429
Journal of Alzheimer’s Disease 40 (2014) 429–442 DOI 10.3233/JAD-131358 IOS Press
Altered Function of Hippocampal CA1 Pyramidal Neurons in the rTg4510 Mouse Model of Tauopathy Nils Ole Dalby1,∗ , Christiane Volbracht, Lone Helboe, Peter Hjørringaard Larsen, Henrik Sindal Jensen, Jan Egebjerg and Anders Brandt Elvang Neuroscience Drug Discovery DK, H. Lundbeck A/S, Valby, Denmark Handling Associate Editor: Ottavio Arancio
Accepted 3 December 2013
Abstract. The formation of neurofibrillary tangles from the assembly of hyperphosphorylated tau leads to dendritic and axonal instability, synaptic degeneration, and neuronal loss. To understand the early physiological consequences of aberrant tau expression, we characterized the physiology of CA1 pyramidal neurons in rTg4510 female mice and non-transgenic (wt) littermate controls. We studied mice at the age of 10–12 weeks where only minimal hyperphosphorylated pretangle tau was present, and 22–24 weeks old mice with significant neurofibrillary tangle pathology. Our electrophysiological analysis included input–output relation, paired-pulse facilitation, and whole cell patch-clamp recordings of neurons to measure action potential threshold and action potential properties, chord-conductance, and characterization of AMPA receptor mediated synaptic transmission. We found that the input–output relation in field (excitatory postsynaptic potentials, EPSP) and whole cell recordings (excitatory postsynaptic currents, EPSC) were impaired in rTg4510 mice compared to wt controls at both ages. We measured a diminished tail current charge after depolarizing voltage input in rTg4510 mice compared to wt in both young and aged mice. Additionally, mini-EPSC properties (peak and decay time) were essentially similar between genotypes and age groups investigated. Surprisingly, in the 22–24 week old group, the mini-EPSC frequency was significantly increased (interevent interval 0.8 ± 0.1 in wt compared to 0.3 ± 0.1 in rTg4510 mice). These data indicate that the developmentally regulated expression of human P301L tau in CA1 pyramidal neurons coincide with changes in neuronal excitability but also that significant presynaptic changes occur late during the progression of tau pathology in this mouse model. Keywords: CA1, excitatory postsynaptic currents, hippocampus, postsynaptic, presynaptic, readily releasable pool, rTg4510 model, tauopathy
INTRODUCTION Conformational changes of the microtubule associated protein tau, induced by either mutations in the MAPT gene encoding tau, hyperphosphoryla1 Department of Pharmacology and Pharmacotherapy, University of Copenhagen, Copenhagen, Denmark. ∗ Correspondence to: Nils Ole Dalby, Department of Pharmacology and Pharmacotherapy, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark. Tel.: +45 2285 9920; Fax: +45 3533 6020; E-mail: [email protected].
