X-linked ubiquitin-specific peptidase 11 increases tauopathy vulnerability in women

Summary

Although women experience significantly higher tau burden and increased risk for Alzheimer’s disease (AD) than men, the underlying mechanism for this vulnerability has not been explained. Here, we demonstrate through in vitro and in vivo models, as well as human AD brain tissue, that X-linked ubiquitin specific peptidase 11 (USP11) augments pathological tau aggregation via tau deubiquitination initiated at lysine-281. Removal of ubiquitin provides access for enzymatic tau acetylation at lysines 281 and 274. USP11 escapes complete X-inactivation, and female mice and people both exhibit higher USP11 levels than males. Genetic elimination of usp11 in a tauopathy mouse model preferentially protects females from acetylated tau accumulation, tau pathology, and cognitive impairment. USP11 levels also strongly associate positively with tau pathology in females but not males. Thus, inhibiting USP11-mediated tau deubiquitination may provide an effective therapeutic opportunity to protect women from increased vulnerability to AD and other tauopathies.

Graphical Abstract

In Brief

The risk of developing Alzheimer’s disease for women is significantly higher than for men. In this study, Yan et al uncovers a molecular mechanism underpinning the heightened susceptibility.

Introduction

Although it is well-known that women are afflicted by Alzheimer’s disease (AD) ~1.7 times more frequently than men (Collaborators, 2022; Rajan et al., 2021), the mechanistic basis for this increased vulnerability has not been established. We reasoned that this could be related to the fact that women endure greater tau burden than men. For example, positron emission tomography (PET) studies show that clinically normal women exhibit significantly higher tau deposition in the brain than men (Buckley et al., 2019; Buckley et al., 2020; Luchsinger et al., 2020; Palta et al., 2021), suggesting that sexual dimorphism in tau burden could be an early foundational event for AD. Furthermore, a recent study of postmortem brain tissue from upwards of 1500 age- and education-matched AD patients revealed significantly higher tau deposition in women compared to men (Oveisgharan et al., 2018).

Tau aggregation and clearance are controlled by various post-translational modifications, including phosphorylation (Augustinack et al., 2002; Bancher et al., 1991; Ulrich et al., 2018), methylation (Balmik and Chinnathambi, 2021; Huseby et al., 2019; Thomas et al., 2012), acetylation (Cohen et al., 2011; Esteves et al., 2018; Min et al., 2015; Min et al., 2010), and ubiquitination (Bancher et al., 1991; Flach et al., 2014; Iqbal and Grundke-Iqbal, 1991; Petrucelli et al., 2004; Shimura et al., 2004). Notably, ubiquitination is the final modification that controls tau clearance via the ubiquitin-proteasome system and autophagy-lysosome pathway (Ciechanover and Kwon, 2015; Ihara et al., 2012; Lee et al., 2013). While tau ubiquitination is associated with increased tau clearance via the E3 ligases C-terminus of Hsc70-interacting protein (CHIP) (Dickey et al., 2008; Petrucelli et al., 2004; Shimura et al., 2004), tumor necrosis factor receptor-associated factor 6 (TRAF6) (Babu et al., 2005), and membrane-associated RING-CH7 (MARCH7) (Flach et al., 2014), comparatively little is known about the role of removal of ubiquitin from tau by deubiquitinases (DUBs) in the brain.

The human genome encodes ~100 DUBs (Komander et al., 2009), ~50% of which are ubiquitin-specific peptidases (USPs) (Clague et al., 2013). Of these, more than 20 are expressed in the central nervous system (CNS) (Clague et al., 2013; Ristic et al., 2014). To investigate their role in regulating tau, we executed a functional siRNA screen against 22 CNS DUBs and identified two positive regulators of tau: USP13 and USP11. Because USP11 is located on chromosome X and implicated in female biology (Li et al., 2016; Orthwein et al., 2015; Schoenfeld et al., 2004; Tukiainen et al., 2017), we hypothesized that it might contribute to female-specific vulnerability to tauopathy. Here, we present both in vitro and in vivo evidence, including in human AD brain tissue, that endogenously increased levels of USP11 in females relative to males results in increased deubiquitination of tau at lysine (K) 281, which drives pathological tau deposition by inhibiting its elimination and promoting its acetylation. This contributes to cognitive impairment, with greater effect in females than males.

Results

USP11 is a tau-stabilizing DUB

To investigate whether DUBs expressed in the CNS impact tau regulation, we conducted a functional siRNA screen in HeLa cells stably expressing wild-type (WT) tau (HeLa-V5-tau, 0N4R). Specifically, we individually silenced 22 DUBs known to be expressed in the CNS (Figure S1A) and found that knockdown of either USP11 or USP13 significantly reduced tau levels (Figures S1A and S1B). Knockdown of USP14 also marginally reduced tau (Figure S1A and S1B), consistent with USP14 aptamers enhancing tau degradation through the proteasome (Lee et al., 2015). However, we excluded USP14 from further consideration, as phospho-tau is increased in usp14^−/− mice (Jin et al., 2012), and USP14 inhibition also impairs autophagy (Kim et al., 2018). Knockdown of both USP11 and USP13 yielded no further reduction in tau levels compared to knockdown of either alone (Figure S1C and S1D). Of these two proteins, USP11 is expressed more abundantly in the CNS and has been reported to regulate DNA damage response and estrogen receptor activity in cancer (Dwane et al., 2020; Orthwein et al., 2015; Schoenfeld et al., 2004; Ting et al., 2019; Wiltshire et al., 2010). While autosomal chromosome-linked USP13 is known to regulate tau levels in the brain (Liu et al., 2019b), X chromosome-linked USP11 has not previously been implicated.

We first focused on validating USP11 as a positive tau regulator. To begin, we replicated our investigation of HeLa-V5-tau cells and observed that USP11 knockdown reduces tau by ~45% (Figures 1A and 1B). We then utilized iHEK cells expressing the frontotemporal dementia with parkinsonism-17 (FTDP-17) MAPT mutation tau^P301L (Shelton et al., 2017) and PS19 primary neurons expressing the FTDP-17 MAPT mutation tau^P301S (Yoshiyama et al., 2007). Knockdown of USP11 by siRNA in iHEK-tau^P301L cells reduced tau^P301L by ~45% compared to control siRNA (Figures 1C and 1D). Likewise, lentivirus-mediated knockdown of USP11 in tau^P301S neurons reduced tau by ~70% (Figures 1E–1H). By contrast, USP11 knockdown had no significant effect on TDP-43 levels in HeLa cells expressing endogenous and exogenous TDP-43 (Ling et al., 2010) (Figures S1E–S1G). Cycloheximide (CHX) chase experiments showed that USP11 siRNA significantly accelerates tau turnover compared to control siRNA (Figures 1I and 1J). However, tau mRNA levels remained constant after USP11 knockdown (Figure 1K), indicating that USP11 stabilizes tau post-translationally.

Figure 1. Deubiquitinase USP11 stabilizes tau.

(A,B) Representative blots of proteins from HELA-V5-tau cells transfected with control or USP11 siRNA. Graph shows quantification of tau (2-tailed t-test, t=6.138, df=4, **P=0.0036, n=3 samples/condition). (C,D) Representative blots of proteins from iHEK-tau^P301L cells transfected with control or USP11 siRNA. Graph shows quantification of tau (2-tailed t-test, t=4.271, df=4, *P=0.0129, n=3 samples/condition). (E,F) Representative blots of proteins from DIV21 tau^P301S cortical neurons transduced with control or USP11 shRNA. Graph shows quantification of tau (2-tailed t-test, t=12.24, df=6, ****P<0.0001, n=4 samples/condition). (G,H) Representative images of tau staining from DIV21 tau^P301S hippocampal neurons transduced with control or USP11 shRNA. Graph shows quantification of tau (2-tailed t-test, t=12.85, df=46, ****P<0.0001, n=24 images/condition from 3 independent experiments). (I,J) Representative blots of proteins from HELA-V5-tau cells transfected with control or USP11 siRNA (CHX, 100μg/mL). Graph shows quantification of tau (2-way ANOVA condition x time: F(3,12)=9.413, P=0.0018; posthoc Sidak, ****P<0.0001, **P<0.01, *P<0.05, n=3 independent experiments). (K) Quantification of tau mRNA by qRT-PCR (2-tailed t-test, not significant, n=6 samples/condition).

USP11 catalytic activity enhances tau aggregation

To determine if USP11 overexpression modifies tau levels, we transfected iHEK-tau^P301L and Hela-V5-tau cells with vector control, WT USP11, or the catalytically inactive USP11^C318S mutant (CS) (Maertens et al., 2010). Compared to control, USP11 overexpression significantly increased insoluble tau without altering soluble tau in iHEK-tau^P301L (Figures 2A and 2B) and Hela-V5-tau cells (Figures S2A and S2B). By contrast, USP11^C318S overexpression did not alter insoluble or soluble tau compared to controls in iHEK-tau^P301L (Figures 2A and 2B) or Hela-V5-tau cells (Figures S2A and S2B). To determine whether USP11-elevated insoluble tau adopted a pathologically aggregated form, we employed filter trap assays to capture aggregated proteins (Nasir et al., 2015). As expected, very little tau was captured from soluble lysates (Figures 2C and 2D), whereas aggregated tau was remarkably increased in the insoluble fraction containing USP11 but not that of USP11^C318S (Figures 2C and 2D). This indicates that elevated USP11 promotes tau insolubility and pathological aggregation by virtue of its DUB catalytic activity.

Figure 2. USP11 catalytic activity mediates tau aggregation and deubiquitination.

(A,B) Representative blots of RIPA-soluble and insoluble proteins from iHEK-tau^P301L cells transfected with control vector, USP11, or USP11^C318S mutant (CS). Graphs show quantification of soluble and insoluble tau (1-way ANOVA; soluble tau=not significant; insoluble tau: 1-way ANOVA, F(2,16)=22.52, P<0.0001; posthoc Tukey: **P=0.0030, ****P<0.0001, n=6-7 samples/condition from 2 independent experiments). (C,D) Filter trap blots of aggregated tau from RIPA-soluble and insoluble fractions of HELA-V5-tau cells transfected with control vector, USP11, or USP11^C318S. Graph shows quantification of insoluble tau aggregates (1-way ANOVA, F(2,9)=3561, P<0.0001; posthoc Tukey: ****P<0.0001, n=4 samples/condition). (E) Representative blots of USP11 and tau in tau complexes (IP: tau) or control immune complex (IgG) in iHEK-tau^P301L cells transfected with control vector, USP11, or USP11^C318S. (F) Representative images of endogenous USP11 complex with tau in HELA-V5-tau cells (red dots) by in situ Duolink® PLA. (G) Representative blots of polyubiquitin conjugates and tau in tau-pulldown samples (IP: tau) from iHEK-tau^P301L cells transfected with HA-ubiquitin plus control vector, USP11, or USP11^C318S. Western blots of input lysates are shown below. (H) Representative blots of ubiquitin-K63 or ubiquitin-K48 conjugates and tau in tau-pulldown samples (IP: tau) from iHEK-tau^P301L cells treated with vehicle or 10μM MG132 for 6h, after which tau-pulldowns were treated ± recombinant USP11 (1.5μg) for 2h. *Asterisk indicates recombinant USP11 protein added to the tau-pulldown. (I) Representative blots showing deubiquitination of CHIP-ubiquitinated recombinant tau by recombinant USP11. (J) Quantification of diGly-ub signatures remaining on tau lysines after deubiquitination of CHIP-ubiquitinated tau with recombinant USP11 (2h and 24h) by tryptic digest and LC-MS/MS (n=3 samples/condition).

USP11 deubiquitinates tau in cells and in vitro

To further characterize the mechanism by which USP11 modifies tau, we next assessed whether USP11 specifically cleaves ubiquitin conjugates from tau. In transfected iHEK-tau^P301L cells, both USP11 and USP11^C318S were readily detected in tau immune complexes, whereas negative control IgG beads did not pull down either protein (Figure 2E). Despite comparable levels in lysates, tau immune complexes contained substantially more USP11^C318S than USP11^WT (Figure 2E), reminiscent of the stabilized enzyme-substrate complexes observed with catalytically dead enzymes (Kurita et al., 2008). We also detected native USP11-tau complexes by proximity ligation assay (PLA) in Hela-V5-tau cells (Figure 2F; Figure S2C, negative controls). In transfected iHEK-tau^P301L cells, USP11 but not USP11^C318S markedly reduced polyubiquitin-conjugated tau in tau immune complexes (Figure 2G), indicative of tau deubiquitination by USP11. In a complementary assessment of tau deubiquitination, MG132 treatment of iHEK-tau^P301L cells resulted in marked accumulation of high molecular weight K48 (ub-K48) and K63 (ub-K63) ubiquitin linkages on tau (Figure 2H), while incubation of tau immune complexes with recombinant USP11 readily cleaved ub-K48 and ub-K63 conjugates from tau to levels similar to no MG132 treatment (Figure 2H).

