzheimer's disease: Tolerability and efficacy in a 24-­week randomized, double blind,
sections were mounted onto coverslips using DPX mounting media placebo-­controlled study. Alzheimers Res. Ther. 4, 43 (2012).
(Thermo Fisher Scientific). The images were collected at ×10 magni- 7. C. M. Wischik, C. R. Harrington, J. M. Storey, Tau-­aggregation inhibitor therapy for
fication with a digital MC170 5 MPixel Leica camera on a Nikon Alzheimer's disease. Biochem. Pharmacol. 88, 529–539 (2014).
8. M. Faiyaz, M. A. Ganayee, S. Akhtar, S. Krishnan, B. Flora, D. Dogra, N. K. Jha,
Eclipse E800M microscope. Image processing and analysis were D. K. Chellappan, P. Negi, K. Dua, Nanomaterials in Alzheimer’s disease treatment: A
performed with the ImageJ software. Macros were developed for comprehensive review. Front. Biosci. 26, 851–865 (2021).
each staining to attain a reproducible semi-­automated quantifica- 9. S. Luo, C. Ma, M. Q. Zhu, W. N. Ju, Y. Yang, X. Wang, Application of iron oxide nanoparticles
tion. The image files were converted to eight-­bit grayscale. A uni- in the diagnosis and treatment of neurodegenerative diseases with emphasis on
Alzheimer's disease. Front. Cell. Neurosci. 14, 21 (2020).
form threshold value of 50 (using the dark background option) was
10. S. A. Sievers, J. Karanicolas, H. W. Chang, A. Zhao, L. Jiang, O. Zirafi, J. T. Stevens, J. Münch,
then applied before performing the “analyze particles” task to deter- D. Baker, D. Eisenberg, Structure-­based design of non-­natural amino-­acid inhibitors of
mine the percent of area covered by positive signal. amyloid fibril formation. Nature 475, 96–100 (2011).
11. P. M. Seidler, K. A. Murray, D. R. Boyer, P. Ge, M. R. Sawaya, C. J. Hu, X. Cheng, R. Abskharon,
Western blot analyses H. Pan, M. A. DeTure, C. K. Williams, D. W. Dickson, H. V. Vinters, D. S. Eisenberg,
Structure-­based discovery of small molecules that disaggregate Alzheimer’s disease
PBS-­perfused unfixed brains were used for biochemical analysis by tissue derived tau fibrils in vitro. Nat. Commun. 13, 5451 (2022).
dissecting the hippocampi separately. Before analysis, the brain sam- 12. K. A. Murray, C. J. Hu, H. Pan, J. Lu, R. Abskharon, J. T. Bowler, G. M. Rosenberg,
ples were sonicated in RIPA buffer (4 vol/g) [50 mM tris, 150 nM C. K. Williams, G. Elezi, M. Balbirnie, K. F. Faull, H. V. Vinters, P. M. Seidler, D. S. Eisenberg,
NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-­40, 5 mM Small molecules disaggregate alpha-­synuclein and prevent seeding from patient
brain-­derived fibrils. Proc. Natl. Acad. Sci. U.S.A. 120, e2217835120 (2023).
EDTA, 1 mM phenylmethylsulfonyl fluoride, 0.1% protease inhibi-
13. P. Chong, C. Sia, B. Tripet, O. James, M. Klein, Comparative immunological properties of
tor mixture, and 0.5% phosphatase inhibitor (pH 8.0)] and centri- enantiomeric peptides. Lett. Pept. Sci. 3, 99–106 (1996).