tion, or other mechanisms, can lead to the formation of intraneuronal accumulation of fibrillar hyperphosphorylated tau protein into neurofibrillary tangles (NFTs). The occurrence of tau species assembling into NFTs has been observed in many neurodegenerative disorders including frontotemporal dementia with Parkinsonism (FTDP-17), Pick’s disease, corticobasal degeneration, progressive supranuclear palsy, and Alzheimer’s disease (AD) [1]. In AD, the presence of NFTs is the postmortem histopathologic hallmark that correlates best with a decline of cognitive function
ISSN 1387-2877/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved
430
N.O. Dalby et al. / Altered Function of Hippocampal CA1 Pyramidal Neurons
[2]. An animal model to study the progression of tau pathology is the rTg4510 mouse that expresses P301L mutant human tau (P301L-htau) under the inducible control of a CaM kinase II promoter system [3] mainly in the forebrain and in excitatory neurons of hippocampus and cortex [4]. The P301L mutation of tau is commonly seen in FTDP-17 and rTg4510 mice closely mimic features of human tauopathies including progressive NFT formation, neuronal death, and cognitive impairment [5]. However, in this animal model, decline of cognitive function and neuronal loss correlated with the presence of pretangle hyperphosphorylated tau, rather than the associated NFTs [5]. For this reason, the biochemical components in the pathway of abnormal tau species en route to formation of NFTs, appear as more important determinants for the neuronal pathophysiology than the NFT end product itself. A causal relationship between presence of aberrant tau species and changes in neuronal function can be assessed by determining the status of each during progression of the disease with age. In the rTg4510 mouse model, hyperphosphorylated tau and NFT deposits emerge first in cortical regions when the mice reach the age of ∼10–12 weeks [3]. However, it is not clear how neuronal anatomy and electrophysiological properties of hippocampal principal neurons are affected in the progression of tauopathy in this mouse model. Because neuronal dendritic branching and morphology influence the excitability and synaptic integration of neurons [6, 7], we hypothesized that changes in active and passive current-voltage properties as well as synaptic input to CA1 pyramidal neurons could be a component of the pathophysiology as a consequence of the presence of aberrant tau species. Here, we used rTg4510 mice at 10–12 and 22–24 weeks of age to investigate the development in time of histological and functional components of pathophysiology. MATERIALS AND METHODS Animals We used female rTg4510 mice aged 10–12 or 22–24 weeks, which conditionally express P301L mutant human tau [5]. Mice expressing the tTA activator transgenes were maintained on 129S6 background strain (Taconic) and mutant tau responder mice were maintained in the FVB/N background strain (Taconic). Mice were screened by PCR using the primer pairs 5 -GATTAACAGCGCATTAGAGCTG-3 and 5 -GC ATATGATCAATTCAAGGCCGATAAG-3 for tTA activator transgenes and 5 -TGAACCAGGATGGC
TGAGCC-3 and 5 -TTGTCATCGCTTCCAGTCCC CG-3 for mutant tau responder transgenes. Agematched non-transgenic littermates were used as wild type (wt) controls (FVB/129 genetic background). Animals were kept in 12-h dark/12-h light controlled room and had access to food and water ad libitum. For experiments involving doxycycline (Dox) treatment, animals were fed chow containing 200 ppm doxycycline (Harlan, Germany) at the indicated ages. When treatment from conception was applied, parents were fed doxycycline chow. All animal experiments were performed in accordance with the European Communities Council Directive #86/609 and the directives of the Danish National Committee on Animal Research Ethics. Slice preparation After decapitation the brain was rapidly dissected out in ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM): NaCl (126), KCl (2.5), CaCl2 (2), MgCl2 (2), NaHCO3 (26), NaH2 PO4 (1.25), D-glucose (10), ascorbate (0.3), pyruvic acid (1.0), and kynurenic acid (3). Osmolality of ACSF was adjusted to 310 ± 5 mOsm. The brain was glued to the platform of a Leica 1200 VS microtome and cut into 350 m thick horizontal slices, hemisected and stored in carbogenated ACSF at 30°C without kynurenate for 1–5 h. Slices were used for field or whole-cell recordings. Field recordings Slices were positioned in an interface chamber at a flow of regular ACSF of 2.5 ml/min at 34◦ C. Stimulation electrode was a bipolar Teflon coated tungsten electrode positioned in the CA2 stratum radiatum for stimulation of the Schaffer collaterals. Stimulations consisted of a 400 s pulse delivered by a World Precision Instruments (WPI) A365 constant current isolation unit. Excitatory postsynaptic potentials (EPSPs) were evoked at 0.0167 Hz (1 per min) and recorded using a glass electrode filled with ACSF (resistance 2–3 M) and positioned 150–200 m from the stimulating electrode in the CA1 stratum radiatum. Recordings were made by an Axon 402A smartprobe/Cyberamp 380, digitized at 5 kHz, and band pass filtered at 0.1–1 kHz (8-pole Bessel). For construction of full input–output (IO) relation, the current was increased in steps of 20 A. Paired pulse experiments were made using an inter pulse interval of 100 ms at stimulation intensities of 0.4 of max.
