Optogenetic Stimulation of a Hippocampal Engram Activates Fear Memory Recall

Study Overview and Theoretical Context

• Objective: The study aims to determine the sufficiency of activating a sparse population of neurons, which were active during learning, to elicit the behavioral output of a specific memory.
• Historical Context:     * Memory engrams are thought to be encoded by sparse populations of neurons.     * Previous research established necessity: selective ablation or inactivation of these neurons results in reduced expression of the memory.     * This research seeks to prove the cellular basis of a specific fear memory engram by conducting a mimicry experiment.
• Target Region: The dentate gyrus (DG) of the hippocampus.     * The DG is known for its role in discriminating between similar contexts (pattern separation).     * Granule cells in the DG show sparse activity during context exposure ($2-4\%$).     * Evidence shows that different environments or tasks activate different populations of DG granule cells.

Genetic and Optogenetic System Design

• Transgenic Model: c-fos-tTA transgenic mice.     * This model couples the promoter of c-fos (an immediate early gene or IEG and marker of recent activity) to the tetracycline transactivator (tTA).
• Viral Construct: AAV9-TRE-ChR2-EYFP (experimental) or AAV9-TRE-EYFP (control).     * TRE: Tetracycline-responsive element.     * ChR2: Channelrhodopsin-2, a light-gated cation channel.     * EYFP: Enhanced Yellow Fluorescent Protein.
• The Doxycycline (Dox) Mechanism:     * Presence of Dox: Prevents tTA from binding to TRE, inhibiting the expression of ChR2-EYFP.     * Absence of Dox: Training-induced neuronal activity (which induces c-fos) allows tTA to bind to TRE, driving the expression of ChR2-EYFP in the active neurons.
• Targeting Logic: The system allows for selective labeling of neurons active during a specific time window (when the mouse is off-Dox) and subsequent light-based reactivation of those specific cells (while the mouse is on-Dox).

Characterization of ChR2-EYFP Expression

• Expression Timeline:     * On Dox: Complete absence of ChR2-EYFP expression in DG neurons.     * Off Dox (Home Cage): Two days off-Dox induced background expression in home-caged mice.     * Off Dox (Fear Conditioning/No Shock): A significant increase in ChR2-EYFP-positive cells was found after two days off-Dox followed by treatment.     * Total Duration: Expression lasted at least $5\,\text{days}$ and was gone by $30\,\text{days}$.
• Labeling Specificity:     * The vast majority of c-fos-positive cells were also ChR2-EYFP-positive.     * Expression was restricted to excitatory neurons; no overlap was found with GABA-positive inhibitory neurons.     * Sparsity: Proportions of labeled cells for both Fear Conditioned (FC) and No Shock (NS) groups were similar ($2-4\%$).     * Kainic Acid Control: Seizure-induced activity resulted in complete labeling of the DG, proving the sparse labeling in behavioral tests was due to natural low activity rather than low viral infection rates.

In Vivo Electrophysiological Validation

• Apparatus: An optrode (tungsten electrode glued to a $200\,\mu m$ optical fiber) coupled to a $473\,nm$ laser.
• Subjects: Anesthetized, head-fixed c-fos-tTA mice.
• Results:     * Identified $10$ responsive DG neurons from $9$ mice in the ChR2 group.     * Observed a reliable increase in spike probability precisely time-locked to the onset of blue light pulses ($473\,nm$, $0.1\,Hz$, $15\,ms$ pulse duration).     * Neurons responded to $20\,Hz$ stimulation trains; spike probability remained significantly higher than baseline despite a slight decrease over time.     * EYFP Control: No light-responsive cells were found in the $10$ control mice injected with EYFP only.     * Downstream Effects: Optical stimulation of DG neurons led to increased c-fos expression in the CA3 region, comparable to levels induced by natural context exposure.