To identify the specific lysines ubiquitinated on tau that were targeted by USP11, we ubiquitinated recombinant tau (0N4R) with the E3 ligase CHIP reaction mix for 6h, after which samples were incubated with or without recombinant USP11 to remove ubiquitin conjugates. Upon confirming the removal of ubiquitin conjugates by USP11 (Figure 2I), we subjected samples to tryptic digest and LC-MS/MS analysis for detection of peptide diGly signatures remaining from tryptic cleavage of conjugated ubiquitin (diGly-ub) (Figure S2D). LC-MS/MS analysis identified 19 tau peptides containing the tryptic diGly signatures (Figure S2E), representing 21 tau lysines ubiquitinated by the CHIP reaction mix (Figure S2E). The addition of USP11 for 24h removed nearly all ubiquitin conjugates from tau except for K254 (Figure 2J). Shorter 2h incubation with USP11 revealed ubiquitinated lysines more sensitive to cleavage by USP11, including K24 and multiple lysines within or near microtubule-binding regions (i.e. K257/K259, K274, K280/K281, K281, K290, K298, K311) (Figure 2J). Mass spectrometry analysis also detected diGly ubiquitin signatures on tryptic peptides from Hsp70 and DNAJB1, which were part of the ubiquitination reaction. Specifically, we detected 22 and 7 diGly tryptic peptides corresponding to Hsp70 and DNAJB1, respectively. However, summation of diGly peptide intensity per protein showed that USP11 removed ubiquitin conjugates from tau but not from Hsp70 or DNAJB1 (Figure S2F), indicating that USP11 is discriminating and directed toward its specific substrate.

USP11 deubiquitinates and insolubilizes WT tau, but not K281Q tau acetylation mimic

Most tau lysines that can be ubiquitinated are also subject to acetylation (Alquezar et al., 2020; Kontaxi et al., 2017; Wesseling et al., 2020), two events that are mutually exclusive on any given lysine. Hence, we tested tau acetylation mimics, which are known to impair microtubule binding, promote tau aggregation, and/or induce pathological conversion of tau (K274Q, K280Q, and K281Q) (Cohen et al., 2011; Sohn et al., 2016; Trzeciakiewicz et al., 2017). K280Q and K281Q mutants exhibited increased insolubility of ~2.2 and ~3-fold, respectively, compared to WT tau, whereas the K274Q mutant showed normal solubility (Figures 3A and 3B). As expected, USP11 overexpression significantly decreased solubility of WT tau; however, it did not significantly affect solubility of the tau acetylation mimics (Figures 3A and 3C). Within tau immune complexes, the tau-K281Q mutant exhibited significantly reduced polyubiquitin, the K280Q mutant showed nonsignificantly reduced polyubiquitin, and the K274Q mutant showed no change (Figures 3D and 3E). USP11 overexpression significantly reduced ubiquitination of WT tau, which was also observed in K274Q and K280Q mutants, while USP11 had no effect on ubiquitination of the K281Q mutant (Figures 3D and 3F). These results collectively indicate that USP11 recapitulates in WT tau the insolubility of tau acetylation mimics K280 and K281, and has no additive effect on tau acetylation mimics. Moreover, reduced ubiquitination of tau-K281Q at steady-state and the inability of USP11 to deubiquitinate this mutant indicates that deubiquitination of tau at K281 is a critical step in USP11-driven tauopathy.

Figure 3. USP11 recapitulates the tau acetylation mimic K281Q and enables tau acetylation at K281/K274 by CBP/p300 or SIRT1 inhibition.

(A) Representative blots of RIPA-soluble and insoluble proteins from HEK293T cells co-transfected with tau variants and vector control or USP11. (B) Quantification of the insoluble/soluble tau (1-way ANOVA, F(3,13)=11.54, P=0.0006, posthoc Dunnett, *P=0.0432, ***P=0.0009, ns=not significant, n=4-5 independent experiments). (C) Quantification of paired fold changes insoluble/soluble tau (paired t-test; tau-WT: **P=0.0013; tau^K274Q, tau^K280Q, ortau^K281Q variants: not significant; n=4-5 independent replicates per tau variant). (D) Representative blots of polyubiquitinated tau and tau in tau pulldowns (IP: tau) from HEK293T cells co-transfected with tau variants and vector control or USP11. (E) Quantification of polyubiquitinated tau/total tau (1-way ANOVA, F(3,20)=5.412, P=0.0068; posthoc Tukey: *P<0.05, n=6 independent experiments). (F) Quantification of paired fold changes per experiment in polyubiquitinated tau/total tau (paired t-test; tau-WT: ***P=0.0003; tau^K274Q: *P=0.016; tau^K280Q: *P=0.0282; tau^K281Q: not significant; n=6 independent experiments). (G) Representative blots of RIPA-soluble proteins from iHEK-Tau^WT cells co-transfected with vector control or p300-HA and control or USP11 siRNA. (H,I) Quantification of Ac281-tau and Ac274-tau (1-way ANOVA, Ac281-tau: F(3,16)=8.893, P=0.001; Ac274-tau: F(3,16)=26.32, P<0.0001; posthoc Tukey: *P<0.05, ***P<0.001, ****P<0.0001; n=5 samples/condition). (J) Representative blots of RIPA-soluble proteins from iHEK-Tau^WT cells transfected with control siRNA or USP11 siRNA for 48h ± SIRT1 inhibitor EX527 treatment (50μM, 6h) prior to lysis. (K,L) Quantification of Ac281-tau and Ac274-tau (1-way ANOVA, Ac281-tau: F(3,12)=26.35, P<0.0001; Ac274-tau: F(3,12)=21.79, P<0.0001; posthoc Tukey: *P<0.05, ***P<0.001, ****P<0.0001; n=4 samples/condition).

USP11 is a molecular switch that provides access for enzymatic regulation of K281/K274 tau acetylation

CREB-binding protein (CBP/p300) acetylates tau on multiple lysines, including K281 and K274 (Min et al., 2010; Shin et al., 2021). We next determined if endogenous USP11 impacts CBP/p300-mediated tau acetylation in iHEK-tau^WT cells using antibodies recognizing tau acetylation at K274 or K281 (Sohn et al., 2016). As expected, CBP/p300 overexpression significantly increased Ac274-tau and Ac281-tau. However, USP11 siRNA significantly prevented tau acetylation by CBP/p300 (Figures 3G–3I; Figure S3A). EX527-mediated inhibition of SIRT1, an enzyme known to deacetylate tau (Min et al., 2010), also strongly increased Ac274-tau and Ac281-tau (Figures 3J–3L). USP11 siRNA fully prevented the increase in Ac281-tau and partially suppressed the increase in Ac274-tau induced by SIRT1 inhibition (Figures 3J–3L). Inhibition of the deacetylase HDAC6 by tubastatin A (TBSTA) increased tau KXGS motif phosphorylation at S262 (Figures S3B–S3E), as previously reported (Trzeciakiewicz et al., 2020), and did not significantly alter tau acetylation at K274 or K281 (Figure S3B–S3E). Interestingly, USP11 knockdown significantly reduced TBSTA-induced tau phosphorylation at S262 (Figure S3B–S3E). These results indicate that USP11-mediated tau deubiquitination acts as a molecular switch that provides access for enzymatic regulation of K281/K274 tau acetylation, and also impacts HDAC6-regulated pS262-tau.

USP11 accumulates in human tauopathies

To determine whether USP11 is deregulated in human disease, we next examined frontal gyrus brain tissue from AD, frontotemporal lobar degeneration with tau pathology (FTLD-tau), and nondementia control subjects (Case information: Figure S4A). Double labeling for USP11 and p-S202/T205-tau (AT8 antibody) showed the expected predominant nuclear pattern of USP11 staining (Hendriks et al., 2015; Ideguchi et al., 2002; Rong et al., 2021; Ting et al., 2019; Zhang et al., 2021) in nondementia controls (Figure 4A), with little to no AT8 signal (Figure 4A). Surprisingly, AD and FTLD-tau brains exhibited much stronger USP11 staining overall, which extended into the cytoplasm and often colocalized with AT8-positive tangles and neuropil threads (Figure 4A; Figure S4B, negative controls). Quantification of USP11 demonstrated highly significant >9.5-fold and >6.5-fold increases in USP11 accumulation in AD (Figure 4B) and FTLD-tau (Figure S4C) brains, respectively, compared to nondementia controls. The ratio of cytoplasmic to nuclear USP11 also significantly increased in AD (Figure 4C) and FTLD-tau (Figure S4D) brains. Manders overlap coefficient for USP11 area overlapping with AT8 area reached ~0.18 in AD (Figure 4D) and FTLD-tau (Figure S4E), compared to near 0 in controls. These in vivo observations thus place USP11 at the scene of pathogenesis and indicate that excessive USP11 is associated with pathological tau aggregation.

Figure 4. Accumulation of USP11 in human tauopathies, strong association of USP11 with tauopathy in women, and higher USP11 expression in disease-free females.

(A) Representative images of USP11 and p-S202/T205-tau (AT8 antibody) in the frontal gyrus (FG) of 3 nondementia controls, 3 AD, and 3 FTLD-tau cases. White boxes are magnified to the right. (B) Quantification of USP11 intensity in nondementia controls and AD (2-tailed t-test, ****p<0.0001, t=5.403, df=130; n=49 images from 4-5 images/case, 2-3 sections/case, and 10 control cases; n=83 images from 4-5 images/case, 2-3 sections/case, and 17 AD cases. (C) Quantification of cytoplasmic/nuclear USP11 ratio in nondementia controls and in AD (2-tailed t-test, ****P<0.0001, t=8.489, df=132; n=50 images from 4-5 images/case, 2-3 section/case, and 10 control cases; n=84 images from 4-5 images/case, 2-3 sections/case, and 17 AD cases. (D) Image J quantification of Manders coefficient of USP11 area overlapping with AT8 area in nondementia controls and AD (2-tailed t-test, t=9.118, df= 141, ****P<0.0001; n=59 images from 4-5 image/case, 2-3 sections/case, and 10 control cases; n=84 images from 4-5 images/case, 2-3 sections/case, and 17 AD cases. (E-H) Correlation between USP11 and p-tau (AT8) intensities in (E) female AD: linear regression test, r²=0.5759, P<0.0001; multiple linear regression AT8 x USP11 main effect adjusted for Thal amyloid score, age at death, and APOE genotypes, adjusted r²=0.5733, F(4,28)=11.75, P<0.0001; n=33 images from 4-5 images/case, 2-3 sections/case, and 7 cases; (F) male AD: linear regression test, r²=0.01183, P=0.0145; multiple linear regression AT8 x USP11 main effect adjusted for Thal amyloid score, age at death, and APOE genotypes, adjusted r²=0.2412, F(4,45)=4.895, P=0.0023; n=50 images from 5 images/case, 2-3 sections/case, and 10 cases; (G) female FTLD-tau: linear regression test, r²=0.8776, P=0.0002, n=9 images from 4-5 images/case, and 2 cases; and (H) male FTLD-tau: linear regression test, r²=0.1944, P=0.0352, n=23 images from 4-5 images/case, 2-3 sections/case, and 5 cases. (I) Representative images of USP11 (red) and DAPI (blue) staining in the frontal gyrus (FG) of 3 female and 3 male nondementia controls. (J) Quantification of USP11 intensity in the FG of female and male nondementia subjects (2-tailed t-test, *P=0.0245, t=2.325, df=47; female: n=30 images from 4-5 images/case, 2-3 sections/case, and 6 control cases; male: n=19 images/case, 2-3 sections/case, and 4 control cases). (K) Representative images of USP11 (red) and DAPI (blue) in the cortex in 7-month-old mice. (L) Quantification of USP11 intensity in the frontal cortex 7-month-old mice (1-way ANOVA, F(4,189)=28.98, P<0.0001, posthoc Tukey: ****P<0.0001, **P<0.01, ns=not significant; n=39-42 images from 5-8 sections/mouse, and 4 mice/genotype/sex, F=female, M=male).

USP11 is more strongly associated with tau pathology in women than in men

USP11 resides on the X chromosome and is reported to escape full X-inactivation (Li et al., 2016; Tukiainen et al., 2017). Notably, estrogen elevates USP11, and USP11 in turn regulates aspects of female-biased biology, such as estrogen-induced estrogen receptor activity (Dwane et al., 2020) and BRCA1/2-mediated DNA damage response (Orthwein et al., 2015; Schoenfeld et al., 2004). Hence, we looked for evidence of USP11 sexual dimorphism in the human brain. USP11 intensity showed a highly significant positive correlation with tau pathology in female AD brains (Figure 4E: r²=0.5759, p<0.0001) compared to male AD brains (Figure 4F: r²=0.1183, p=0.0145), and multiple linear regression analysis for AT8 and USP11 adjusting for Thal amyloid score, age at death, and APOE genotype did not alter this finding (Figures 4E and 4F). Similar female-favored association was also observed in FTLD-tau, in which USP11 accumulation positively correlated with tau pathology with an r²=0.8776 (p=0.0002) in females (Figure 4G) and r²=0.1944 (p=0.0352) in males (Figure 4H), suggesting that USP11 accumulation is a more significant driver of tau pathology in females than males.

USP11 expression is elevated in disease-free female humans and mice

Next, we looked for evidence of sexual dimorphism in USP11 expression in nondementia subjects. Indeed, USP11 intensity was ~3-fold higher in women compared to men among nondementia subjects (Figures 4I and 4J). Hence, a female-dependent association of USP11 with tau pathology appears to be a foundational event occurring before AD onset in women.