N.O. Dalby et al. / Altered Function of Hippocampal CA1 Pyramidal Neurons
Whole cell recordings For recordings of excitatory postsynaptic currents (EPSCs), slices were maintained submerged in a 2 ml chamber perfused with ACSF which were modified to contain Ca/Mg ratio at 3/1 at a flow of 2.5 ml/min at 33–34◦ C and which included 30 M picrotoxin and 50 M D-APV and for mini-EPSCs (mEPSCs) were also included 1 M TTX. Cells were visualized using a custom-built infra-red (IR) video-microscopy system. Somatic whole cell recordings in current- or voltageclamp (CC or VC, respectively) of CA1 pyramidal neurons were made using a Multiclamp 700A amplifier (Axon, Molecular Devices). Recordings were digitized at 20 kHz on a 1320 digidata digitizer (Axon, Molecular Devices) and lowpass filtered (8-pole Bessel) at 3 kHz. For recording of EPSCs or mEPSCs, the intracellular solution contained (in mM): CsCH3 SO3 (135), CsCl (8), NaCl (4), MgCl2 (2), EGTA (0.05), HEPES (10), ATP (2), GTP (0.5), TEA (5), and QX-314 (5), pH adjusted to 7.3 (at 4ºC) and osmolality to 292 mOsm. Patch pipettes had a resistance of 4–5 M using this solution. Recordings in voltage clamp were 70% compensated for series resistance and discarded for further analysis if values for capacitance and series resistance during recordings deviated more than 30% from initial values. When in CC mode, recordings were bridge-balanced and the electrode capacitance compensated. Recordings were not corrected for junction potential. Stimulation of the Schaffer collaterals was done via a bipolar Teflon coated tungsten electrode positioned in the middle of stratum radiatum (between stratum pyramidale and stratum lacunosummoleculare) approximately 100 m from the apical dendrite of the recorded neuron. IO relation was done by varying the duration of the digidata 1320 voltage input to a A365 constant current stimulus isolation unit (WPI) from 60–240 s (20 s steps). When in CC mode for studying action potential characteristics or in VC mode for tail currents, the intracellular solution had CsCH3 SO3 and CsCl replaced with equimolar KCH3 SO3 and KCl, respectively, and TEA and QX314 were omitted. For ramp studies determining action potential (AP) threshold and chord conductance, current was injected from 0 to 300 pA over 2500 ms. For AP characteristics, neurons were depolarized for 200 ms with a rectangular current twice the magnitude of the neuron’s individual AP-threshold value. Tail currents were studied in VC by a protocol jumping from −70 mV (Vh ) to −110 mV for 400 ms followed by a series of jumps from −70 to +40 mV ( 10 mV) for 400 ms before returning to Vh (inter-sweep interval
431
8 s). Passive conductances were also characterized in VC using a ramp moving from −110 to −45 mV in 2800 ms. Data analysis All whole cell recordings were made in pClamp 10 and further analyzed in either Clampfit 10 (Axon, Molecular Devices) or for spontaneous events, in mini Analysis 6.03 (Justin Lee, Synaptosoft). Unless where indicated, we have used two-sample t-test with equal variance assumed and p < 0.05 as level of significance for assessment of genotype and age differences. For detection of EPSCs, an algorithm in mini-analysis was made detecting baseline deflections of 5 * SD of the trace and sampling 8 ms before and 50 ms after detection. Exponential fits to log-binned histograms of all detected events were used to assess the mean inter-event interval (mean inter-event interval = 1/mean frequency) by fitting the data to N = k*exp[ln(x/)–(x/)]. For construction of an average waveform of the synaptic current, events with an inter-event interval