fuged at 100,000g for 30 min at 4°C. The supernatants were saved 14. P. M. Seidler, D. R. Boyer, K. A. Murray, T. X. P. Yang, M. Bentzel, M. R. Sawaya, G. Rosenberg,
as RIPA-­soluble fractions, whereas the RIPA-­insoluble pellets were D. Cascio, C. K. Williams, K. L. Newell, B. Ghetti, M. A. DeTure, D. W. Dickson, H. V. Vinters,
washed with 1 M sucrose in RIPA buffer to remove myelin and as- D. S. Eisenberg, Structure-­based inhibitors halt prion-­like seeding by Alzheimer's
disease–and tauopathy–derived brain tissue samples. J. Biol. Chem. 294, 16451–16464
sociated lipids and centrifuged at 100,000g for 30 min at 4°C. The (2019).
RIPA-­insoluble pellets were then extracted in tissue (1 vol/g) with 15. O. A. Morozova, Z. M. March, A. S. Robinson, D. W. Colby, Conformational features of tau
2% SDS buffer [50 mM tris-­HCl (pH 7.6)]. The protein contents of fibrils from Alzheimer's disease brain are faithfully propagated by unmodified
the samples were measured using the BCA Protein Assay Kit (Pierce, recombinant protein. Biochemistry 52, 6960–6967 (2013).
Bonn, Germany) and diluted to the same concentration (500 μg/ml). 16. W. Zhang, B. Falcon, A. G. Murzin, J. Fan, R. A. Crowther, M. Goedert, S. H. W. Scheres,
Heparin-­induced tau filaments are polymorphic and differ from those in Alzheimer’s and
Soluble and insoluble fractions were analyzed by SDS-­PAGE, fol- Pick’s diseases. eLife 8, e43584 (2019).
lowed by Western blotting using AT8 antibody (1:200; MN1020, 17. A. W. P. Fitzpatrick, B. Falcon, S. He, A. G. Murzin, G. Murshudov, H. J. Garringer,
Thermo Fisher Scientific). β-­Actin was used as a loading control R. A. Crowther, B. Ghetti, M. Goedert, S. H. W. Scheres, Cryo-­EM structures of tau filaments
(1:2000; sc-­47778, Santa Cruz Biotechnology). Bands were quanti- from Alzheimer’s disease. Nature 547, 185–190 (2017).
18. T. Hamano, T. F. Gendron, L. W. Ko, S. H. Yen, Concentration-­dependent effects of
fied using ImageJ software. proteasomal inhibition on tau processing in a cellular model of tauopathy. Int. J. Clin. Exp.
Pathol. 2, 561–573 (2009).
Statistical analysis 19. S. B. Prusiner, A unifying role for prions in neurodegenerative diseases. Science 336,
Graphs are expressed as means + SD, and data were analyzed using 1511–1513 (2012).
20. M. Goedert, Alzheimer’s and Parkinson’s diseases: The prion concept in relation to
SPSS 19.0 statistical analysis software (SPSS, Chicago, IL, USA).
assembled Aβ, tau, and α-­synuclein. Science 349, 1255555 (2015).
Two-­way analysis of variance (ANOVA) followed by a least signifi- 21. B. B. Holmes, J. L. Furman, T. E. Mahan, T. R. Yamasaki, H. Mirbaha, W. C. Eades,
cant difference posttest were used to analyze difference among L. Belaygorod, N. J. Cairns, D. M. Holtzman, M. I. Diamond, Proteopathic tau seeding
multiple groups. Statistical differences for all tests were considered predicts tauopathy in vivo. Proc. Natl. Acad. Sci. 111, E4376–E4385 (2014).

