Caffeine intake exerts dual genome-wide effects on hippocampal metabolism and learning-dependent transcription

Caffeine is the world's most widely consumed psychoactive substance, with approximately 80% of the global population consuming it through coffee, tea, and soda. Its popularity stems from its ability to enhance well-being and cognitive functions like attention and alertness. Epidemiological studies have linked moderate caffeine consumption (200-400 mg/day in humans, equivalent to 0.3 g/L in rodents) to a reduced risk of all-cause mortality. Further evidence suggests that chronic caffeine consumption may normalize synaptic plasticity and mitigate cognitive decline associated with aging, Alzheimer's disease, and other neuropsychiatric conditions. While acute caffeine administration has been shown to enhance memory performance in both animals and humans, the molecular mechanisms underlying the effects of chronic caffeine intake remain largely unknown. This study aimed to address this gap by utilizing a multi-omics approach to analyze the epigenome, transcriptome, proteome, and metabolome of the mouse hippocampus following chronic caffeine exposure, focusing on its impact on neuronal processing during learning.

Existing research demonstrates caffeine's interaction with the adenosinergic system, acting as an antagonist. However, the adaptive downstream pathways triggered by chronic caffeine consumption require further investigation. Studies using acute caffeine administration have shown its effects on hippocampal and cortical excitability, enhancing basal synaptic transmission and modulating long-term potentiation (LTP) and sharp wave-ripple complexes, which are crucial for memory consolidation. However, these studies offer limited insight into the effects of habitual caffeine consumption. Studies in honeybees and rats have also demonstrated caffeine's cognition-enhancing properties, suggesting a potential role beyond simple arousal and attention.

Male C57BL6/J mice were divided into water (control) and caffeine (0.3 g/L in drinking water) groups for two weeks, mimicking moderate human consumption. Brain caffeine and metabolite concentrations were measured using LC-MS. Untargeted orthogonal omics techniques were employed. Histone acetylation (H3K27ac and H3K9/14ac) was assessed using ChIP-seq in bulk hippocampal tissue. RNA-seq was used to analyze the transcriptome. Proteomic analysis via mass spectrometry (MS) examined protein changes in bulk hippocampal tissue. Metabolomic changes were investigated using MALDI mass spectrometry imaging (MALDI-MSI). To examine cell-specific effects, CUT&Tag was performed on neuron-enriched hippocampal cell populations to analyze H3K27ac and H3K27me3. Finally, a Morris water maze (MWM) was used to assess spatial memory and RNA-seq was conducted to analyze learning-induced transcriptional changes in both water and caffeine groups.

Chronic caffeine consumption significantly decreased histone acetylation (H3K27ac and H3K9/14ac) at numerous genomic loci in bulk hippocampal tissue, primarily affecting genes involved in metabolic processes (amide, lipid metabolism; mRNA transport; translation; dendritic spine development). Metabolomic analysis revealed a substantial decrease in metabolites and lipids in the hippocampus of caffeine-treated mice. Proteomic analysis showed decreased levels of metabolism-related proteins and increased levels of neuron/synapse-associated proteins. In contrast to bulk tissue analysis, CUT&Tag analysis of neuron-enriched populations revealed a significant increase in H3K27ac at genes related to synaptic transmission and plasticity, indicating a neuron-autonomous effect. H3K27me3 was depleted at genes related to ion transport and learning and memory. The integration of epigenomic and proteomic data identified 28 proteins increased by caffeine with H3K27ac enrichment at their coding sequences. Finally, caffeine significantly increased the number of genes differentially regulated by learning in the MWM, suggesting an enhancement of learning-dependent hippocampal transcription. Specifically, caffeine increased the expression of genes related to metabolic processes that were deacetylated in resting conditions, implying a resetting of histone acetylation profiles to enhance inducibility during learning.

The study's findings reveal a dual role for chronic caffeine intake in the hippocampus. In bulk tissue (likely glial cells), caffeine reduces metabolic processes, potentially improving the signal-to-noise ratio in information encoding. Simultaneously, caffeine triggers a neuron-autonomous positive modulation of synaptic transmission and plasticity, enhancing the salience of information processing during learning. The increased number of differentially regulated genes in response to learning in the caffeine group suggests a priming effect, making neuronal networks more responsive to stimuli. The opposing effects of caffeine on resting and activated brain states align with human brain imaging studies, which show decreased functional connectivity at rest and increased activation during cognitive tasks.

This study demonstrates that chronic caffeine consumption exerts distinct effects on hippocampal function. It fine-tunes metabolic processes in non-neuronal cells, while simultaneously promoting synaptic plasticity and enhancing learning-dependent transcription in neurons. This dual action suggests that moderate caffeine intake may optimize brain function by improving the efficiency of information encoding and processing. Future research should explore the cell-type specific mechanisms underlying these effects and their implications for neurodegenerative diseases and brain development.

The study was conducted in male mice, limiting the generalizability to other sexes. The specific glial cell types involved in the metabolic changes observed in bulk tissue remain unidentified. While the study used several orthogonal omics approaches, other molecular mechanisms may also be involved.