Behavioral Experiments: Sufficiency of Engram Reactivation

• General Procedure:     1. Habituation: Five days in Context A (white plastic floor, $0.25\%$ benzaldehyde scent) while on-Dox to record basal freezing with light-on/off epochs.     2. Labeling: Two days off-Dox followed by training in Context B (gridded floor, $1\%$ acetic acid scent, triangular roof).     3. Testing: Five days in Context A while on-Dox with alternating $3\text{-minute}$ light-off and light-on epochs.
• Light Stimulation Parameters: $473\,nm$, $20\,Hz$, $15\,ms$ pulses, $9\,mW$ intensity.
• Experimental (Exp) Group Results:     * Showed significantly higher freezing during light-on epochs in Context A compared to light-off epochs.     * This indicates light-induced recall of the fear memory associated with Context B.
• Control Groups:     * No Shock (NS) Group: Labeled in Context B without foot shocks; light did not induce freezing.     * EYFP Group: Labeled during FC but lacked ChR2; light did not induce freezing.
• Optimization of Freezing Levels:     * Base Exp Group: $∼ 15\%$ freezing.     * Exp-1day: Limiting the off-Dox period to one day reduced background expression twofold and increased light-induced freezing to $∼ 25\%$.     * Exp-Bi (Bilateral): Bilateral viral injection and fiber implants increased light-induced freezing to $∼ 35\%$, nearly as high as natural context-induced freezing ($∼ 60\%$).

Context-Specificity and Dual Engram Hypothesis

• Context Discrimination Testing:     * Mice were off-Dox for Context C (open field) to label cells with ChR2-EYFP.     * Mice were put back on-Dox and fear conditioned in Context B (monitored via endogenous c-fos).     * Result: Overlap between ChR2-EYFP (Context C) and c-fos (Context B) was at chance level, indicating independent DG cell populations represent different contexts.
• Cross-Context Reactivation:     * OF-FC Group: Labeled neurons in Context C (no shock); then fear conditioned in Context B. Light reactivation of Context C neurons did not induce freezing.     * FC-OF Group: Fear conditioned in Context B; then labeled neurons in Context C. Light reactivation of Context C neurons did not induce freezing.     * Conclusion: Light-induced recall is specific to the context associated with the shock.
• Dual Memory Engram Hypothesis:     * Contextual engrams are formed in the hippocampus.     * For a fear memory to be complete, context information must be transferred to the basolateral amygdala (BLA) and paired with shock information.     * Because Context C was never paired with a shock, reactivating its hippocampal engram cannot trigger the BLA fear response.

Discussion and Implications

• Summation of Sufficiency: This study is the first to demonstrate that direct activation of a sparse subset of cells involved in memory formation is sufficient to induce the behavioral expression of that memory.
• Necessity vs. Sufficiency: While DG engrams are sufficient for recall, the study suggests they might not be strictly necessary due to redundancy in the hippocampal circuit (e.g., direct input from entorhinal cortex to CA1 can circumvent the DG-CA3 path).
• Redundancy Examples: Mutant mice with disrupted DG-to-CA3 input can still perform contextual fear memory retrieval.
• Future Scope: This methodology provides a tool for mapping multi-component engrams and studying the interplay between different brain regions during memory storage and recall.

Surgical and Quantitative Specifics

• Stereotaxic Coordinates: $-2.2\,mm AP$; $\pm1.3\,mm ML$; $-2.0\,mm DV$ for injections and $-1.6\,mm DV$ for fiber placement.
• Viral Titer: $1 imes 10^{13} GC/ml$ for AAV9-TRE-ChR2-EYFP and $1.5 imes 10^{13} GC/ml$ for AAV9-TRE-EYFP.
• Injection Volume: $0.15\,ΐl$ at a rate of $0.1\,ΐl/min$.
• Training Trial Specs: $500\,s$ session; tone ($20\,s$, $75\,dB$, $2000\,Hz$); foot shock ($2\,s$, $0.75\,mA$).
• Dox Dosing: $40\,mg/kg$ for maintenance; $1\,g/kg$ overnight to rapidly suppress expression after training.