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 13 of 14

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

22. M. Lu, M. H. Cohen, D. Rieves, R. Pazdur, FDA report: Ferumoxytol for intravenous iron 32. M. R. Sawaya, M. P. Hughes, J. A. Rodriguez, R. Riek, D. S. Eisenberg, The expanding amyloid
therapy in adult patients with chronic kidney disease. Am. J. Hematol. 85, 315–319 (2010). family: Structure, stability, function, and pathogenesis. Cell 184, 4857–4873 (2021).
23. H. Arami, A. Khandhar, D. Liggitt, K. M. Krishnan, In vivo delivery, pharmacokinetics,
biodistribution and toxicity of iron oxide nanoparticles. Chem. Soc. Rev. 44, 8576–8607 Acknowledgments: We thank the organ donors and their families; without whom this work
(2015). would not have been possible. The authors thank M. Diamond for sharing the YFP-­labeled tau
24. S. Belbekhouche, M. Guerrouache, B. Carbonnier, Thiol-­maleimide michael addition click K18 expressing HEK293 cells. The behavior tests were conducted in the BTC facility under the
reaction: A new route to surface modification of porous polymeric monolith. Macromol. supervision of L. M. Lueptow and I. Zhuravka. We acknowledge UCLA Tissue Pathology Core
Chem. Phys. 217, 997–1006 (2016). Laboratory (TPCL) for H&E staining. We thank C. H. Chang at the ICP-­MS facility of UCLA for
25. A. M. Jackson, J. W. Myerson, F. Stellacci, Spontaneous assembly of subnanometre-­ analysis of the Fe concentration in brain tissues. We acknowledge Y. Jiang for discussion, M.
ordered domains in the ligand shell of monolayer-­protected nanoparticles. Nat. Mater. 3, Goedert for the pcDNA3.1 for 1N4R tau with a P301S mutation, and V. Lee for GT38 antibody.
330–336 (2004). Funding: This work was supported by National Institutes of Health 1R01AG070895 (D.S.E.),
26. M. Schweighauser, A. G. Murzin, J. Macdonald, I. Lavenir, R. A. Crowther, H. S. Scheres, National Institutes of Health RF1AG065407 (D.S.E.), National Institutes of Health grant
M. Goedert, Cryo-­EM structures of tau filaments from the brains of mice transgenic for R01AG066212 (S.F.), National Institutes of Health grant R35 GM145286 (J.A.L.), Alzheimer’s
human mutant P301S Tau. Acta Neuropathol. Commun. 11, 160 (2023). Association Research Fellowship AARF-­21-­848751 (K.H.), VA Merit BX004332 (G.C.), and VA
27. M. Iba, J. L. Guo, J. D. McBride, B. Zhang, J. Q. Trojanowski, V. M. Lee, Synthetic tau fibrils Merit 1 I01 BX005919-­01 (S.F.). Author contributions: Conceptualization: K.H. and D.S.E.
mediate transmission of neurofibrillary tangles in a transgenic mouse model of Protein purification: R.A. and P.S. Nanoparticle synthesis: K.H. and M.B. In vitro and cell
Alzheimer's-­like tauopathy. J. Neurosci. 33, 1024–1037 (2013). experiments: K.H. and J.L.D. Mouse experiments: K.H., H.P., H.S.-­K., C.H., M.M., G.C., and S.F.
28. K. Gawel, E. Gibula, M. Marszalek-­Grabska, J. Filarowska, J. H. Kotlinska, Assessment of Immunohistochemistry staining and Western blot: K.H., M.J., X.Z., S.F., and G.C. Mass
spatial learning and memory in the Barnes maze task in rodents-­methodological spectrometry: C.L. and J.A.L. Supervision: G.C., S.F., and D.S.E. Writing—original draft: K.H. and
consideration. Naunyn Schmiedebergs Arch. Pharmacol. 392, 1–18 (2019). D.S.E. Writing—review and editing: K.H., M.R.S., and D.S.E. Competing interests: D.S.E. is SAB
29. M. Leger, A. Quiedeville, V. Bouet, B. Haelewyn, M. Boulouard, P. Schumann-­Bard, T. Freret, chair and equity holder of ADRx Inc. All other authors declare they have no competing
Object recognition test in mice. Nat. Protoc. 8, 2531–2537 (2013). interests. We acknowledge our provisional patent application (Serial No. 63/510,194). Data
30. T. K. Jain, M. K. Reddy, M. A. Morales, D. L. Leslie-­Pelecky, V. Labhasetwar, Biodistribution, and materials availability: All data needed to evaluate the conclusions in the paper are
clearance, and biocompatibility of iron oxide magnetic nanoparticles in rats. Mol. Pharm. present in the paper and/or the Supplementary Materials.
5, 316–327 (2008).

Downloaded from https://www.science.org on May 06, 2024

31. E. Nachman, A. S. Wentink, K. Madiona, L. Bousset, T. Katsinelos, K. Allinson, H. Kampinga, Submitted 12 October 2023
W. A. McEwan, T. R. Jahn, R. Melki, Disassembly of Tau fibrils by the human Hsp70 Accepted 29 March 2024
disaggregation machinery generates small seeding-­competent species. J. Biol. Chem. Published 1 May 2024
295, 9676–9690 (2020). 10.1126/sciadv.adl2991

Hou et al., Sci. Adv. 10, eadl2991 (2024) 1 May 2024 14 of 14