Journal of Alzheimer’s Disease 40 (2014) 429–442 DOI 10.3233/JAD-131358 IOS Press
Altered Function of Hippocampal CA1 Pyramidal Neurons in the rTg4510 Mouse Model of Tauopathy Nils Ole Dalby1,∗ , Christiane Volbracht, Lone Helboe, Peter Hjørringaard Larsen, Henrik Sindal Jensen, Jan Egebjerg and Anders Brandt Elvang Neuroscience Drug Discovery DK, H. Lundbeck A/S, Valby, Denmark Handling Associate Editor: Ottavio Arancio
Accepted 3 December 2013
Abstract. The formation of neurofibrillary tangles from the assembly of hyperphosphorylated tau leads to dendritic and axonal instability, synaptic degeneration, and neuronal loss. To understand the early physiological consequences of aberrant tau expression, we characterized the physiology of CA1 pyramidal neurons in rTg4510 female mice and non-transgenic (wt) littermate controls. We studied mice at the age of 10–12 weeks where only minimal hyperphosphorylated pretangle tau was present, and 22–24 weeks old mice with significant neurofibrillary tangle pathology. Our electrophysiological analysis included input–output relation, paired-pulse facilitation, and whole cell patch-clamp recordings of neurons to measure action potential threshold and action potential properties, chord-conductance, and characterization of AMPA receptor mediated synaptic transmission. We found that the input–output relation in field (excitatory postsynaptic potentials, EPSP) and whole cell recordings (excitatory postsynaptic currents, EPSC) were impaired in rTg4510 mice compared to wt controls at both ages. We measured a diminished tail current charge after depolarizing voltage input in rTg4510 mice compared to wt in both young and aged mice. Additionally, mini-EPSC properties (peak and decay time) were essentially similar between genotypes and age groups investigated. Surprisingly, in the 22–24 week old group, the mini-EPSC frequency was significantly increased (interevent interval 0.8 ± 0.1 in wt compared to 0.3 ± 0.1 in rTg4510 mice). These data indicate that the developmentally regulated expression of human P301L tau in CA1 pyramidal neurons coincide with changes in neuronal excitability but also that significant presynaptic changes occur late during the progression of tau pathology in this mouse model. Keywords: CA1, excitatory postsynaptic currents, hippocampus, postsynaptic, presynaptic, readily releasable pool, rTg4510 model, tauopathy
INTRODUCTION Conformational changes of the microtubule associated protein tau, induced by either mutations in the MAPT gene encoding tau, hyperphosphoryla1 Department of Pharmacology and Pharmacotherapy, University of Copenhagen, Copenhagen, Denmark. ∗ Correspondence to: Nils Ole Dalby, Department of Pharmacology and Pharmacotherapy, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark. Tel.: +45 2285 9920; Fax: +45 3533 6020; E-mail: [email protected].
