Optogenetic activation of SST-positive interneurons restores hippocampal theta oscillation impairment induced by soluble amyloid beta oligomers in vivo

Abnormal accumulation of amyloid β oligomers (AβO) is a hallmark of Alzheimer’s disease (AD), which leads to learning and memory deficits. Hippocampal theta oscillations that are critical in spatial navigation, learning and memory are impaired in AD. Since GABAergic interneurons, such as somatostatin-positive (SST+) and parvalbumin-positive (PV+) interneurons, are believed to play key roles in the hippocampal oscillogenesis, we asked whether AβO selectively impairs these SST+ and PV+ interneurons. To selectively manipulate SST+ or PV+ interneuron activity in mice with AβO pathology in vivo, we co-injected AβO and adeno-associated virus (AAV) for expressing floxed channelrhodopsin-2 (ChR2) into the hippocampus of SST-Cre or PV-Cre mice. Local field potential (LFP) recordings in vivo in these AβO–injected mice showed a reduction in the peak power of theta oscillations and desynchronization of spikes from CA1 pyramidal neurons relative to theta oscillations compared to those in control mice. Optogenetic-activation of SST+ but not PV+ interneurons in AβO–injected mice fully restored the peak power of theta oscillations and resynchronized the theta spike phases to a level observed in control mice. In vitro whole-cell voltage-clamp recordings in CA1 pyramidal neurons in hippocampal slices treated with AβO revealed that short-term plasticity of SST+ interneuron inhibitory inputs to CA1 pyramidal neurons at theta frequency were selectively disrupted while that of PV+ interneuron inputs were unaffected. Together, our results suggest that dysfunction in inputs from SST+ interneurons to CA1 pyramidal neurons may underlie the impairment of theta oscillations observed in AβO-injected mice in vivo. Our findings identify SST+ interneurons as a target for restoring theta-frequency oscillations in early AD.

To address this, we selectively manipulated the activity in SST+ or PV+ interneurons using optogenetics in mice with AβO pathology in vivo. We co-injected AβO (Brouillette et al., 2012;Cetin and Dincer, 2007;Faucher et al., 2015;Kim et al., 2014b) and adenoassociated virus, AAV (AAV5-EF1a-DIO-hChR2-mCherry) for expressing floxed channelrhodopsin2 (ChR2) in SST+ and PV+ interneurons in SST-Cre mice or PV-Cre mice, respectively (Yizhar et al., 2011). Local field potential (LFP) recordings in vivo revealed that reduction of theta oscillation power and desynchronized spikes of CA1 pyramidal neurons relative to theta oscillation in AβO-injected mice were restored to the level observed in the control mice by optogenetic activation of SST+ interneurons with blue light but not by PV+ interneurons. In vitro whole-cell voltage-clamp recordings in CA1 pyramidal neurons in hippocampal slices treated with AβO revealed disruption of short-term plasticity of SST's inhibitory inputs to pyramidal neurons at theta frequency.
These results suggest a synaptic dysfunction of SST+ interneuron inputs to CA1 pyramidal neurons may underpin impairment of theta oscillation in AβO-injected mice in vivo.

Animals
SST-IRES-Cre (Jackson Laboratory, stock #013044, (Taniguchi et al., 2011)) knock-in mice and B6 PV-Cre (Jackson Laboratory, stock #017320) knock-in mice were housed in a temperature-controlled environment under a 12h light/dark cycle with the guideline of the Gyerim Experimental Resource Center of Korea University. Food and water were provided ad libitum. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Korea University (KUIACUC-2017-112).