As usp11 likewise resides on the X chromosome in rodents, we also examined mouse brains for sexual dimorphism in usp11. In the absence of sex-specific USP11 regulation and normal X-inactivation in female mice, wild-type female mice (usp11^+/+) would be expected to demonstrate the same level of USP11 expression as wild-type male mice (usp11⁺). However, 7-month-old wild-type female mice (usp11^+/+) showed significantly higher USP11 in the frontal cortex, compared to age-matched usp11⁺ males or usp11^+/− females (Figures 4K and 4L), indicating conservation of USP11 sexual dimorphism between mice and humans. Female usp11^+/− mice did not show significant differences in USP11 intensity compared to wild type usp11⁺ male mice (Figures 4K and 4L), while neither female nor male usp11 knockout mice showed detectable levels of USP11 (Figures 4K and 4L). These results in mice corroborate the findings in humans and suggest a prominent role of USP11 in driving female tau pathology.

Mitigation of tau pathology and associated gliosis by loss of usp11 in tau^P301S mice

Usp11 knockout mice exhibit a mild hematopoietic phenotype (increased T cells, LDL cholesterol, and hemoglobin) (IMPC, mousephenotype.org) (Skarnes et al., 2011). However, both male and female usp11 knockout mice are adult viable, grossly normal, and fertile. Moreover, IHC examination of male and female usp11 knockout mice and WT littermates showed no differences in synaptic integrity (synaptophysin) or astrogliosis (GFAP) at seven months of age (Figures S5A and S5B). Likewise, usp11 knockout and WT littermates showed no differences in spatial memory as assessed by Morris water maze (Figures S5C–S5F).

We next sought to determine whether USP11 contributes to tau pathogenesis in vivo by crossing usp11^−/− mice with PS19 mice expressing the FTDP17-linked tau^P301S mutation. Since usp11 resides on the X chromosome, we assessed male and female mice separately. Among 7-month-old female mice, loss of usp11 in tau^P301S;usp11^−/− mice modestly but significantly reduced soluble tau compared to tau^P301S mice (Figures 5A and 5B; Figures S5G and S5H), while the reduction in soluble pS199/202-tau was not significant (Figures 5A and 5C). In the insoluble fraction, tau^P301S;usp11^−/− mice exhibited a significant ~60% reduction in total tau (Figures 5A and 5B; Figures S5G and S5I) and ~50% reduction in pS199/202-tau (Figures 5A and 5C) compared to those in tau^P301S mice. Furthermore, IHC staining of female mouse brains revealed ~70% reduction in pS199/202-tau (Figures 5D and 5E) and ~50% reduction in pS262-tau (Figures 5F and 5G) in both cortex and hippocampus of tau^P301S;usp11^−/− mice, compared to tau^P301S mice.

Figure 5. Loss of usp11 mitigates Ac-tau accumulation, tau pathology, and gliosis in female tau^P301S mice.

(A) Representative blots of RIPA-soluble and insoluble proteins in the cortex of 7-month-old female mice. (B,C) Quantification of soluble and insoluble tau and pS199/202-tau in the cortex of female mice. (B) Total tau (soluble): 2-tailed t-test, t=3.177, df=22, **P=0.0044; total tau (insoluble): 2-tailed t-test, t=9.186, df=22, ****P<0.0001, n=12 mice/genotype. (C) pS199/202-tau (soluble): 2-tailed t-test, not significant; pS199/202-tau (insoluble): 2-tailed t-test, t=3.647, df=14, **P=0.0026, n=8 mice/genotype. (D) Representative images of pS199/202-tau (green) and DAPI (blue) in the hippocampus CA1 and cortex of 7-month-old female mice. (E) Quantification of pS199/202-tau intensity in 7-month-old female mice (2-tailed t-test: t=15.02, df=59 (Hippo); t=16.57, df=70 (Cortex), ****P<0.0001, n=25-40 images from 5-8 sections/mouse and 4 mice/genotype). (F) Representative images of pS262-tau (red) and DAPI (blue) in the hippocampus and cortex of 7-month-old female mice. (G) Quantification of pS262-tau intensity in 7-month-old female mice (2-tailed t-test: t=4.564, df=98 (Hippo); t=4.908, df=98 (cortex), ****P<0.0001, n=50 images from 10 sections/mouse and 5 mice/genotype). (H) Representative images of GFAP (green), Iba1 (red), and DAPI (blue) in the hippocampus of 7-month-old female mice. (I) Quantification of GFAP or Iba1 intensity in female mice (GFAP: 1-way ANOVA, F(2,110)=26.62, P<0.0001, posthoc Tukey: ****P<0.0001, n=33-45 images from 5-8 sections/mouse and 5 mice/genotype; Iba1: 1-way ANOVA, F(2,95)=29.16, P<0.0001, posthoc Tukey: ***P<0.001, ****P<0.0001, n=29-36 images from 5-8 sections/mouse and 6 mice/genotype). (J) Blots of RIPA-soluble and insoluble USP11, ac-tau (Ac274 or Ac281), and actin from the cortex of 7-month-old mice. F=female, M=male. *Asterix indicates the same male mice lysates in Figure S6L used for normalization between gels. (K,L) Quantification of soluble and insoluble Ac-tau in female mice from Figure 5J (2-tailed t-test, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n=4 mice/genotype). (M-O) Quantification of soluble and insoluble Ac-tau and USP11 in female (F) and male (M) mice from Figures 5J and S6L (2-tailed t-test, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, n=4 mice/sex).

Among 7-month-old male mice, loss of usp11 in tau^P301S;usp11⁻ mice did not significantly reduce soluble tau or soluble pS199/202-tau, although modest but significant reductions in insoluble tau (−25%) and insoluble pS199/202-tau (−40%) were seen (Figures S5J–S5L). Hence, the magnitude and significance of these effects in male mice were markedly less pronounced than in females. Likewise, IHC staining for pS199/202-tau or pS262-tau showed significantly reduced p-tau in the hippocampus (~25%) and cortex (~20-40%) in tau^P301S;usp11⁻ mice compared to WT tau^P301S mice (Figures S5M–S5O), with lesser magnitude than in female mice.

We next examined astrogliosis and microgliosis using the established markers GFAP and Iba1, respectively. As expected, 7-month-old female tau^P301S mice displayed highly significant increases in GFAP and Iba1 in the hippocampus, both of which were nearly entirely prevented by loss of usp11 in tau^P301S;usp11^−/− mice (Figures 5H and 5I). Likewise, the neuroinflammation marker IL-1β was highly elevated in tau^P301S mice compared to WT mice, and significantly reduced in tau^P301S;usp11^−/− mice (Figures S6A and S6C). Similar results were obtained in 7-month-old male mice. GFAP, Iba1, and IL-1β intensities were significantly elevated in tau^P301S mice compared to WT littermate mice, and significantly reversed in tau^P301S;usp11⁻ mice (Figures S6B, S6D, S6E, and S6F).

Elevated acetylated tau in female tau^P301S mice is reversed by loss of usp11

We next determined whether loss of usp11 alters acetylated tau (Ac-tau) in the cortex of 7-month-old tau^P301S mice. Indeed, insoluble Ac274-tau and Ac281-tau were significantly reduced by ~80% in female tau^P301S;usp11^−/− mice compared to female tau^P301S mice (Figures 5J–5L). A modest but significant ~35% reduction in soluble Ac281-tau was also seen in female tau^P301S;usp11^−/− mice compared to female tau^P301S mice (Figures 5J and 5L). Remarkably, female tau^P301S mice exhibited significant ~3.5-fold and ~2.4-fold increases in insoluble Ac274-tau and Ac281-tau, respectively, compared to male tau^P301S mice of the same age, whereas no changes in soluble Ac-tau were seen between female and male tau^P301S mice (Figures 5J, 5M, and 5N; Figure S6G). While insoluble total tau was also significantly elevated in female tau^P301S mice (Figures S6G and S6I), the increase in insoluble Ac-tau in female tau^P301S mice and its reduction by usp11 elimination remained significant after normalization to insoluble total tau (Figures S6J and S6K). As expected, both soluble and insoluble USP11 levels were significantly elevated (~1.5-fold and ~2.2-fold, respectively) in female tau^P301S mice compared to male tau^P301S mice (Figures 5J and 5O). In male tau^P301S mice, loss of usp11 drove a modest but significant ~25-35% reduction in insoluble Ac274-tau and Ac281-tau (Figures S6L–S60), whereas only Ac281-tau remained significantly reduced after normalization to insoluble total tau (Figure S6O). Soluble Ac-tau was not altered by loss of usp11 in male tau^P301S mice (Figures S6L–S6N).

We wondered if the blunted reduction in Ac-tau by usp11 elimination in male tau^P301S mice might in part be explained through sex-biased genetic compensation by related USPs. To address this question, we conducted qRT-PCR for the 22 CNS-expressed USPs originally screened by siRNA. Other than USP11, which was knocked out, 17 out of 21 USPs were significantly upregulated in male usp11⁻ brains compared to male WT brains, including USP13 (Figure S6P). By contrast, none of the 21 USPs were significantly upregulated in the brains of 7-month-old female usp11^−/− mice, compared to female WT mice of the same age (Figure S6Q). These results indicate that endogenous USP11 contributes to tau pathology in vivo, with far greater effect in female mice, by promoting aggregation-prone tau acetylation at K281 and K274. Male tau^P301S mice, by contrast, exhibit lower Ac-tau pathology than females, and loss of usp11 in males but not females shows significant genetic compensation through upregulation of related USPs.

Loss of usp11 rescues tau^P301S-induced synaptic plasticity deficits

To assess any changes in short- and long-term synaptic plasticity as a function of USP11, we conducted electrophysiological recordings of acute brain slices from 5-month-old female mice (Figure 6A). In determining input/output (I/O) curves of the field excitatory postsynaptic potential (fEPSP), we did not detect a significant difference between WT, tau^P301S, tau^P301S;usp11^+/−, and tau^P301S;usp11^−/− mice (Figure 6B), indicating normal short-term basal synaptic efficacy in all genotypes. We next determined paired-pulse facilitation (PPF), a form of presynaptic efficacy. The ratio of the second versus the first fEPSP overall was significantly reduced in tau^P301S slices compared to WT slices, indicating impaired presynaptic facilitation. By contrast, the PPF fEPSP slope in tau^P301S;usp11^+/− and tau^P301S;usp11^−/− slices recovered to levels indistinguishable from WT slices (Figure 6C). For induction of long-term potentiation (LTP), we applied theta-burst stimulation. As previously documented (Fang et al., 2021; Woo et al., 2019; Woo et al., 2017; Yoshiyama et al., 2007), tau^P301S slices exhibited severely blunted LTP over the 1h period compared to WT slices (Figure 6D). However, tau^P301S;usp11^−/− and tau^P301S;usp11^−/− slices exhibited LTP induction and maintenance indistinguishable from LTP of WT slices (Figure 6D), indicating that usp11 reduction or loss recovers long-term synaptic plasticity in tau^P301S mice. Per these functional results ex vivo, we detected significantly reduced synaptophysin immunoreactivity in the CA3 stratum radium (SR) of 7-month-old female tau^P301S mice, which was restored to WT levels in tau^P301S;usp11^−/− brains (Figures 6E and 6F).

Figure 6. Loss of usp11 rescues synaptic plasticity deficits in female tau^P301S mice.

(A) Schematic of electrophysiological recordings from ex vivo brain slices (Biorender.com). (B) Quantification of I/O curves from 5-month-old female mice (2-way ANOVA, not significant, n=31 slices from 5 WT, n=22 slices from 5 tau^P301S, n=43 slices from 8 tau^P301S;usp11^+/−, n=14 slices from 3 tau^P301S;usp11^−/−). (C) Quantification of PPF from 5-month-old female mice (2-way ANOVA: F(3,2475)=44.28, P<0.0001; posthoc Tukey main genotype effects - WT vs tau^P301S: P<0.0001; tau^P301S vs tau^P301S;usp11^+/− mice: P<0.01; tau^P301S vs tau^P301S;usp11^−/− mice: P<0.0001; n=40 slices from 5 WT, n=42 slices from 5 tau^P301S, n=63 slices from 8 tau^P301S;usp11^+/−, n=24 slices from 3 tau^P301S; usp11^−/−). (D) Quantification of LTP after theta-burst stimulation from 5-month-old female mice (2-way ANOVA: F(3,12240)=823.3, P<0.0001; posthoc Tukey main genotype effects: ****P<0.0001; n=35 slices from 5 WT, n=37 slices from 5 tau^P301S, n=62 slices from 8 tauP301;usp11^+/−, n=23 slices from 3 tau^P301S;usp11^−/−). (E) Representative images of synaptophysin (red) and DAPI (blue) in the stratum radium (SR) of CA3 from 7-month-old female mice. (F) Quantification of synaptophysin intensity in the SR of CA3 (1-way ANOVA, F(2,115)=13.43, P<0.0001; posthoc Tukey: ****P<0.0001, ***P=0.0007, n=34-43 images from 5-8 sections/mouse and 6 mice/genotype).

