Altered ubiquitin signaling induces Alzheimer's disease-like hallmarks in a three-dimensional human neural cell culture model

Abstract

Alzheimer's disease (AD) is characterized by toxic protein accumulation in the brain. Ubiquitination is essential for protein clearance in cells, making altered ubiquitin signaling crucial in AD development. A defective variant, ubiquitin B + 1 (UBB⁺¹), created by a non-hereditary RNA frameshift mutation, is found in all AD patient brains post-mortem. We now detect UBB⁺¹ in human brains during early AD stages. Our study employs a 3D neural culture platform derived from human neural progenitors, demonstrating that UBB⁺¹ alone induces extracellular amyloid-β (Aβ) deposits and insoluble hyperphosphorylated tau aggregates. UBB⁺¹ competes with ubiquitin for binding to the deubiquitinating enzyme UCHL1, leading to elevated levels of amyloid precursor protein (APP), secreted Aβ peptides, and Aβ build-up. Crucially, silencing UBB⁺¹ expression impedes the emergence of AD hallmarks in this model system. Our findings highlight the significance of ubiquitin signalling as a variable contributing to AD pathology and present a nonclinical platform for testing potential therapeutics.

Fig. 1. UBB⁺¹ accumulates in the early stages of AD postmortem tissue.

a Immunofluorescence staining of human tissue (ER = hippocampal complex, including entorhinal cortex) with an antibody against UBB⁺¹ (red). b Quantification of UBB⁺¹-positive aggregates counted in different biological samples taken from dentate gyrus (DG) of patients diagnosed at different Braak stages [n = 3 biologically independent samples] (Supplementary information Table 1) (p = 0.038). c Immunohistochemistry staining of human postmortem tissue (ER) with anti-UBB⁺¹. d Representative immunoblots of proteins isolated from the hippocampus of Apolipoprotein E (ApoE3/4)-TR mice using an anti-UBB⁺¹ antibody. e Quantification of (d) by densitometry and normalized to β-actin [n = 5 mice] (p = 0.00042) f, g Immunofluorescence staining of hippocampal sections of one (f) and 4-month-old (g) ApoE (ApoE3/4)-TR mice, showing UBB⁺¹ is specifically expressed in neurons (MAP2) and not glial cells (GFAP). P-values were determined by unpaired two-tailed Student’s t-test. Error bars represent ± s.d. Images are representative of three independent wells. All experiments were repeated at least twice.Scale bars: 20 µm (a), 50 μm (c), 100 μm (f, g).

Fig. 2. Expression of UBB⁺¹ induces Amyloid beta (Aβ) and phosphorylated tau (p-tau) aggregates in a 3D human neural culture system.

a Schematic of UBB⁺¹ expressing 3D system derived from ReN VM cells. Image was created using BioRender software. b Representative immunofluorescence images of 6-week-old differentiated control, APP^SL-PSNE1^ΔE9 (FAD) and UBB⁺¹ expressing 3D cultures, showing Aβ-positive-aggregates in white arrows, indicating that they are extracellular deposits. c Quantification of the mean total number of extracellular Aβ deposits in 6-week-old control, FAD, and UBB⁺¹ cultures [n = 3 biologically independent samples, the whole well was counted] (FAD/ Ctrl. p = 0.000029; UBB⁺¹ /Ctrl. p = 0.00016). d Immunoblot analysis of SDS-solubilized aggregates isolated from 6-week-old -differentiated control, FAD, and UBB⁺¹ cultures, showing Aβ oligomers. β-tubulin III levels were used as loading control. e ELISA of Aβ42 and Aβ40 in conditioned media collected from 3D 6-week-old control, FAD, and UBB⁺¹ cultures. Aβ42/Aβ40 ratio is displayed [n = 3 biologically independent samples] (FAD/Ctrl. p = 0.007; UBB⁺¹ /Ctrl. p = 0.04). f ELISA of Aβ42 in whole lysates of 3D 8-week-old control, FAD, and UBB⁺¹ cultures. Results are displayed as Aβ42 normalized to total protein [n = 3 biologically independent samples] (FAD/ Ctrl. p = 0.014; UBB⁺¹ /Ctrl. p = 0.043). g Representative immunohistochemistry images of 3D 8-week-old FAD and UBB⁺¹ cultures stained with anti- p-tau. Cells displaying increased p-tau staining are marked by black arrowheads. h Quantification of the mean number of p-tau positive-stained cells [n = 3 biologically independent samples, whole well was counted] (FAD/ Ctrl. p = 0.004; UBB⁺¹ /Ctrl. p = 0.007). i Sarkosyl-soluble fractions extracted from 3D FAD and UBB⁺¹ -8-week-old thick cultures immunoblotted for tau or p-tau. Calnexin levels were used as loading control. j Representative image of Cryo-TEM of 14-week-old UBB⁺¹ Sarkosyl-insoluble fraction showing specific binding of tau antibody to aggregates, visualized with 5 nm nano-gold particles. P-values were determined by unpaired two-tailed Student’s t-test. Error bars represent ± s.d. Images are representative of three independent samples. All experiments were repeated at least twice. Scale bars: 20 µm (b), 50 µm (g), and 100 nm (j).