Soluble AβO preparation
Soluble AβO was prepared following methods described by Lambert and colleagues (Lambert et al., 1998) with a slight modification (Wang et al., 2008). In brief, soluble
SST-Cre and PV-Cre mice were deeply anesthetized with isoflurane and head-fixed into a stereotaxic frame (Stoelting Co.). The position of the hippocampal CA1 region (2.7 mm posterior, 2.7 mm lateral and 1.85 mm ventral to the bregma) was marked on the skull and thinned using a high-speed drill (K.1070, Foredom Electric Co.) to make a small craniotomy for injection. 3 μ L of AβO 1-42 (10 μ M) (Kim et al., 2014a) and 1 μ L of AAV5-EF1a-DIO-hChR2-mCherry were delivered into the hippocampal CA1 region through 5 μ L micro-needle (Hamilton Company) connected to a motorized stereotaxic injector (Stoelting Co.) at the rate of 0.3 μ L/min and 0.1 μ L/min, respectively. Body temperature was maintained using a heating pad during surgery. After the injection, the needle was left in the brain for > 5 min to minimize backflow of the injection. At least three weeks of recovery period was allowed after surgery before in vivo recording to ensure the proper expression of the virus as well as induction of AβO-mediated oscillogenesis as observed in other study (Villette et al., 2010).

Hippocampal in vivo recording
Mice were deeply anesthetized with a mixture of ketamine (75-100 mg/kg) and medetomidine (1 mg/kg) and head-fixed into a stereotaxic frame (Stoelting Co.). In vivo local field potential (LFP) recordings were performed using a 32-channel silicon probe (A1x32-Poly2-5mm-50s-177, Neuronexus) inserted into the hippocampal CA1 region (2.7 mm posterior, 2.7 mm lateral to bregma, and 1.85 mm ventral to the pia). A reference metal wire was inserted into cerebellum. Body temperature was monitored and maintained at ~37 using a DC temperature control system (FHC Inc.  2B). To verify the position of the implanted probe in the hippocampus and identify the location of each electrode of the probe in each hippocampal layer, the electrode was removed after in vivo recording and coated with a fluorescent dye (Alexa 594, Fig. 1E), after which it was re-implanted at the same in vivo recording coordination for fluorescent imaging.

Data Analysis of in vivo LFPs and spikes
Acquired spontaneous LFP signals were analyzed using customized protocols in MATLAB (R2018a). LFP signals were analyzed using a common average filter and 3 rd order butterworth bandpass filter (3 -12 Hz, Fig. 1F and Fig. 2D), after which fast To investigate the phase relationship between CA1 pyramidal neurons and thetafrequency oscillation in each AβO-injected SST-Cre and PV-Cre mice, we first extracted single unit spikes by performing spike detection and sorting using the Klusta-suite software (Rossant et al., 2016). Spike events were detected when the 300-5,000 Hz bandpass filtered signal exceeds 4.5 times the standard deviation. Spike waveform, inter-spike intervals (ISIs), and the shape of auto-correlogram of spike times were examined to refine single unit identification (Hill et al., 2011). We classified putative CA1 pyramidal neurons and putative CA1 interneurons by calculating the asymmetry index ([(b -a)/(b + a)]) from the spike waveform (Fig. 3A). Neurons that were located on the right of the decision boundary ([(b -a)/(b + a)] = 2*c -1.2) were considered putative CA1 pyramidal neurons (Fig. 3A). To confirm successful optogenetic activation of SST+ or PV+ interneurons with blue light, we calculated the peri-stimulus time histogram (PSTH, 100 ms bin) before and after light stimulation (Fig. 3B). To investigate the spike phases of CA1 pyramidal neurons relative to theta oscillation, the spike timings of CA1 pyramidal neurons were extracted and the corresponding phase of the theta oscillation for each spike time was analyzed using a Hilbert transform (Khodagholy et al., 2017;Nakazono et al., 2017;Tort et al., 2010). The probability distribution of spike phases was normalized by dividing the bins (10 degree) by the total number of spikes (Fig. 3D, G). Each spike phase probability was vectorized on the polar coordinate. Average of spike phases was analyzed using the Circular Statistics Toolbox in MATLAB (Berens, 2009) (Fig. 3D, G).

1
To quantify the strength of phase-locking, we obtained the vector of each spike phases of each CA1 pyramidal neurons and calculated the mean vector length (Fig. 3E, H).