Loss of usp11 reverses tau^P301S-induced spatial memory impairment in female mice

We evaluated 6-7-month-old female mice for spatial learning and memory using the Morris water maze (MWM) (Figure 7A), a cognitive test that correlates with hippocampal synaptic plasticity (Burgess et al., 2002; O’Keefe and Dostrovsky, 1971; Schmidt-Hieber and Nolan, 2017). Compared to WT littermates, latency to find the hidden platform was significantly longer in tau^P301S mice over the 5 days of training, whereas tau^P301S;usp11^+/− or tau^P301S;usp11^−/− mice did not differ significantly from WT mice (Figure 7B). For the probe test on day 6, which evaluates spatial reference memory upon removal of the hidden platform, latency to enter the target zone was significantly longer in tau^P301S mice compared to WT mice, whereas tau^P301S;usp11^−/− or tau^P301S;usp11^+/− mice were indistinguishable from WT mice (Figure 7C). Average swim speeds over 60 seconds did not differ significantly among the genotypes (Figure 7C), indicating that swimming ability did not affect MWM performance. We also conducted MWM test in 6-7-month-old male mice. Male tau^P301S mice at this age did not differ from WT mice in the training phase (Figure S7A), although male tau^P301S mice exhibited a nonsignificant trend of impairment in the probe test (Figures S7B). This is similar to previous studies in which female-biased MWM impairment in tau^P301S mice was observed (Sun et al., 2020; Takeuchi et al., 2011). Loss of usp11 slightly but not significantly improved probe test performance in male tau^P301S mice (Figure S7B), and no differences in swim speeds were seen across genotypes (Figure S7B). Collectively, impaired spatial memory in female tau^P301S mice and rescue by usp11 elimination is consistent with corresponding changes in Ac-tau and the greater effect of USP11 in driving tau pathogenesis in females versus males.

Figure 7. Loss of usp11 reverses spatial memory impairment in female tau^P301S mice.

(A) Schematic of training and probe test in Morris water maze (MWM). (B) Quantification of time to find the hidden platform during the training phase in female mice (RM 1-way ANOVA: genotype F(3,12)=7.742, P=0.0039; posthoc Tukey: **P<0.01 WT vs. tau^P301S, ns=not significant WT vs. tau^P301S;usp11^+/− or tau^P301S;usp11^−/−; n=17 WT, n=16 tau^P301S, n=17 tau^P301S;usp11^+/−, n=12 tau^P301S;usp11^−/−). (C) Quantification of time to target zone (sec) and swim speed (m/s for 1 min) during the probe test on day 6 with the hidden platform removed (1-way ANOVA genotype effect for time to target: F(3,58)=8.202, P=0.0001; genotype effect for swim speed: F(3,58)=2.131, P=0.1061, posthoc Dunnett: ***P<0.001; ns=not significant; n=17 WT, n=16 tau^P301S, n=17 tau^P301S;usp11^+/−, n=12 tau^P301S;usp11^−/−).

Discussion

Among the triad of age, APOE ε4, and sex, female sex represents one of the most significant risk factors for AD (Collaborators, 2022; Rajan et al., 2021). Possible factors that have been previously proposed to contribute to this sexual dimorphism include differences in sex hormones and biological responses (Xiong et al., 2022) (Udeh-Momoh et al., 2021), discrepancies in the level of education (Hasselgren et al., 2020), and multiple pregnancies (Colucci et al., 2006). While female menopause is associated with significant risk for declining brain health (Jett et al., 2022), hormone replacement therapy has shown mixed, inconclusive, and sometimes deleterious results for dementia prevention (Espeland et al., 2004; LeBlanc et al., 2001; Maki, 2013; Shumaker et al., 2004; Shumaker et al., 2003). Here, we show a mechanism of sex-based vulnerability to greater tau burden by virtue of physiologically higher expression of USP11 in the female brain compared to males. Higher tauopathy signatures are clearly seen in women, as women exhibit higher tau tangle density compared to men (Oveisgharan et al., 2018). Higher tau deposition has also been observed in PET imaging studies of cognitively normal women (Buckley et al., 2019; Buckley et al., 2020; Luchsinger et al., 2020; Palta et al., 2021), and higher CSF p-tau levels are seen in women with mild cognitive impairment and carrying the APOE ε4 allele (Altmann et al., 2014; Babapour Mofrad et al., 2020; Liu et al., 2019a).

Through conducting an unbiased siRNA screen of CNS-expressed USPs, we identified USP11 as a robust driver of tau pathology in vitro and in vivo. USP11 resides on chromosome X in mammals, yet can escape complete X inactivation (Li et al., 2016; Loda et al., 2022; Tukiainen et al., 2017) and also plays a role in female-biased biology (Dwane et al., 2020; Orthwein et al., 2015; Schoenfeld et al., 2004). Hence, USP11 is significantly elevated in females compared to males in cognitively normal humans and mice. Regardless of sex, USP11 drastically accumulates in the brains of AD and FTLD-tau, and frequently colocalizes with AT8-positive tangles and neuropil threads. These results are consistent with the capacity of excess USP11 to promote tau aggregation in vitro, and for loss of usp11 to mitigate tau pathology in vivo. However, the positive association of USP11 level with tau pathology in AD and FTLD-tau is remarkably stronger in women versus men, consistent with the elevated USP11 seen in women at a preclinical stage and strengthening the thesis that USP11 plays a more significant role in women starting from an early stage of tauopathy. This association is consistent with experimental observations, in which female tau^P301S mice exhibit greater accumulation of aggregation-prone Ac-tau species Ac274 and Ac281, and genetic loss of usp11 demonstrates strong female-specific mitigation of Ac-tau pathology and spatial memory deficits. These observations in humans and mice, therefore, form an intriguing thesis for increased vulnerability of females to Ac-tau accumulation.

The major ub-K48 modification results in rapid proteasomal clearance of soluble proteins, whereas ub-K63 modification is associated with slower clearance of often misfolded or insoluble proteins through the autophagy-lysosome system (Nathan et al., 2013; Tan et al., 2008). Pathological tau is cleared through a process requiring the autophagy cargo receptor SQSTM1/p62 (Fang et al., 2021; Ramesh Babu et al., 2008; Woo et al., 2020; Xu et al., 2019). Our studies in cultured cells show that USP11 knockdown promotes tau clearance, while excess USP11 interferes with tau clearance and promotes aggregation through the removal of both ub-K48 and ub-K63 linkages on tau. This suggests that USP11 impacts tau clearance through both proteasomal and autophagy pathways. This differs from the Otub1, which upon overexpression removes ub-K48 but not ub-K63 from tau (Wang et al., 2017). While another DUB, USP13, regulates p-tau and ub-tau levels and alters proteasome and autophagy machineries, it is unclear if USP13 directly cleaves ubiquitin conjugates from tau (Liu et al., 2019b). By contrast, the tauopathy-driving effect of excess USP11 requires its enzymatic activity, as the catalytically dead USP11^C318S is inactive in this regard.

This study showed that endogenous USP11-mediated tau deubiquitination acts as an enabling switch to enhance tau acetylation at K281 and K274. Acetylation events specifically at K274, K280, or K281 are associated with impaired microtubule binding and increased tau aggregation, probably due to charge neutralization (Cohen et al., 2011; Rane et al., 2019; Trzeciakiewicz et al., 2017). The double K274Q/K281Q mutant disrupts the neuronal axonal initial segment, and ac-K274 and ac-K281 tau species accumulate across increasing Braak stages in AD (Sohn et al., 2016) and after brain injury (Shin et al., 2021). Acetylation of the double lysines K280 and K281, both of which are elevated in AD brains, typically occurs together on the same tau molecule in AD (Cohen et al., 2011). However, acetylation and ubiquitination are seen on K281, but not K280 or K274, in normal mouse (Morris et al., 2015) and human tau (Wesseling et al., 2020), suggesting that K281 is a significant physiological site for molecular switching between ub-tau and ac-tau. This agrees with our observation that K281 is an important initiation site for the removal of ubiquitin from tau by USP11.

PHF-tau isolated from AD brains is primarily ubiquitinated at K254, K311, K317, K321, and K353, which are common to both 3-repeat (3R) and 4-repeat (4R) tau (Arakhamia et al., 2020; Cripps et al., 2006), and two of which (K254 & K353) are relatively insensitive to USP11 in vitro. This raises the intriguing possibility that USP11 preferentially removes ubiquitin from pathogenic tau acetylation sites. Seed-incompetent sarkosyl-soluble tau monomers are not modified by acetylation or ubiquitination, whereas seed-competent AD tau oligomers are acetylated at K281, K343, and K353, and also ubiquitinated at non-overlapping residues (K254, K257, K290, K311, K317, K385, and K395) (Puangmalai et al., 2022; Wesseling et al., 2020). Interestingly, K63-linked, but not K48-linked, ubiquitination of AD-derived tau oligomers at these sites is associated with tau seeding activity and propagation (Puangmalai et al., 2022).

It is notable that hemizygous loss of usp11 improves spatial memory and LTP deficits and reduces insoluble tau to a similar extent as complete loss of usp11 in female PS19 mice. This might be explained by the way X-inactivation (Xi) occurs in females, in which ~50% of cells in usp11^+/− mice are fully knocked out for USP11, assuming random Xi and no Xi-escape. Hence, near-complete loss of USP11 in ~50% of neurons may be sufficient to achieve close to the full effect of usp11 KO on measures that rely on neuronal connectivity, such as synaptic plasticity and pathogenic tau spreading. However, female usp11^+/− cortex exhibited an average of ~65% reduction in USP11 compared to littermate female usp11^+/+ mice, while female usp11^+/+ mice showed significantly higher USP11 levels compared to male usp11⁺ mice. This, together with significantly higher USP11 levels in women, suggest that USP11 partially escapes Xi in the aging brain, consistent with USP11 escape seen in interspecific hybrid cellular models (Li et al., 2016; Tukiainen et al., 2017). A recent study showed that estrogen increases USP11 at the protein level (Dwane et al., 2020), which may also contribute to increased USP11 protein in females. Regardless of sex, however, USP11 accumulates in the brains of AD patients. Such elevation of USP11 is likely related to increased DNA damage observed in AD starting from the early stage of mild cognitive impairment (Adamec et al., 1999; Shanbhag et al., 2019; Sheng et al., 1998a; Sheng et al., 1998b), as USP11 is strongly increased by DNA damage (Ting et al., 2019; Yu et al., 2016) or apoptotic stimuli (Rong et al., 2021; Zhang et al., 2021). This disease-associated USP11 accumulation may fuel a maladaptive feed-forward cycle to exacerbate tauopathy further.

Another intriguing finding is that male usp11 KO mice exhibited significant genetic compensation through upregulation of related USPs, while female usp11 KO mice did not. This is reminiscent of sex differences observed in microglial gene expression in response to microglial miRNA depletion in PS19 mice (Kodama et al., 2020). While genetic compensation in response to gene knockout is a widespread phenomenon (El-Brolosy and Stainier, 2017), and sex differences in genetic compensation have been observed (Eva et al., 2020; Shay et al., 2018), the mechanistic basis is not well understood. Upregulation of related USPs may contribute to the blunted tauopathy mitigation by loss of usp11 in male tau^P301S mice. By contrast, the effects of USP11 in females are not encumbered by genetic compensation. In this regard, it is notable that histone demethylases KDM5C and UTX are located on the X-chromosome and escape Xi, resulting in higher histone demethylase activity and an increased threshold for transcriptional activation of multiple genes in females (Keiser and Wood, 2019).

Overall, our unbiased identification of USP11 as a DUB driving strong female-biased effects on tauopathy in humans and mice underpins a new mechanistic basis for greater vulnerability to tauopathy in women from a preclinical stage. In addition to presenting novel insights for preventing human tauopathies in women, this study also sets a framework for identifying other X-linked factors that could confer increased susceptibility to tauopathy in women.

Limitations of the study

It is important to note the limitations of our study. First, mouse models of tauopathy may not fully capture the sexual dimorphism in tau pathology seen in humans. In PS19 mice, females but not males exhibit deficits in spatial memory (Sun et al., 2020), which we also observed. However, male but not female PS19 mice exhibit substantial motor deficits (Sun et al., 2020). This type of sexual dimorphism has not been reported in human FTDP-17 MAPT^P301S carriers (Lossos et al., 2003; Sperfeld et al., 1999). Second, our in vitro, cellular, and animal models all employed the 4-repeat (4R) form of tau. As K274 is common to 3R and 4R tau, while K280 and K281 are exclusive to 4R tau, it is possible that USP11 could impact 3R tau in a manner distinct from 4R tau. Third, although the most conservative mechanistic basis of our findings is through direct USP11 DUB activity on tau, other activities downstream of USP11 may indirectly impinge on tauopathy or brain injury through additional mechanisms, such as the beneficial effects of USP11 silencing seen during intracerebral hemorrhage (Zhang et al., 2021) and ischemia-reperfusion (Rong et al., 2021), or changes in proteostasis balance (Basic et al., 2021; Rong et al., 2021). Fourth, slower neuronal migration during embryonic development and behavioral abnormalities in 3-month-old male usp11 KO mice (usp11−) have been reported (Chiang et al., 2021). However, the concerns surrounding this limitation are largely mitigated by our findings that spatial memory, synaptic integrity, and gliosis are unaffected in usp11 KO mice of either sex in the C57BL/6 background at 7 months of age. This might be explained by the differences in genetic background, age, sex, and assays examined. Finally, while USP11 DUB activity promoted tauopathy, our analysis was confined to K274, K280, and K281 sites. Hence, tau ubiquitination and acetylation events outside of these residues may impact tau in different ways.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources or reagents should be directed to and will be fulfilled by the lead contact, David E. Kang (dek94@case.edu).

Material availability

This study did not generate new unique reagents.