Fig. 3. UBB⁺¹ binds UCH-L1 and increases APP protein levels.

a Representative immunoblot of Aβ from conditioned media collected from HEK293FT cells transfected with either APP^SL-PSNE1^ΔE9 (FAD), UBB⁺¹, or both (top panel). α-tubulin immunoblot in whole cell lysate (WCL) from the same culture (bottom panel). b Representative immunoblot of Aβ from conditioned media collected from HEK293FT cells expressing FAD or UBB⁺¹ or both, with or without overexpression of ubiquitin (3xUb). GAPDH immunoblot in WCL from the same culture (bottom panel). c Co-IP was performed with an anti-UBB⁺¹ antibody from HEK293FT WCL expressing UBB⁺¹ or an empty vector, immunoblotted for UCHL-1 and UBB⁺¹. Note the higher migrating form of UBB⁺¹ that has been confirmed to be a ubiquitinated form of UBB⁺¹. d Co-immunoprecipitation of UBB⁺¹ from HEK293FT cells expressing one of three UBB⁺¹ constructs - UBB⁺¹, STREP-UBB⁺¹, or Myc-UBB⁺¹ - and co-transfected with either one copy of ubiquitin or 3 copies of ubiquitin (3xUb). e Immunoblot of APP in HEK293FT cells transfected with FAD and UCHL1 or UBB⁺¹ as noted. Some samples were treated with the UCHL1-specific chemical inhibitor LDN57444. f Ubiquitin immunoblot of APP IP from HEK293FT cells transfected with FAD and either UBB⁺¹ or UCHL1 as noted. WCL was immunoblotted as indicated. GAPDH was used as a loading control. All blots in a–f were repeated at least twice. g Representative immunofluorescence images of UCH-L1 expression (top panel) and Aβ staining (3D6; white arrows, bottom panel) in 4-week-old 3D UBB⁺¹ and UBB⁺¹/UCH-L1 cultures. At 3 weeks, half of the UBB⁺¹ cultures were transduced with UCHL1 lentiviral particles (right) and the other half with mock particles (left). h Quantification of the mean total number of extracellular Aβ+ deposits per well in UBB⁺¹ and UBB⁺¹/UCH-L1 cultures [n = 3 independent wells, whole well counted] (Ctrl./UBB⁺¹ + UCHL1 p = 0.016; UBB⁺¹ /UBB⁺¹ + UCHL1 p = 0.0004 Ctrl./UBB⁺¹ p = 0.0001). P-values were determined by one way ANOVA test. Error bars represent ± s.d. Scale bars: 50 µm (g, top panel), 20 µm (g, bottom panel).

Fig. 4. Silencing UBB⁺¹ expression decreases Aβ and p-tau aggregates in a 3D human neuronal culture.

a Immunofluorescence of Aβ deposits in 6-week- old FAD+sh^SCR or FAD+sh^shUBB+1 (green, MAP2; red, 3D6; arrowheads, extracellular Aβ deposits). b Quantification of extracellular Aβ deposits in FAD+sh^SCR or FAD+sh^shUBB+1 cultures [n = 3 biologically independent samples, the whole well was counted] (p = 0.0006)c Immunoblot of Aβ aggregates from 6-week-old FAD+sh^SCR or FAD+sh^shUBB+1 cultures. d Quantification of Aβ42/Aβ40 by ELISA of conditioned media from 6-week-old FAD+sh^SCR or FAD+sh^shUBB+1 cultures [n = 4 biologically independent samples] (p = 0.00018). e Quantification of Aβ42/Aβ40 by ELISA of lysates from 6-week-old FAD+sh^SCR or FAD+sh^shUBB+1 cultures [n = 4 biologically independent samples] (p = 0.0373). f Immunohistochemistry of 8-week-old FAD+sh^SCR or FAD+sh^shUBB+1 cultures showing p-tau staining (brown, p-tau; arrows indicate cells with high levels of p-tau). g Quantification of p-tau deposits in 8-week-old FAD^SCR or FAD^shUBB+1 cultures. [n = 3 biologically independent samples, the whole well was counted] (p = 0.00007). h Immunoblot of 8-week-old FAD+sh^SCR or FAD+sh^shUBB+1 cultures using anti-p-tau, and anti-tau, Calnexin was used as housekeeping gene. P-values were determined by unpaired two-tailed Student’s t-test where. Error bars represent ± s.d. Images are representative of three independent wells. All experiments were repeated at least twice. Scale bars: 20 µm (a), 50 µm (f).