tion, or other mechanisms, can lead to the formation of intraneuronal accumulation of fibrillar hyperphosphorylated tau protein into neurofibrillary tangles (NFTs). The occurrence of tau species assembling into NFTs has been observed in many neurodegenerative disorders including frontotemporal dementia with Parkinsonism (FTDP-17), Pick’s disease, corticobasal degeneration, progressive supranuclear palsy, and Alzheimer’s disease (AD) [1]. In AD, the presence of NFTs is the postmortem histopathologic hallmark that correlates best with a decline of cognitive function
ISSN 1387-2877/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved
430
N.O. Dalby et al. / Altered Function of Hippocampal CA1 Pyramidal Neurons
[2]. An animal model to study the progression of tau pathology is the rTg4510 mouse that expresses P301L mutant human tau (P301L-htau) under the inducible control of a CaM kinase II promoter system [3] mainly in the forebrain and in excitatory neurons of hippocampus and cortex [4]. The P301L mutation of tau is commonly seen in FTDP-17 and rTg4510 mice closely mimic features of human tauopathies including progressive NFT formation, neuronal death, and cognitive impairment [5]. However, in this animal model, decline of cognitive function and neuronal loss correlated with the presence of pretangle hyperphosphorylated tau, rather than the associated NFTs [5]. For this reason, the biochemical components in the pathway of abnormal tau species en route to formation of NFTs, appear as more important determinants for the neuronal pathophysiology than the NFT end product itself. A causal relationship between presence of aberrant tau species and changes in neuronal function can be assessed by determining the status of each during progression of the disease with age. In the rTg4510 mouse model, hyperphosphorylated tau and NFT deposits emerge first in cortical regions when the mice reach the age of ∼10–12 weeks [3]. However, it is not clear how neuronal anatomy and electrophysiological properties of hippocampal principal neurons are affected in the progression of tauopathy in this mouse model. Because neuronal dendritic branching and morphology influence the excitability and synaptic integration of neurons [6, 7], we hypothesized that changes in active and passive current-voltage properties as well as synaptic input to CA1 pyramidal neurons could be a component of the pathophysiology as a consequence of the presence of aberrant tau species. Here, we used rTg4510 mice at 10–12 and 22–24 weeks of age to investigate the development in time of histological and functional components of pathophysiology. MATERIALS AND METHODS Animals We used female rTg4510 mice aged 10–12 or 22–24 weeks, which conditionally express P301L mutant human tau [5]. Mice expressing the tTA activator transgenes were maintained on 129S6 background strain (Taconic) and mutant tau responder mice were maintained in the FVB/N background strain (Taconic). Mice were screened by PCR using the primer pairs 5 -GATTAACAGCGCATTAGAGCTG-3 and 5 -GC ATATGATCAATTCAAGGCCGATAAG-3 for tTA activator transgenes and 5 -TGAACCAGGATGGC
TGAGCC-3 and 5 -TTGTCATCGCTTCCAGTCCC CG-3 for mutant tau responder transgenes. Agematched non-transgenic littermates were used as wild type (wt) controls (FVB/129 genetic background). Animals were kept in 12-h dark/12-h light controlled room and had access to food and water ad libitum. For experiments involving doxycycline (Dox) treatment, animals were fed chow containing 200 ppm doxycycline (Harlan, Germany) at the indicated ages. When treatment from conception was applied, parents were fed doxycycline chow. All animal experiments were performed in accordance with the European Communities Council Directive #86/609 and the directives of the Danish National Committee on Animal Research Ethics. Slice preparation After decapitation the brain was rapidly dissected out in ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM): NaCl (126), KCl (2.5), CaCl2 (2), MgCl2 (2), NaHCO3 (26), NaH2 PO4 (1.25), D-glucose (10), ascorbate (0.3), pyruvic acid (1.0), and kynurenic acid (3). Osmolality of ACSF was adjusted to 310 ± 5 mOsm. The brain was glued to the platform of a Leica 1200 VS microtome and cut into 350 m thick horizontal slices, hemisected and stored in carbogenated ACSF at 30°C without kynurenate for 1–5 h. Slices were used for field or whole-cell recordings. Field recordings Slices were positioned in an interface chamber at a flow of regular ACSF of 2.5 ml/min at 34◦ C. Stimulation electrode was a bipolar Teflon coated tungsten electrode positioned in the CA2 stratum radiatum for stimulation of the Schaffer collaterals. Stimulations consisted of a 400 s pulse delivered by a World Precision Instruments (WPI) A365 constant current isolation unit. Excitatory postsynaptic potentials (EPSPs) were evoked at 0.0167 Hz (1 per min) and recorded using a glass electrode filled with ACSF (resistance 2–3 M) and positioned 150–200 m from the stimulating electrode in the CA1 stratum radiatum. Recordings were made by an Axon 402A smartprobe/Cyberamp 380, digitized at 5 kHz, and band pass filtered at 0.1–1 kHz (8-pole Bessel). For construction of full input–output (IO) relation, the current was increased in steps of 20 A. Paired pulse experiments were made using an inter pulse interval of 100 ms at stimulation intensities of 0.4 of max.