Data and code availability

• Original western blotting data have been deposited at Mendeley and are publicly available as of the date of publication. The DOI is listed in the key resources table.

• This manuscript does not report original code.

• Any additional information required to reanalyze the data reported in this work paper is available from the Lead Contact upon request.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-tau antibody (A-10)	Santa Cruz Biotechnology	Cat# sc-390476
Rabbit polyclonal anti-phospho-tau (Ser199, Ser202) antibody	Invitrogen	Cat# 44-768G
Rabbit polyclonal anti-ac-tau (K274)	Gan lab, Cornell	N/A
Rabbit polyclonal anti-ac-tau (K281)	Gan lab, Cornell	N/A
Rabbit monoclonal recombinant anti-synaptophysin antibody [YE269]	Abcam	Cat# ab32127
Mouse monoclonal anti-β-actin antibody (C4)	Santa Cruz Biotechnology	Cat# sc-47778
Mouse monoclonal anti-USP11 antibody (C-6)	Santa Cruz Biotechnology	Cat# sc-365528
Rabbit monoclonal recombinant anti-ubiquitin (linkage specific K48) antibody [EP8589]	Abcam	Cat# ab140601
Rabbit monoclonal recombinant anti-ubiquitin (linkage specific K63) antibody [EPR8590-448]	Abcam	Cat# ab179434
Rabbit polyclonal anti-ubiquitin antibody	Abcam	Cat# ab19247
Rabbit polyclonal anti-USP11 antibody	Protein Tech	Cat# 22340-1-AP
Rabbit polyclonal anti-USP11 antibody	Abnova	Cat# PAB4191
Rat monoclonal anti-glial fibrillary acidic protein (GFAP) antibody (2.2B10)	Invitrogen	Cat# 13-0300
Rabbit polyclonal anti-Iba1 antibody	FUJIFILM Wako Pure Chemical Corporation	Cat# 019-19741
Rabbit polyclonal anti-phospho-tau (Ser 262) antibody	Invitrogen	Cat# 44-750G
Rabbit polyclonal anti-IL-1 beta antibody	Abcam	Cat# ab9722
Mouse monoclonal anti-phospho-tau (Ser202, Thr205) antibody (AT8)	Invitrogen	Cat# MN1020
Mouse polyclonal anti-horseradish peroxidase	Jackson ImmunoResearch Inc.,	Code: 223–005-024, RRID: AB_2339261
Rabbit polyclonal anti-horseradish peroxidase	Jackson ImmunoResearch Inc.,	Code: 323–005-024, RRID: AB_2315781
Goat anti-rabbit IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 488	Invitrogen	Cat# A11034
Goat anti-rabbit IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 594	Invitrogen	Cat# A11037
Goat anti-mouse IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 594	Invitrogen	Cat# A11032
Goat anti-rat IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 488	Invitrogen	Cat# A11006
Goat anti-mouse IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 488	Invitrogen	Cat# A11029
Bacterial and virus strains
Rosetta 2(DE3) singles competent cells - Novagen	Sigma-Aldrich	Cat# 71400-M
USP11 shRNA lentivirus (mouse)	ABM	Cat #: iV048269
Biological samples
Human cortical brain tissue	Alzheimer’s Disease Research Center at Emory University	N/A
Chemicals, peptides, and recombinant proteins
Recombinant human USP11 protein	Abcam	RRID ab206019
Recombinant tau (human) protein	Kang lab, CWRU	N/A
Tetracycline	Gibco™	Cat# A39246
MG 132	Thermo Scientific™	J63250-M^
Dulbecco’s modified Eagle’s medium (DMEM)	Corning	10–013-CV
Fetal bovine serum	Sigma-Aldrich	12306 C
Penicillin/streptomycin	Gibco	15140–122
HBSS buffer	Gibco	14175-095
Trypsin-EDTA	Gibco	25200056
Neurobasal media medium	Invitrogen	10888022
Opti-MEM	Gibco	31985-070
Fugene HD transfection reagent	Promega	E2312
Zeocin selection reagent	Thermofisher	R25001
Lipofectamine 2000	Invitrogen	116688500
GlutaMAX	Invitrogen	35,050,061
Ni Sepharose	Millipore Sigma	GE17-5268-01
Protease inhibitors cocktail (Thermo Scientific, 78425),	Thermo Scientific	78425
Tris base	Thermofisher	BP152-10
Protease inhibitors	GeneDEPOT	P3100-010
Phosphatase inhibitors	GeneDEPOT	P3200-005
Deacetylase inhibitor	APExBIO	Cat# K1017
Nicotinamide	Sigma-Aldrich	72340
Trichostatin A	Sigma-Aldrich	T8552
SuperSignal™ west pico PLUS chemiluminescent Substrate	Thermo Scientific™	Cat# 34580
SuperSignal™ west Atto ultimate sensitivity chemiluminescent substrate	Thermo Scientific™	A38556
Cycloheximide	Millipore Sigma	239763-M
Dimethyl sulfoxide	VWR	N182
DL-1,4-dithiothreitol	Millipore Sigma	Cat# 3154
Duolink in situ mounting medium with DAPI	Sigma	DUO82040
Anti-mouse IgG agarose beads	Bional Transduction Products	G1360B
Triton X-100	Fluka	93426
SDS, 20% solution	RPI	L23100
Paraformaldehyde	Fisherscientific	AC416785000
Normal goat serum	KPL, product code	code 71-00-27
DAPI	Thermo Scientific	62248
Trizol	Ambion	15596026
Citrate buffer	Alfa Aesar	J63950
TrueBlack® Lipofuscin Autofluorescence Quencher	Biotium	#23007
Fluromount-G	Invitrogen	00-4958-02
Geneticin	Gibco	10131-027
LDS-sample buffer (4X), non-reducing	Alfa Aesar	J61894
Bromophenol blue	FisherBiotech	BP115-25
Zeocin	ThermoFisher	R25001
Blasticidin	Gibco	Cat# 1113903
Tubastatin A	SIGMA	SML0044
EX527	SIGMA	E7034
Critical commercial assays
Human CHIP ubiquitin ligase kit	BostonBiotech	Cat# K-280
BCA protein assay kit	Pierce	Cat# 23225
Monarch total RNA miniprep kit	New England Biolabs Inc.	Cat# T2021
Brilliant III SYBR green QRT-PCR mix	Agilent Technologies	Cat # 600886
iScript cDNA Synthesis Kit	Bio-rad	Cat# 1708841
Fast Taqman Master Mix on a Step One Plus Real-time PCR System	Thermo Scientific	Cat# 4444557
Duolink® In Situ PLA® proximity ligation assay	Sigma-Aldrich	DUO92101
Deposited data
Raw data of all western blots	Mendeley Data	doi: 10.17632/8pjj854kgh.1
Experimental models: Cell lines
HELA-V5-tau cells	Laura Blair lab, USF	N/A
iHEK-tauP301L cells	Laura Blair lab, USF	N/A
iHEK-tauWT cells	Laura Blair lab, USF	N/A
HEK293T cells	ATCC	Cat# CRL-3216, RRID: CVCL_0063
TauP301L primary neurons	Kang lab, CWRU	N/A
Experimental models: Organisms/strains
Usp11+/− mice (C57BL/6N-Atm1Brd Usp11tm1a(KOMP)Wtsi/WtsiBiat)	INFRARONTIER (https://www.infrafrontier.eu/Mice) European Mouse Mutant Archive (EMMA)	Strain ID EM:08053
C57BL/6J mice	The Jackson Laboratory	Cat# JAX:000664, RRID:IMSR_JAX:000664
Tau P301S (PS19) mice	The Jackson Laboratory	Cat# JAX:008169, RRID:IMSR_JAX:008169
Oligonucleotides
Primer RT-qPCR mouse USP11 Forward GCACCTTTCCTGGCTGTATC	IDT	N/A
Primer RT-qPCR mouse USP11 Reverse (WT) TCCCTGAGATGCCAGCTTAT	IDT	N/A
Primer RT-qPCR mouse USP11 Reverse (WT) TCCCTGAGATGCCAGCTTAT	IDT	N/A
Primer RT-qPCR mouse USP11 Reverse (knock down or knock out) TCGTGGTATCGTTATGCGCC	IDT	N/A
Primer RT-qPCR mouse tau Forward GGGGACACGTCTCCA CGGCAT	IDT	N/A
Primer RT-qPCR mouse tau Reserve 5′ TCCCCCAGCCTAGACCACGAG 3′	IDT	N/A
Human tau forward primer: CTCCAAAATCAGGGGATCGC	IDT	N/A
Human tau reverse primer: CCTTGCTCAGGTCAACTGGT	IDT	N/A
Human GAPDH forward primer: AAGGTCGGAGTCAACGGA	IDT	N/A
Human GAPDH reverse primer: CCATGGGTGGAATCATATTGG	IDT	N/A
siRNAs targeting USPs (related to Figure S1)	Table S1 (this paper)	N/A
RT-qPCR primers for USPs (related to Figure S6)	Table S2 (this paper)	N/A
Recombinant DNA
pQHA-USP11 WT puro	Addgene	Plasmid #46749 (Maertens et al., 2010)
pQHA-USP11 CS puroR	Addgene	Plasmid #46750 (Maertens et al., 2010)
pcDNA™3.1 (+)	Invitrogen	Cat# V79020
HA-ubiquitin	Addgene	Plasmid #18712
Tau-WT	Pieper lab, CWRU	N/A
Tau-K274Q	Pieper lab, CWRU	N/A
Tau-K280Q	Pieper lab, CWRU	N/A
Tau-K281Q	Pieper lab, CWRU	N/A
P300-HA	Gan lab, Cornell	N/A
Software and algorithms
GraphPad Prism 8	GraphPad	https://www.graphpad.com/scientific-software/prism/
Image J	National Institutes of Health, Bethesda, MD, US	https://imagej.nih.gov/ij/index.html
ANY-Maze behavioral tracking software	ANY-maze, UK	https://www.anymaze.co.uk/index.htm
FV10-ASW 4.2	Olympus
ZEN	ZEISS	https://www.zeiss.com/microscopy/us/prodcts/microscope-software/zen.html
pClamp 11.3 software	Molecular Devices	https://www.moleculardevices.com/
MaxQuant	MaxQuant	https://www.maxquant.org/
Other
Olympus FV10i confocal microscope	Olympus, PA, USA	N/A
ZEISS LSM880 confocal microscope	ZEISS	https://www.zeiss.com/corporate/int/home.html
Fuji LAS-4000 imager	LAS-4000, Pittsburgh, PA, USA	N/A
Keyence digital microscope	Keyence	BZ-X810
Amersham ImageQuant 800	Cytiva	https://www.cytivalifesciences.com/en/us
QuantStudio 3 Real-Time PCR system	ThermoFisher	https://www.thermofisher.com/us/en/home.html
Q Exactive™ plus hybrid quadrupole-orbitrap™ mass spectrometer	Thermo Scientific	https://www.thermofisher.com/order/catalog/product/IQLAAEGAAPFALGMBDK
Axon Digidata 1550B Data Acquisition System	Molecular Devices	https://www.moleculardevices.com/
Analog stimulus Isolator model 2200	A-M Systems	https://www.a-msystems.com/
PTC03 proportional Temperature Controller	Scientific Systems Design, Inc.	https://scisys.info/
BSC1 slice chamber	Scientific Systems Design, Inc.)	https://scisys.info/
P-97 micropipette puller	Sutter Instrument	https://www.sutter.com/
Microelectrode AC amplifier model 1800	A-M Systems	https://www.a-msystems.com/
Q125 sonicator (Homogenizers)	QSONICA	https://www.sonicator.com/products/q125-sonicator
Q800r sonicator (Homogenizers)	QSONICA	https://www.sonicator.com/pages/chromatin-dna-shearing
Leica CM 1860	Leica	https://www.leicabiosystems.com/us/histology-equipment/cryostats/leica-cm1860/
Synergy Neo 2 multi-mode reader	Biotek	https://www.biotek.com/products/detection-multi-mode-microplate-readers/synergy-neo2-hybrid-multi-mode-reader/
Poly-D-lysine/laminin-coated cover glasses	Corning	354087
Nitrocellulose membrane	GE Healthcare	10600002
Dot blot apparatus	GE Whatman	10447941
Cellulose acetate membrane	Sterlitech	CA023001

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

HEK293T cells (ATCC, Cat# CRL-3216); Hela-V5-tau cells (gift from Dr. Laura Blair lab, USF, Tampa, FL), stably expressing wild-type 4R0N human tau; iHEK-tau^P301L and iHEK-tau^WT cells (gift from Dr. Laura Blair lab, USF) with tetracycline induced 4R0N tauP301L or WT tau expression (Shelton et al., 2017) were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Corning, 10–013-CV) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich, 12306 C) and 1% penicillin/streptomycin (P/S, Gibco, 15140–122). Zeocin (400 μg/ml) and blasticidin (5 μg/ml) were added to the medium to iHEK-tauP301L cells, and geneticin (50 μg/ml) was added to the medium of HeLa-V5-tau cells to maintain selection.

