Tau in cerebrospinal fluid induces neuronal hyperexcitability and alters hippocampal theta oscillations

Abstract

Alzheimer's disease (AD) and other tauopathies are characterized by the aggregation of tau into soluble and insoluble forms (including tangles and neuropil threads). In humans, a fraction of both phosphorylated and non-phosphorylated N-terminal to mid-domain tau species, are secreted into cerebrospinal fluid (CSF). Some of these CSF tau species can be measured as diagnostic and prognostic biomarkers, starting from early stages of disease. While in animal models of AD pathology, soluble tau aggregates have been shown to disrupt neuronal function, it is unclear whether the tau species present in CSF will modulate neural activity. Here, we have developed and applied a novel approach to examine the electrophysiological effects of CSF from patients with a tau-positive biomarker profile. The method involves incubation of acutely-isolated wild-type mouse hippocampal brain slices with small volumes of diluted human CSF, followed by a suite of electrophysiological recording methods to evaluate their effects on neuronal function, from single cells through to the network level. Comparison of the toxicity profiles of the same CSF samples, with and without immuno-depletion for tau, has enabled a pioneering demonstration that CSF-tau potently modulates neuronal function. We demonstrate that CSF-tau mediates an increase in neuronal excitability in single cells. We then observed, at the network level, increased input-output responses and enhanced paired-pulse facilitation as well as an increase in long-term potentiation. Finally, we show that CSF-tau modifies the generation and maintenance of hippocampal theta oscillations, which have important roles in learning and memory and are known to be altered in AD patients. Together, we describe a novel method for screening human CSF-tau to understand functional effects on neuron and network activity, which could have far-reaching benefits in understanding tau pathology, thus allowing for the development of better targeted treatments for tauopathies in the future.

Keywords: Cerebrospinal fluid; Electrophysiology; Tau; Tauopathy; Theta oscillations.

Fig. 1

Immunodepletion of pooled CSF to remove tau using a combination of antibodies targeting the N terminus, mid region, and C terminus of tau. a. Pooled human CSF from multiple de-identified patients (CSF-tau) was divided into 3 aliquots. An aliquot of the pooled CSF-tau sample was immuno-depleted for tau by the simultaneous application of the monoclonal antibodies Tau12, HT7 and TauAB (epitopes: amino acids 6–18, 159–163 and 425–441, respectively) that together cover nearly the entire tau-441 protein sequence using published protocols [79], this is referred to herein as CSF-tau-depleted. The second aliquot remained the full sample, referred to as CSF-tau. The final aliquot underwent the same immunodepletion protocol with different antibodies (anti-mouse IgG and neurogranin (H-6) antibody) to provide a control for the tau immunodepletion, referred to as CSF-mock-depletion. b, To verify the removal of tau after the immuno-depletion in aliquot 1, t-tau assay from Quanterix, p-tau181 [75] and p-tau231 [8] in house assays were used for tau measurements in the CSF-depleted and non-depleted aliquots and compared in parallel against the CSF-mock-depleted sample on a Simoa HD-X instrument (Quanterix, Billerica, MA, USA) following methods originally described in the cited publications. c, The immunodepletion of tau was further validated by comparison to the full CSF sample using fully automated Lumipulse (Fujirebio) platform following published protocols [49]

Fig. 2

CSF-tau enhances hippocampal pyramidal neuronal excitability. a. Representative examples of standard current–voltage responses for slices that have been incubated in control aCSF (light blue; n = 10), CSF-tau (CSF; dark blue; n = 11) or CSF-tau immunodepleted for tau (CSF-depleted; Pink; n = 10). See Materials and Methods for incubation protocols. CSF-tau depolarised the resting membrane potential of the recorded neurons. b, the most negative step from traces in (a), clearly highlighting that CSF-tau also mediates an increase in input resistance. c, CSF-tau incubation resulted in a significant depolarisation of the resting membrane potential (p < 0.0006), which was not observed in the CSF sample immuno-depleted for tau. d. CSF-tau incubation also significantly increased input resistance (p < 0.0020), an effect which was also not observed in the CSF sample that was immunodepleted for tau. e. Representative example of membrane-potential responses to naturalistic current injection for each of the three conditions. f. CSF-tau significantly increased the firing rate (a correlate of neuronal excitability; p < 0.0055). No change was observed with the CSF-tau sample immunodepleted for tau compared to control. g. The rheobase current (minimal current to evoke an AP) was determined by injecting a current ramp (− 50 to 200 pA) and measuring the minimum current required to fire an action potential. h. CSF-tau incubation significantly decreased the rheobase (p = 0.0093). Panels a, b, e and g show representative example traces and  c, d, f and h show the mean data and SEM, with individual datapoints overlaid

Fig. 3

CSF-tau enhances basal synaptic transmission and synaptic plasticity. a. Graph plotting mean fEPSP slope against stimulus strength for control (n = 10 slices), CSF-tau (n = 11 slices) and CSF-tau-depleted (n = 10 slices). b, Superimposed fEPSP waveforms at increasing stimulus strengths (0.5 to 5 V) for each of the three conditions. CSF-tau significantly enhanced fEPSP slope. c, Graph plotting mean paired pulse ratio against interval for all three conditions. CSF-tau significantly enhanced paired pulse facilitation. d, Representative example traces of fEPSP waveforms for the 200 ms interval for all three conditions. The first fEPSPs have been normalised in this figure so that facilitation can be compared across conditions e, Graph plotting mean normalised (to the baseline) fEPSP slope against time for all three conditions. After a 20-min baseline, LTP was induced by HFS. Inset, example fEPSP waveforms before and after LTP induction (average of waveforms at 75–80 min). The mean potentiation was enhanced in CSF-tau. All Data is represented as Mean ± SEM

Fig. 4

CSF-tau increases the frequency and amplitude of AMPA receptor -mediated mEPSCs. a, Representative traces (10 s duration) from control, CSF-tau and CSF-tau depleted incubated slices, demonstrating the significant difference in mEPSC frequency and amplitude. b, Cumulative frequency plot of mEPSC interval (between events). c, Mean mEPSC interval for each slice is plotted. CSF-tau decreases the interval between mEPSCs, representing an increase in frequency, compared to control (p = 0.0004). d, Cumulative frequency plot of mEPSC amplitude. e, Mean mEPSC amplitude for each slice is plotted. CSF-tau significantly enhances the amplitude of mEPSCs compared to control (p = 0.0225) and CSF-tau-depleted (p = 0.0026). f, g, The average mEPSC waveforms for each recording were analysed for kinetics (10–90% for the rise time and exponential fit for the decay time). An example is shown for control (f) and CSF-tau (g), and the exponential fit is overlaid in black. Inset, higher magnification to show the decay fit. Data is represented as Mean ± SEM

Fig. 5

CSF-tau alters the generation and amplitude of theta oscillations. Theta oscillations in the CA3 region of the hippocampus were recorded in an interface chamber (see methods for details). Bath application of carbachol (50 µM) in this study induced robust, reliable theta oscillations across all three experimental conditions (a, b, c). Left, Representative examples of baseline and theta oscillations for each of the three conditions. Right, Power spectrum from the representative example confirming the oscillations to be in the theta range (4–7 Hz). d, mean power spectrums for each of the conditions. CSF-tau incubated slices had significantly stronger oscillatory power compared to control or CSF-tau-depleted incubated slices (P < 0.0001). Inset, higher magnification of control and CSF-tau-depleted traces. e, CSF-tau incubated slices also showed a significantly quicker onset of oscillatory activity relative to control (p = 0.005). Data is presented as Mean ± SEM