pS396/pS404 (PHF1) tau vaccine outperforms pS199/pS202 (AT8) in rTg4510 tauopathy model

Abstract

Tauopathies, including Alzheimer's disease (AD) and Frontotemporal Dementia (FTD), are histopathologically defined by the aggregation of hyperphosphorylated pathological tau (pTau) as neurofibrillary tangles in the brain. Site-specific phosphorylation of tau occurs early in the disease process and correlates with progressive cognitive decline, thus serving as targetable pathological epitopes for immunotherapy development. Previously, we developed a vaccine (Qβ-pT181) displaying phosphorylated Thr181 tau peptides on the surface of a Qβ bacteriophage virus-like particle (VLP) that induced robust antibody responses, cleared pathological tau, and rescued memory deficits in a transgenic mouse model of tauopathy. Here we report the characterization and comparison of two additional Qβ VLP-based vaccines targeting the dual phosphorylation sites Ser199/Ser202 (Qβ-AT8) and Ser396/Ser404 (Qβ-PHF1). Both Qβ-AT8 and Qβ-PHF1 vaccines elicited high-titer antibody responses against their pTau epitopes. However, only Qβ-PHF1 rescued cognitive deficits, reduced soluble and insoluble pathological tau, and inflammatory microgliosis in a 4.5-month rTg4510 model of FTD. Both sera from Qβ-AT8 and Qβ-PHF1 vaccinated mice were specifically reactive to tau pathology in human AD post-mortem brain sections. These studies further support the use of VLP-based immunotherapies to target pTau in AD and related tauopathies and provide potential insight into the clinical efficacy of various pTau epitopes in the development of immunotherapies.

Fig. 1. Qβ-AT8 and Qβ-PHF1 VLP conjugation, treatment timeline, and antibody responses.

a A 21-mer pTau peptide containing the Ser199/Ser202 phosphorylation sites (AT8 peptide) or a 25-mer pTau peptide containing the Ser396/Ser404 phosphorylation sites (PHF1 peptide) and a modified Gly-Gly-Cys (GGC) C-terminal sequence were individually conjugated to surface-exposed Lys residues on Qβ bacteriophage VLPs using SMPH crosslinker. b 2-month-old rTg4510 mice received three bi-weekly intramuscular vaccinations with either unconjugated Qβ Control VLP, Qβ-AT8 VLP, or Qβ-PHF1 VLP and were evaluated at 4.5-months for vaccine efficacy. c Upward mobility shift on a denaturing gel shows successful conjugation of AT8 or PHF1 peptides per Qβ coat protein monomer. d Qβ-AT8 and Qβ-PHF1 VLP vaccines elicited significantly elevated serum IgG antibody titers at 6-weeks and e 10-weeks post-vaccination compared to Qβ Control. f Serum IgG antibody titers remain unchanged between the 6- and 10-week timepoints. g Qβ-AT8 and Qβ-PHF1 VLP elicited antibody responses are detectable in hippocampal lysates. Graph displays mean ± SEM. p < 0.05*; p < 0.001***; p < 0.0001****; one-way ANOVA with Dunnett’s multiple comparisons. Qβ Control VLP (n = 5), Qβ-AT8 VLP (n = 7), or Qβ-PHF1 VLP (n = 8). a, b Were created by the authors using BioRender.com.

Fig. 2. Qβ-PHF1 vaccination rescues delay-dependent memory deficits while Qβ-AT8 vaccination does not.

a Novel Object Recognition (NOR) test sample day showed no difference in the time spent with each object for all groups. b. Twenty-four hours later (Test Day), non-transgenic (B6) mice spent significantly more time with the novel object. There was no difference in time spent with the novel object and familiar object for Qβ Control and Qβ-AT8 vaccinated rTg4510 mice, but this impairment was significantly rescued in Qβ-PHF1 vaccinated rTg4510 mice. c Hidden platform trials of the Morris Water Maze (MWM) test show that B6 mice had a shorter latency to find the platform than all rTg4510 vaccine groups. All rTg4510 vaccine groups exhibited hyperactivity based on distance traveled and velocity compared to B6 mice. d-e During the probe trial (platform removed), B6 mice spent significantly more time exploring the target quadrant. Qβ Control rTg4510 mice spent significantly more time in the wrong quadrants. Qβ-AT8 and Qβ-PHF1 vaccinated rTg4510 spent equal time in all quadrants, indicating a potentially milder impairment than Qβ Control mice but incomplete rescue of spatial memory. All graphs display mean ± SEM. Two-way ANOVA with Šidák correction (a, b). Two-way ANOVA with Dunnett’s correction comparing each group to the B6 group on each day (c). Student’s t-test (e). p < 0.05*, p < 0.01**, p < 0.001***. B6 (n = 9), Qβ Control (n = 7), Qβ-AT8 (n = 7), Qβ-PHF1 (n = 8). d Was created by the authors using BioRender.com.