Primary neuronal culture

Primary mouse cortical and hippocampal neurons were prepared from P0 mice. Briefly, both cortex and hippocampus were dissected in ice-cold HBSS buffer (Gibco, 14175-095) with 0.6% glucose, then digested with 0.25% trypsin-EDTA (Gibco, 25200056) at 37°C. The tissues were homogenized and then gently centrifuged at 2000 RPM, 4°C, for 2 minutes. Cells were seeded on poly-D-lysine/laminin-coated cover glasses (Corning, 354087) or plates with neurobasal media medium (Invitrogen, 10888022) containing 2% B27 (Invitrogen, 17,504,044), 2% GlutaMAX (Invitrogen, 35,050,061), and 1% penicillin/strep (Gibco, 15140-122).

Mice

Usp11^−/− mice were obtained from the European Mouse Mutant Archive (EMMA). The Strain ID is EM:08053, and the strain name is C57BL/6N-A^tm1Brd Usp11^{tm1a(KOMP)Wtsi}/WtsiBiat. Genomic DNA isolated from tail snips was used for genotyping by PCR with the following primers: usp11 Forward primer: 5′ GCACCTTTCCTGGCTGTATC 3′; usp11 reverse primer: 5′ TCCCTGAGATGCCAGCTTAT 3′. Cas reverse primer (usp11 knock out): 5′ TCGTGGTATCGTTATGCGCC 3′.

Previously described Tau^P301S transgenic mice (PS19) (Yoshiyama et al., 2007) were obtained from Jackson Laboratories (B6; C3-Tg (Prnp-MAPT*P301S) PS19Vle/J, JAX: 008169). Genomic DNA isolated from tail snips was used for genotyping by PCR with the following primers: forward primer: 5′ GGGGACACGTCTCCA 3′, reverse primer: 5′ TCCCCCAGCCTAGACCACGAG 3′.

Usp11^−/− mice were crossed with tau^P301S transgenic mice to generate tau^P301S;usp11^−/−, tau^P301S;usp11^+/−, tau^P301S;usp11⁻ mice. All mice were maintained in the C57BL/6 background for at least three generations prior to breeding. All experimental procedures on mice study were approved by the IACUC.

Human brain tissue

Paraformaldehyde-fixed floating frontal cortex tissues were obtained from the Alzheimer’s Disease Research Center at Emory University (Drs. Levey and Gearing). The primary neuropathologic diagnosis confirms AD and FTLD-tau patients (Montine et al., 2012). PMI, age at onset, duration of illness, age at death, APOE genotype, Thal amyloid score, Braak stage, CERAD score, race, and gender are summarized in Figure S4A.

Recombinant proteins

pET-28a-Tau WT (pET-28a-WT 0N4R Tau plasmid was a gift from Dr. Laura Blair, USF, Tampa, FL) was transformed into Escherichia coli BL21 Rosetta 2(DE3) (Sigma-Aldrich, 71400-M) to express recombinant His-tagged tau protein. Protein expression was induced by 0.15 mM IPTG overnight at 18°C. Cells were harvested and homogenized with the Q125 Sonicator (Homogenizers, Atkinson, NH, USA). After centrifugation at 16,000g for 30 minutes, the supernatant was collected and shaken for 1h at 4°C. Recombinant proteins were purified by Ni Sepharose (Millipore Sigma, GE17-5268-01), eluted with 500mM imidazole in 50mM HEPES-KOH, pH 8.5, 50mM KCl, and 5mM β-mercaptoethanol. Purified protein was stored in 25 mM HEPES-KOH pH 7.5, protease inhibitors cocktail (Thermo Scientific, 78425), 75mM KCl, 75mM NaCl and 10mM DTT. USP11 recombinant protein was purchased from Abcam (ab206019).

METHOD DETAILS

DNA and siRNA transfections and shRNA transduction

DNA plasmids or siRNA (100nM) were transiently transfected in Hela-V5-tau, iHEK-tau^P301L cells, or HEK293T cells using lipofectamine 2000 (Invitrogen, 116688500). Four to 6 h post-transfection, the medium was replaced with a new complete medium. Control or USP11 shRNA lentivirus (ABM, iV048269) were used to transduce primary mouse cortical and hippocampal neurons.

Protein extraction, immunoblotting, quantification, and filter trap assay

Cells and brain tissues were lysed with RIPA buffer (50 mM Tris pH 7.4, 0.1% SDS, 2mM ethylenediaminetetraacetic acid, 150mM NaCl, 1% Triton-100) with 1% protease inhibitors (GeneDEPOT, P3100-010) and phosphatase inhibitors (GeneDEPOT, P3200-005) with sonication on ice until brain particulates became invisible. To detect acetylated tau, deacetylase inhibitors (APExBIO, K1017) or 5 mM nicotinamide (Sigma-Aldrich, 72340) and 1 μM trichostatin A (Sigma-Aldrich, T8552) were also added into the lysis buffer. After centrifugation at 16,000g for 15 minutes at 4°C, the supernatant was transferred to new tubes. The pellet was washed with ice-cold RIPA buffer and then lysed with 10% SDS buffer (10% SDS, 250 mM Tris pH 6.8) and sonicated in a water bath sonicator until no particulate material was visible. Protein concentration was measured by bicinchoninic acid (Pierce™ BCA protein assay kit, 23225) for both RIPA-soluble and RIPA-insoluble extracts. Equal amounts of total protein were subjected to SDS-PAGE. After transferring the proteins to a nitrocellulose membrane (GE Healthcare, 10600002), the membrane was blocked with 5% skim milk for 1 hour at room temperature (RT). Indicated primary antibodies were applied overnight at 4°C. The corresponding peroxidase-conjugated secondary antibody was detected by ECL western blot reagents (Thermo Scientific, 34580). ECL images were obtained with the Fuji LAS-4000 imager (LAS-4000, Pittsburgh, PA, USA) or Amersham ImageQuant 800 (Cytiva) and analyzed with Image J software (NIH, https://imagej.nih.gov/ij/index.html).

For filter trap assays, dot blot apparatus (GE Whatman, 10447941) with 0.22 μm pore size cellulose acetate membrane was used. Five μg of protein were loaded onto the dot blot apparatus. The membrane was washed twice with 200μl PBS and fixed with 200μl 20% methanol. The vacuum was maintained during the entire procedure. Membranes were then processed by immunoblotting as described above.

Quantified bands in Western blots were normalized to the loading control (actin) prior to comparisons between experimental conditions. For normalization between 2 gels, at least 2 identical samples were run in both gels, and the average intensity of the bands normalized to actin was used for normalization between gels.

The following antibodies were used to probe target proteins by immunoblotting: mouse monoclonal anti-tau antibody (A-10) (Santa Cruz Biotechnology, sc-390476); rabbit polyclonal anti-phospho-tau (Ser199/pSer202) antibody (Invitrogen, 44-768G); rabbit polyclonal anti-ac-tau (K274) (Gan lab, Cornell); rabbit polyclonal anti-ac-tau (K280) (Gan lab, Cornell);mouse monoclonal antβ-actin antibody (C4) (Santa Cruz Biotechnology, sc-47778); mouse monoclonal anti-USP11 antibody (C-6) (Santa Cruz Biotechnology, sc-365528); rabbit monoclonal recombinant anti-ubiquitin (linkage-specific K48) antibody [EP8589] (Abcam, ab140601); rabbit monoclonal recombinant anti-ubiquitin (linkage-specific K63) antibody [EPR8590-448] (Abcam, ab179434); rabbit polyclonal anti-ubiquitin antibody (Abcam, ab19247); rabbit polyclonal anti-USP11 antibody (Protein Tech, 22340-1-AP); rabbit anti-phospho-tau (Ser 262) polyclonal antibody (Invitrogen, 44-750G); mouse polyclonal anti-horseradish peroxidase (Jackson ImmunoResearch Inc., 223–005-024); rabbit polyclonal anti-horseradish peroxidase (Jackson ImmunoResearch Inc.,323–005-024). Recombinant human USP11 protein was purchased from Abcam.

Cycloheximide protein turnover assay

HELA-V5-tau cells were transfected with either control siRNA or USP11 siRNA for 48 hours. Cells were treated with DMSO or 100μg/mL cycloheximide (Milipore Sigma, 239763-M) for the indicated time points and subjected to immunoblotting for tau, USP11, and actin.

In vitro tau ubiquitination and deubiquitination assays

Recombinant tau was ubiquitinated by E1, E2 UBE2D3, and E3 CHIP for 6 hours with the CHIP ubiquitin ligase kit (human CHIP ubiquitin ligase kit K280, Boston Biochem). In brief, 6μl (1μg/μl) recombinant tau protein, or 6 ul luciferase (positive control), or 6 ul water (negative control) was incubated with 6μl 10x reaction buffer, 6ul 10x Mg²⁺-ATP, 6μl 10x HSP70/HSP40 mix, and incubated for 7 min at 43 °C. Reaction tubes were then cooled on ice for 10 minutes, followed by addition of the following reagents to initiate the ubiquitination reaction: 6μl of the 10x E1 enzyme, 10x E2 UBE2D3 enzyme, and 10x CHIP. Mixtures were vortexed, spun, and then incubated at 37°C for 6h. Reaction products were then inactivated for 15min at 70 °C following ice incubation for 10min. Equal amount of deubiquitination reaction buffer was added to the reaction products. Then, 3μg USP11 or an equal amount of purified water was incubated with reaction products in equal amounts of deubiquitination reaction buffer containing 100 mM HEPES (pH=7.4, 500mM NaCl, 100 mM MgCl₂, 100 mM dithiothreitol, and 10 mM ATP at 37 °C for 2-24h (Wu et al., 2014). Reaction products were analyzed by Western blotting and mass spectrometry.

Cell-based ubiquitin detection and deubiquitination

For detection of ubiquitin conjugates in tau immunoprecipitated from cells, iHEK-Tau^P301L cells were transfected with HA-Ubiquitin together with USP11^WT or USP11^CS or empty vector control using Lipofectamine 2000 (Invitrogen, REF 11668-500). After 48h, cells were harvested with RIPA buffer with 1% protease inhibitors (GeneDEPOT, P3100-010) and phosphatase inhibitors (GeneDEPOT, P3200-005). Lysates were precleared with anti-mouse IgG agarose beads (Bional Transduction Products G1360B), while anti-mouse IgG beads were blocked with 5% BSA in RIPA buffer for 6h. Tau A10 antibody was then added to precleared lysates with blocked anti-mouse IgG beads nutating overnight at 4°C. Beads were washed with RIPA buffer 5 times, followed by Westerning blotting. To detect ub-K48 and Ub-K63 conjugates and assess the ability for USP11 to deubiquitinate them, iHEK-Tau^P301L cells were treated with vehicle or MG132 (10μM) for 6h before lysis. Tau was pulled down as described above, and washed tau pulldowns were incubated in 1x deubiquitination buffer ± recombinant USP11 (1.5μg) for 2h with gentle tapping every 15 min.

LC-MS/MS detection of diGly ubiquitin signatures

In vitro ubiquitinated recombinant tau samples were processed for LC-MS/MS using s-traps (Protifi)(HaileMariam et al., 2018; Zougman et al., 2014). Briefly, SDS was added to the sample for a final concentration of 5%, after which proteins were reduced with dithiothreitol (DTT), alkylated with iodoacetamide (IAA), acidified using phosphoric acid, and combined with s-trap loading buffer (90% MeOH, 100mM TEAB). Proteins were loaded onto s-traps, washed, and digested with Trypsin/Lys-C (1:100, w:w; enzyme:protein) overnight at 37°C. Peptides were eluted and dried with a vacuum concentrator, and then resuspended in H₂O/1% acetonitrile/0.1% formic acid for LC-MS/MS analysis. Peptides were separated using a 75 μm x 50 cm C18 reversed-phase-HPLC column (Thermo Scientific) on an Ultimate 3000 UHPLC (Thermo Scientific) with a 120-minute gradient (2-32% ACN with 0.1% formic acid) and analyzed on a hybrid quadrupole-Orbitrap instrument (Q Exactive Plus, Thermo Fisher Scientific). Full MS survey scans were acquired at 70,000 resolutions. The top 10 most abundant ions were selected for MS/MS analysis.

Raw data files were processed in MaxQuant (v 1.6.17.0, www.maxquant.org) and searched against the current Uniprot Homo sapiens protein sequences database. Search parameters included constant modification of cysteine by carbamidomethylation and the variable modifications, methionine oxidation, protein N-term acetylation, and the addition of a GlyGly residue on Lysine. While trypsin is unable to cleave lysine residues modified by ubiquitin, trypsin cleaves the ubiquitin moiety at its C-terminus (sequence KESTLHLVLRLRGG), yielding a characteristic diGly residue that results in addition of +114.1 Da, enabling detection by mass spectrometry (Peng et al., 2003). Proteins were identified using the filtering criteria of 1% protein and peptide false discovery rate. Protein intensity values were normalized using the MaxQuant LFQ function (Cox et al., 2014).

Immunocytochemistry and proximity Ligation Assay (PLA)

Hela-V5-tau cells were washed with PBS and fixed with 4% paraformaldehyde (PFA) for 15 min at room temperature. Fixed cells were blocked with 3% NGS at RT for 1 hr. Indicated primary antibodies were applied overnight at 4°C followed by secondary antibodies (Alexa Fluor 488, Alexa Fluor 594) and DAPI (Thermo Scientific, 62248) or the Duolink® In Situ PLA® secondary antibody probes (Sigma-Aldrich) incubation as previously shown (Fang etai, 2021).