N.O. Dalby et al. / Altered Function of Hippocampal CA1 Pyramidal Neurons
Whole cell recordings For recordings of excitatory postsynaptic currents (EPSCs), slices were maintained submerged in a 2 ml chamber perfused with ACSF which were modified to contain Ca/Mg ratio at 3/1 at a flow of 2.5 ml/min at 33–34◦ C and which included 30 M picrotoxin and 50 M D-APV and for mini-EPSCs (mEPSCs) were also included 1 M TTX. Cells were visualized using a custom-built infra-red (IR) video-microscopy system. Somatic whole cell recordings in current- or voltageclamp (CC or VC, respectively) of CA1 pyramidal neurons were made using a Multiclamp 700A amplifier (Axon, Molecular Devices). Recordings were digitized at 20 kHz on a 1320 digidata digitizer (Axon, Molecular Devices) and lowpass filtered (8-pole Bessel) at 3 kHz. For recording of EPSCs or mEPSCs, the intracellular solution contained (in mM): CsCH3 SO3 (135), CsCl (8), NaCl (4), MgCl2 (2), EGTA (0.05), HEPES (10), ATP (2), GTP (0.5), TEA (5), and QX-314 (5), pH adjusted to 7.3 (at 4ºC) and osmolality to 292 mOsm. Patch pipettes had a resistance of 4–5 M using this solution. Recordings in voltage clamp were 70% compensated for series resistance and discarded for further analysis if values for capacitance and series resistance during recordings deviated more than 30% from initial values. When in CC mode, recordings were bridge-balanced and the electrode capacitance compensated. Recordings were not corrected for junction potential. Stimulation of the Schaffer collaterals was done via a bipolar Teflon coated tungsten electrode positioned in the middle of stratum radiatum (between stratum pyramidale and stratum lacunosummoleculare) approximately 100 m from the apical dendrite of the recorded neuron. IO relation was done by varying the duration of the digidata 1320 voltage input to a A365 constant current stimulus isolation unit (WPI) from 60–240 s (20 s steps). When in CC mode for studying action potential characteristics or in VC mode for tail currents, the intracellular solution had CsCH3 SO3 and CsCl replaced with equimolar KCH3 SO3 and KCl, respectively, and TEA and QX314 were omitted. For ramp studies determining action potential (AP) threshold and chord conductance, current was injected from 0 to 300 pA over 2500 ms. For AP characteristics, neurons were depolarized for 200 ms with a rectangular current twice the magnitude of the neuron’s individual AP-threshold value. Tail currents were studied in VC by a protocol jumping from −70 mV (Vh ) to −110 mV for 400 ms followed by a series of jumps from −70 to +40 mV ( 10 mV) for 400 ms before returning to Vh (inter-sweep interval
431
8 s). Passive conductances were also characterized in VC using a ramp moving from −110 to −45 mV in 2800 ms. Data analysis All whole cell recordings were made in pClamp 10 and further analyzed in either Clampfit 10 (Axon, Molecular Devices) or for spontaneous events, in mini Analysis 6.03 (Justin Lee, Synaptosoft). Unless where indicated, we have used two-sample t-test with equal variance assumed and p < 0.05 as level of significance for assessment of genotype and age differences. For detection of EPSCs, an algorithm in mini-analysis was made detecting baseline deflections of 5 * SD of the trace and sampling 8 ms before and 50 ms after detection. Exponential fits to log-binned histograms of all detected events were used to assess the mean inter-event interval (mean inter-event interval = 1/mean frequency) by fitting the data to N = k*exp[ln(x/)–(x/)]. For construction of an average waveform of the synaptic current, events with an inter-event interval