Statistical analysis
All data are represented as mean ± standard error of the mean (SEM). Statistical analysis was performed by one-way and two-way ANOVA followed by Tukey's post hoc test. p values less than 0.05 were considered statistically significant. Statistical significance of spike phases were tested using Watson-Williams multi-sample circular test (Zar, 2009).
What could be the cellular mechanism underlying this restoration of synchrony? Since light stimulation of SST+ interneurons could selectively restore theta oscillation as well as the synchronization of theta spike phases in AβO-injected SST-Cre mice (Fig. 3) shown in statistical analysis (Control (n = 4) vs. AβO (n = 7), N.S. p > 0.05, two-way ANOVA; Fig. 4F). These results demonstrate for the first time that enhancement of STD in SST+ interneuronal input to CA1 pyramidal neurons in AβO-injected mice may underlie the desynchronization of CA1 pyramidal neuron output, leading to a reduction in theta oscillation power in vivo.

Discussion
Here we show that optogenetic-activation of SST+ interneurons can selectively restore AβOinduced impairment of hippocampal theta oscillations and aid in resynchronizing theta spike phases of CA1 pyramidal neurons in mice. Such interneuron subtype-specific restoration of theta oscillation was possible since AβO caused synaptic dysfunction specifically to SST+ interneuron's theta-frequency inhibitory inputs to CA1 pyramidal neurons, identifying SST+ interneurons as a target for restoring theta-frequency oscillations in early AD.
In our study, we injected AβO into the hippocampus to create AβO pathology in vivo, a method adopted in many studies investigating the impact of AβO physiological and cognitive in SST-Cre and PV-Cre mice, we observed a significant reduction in peak power of theta oscillations across all hippocampal sublayers compared to that in control SST-Cre and PV-Cre mice ( Fig. 1-2). These results are consistent with the reduction of theta oscillation peak power observed in AβO-injected mice after three weeks of incubation (Villette et al., 2010).
The major advantage of using the AβO-injection model of AD in mice was that we were able to combined this with optogenetic activation of SST+ (Fig. 1F) and PV+ interneurons ( Fig.   2D) using cre-transgenic mice. Using these mice, we were able to show, for the first time, We found that the spike phase of CA1 pyramidal neurons relative to theta oscillations, which normally occurs before the trough of the theta cycle, was significantly shifted to a later phase after the trough (   o  s  c  i  l  l  a  t  i  o  n  s  a  n  d  r  e  d  u  c  e  s  h  i  p  p  o  c  a  m  p  a  l  e  x  c  i  t  a  b  i  l  i  t  y  i  n  A  l  z  h  e  i  m  e  r  '  s  m  o  d  e  l  .  J  B  i  o  l  C  h  e  m  2  9  3  ,  8  4  6  2  -8  4  7  2  .  N  a  k  a  z  o  n  o  ,  T  .  ,  L  a  m  ,  T  .  N  .  ,  P  a  t  e  l  ,  A  .  Y  .  ,  K  i  t  a  z  a  w  a  ,  M  .  ,  S  a  i  t  o  ,  T  .  ,  S  a  i  d  o  ,  T  .  C  .  ,  a  n  d  I  g  a  r  a  s  h  i  ,  K  .  M  .  (  2  0  1  7  ) . , e t a l .
( 2  0  1  7  )  .  G  a  m  m  a  r  h  y  t  h  m  l  o  w  f  i  e  l  d  m  a  g  n  e  t  i  c  s  t  i  m  u  l  a  t  i  o  n  a  l  l  e  v  i  a  t  e  s  n  e  u  r  o  p  a  t  h  o  l  o  g  i  c  c  h  a  n  g  e  s  a  n  d  r  e  s  c  u  e  s  m  e  m  o  r  y  a  n  d  c  o  g  n  i  t  i  v  e  i  m  p  a  i  r  m  e  n  t  s  i  n  a  m  o  u  s  e  m  o  d  e  l  o  f  A  l  z  h  e  i  m  e  r  '  s  d  i  s  e  a  s  e  .  A  l  z  h  e  i  m  e  r  s  D  e  m  e  n  t  (  N  Y  )   (red, n = 5) and AβO-injected PV-Cre mice with blue light (blue, n = 5). All data represent mean ± SEM. Inset: N.S. p > 0.05, * p < 0.05, ** p < 0.01 and *** p < 0.001, one-way ANOVA followed by Tukey's post hoc test.