Real-time PCR

For cells: Hela-V5-tau cells were transfected with 100nM USP11 siRNA or control siRNA. Cells were harvested with 250μl trizol (Ambion, 15596026). After adding 62.5μl chloroform, samples were centrifuged at 13,000 RPM for 20 minutes at 4°C. RNA was isolated and purified with Monarch total RNA miniprep kit (NEB, T2010S) and then subjected to real-time PCR analysis using Brilliant III SYBR Green QRT-PCR Mix (Agilent Technologies, 600886). The comparative threshold cycle (Ct) value was used to calculate the amplification factor. GAPDH was used as a control. The primer sequences used in PCR are as follows:

Human tau forward primer: CTCCAAAATCAGGGGATCGC.

Human tau reverse primer: CCTTGCTCAGGTCAACTGGT.

Human GAPDH forward primer: AAGGTCGGAGTCAACGGA.

Human GAPDH reverse primer: CCATGGGTGGAATCATATTGG.

For mouse brain cortex: Total RNA was isolated from frozen brains using the High Pure RNA Isolation Kit (Roche Life Science, Indianapolis, IN, USA) according to the manufacturer’s instructions. RNA concentrations and purity were determined by UV-visible absorption spectra, using Nanodrop 2000 (Thermo Scientific, Brookfield, WI, USA). First-strand cDNA was synthesized with 500ng of total RNA, using iScript cDNA Synthesis Kit (Bio-rad, Cat# 1708841) according to the manufacturer’s protocol. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed in technical duplicate by using Fast Taqman Master Mix on a Step One Plus Real-time PCR System (Thermo Scientific, Cat# 4444557). Fold change of gene expression was calculated by comparative CT quantification method (Schmittgen and Livak, 2008) and normalized to the expression of internal control gene, GAPDH. Primer information is listed in key resources table.

Ex vivo electrophysiological recordings

Acute hippocampal slices were prepared as previously described (Woo et al., 2015). Briefly, 400 micron parasagittal hippocampal slices were generated in oxygenated cutting solution (110mM sucrose, 60mM NaCl, 3mM KCl, 28mM NaHCO₃, 1.25mM NaH₂PO₄, 5mM glucose, 0.6mM ascorbate, 7mM MgCl₂, and 0.5mM CaCl₂). Slices were then transferred to room temperature cutting solution diluted 1:1 with artificial cerebrospinal fluid (ACSF): 125mM NaCl, 2.5mM KCl, 26mM NaHCO₃, 1.25mM NaH₂PO₄, 25mM glucose, 1mM MgCl₂, and 2mM CaCl₂ for 10 min before incubating in ACSF with constant 95% O₂/5% CO₂ perfusion for 30 min before placing into the brain slice recording chamber (Scientific Systems Design Inc.).

Slices were placed in the recording chamber, with ACSF flow rate at 1 mL/min at 30° ± 0.5°C. A stimulating electrode was positioned on the Schaffer collaterals from the CA3 region. A glass electrode (1–4 MΩ) was placed in hippocampal CA1. The electric stimulation, controlled by pClamp 11 software (Molecular Devices), was delivered via a stimulus isolator (model 2200; A-M Systems). The evoked field potentials were amplified using a differential amplifier (model 1800; A-M Systems), filtered at 1 kHz, and digitized at 10 kHz.

Input-output (I-O) curves were generated by stepping up stimulation amplitude that elicited half-maximal fEPSP from 1 to 15 mV at the rate of 0.05 Hz. Paired-pulse facilitation (PPF) was evoked by two pulses with interpulse intervals from 20 to 300ms. The percentage of the facilitation was calculated by dividing the fEPSP slope elicited by the second pulse with the fEPSP slope elicited by the first pulse. LTP was induced by theta-burst stimulation (five trains of four pulses at 200Hz separated by 200ms, repeated six times with an inter-train interval of 10s) and sampled for 60 min after the induction. LTP was calculated by dividing the slope of 60 min post-induction responses with the average slope of 20 min baseline responses.

Mouse tissue Immunohistochemistry

Mice were perfused with PBS, and half brains were frozen on dry ice and stored at −80°C for subsequent biochemical analyses. The other half brains were fixed with 4% paraformaldehyde (PFA) at 4°C for 48h followed by cryoprotection in 30% sucrose. 25μm sections were washed 3 times with 0.2% triton in tris-buffered saline (TBS). After washing, the tissues were blocked with 3% normal goat serum (NGS) (KPL, product code 71-00-27) at room temperature for 1h. Indicated primary antibodies were applied overnight at 4°C, followed by secondary fluorescence antibodies (Alexa Fluor 488, Alexa Fluor 594) and DAPI (Thermo Scientific, 62248) incubation at RT for 1h. Each staining procedure included samples processed without primary antibody and only with secondary antibody. Confocal images were captured with the Olympus FV10i confocal microscope (Olympus, PA, USA) or ZEISS LSM880 confocal microscope (ZEISS). All comparison mages were obtained with the same light intensity, magnification, and exposure time. Image J software was utilized to quantify the immunoreactive signals from images, with background secondary antibody-only signals subtracted. Adjustments to the brightness, contrast, and threshold were applied in the same way in all comparison images, including secondary only controls. Investigators were blinded to genotype during staining, image acquisition, and quantification.

The following antibodies were applied to probe target proteins: rabbit polyclonal anti-USP11 antibody (Abnova, PAB4191); rat monoclonal anti-glial fibrillary acidic protein (GFAP) antibody 1:200 (2.2B10) (Invitrogen, 13-0300); rabbit polyclonal anti-Iba1 antibody 1:500 (FUJIFILM Wako, 019-19741); ); rabbit monoclonal recombinant anti-synaptophysin antibody 1:100 [YE269] (Abcam, ab32127); rabbit polyclonal anti-phospho-tau (Ser199, Ser202) antibody 1:100 (Invitrogen, 44-768G); rabbit anti-phospho-tau (Ser 262) polyclonal antibody (Invitrogen, 44-750G); rabbit anti-IL-1 beta polyclonal antibody (Abcam ab9722); goat anti-rabbit IgG (H+L) cross-adsorbed secondary antibody, Alexa Fluor 594 1: 1000 (Invitrogen, A11037); goat anti-rat IgG (H+L) cross-adsorbed secondary antibody, Alexa Fluor 488 1: 1000 (Invitrogen, A11006); goat anti-rabbit IgG (H+L) cross-adsorbed secondary antibody, Alexa Fluor 594 1: 1000 (Invitrogen, A11037); goat anti-rabbit IgG (H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 488 1: 1000 (Invitrogen, A11034). The nucleus was probed by DAPI 1:1000 (Thermo Scientific, 62248).

Human brain immunohistochemistry

Floating 30-micron paraformaldehyde-fixed human brain sections were washed 3 times with 0.2% triton in TBS. Antigen retrieval was performed by utilizing citrate buffer, pH=6. After antigen retrieval, tissues were washed and blocked with 3% NGS at RT for 1h. Indicated primary antibodies were applied overnight at 4°C, followed by secondary fluorescence antibodies (Alexa Fluor 488, Alexa Fluor 594) and DAPI (Thermo Scientific, 62248) for 1h at room temperature. TrueBlack® Lipofuscin Autofluorescence Quencher (Biotium, #23007) was used to eliminate autofluorescence and mitigate nonspecific background signals. Each staining procedure included samples processed without primary antibody and only with secondary antibody. Images were captured by Keyence digital microscope (Keyence, BZ-X810) with the same intensity, magnification, and exposure time for all comparison images. Image J software was utilized to quantify the immunoreactive signals from images, with background secondary antibody-only signals subtracted. Adjustments to the brightness, contrast, and threshold were applied in the same way across all comparison images, including secondary only controls. Quantification of colocalization was performed using the Image J JACoP plug-in with in-program threshold. Investigators were blinded to information about the samples during staining, image acquisition, and quantification.

The following antibodies were applied to probe target proteins: rabbit polyclonal anti-USP11 antibody (Protein Tech, 22340-1-AP); mouse monoclonal anti-phospho-tau (Ser202, Thr205) antibody (AT8) (Invitrogen, MN1020); goat anti-rabbit IgG (H+L) cross-adsorbed secondary antibody, Alexa Fluor 594 (Invitrogen, A11037); goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (Invitrogen, A11029). The nucleus was probed by DAPI (Thermo Scientific, 62248).

Morris Water Maze

Six- to 7-month-old mice were individually housed one week prior to behavioral testing. Each mouse was handled for at least 2 minutes every day. Morris water maze (MWM) test was performed on littermate mice at 6-7 months of age, as previously described (Vorhees and Williams, 2006). During the 5 days of training, mice were trained to find a hidden platform for 1 minute in a room temperature water-filled pool. In 4 training trials/day, mice were placed into the pool from 4 different directions per trial. The pool was black, circular, and 120cm in diameter, and filled with water containing a non-toxic white paint for increasing contrast. Experimental mice were black or brown. A transparent square platform with an area of 58 cm² was placed at the center of one quadrant and submerged 1cm below the water surface. Visual signs with different shapes and colors were hung on the wall in each quadrant. One hour break was allowed between every trial. Once mice found the platform, they were allowed to sit on the platform for 20 seconds before being removed. If mice failed to find the platform in 60 seconds, they were gently guided to the platform by the tail and allowed to sit on the platform for 20 seconds before being removed. Floating mice that showed no motivation to find the hidden platform during 5 days of training were excluded. Clean towels and heating pads were prepared to dry and warm mice after each trial. The probe test was performed 24 hours after the last day of training on day 6, in which mice were allowed to swim freely for 1 minute in the absence of the hidden platform. Each trial and probe test were recorded by an overhead camera connected to a PC, and each parameter was analyzed by ANY-maze 6.3 tracking and analysis software. The investigator performing testing was blinded to the genotype. White noise was played during the test.

QUANTIFICATION AND STATISTICAL ANALYSIS

Images and Western blots were analyzed and organized using Image J (https://imagej.nih.gov/ij/index.html) and Microsoft Excel. Statistical analysis was performed with the Prism 8.0 software (GraphPad Software, San Diego, CA, USA) using paired or unpaired two-tailed t-tests, 1-way ANOVA, repeated measures 1-way ANOVA, 2-way ANOVA, linear regression, or multiple linear regression. ANOVA (both 1-way and 2-way) was followed by the indicated posthoc tests. P values ≤ 0.05 were considered significant. All quantitative graphs with error bars are presented as means ± SEM (standard error of the mean).

Supplementary Material

1

Figure S1 USP siRNA screen identifies USP11 and USP13 as positive tau regulators, related to Figure 1

(A) Representative blots of tau and actin from HELA-V5-tau cells transiently transfected with ON-TARGETplus siRNA (Horizon) of individual USPs. (B) Graph shows quantification of tau levels with means ± SEM (2-tailed t-test compared to control siRNA, t=2.843, df=15, *P=0.0123, ***P=0.0001, n≥3 independent experiments). (C) Representative blots of USP11, USP13, tau, and actin from HELA-V5-tau cells co-transfected with control siRNA or USP11 siRNA and control siRNA or USP13 siRNA. (D) Graph shows quantification of tau levels with means ± SEM (1-way ANOVA, F(3,12)=11.34, P=0.0008; posthoc Tukey: **P<0.01, ns=not significant, n=4 samples/condition). (E) Representative blots of USP11, GFP-TDP-43, endogenous TDP-43, and actin from HELA tet-inducible GFP-TDP-43^WT cells transfected with control siRNA or USP11 siRNA. (F,G) Graphs show quantification of endogenous TDP-43 and exogenous GFP-TDP-43 levels with means ± SEM (2-tailed t-test, ns=not significant, n=4 samples/condition).

2

Figure S2 USP11 catalytic activity mediates tau aggregation and removes ubiquitin conjugates from tau but not Hsp70 or DNAJB1, related to Figure 2

(A-B) Representative blots of RIPA-soluble and RIPA-insoluble USP11, tau, and actin from HELA-V5-tau cells transiently transfected with control vector, wild type USP11, or catalytic dead USP11^C318S mutant (CS). Graphs show quantification of soluble and insoluble tau, represented as means ± SEM normalized to controls (1-way ANOVA; soluble tau=not significant (ns); insoluble tau: 1-way ANOVA, F(2,12)=7.403, P=0.008; posthoc Tukey: *P<0.05; n=5 samples/condition from 2 independent experiments). (C) Representative images of negative controls for in situ PLA immunostaining of HELA-V5-tau cell (main Figure 2F) with PLA probes only (no primary antibody), or with primary antibody but only 1 PLA probe. The nucleus was stained with DAPI (dark blue). Scale bar=5μm. (D) Schematic of recombinant tau ubiquitination by CHIP, deubiquitination by USP11, and detection of diGly-ub peptide signatures by LC-MS/MS. (E) The sequence of CHIP-ubiquitinated diGly-ub peptides and corresponding probable lysine position(s) on the hTau441 isoform based on MaxQuant v1.6.17.0. Highest probability=1. The diGly-lysines and probabilities in parenthesis are in bold. (F) Graph shows quantification of diGly-ub signatures remaining on lysines of CHIP-ubiquitinated tau, Hsp70 or DNAJB1 after deubiquitination with or without recombinant USP11 for 2h or 24h, normalized to CHIP-ubiquitinated indicated protein without USP11. Samples were subjected to trypsin digest and LC-MS/MS to detect diGly-ub ubiquitinated peptide signatures. Data are represented as means ± SEM (Tau protein: 1-way ANOVA, F(2,9)=18.02, P=0.0007, posthoc Tukey: **P=0.0077, ***P=0.0009; n=3 samples/condition; Hsp70 or DNAJB1 proteins: 1-way ANOVA, not significant).