Fig. 3. Qβ-PHF1 vaccination reduces soluble pathological tau in the brains of rTg4510 mice, while Qβ-AT8 vaccination does not.

a Western blot of soluble hippocampal lysates from Qβ Control and Qβ-PHF1 vaccinated rTg4510 mice were evaluated for markers of phosphorylated (AT8, AT180, PHF1) and total tau (Tau5). All samples were run on the same gel. b Compared to Qβ Control vaccinated mice, Qβ-PHF1 vaccinated mice showed a significant reduction in pathological AT180⁺ tau without any reduction in total physiologic tau. c. Immunohistochemistry of brain sections from Qβ Control and Qβ-PHF1 vaccinated rTg4510 mice staining for AT180 tau pathology in the cerebral cortex (CTX) and CA1 hippocampus. d There was a significant reduction in AT180⁺ tau pathology in both the cortex and hippocampus of Qβ-PHF1 vaccinated mice compared to Qβ Control mice (e, f). Western blot of soluble hippocampal lysates from Qβ Control and Qβ-AT8 vaccinated rTg4510 mice showed no significant reductions in pathological tau. All graphs display mean ± SEM. Student’s t-test (b, d, f). p < 0.05*; p < 0.001***; p < 0.0001****. Qβ Control (n = 5), Qβ-AT8 (n = 7), Qβ-PHF1 (n = 8).

Fig. 4. Qβ-PHF1 vaccination preferentially reduces Sarkosyl insoluble tau in rTg4510 mice and elicits antibodies specific to human pathological tau.

a Western blot of Sarkosyl soluble and insoluble tau from Qβ-PHF1 and Qβ Control vaccinated mouse hippocampal samples b showing a significant reduction in the ratio of Sarkosyl insoluble/soluble AT8⁺ and Tau12⁺ tau in Qβ-PHF1 vaccinated mice compared to Qβ Control. c, d Gallyas silver impregnation of the cortex and CA1 hippocampal sections from Qβ-PHF1 and Qβ Control vaccinated mice showed a significant reduction in insoluble tau and dramatic clearance of NFTs of Qβ-PHF1 vaccinated mice compared to the Qβ Control group. e Qβ-PHF1 immune sera stained somatodendritic NFTs, pre-tangles, neuritic plaques, neuropil threads, and ghost tangles in human autopsy AD, but not non-AD, brain tissue. Qβ Control sera did not show any specific staining. AT8 antibody staining was used as a positive control. All graphs show mean ± SEM. Student’s t-test (b, d). p < 0.05*, p < 0.01**. b Qβ Control (n = 5), Qβ-PHF1 (n = 8). d Qβ Control (n = 5), Qβ-PHF1 (n = 5).

Fig. 5. Qβ-PHF1 vaccination reduces inflammatory microgliosis in rTg4510 mice.

a, b IHC of Qβ Control and Qβ-PHF1 vaccinated rTg4510 mouse CA1 hippocampal brain sections using Iba1 and CD45 microglial markers showed a reduction in ameboid microglial morphology and overall Iba1 and CD45 staining in Qβ-PHF1 vaccinated mice compared to the Qβ Control group. c, d. Western blot of hippocampal lysates from Qβ Control and Qβ-PHF1 vaccinated mice showed a significant increase in IκB-α levels and pro-caspase-1 with a concomitant decrease in active caspase-1 and mature interleukin-1β cytokine levels in Qβ-PHF1 vaccinated mice compared to Qβ Control. e Gene expression of NF-κB-regulated genes (Nfkb1, Il1b, Il6, and Tnfa) and the microglial activation marker CD45 are reduced by Qβ-PHF1 vaccination compared to Qβ Control while the general microglial marker TMEM119 is unchanged. All graphs show mean ± SEM. Student’s t-test (b, d, e). p < 0.05*, p < 0.01**, p < 0.001***. b Qβ Control (n = 5), Qβ-PHF1 (n = 6,4). d Qβ Control (n = 5), Qβ-PHF1 (n = 8). E. Qβ Control (n = 4), Qβ-PHF1 (n = 7).