3

Figure S3 USP11 knockdown prevents tau phosphorylation at S262 induced by HDAC6 inhibition, related to Figure 3

(A) Representative blots of tau, Ac-tau species, and actin in the RIPA-insoluble fraction of iHEK-Tau^WT cells co-transfected with vector control or p300-HA and control siRNA or USP11 siRNA from Figure 3G. (B) Representative blots of USP11, tau, pS262-tau, Ac-tau species, and actin from RIPA-soluble and RIPA-insoluble fractions of iHEK-Tau^WT cells transfected with control siRNA or USP11 siRNA for 48h ± TBSTA treatment (5μM, 6h) prior to lysis. (C-E) Graphs show quantification of RIPA-soluble pS262-tau, Ac281-tau, and Ac274-tau levels with means ± SEM (1-way ANOVA, pS262-tau: F(3,12)=6.533, P=0.0072; Ac281-tau: F(3,12)=11.67, P=0.0007; Ac274-tau: F(3,12)=9.817, P=0.0015; posthoc Tukey: *P<0.05, **P<0.01, ***P<0.001, n=not significant, n=4 samples/condition).

4

Figure S4 Accumulation of USP11 in human tauopathies and strong correlation of higher USP11 level with elevated tauopathy in women, related to Figure 4

All data on bar graphs are represented as means ± SEM. (A) Case information of AD, FTLD-tau, and nondementia control brains. Floating sections of paraformaldehyde-fixed frontal gyrus cortex brain tissues from AD, FTLD-tau, or nondementia control cases were obtained from the Alzheimer’s Disease Research Center at Emory University. Available information on APOE genotypes, ethnicity (w=White, b=Black, h=Hispanic), sex (m=male, f=female), post-mortem interval (PMI), age at onset of disease, age at death, Braak stage, Thal amyloid score, CERAD score, and ABC score are indicated. (B) Representative negative control images of the frontal gyrus of an AD case immunostained with only secondary antibodies (red) and DAPI (blue). Scale bar=20μm. (C) Graph shows quantification of USP11 intensity in nondementia controls and FTLD-tau normalized to controls (2-tailed t-test, t=4.649, df=79, ****P<0.0001; n=49 images from 4-5 images/case, 2-3 sections/case, and 10 control cases; n=32 images from 4-5 images/case, 2-3 sections/case, and 7 FTLD-tau cases. (D) Graph shows quantification of cytoplasmic/nuclear USP11 ratio in nondementia controls and FTLD-tau (2-tailed t-test, t=4.607, df=81, ****P<0.0001; n=50 images from 4-5 images/case, 2-3 sections/case, and 10 control cases; n=34 images from 4-5 images/case, 2-3 sections/case, and 7 FTLD-tau cases. (E) Graph shows Image J quantification of Manders coefficient of USP11 colocalization with AT8 in nondementia controls and FTLD-tau (2-tailed t-test, t=6.486, df=91, ****P<0.0001; n=59 images from 4-5 image/case from 2-3 sections/case, and 10 control cases; n=34 images from 4-5 images/case, 2-3 sections/case, and 7 FTLD-tau cases. (F,G) Graphs show the correlation between USP11 and p-tau (AT8) intensities in (F) AD case (female & male): linear regression test, r²=0.2602, P<0.0001; multiple regression AT8 x USP11 main effect adjusting for Thal amyloid score, age at death, and APOE genotypes, adjusted r²=0.3984, F(4,78)=14.58, P<0.0001; n=83 images from 4-5 images/case, 2-3 sections/case, and 17 AD cases; and (G) FTLD-tau cases (female & male): linear regression test, r²=0.8087, P<0.0001, n=32 images from 4-5 images/case, 2-3 sections/case, and 7 FTLD-tau cases.

5

Figure S5 Usp11 KO mice exhibit normal gliosis, synaptic integrity, and spatial memory; ablating one copy of usp11 in tau^P301S mice partially mitigates tau pathology, related to Figure 5

All data in graphs are represented as means ± SEM. (A1) Representative images of the hippocampus from usp11 KO mice and WT littermates immunostained with GFAP antibody (green) and DAPI (blue). Scale bar=100μm. (A2,A3) Graphs show quantification of GFAP intensity (green) in the hippocampus of (A2) female WT and usp11^−/− mice: 2-tailed t-test, not significant; and (A3) male WT and usp11⁻ mice: 2-tailed t-test, not significant; female WT: n=27 images from 4-5 sections/mouse and 3 mice/genotype; female usp11^−/−: n=29 images from 4-5 sections/mouse and 5 mice/genotype; male WT: n=13 images from 4-5 sections/mouse and 3 mice/genotype; male usp11⁻ : n=22 images from 4-5 sections/mouse and 4 mice/genotype. (B1) Representative images of the hippocampus from WT and usp11 KO littermates immunostained with synaptophysin antibody (red) and DAPI (blue). Scale bar=100μm. (B2,B3) Graphs show quantification of synaptophysin intensity (red) in stratum lucidum (SR) of the hippocampus of (B2) female WT and female usp11^−/− mice: 2-tailed t-test, not significant; and (B3) male usp11⁺ (wild type) and male usp11⁻ mice: 2-tailed t-test, not significant; female WT: n=12 images from 4-5 sections/mouse and 3 mice/genotype; female usp11^−/−: n=12 images from 3 sections/mouse and 4 mice/genotype, male usp11⁺ (wild type): n=13 images from 4-5 sections/mouse and 3 mice/genotype; male usp11⁻: n=12 images from 3 sections/mouse and 4 mice/genotype. (C) Graph shows time to find the hidden platform during the training phase in female mice (RM 1-way ANOVA: not significant; n=17 WT, n=9 usp11^+/−, n=8 usp11^−/−). (D) Graph shows time to target zone (sec) during the probe test on day 6 with the hidden platform removed in female mice (1-way ANOVA: not significant; n=17 WT, n=9 usp11^+/−, n=8 usp11^−/−). (E) Graph shows time to find the hidden platform during the training phase in male mice (RM 1-way ANOVA: not significant, n=13 WT mice, n=4 usp11⁻ mice). (F) Graph shows time to target zone (sec) during the probe test on day 6 with the hidden platform removed in male mice (2-tailed t-test, not significant. n=13 male WT mice, n=4 usp11⁻ mice). (G) Representative blots of RIPA-soluble and RIPA-insoluble tau, USP11, and actin in the cortex of 7-month-old female tau^P301S, tau^P301S;usp11^+/−, tau^P301S;usp11^−/−, and WT mice. (H,I) Graphs show quantification of soluble and insoluble tau in the cortex of 7-month-old female mice (soluble tau: 1-way ANOVA: F(2,27)=5.207, P=0.0122, posthoc Tukey: *P<0.05; insoluble tau: 1-way ANOVA: F(2,27)=30.65, P<0.0001, posthoc Tukey: ****P<0.0001; n=12 tau^P301S, n=6 tau^P301S;usp11^+/−, n=12 tau^P301S;usp11^−/−). (J) Representative blots of RIPA-soluble and RIPA-insoluble USP11, tau, pS199/202-tau, and actin in the cortex of 7-month-old male tau^P301S and tau^P301S;usp11⁻ mice. (K-L) Graphs show quantification of soluble and insoluble tau and pS199/202-tau in the cortex of male tau^P301S, tau^P301S;usp11⁻, and WT littermates. (K) Soluble tau: 2-tailed t-test, not significant; insoluble tau: 2-tailed t-test, t=2.561, df=18, *P=0.0196; n=9 tau^P301S, n=11 tau^P301S;usp11⁻. (L) Soluble pS199/202-tau: 2-tailed t-test, not significant; insoluble pS199/pS202-tau: 2-tailed t-test, t=2.565, df=10, *P=0.0281; n=5 tau^P301S, n=7 tau^P301S;usp11⁻. (M) Representative confocal images of pS199/202-tau staining (green), scale bar=20μm; pS262-tau (red), scale bar=50μm. and DAPI (blue) in the hippocampus CA1 and cortex of 7-month-old male tau^P301S and tau^P301S;usp11⁻ mice. (N) Graphs show quantification of pS199/202-tau intensity in the hippocampus and cortex of male mice (2-tailed t-test, Hippo: t=6.458, df=107, ****P<0.0001; Cortex: 2-tailed t-test, t=11.99, df=99, ****P<0.0001; n=40-63 images from 5-8 sections/mouse and 5 mice/genotype). (O) Graphs show quantification of pS262-tau intensity in the hippocampus and cortex of 7-month-old male mice (2-tailed t-test: t=5.058, df=98, ****P<0.0001 (Hippo); t=2.371, df=98, *P<0.05 (Cortex), n=50 images from 10 sections/mouse and 5 mice/genotype).

6

Figure S6 Usp11 elimination mitigates neuroinflammation in tau^P301S mice and causes sex-biased Ac-tau reduction and genetic compensation, related to Figure 5

All data in graphs are represented as means ± SEM normalized to corresponding controls. (A,B) Representative confocal images of IL-1β staining (red) and DAPI (blue) in female (A) and male (B) mouse hippocampus. DG=dentate gyrus. (C,D) Graphs show quantification of IL-1β intensity in DG-CA2 hippocampal region of 7-month old female and male mice (1-way ANOVA, Female: F(2,117)=60.2, P<0.0001; Male: F(2,123)=18.83, P<0.0001; posthoc Tukey: **P<0.01, ***P<0.001, ****P<0.0001, n=40-42 images from 10-11 sections/mouse and 4 mice/genotype). (E) Representative images of GFAP (green), Iba1 (red), and DAPI (blue) staining in the hippocampus of 7-month male mice. Scale bar=100μm. (F) Graphs show quantification of GFAP and Iba-1 intensity in the hippocampus of male mice (GFAP: 1-way ANOVA, F(2,109)=10.27, P<0.0001, posthoc Tukey, **P=0.0088, ****P<0.0001; n=32-42 images, 3-5 sections/mouse and 4-5 mice/genotype; Iba1: 1-way ANOVA, F(2,84)=13.70, P<0.0001, posthoc Tukey, ***P=0.0008, ****P<0.0001; n=27-31 images, 3-5 sections/mouse and 4-5 mice/genotype). (G) Blots of RIPA-soluble (sol) and insoluble (insol) tau, Ac-tau, and actin in the cortex of the indicated mice. M=male, F=female. *Asterix indicates the same male mice lysates in Figure 5J used for normalization between gels. (H) Graphs show quantification of soluble and insoluble tau in female mice from Figure S6G (2-tailed t-test, *P<0.05, ***P<0.001, n=4 mice/genotype). (I) Graphs show quantification of soluble and insoluble tau in female and male mice from Figures S6G and S6L (2-tailed t-test, **P<0.01, ***P<0.001, n=4 mice/sex, M=male, F=female). (J) Graphs show insoluble Ac-tau normalized to insoluble total tau in male and female mice from Figures 5J, S6G, and S6L. (K) Graphs show insoluble Ac-tau normalized to insoluble total tau in female mice from Figures 5J and S6G. (L) Blots of RIPA-soluble and insoluble USP11, tau, Ac-tau (Ac274 or Ac281), and actin from the cortex of the indicated male mice. *Asterisk indicates the same male mice lysates in Figures 5J and S6G used for normalization between gels. (M,N) Graphs show quantification of soluble and insoluble Ac-tau in male mice from Figure S6L (2-tailed t-test, *P<0.05). (O) Graphs show insoluble Ac-tau normalized to insoluble total tau in male mice from Figure S6L (2-tailed t-test, **P<0.01). (P,Q) Graphs show the expression of USP transcripts normalized to GAPDH mRNA determined by qRT-PCR from RNA isolated from the frontal cortex of (P) 7-month-old male WT and usp11⁻ mice and (Q) 7-month-old female WT and usp11^−/− mice. Multiple t-tests are corrected by 2-stage Benjamini, Krieger and Yekuteli method with 1% false discovery rate (FDR) (*q<0.01, **q<0.005, ***q<0.0005, #usp11 KO, unmarked bars = not significant after correction for multiple t-tests, n=4 mice/genotype).

7

Figure S7 Genetic loss of usp11 does not improve spatial memory in male tau^P301S mice, related to Figure 7

All data in graphs are represented as means ± SEM. (A) Graph shows time to find the hidden platform during the training phase in male mice (RM 1-way ANOVA: not significant; n=13 WT, n=19 tau^P301S, n=12 tau^P301S;usp11⁻). (B) Graphs show time to target zone (sec) and swim speed (m/s for 1 min) during the probe test on day 6 with the hidden platform removed (time to target zone: 1-way ANOVA: ns=not significant; swim speed: 1-way ANOVA: not significant; n=13 WT, n=19 tau^P301S, n=12 tau^P301S;usp11⁻).

Highlights.

• X-linked USP11 deubiquitinates tau, enhancing its acetylation and aggregation

• USP11 escapes X-inactivation leading to elevated expression in females than in males

• USP11 levels correlate strongly with tau brain pathology in females but not males

• Elimination of usp11 protects females from tau pathology and cognitive impairment