Meta-analysis of Cannabis Use Identifies Shared Genetic Loci with Sleep and Circadian Rhythms

This meta-analysis of over 370,000 individuals identifies 39 genetic loci associated with cannabis use, revealing significant shared genetic risk with sleep disorders and circadian rhythms, particularly enriched in the central nervous system.

The Gist

Imagine your DNA as a massive, ancient library containing billions of books. Each book holds instructions for how your body and mind work. For a long time, scientists knew that some people were more likely to use cannabis, and they knew some people struggled with sleep, but they didn't fully understand if the same books in the library were responsible for both.

This study is like sending a team of detectives into that library to find the specific pages where the stories of "cannabis use" and "sleep trouble" overlap.

Here is the story of what they found, broken down into simple concepts:

1. The Big Search (The Meta-Analysis)

The researchers didn't just look at a small group of people; they combined data from two massive "libraries" of human information: the UK Biobank (people from the UK) and All of Us (people from the US).

• The Crowd: They looked at nearly 400,000 people.
• The Question: They asked, "Have you ever used cannabis?"
• The Result: They found 39 specific spots in the DNA library that are linked to cannabis use. Before this, we only knew about a few. It's like finding 39 new "chapters" in the story of why people use weed.

2. The "Brain Control Center"

When they looked at where these 39 spots were located in the body, they found something very clear: They are almost entirely in the brain.

Think of your body as a giant city. Most of the "cannabis use" instructions are written in the City Hall (the brain), specifically in the neighborhoods that control:

• How neurons (brain cells) talk to each other.
• How you feel emotions and pain.
• How you sleep and wake up.

This confirms that using cannabis isn't just a "habit"; it's deeply wired into how our brains are built.

3. The Sleep Connection (The "Double-Edged Sword")

This is the most interesting part. The study found that the genetic instructions for using cannabis are tightly linked to the instructions for sleep problems.

• The "Night Owl" Gene: People with these specific genetic traits are more likely to be "night owls" (staying up late) and less likely to be "early birds."
• The Insomnia Link: There is a strong genetic handshake between cannabis use and insomnia.
  • The Analogy: Imagine two neighbors who always borrow tools from each other. If you have the "genetic tool" for insomnia, you are more likely to have the "genetic tool" for cannabis use.
• The Medication Clue: The study found that people genetically prone to cannabis use are also genetically prone to taking sleep medication.

4. The "REM" Mystery (Why We Dream)

You might have heard that weed stops you from dreaming. That's because it suppresses REM sleep (the stage where vivid dreams happen).

• The Catch-22: When you stop using weed, your brain goes into "rebound mode." It suddenly floods with REM sleep, leading to intense, scary, or vivid dreams (nightmares).
• The Study's Twist: The researchers found that people genetically prone to cannabis use also have a genetic tendency to remember their dreams more vividly and have more nightmares.
• The Theory: It's possible that some people use cannabis because they have bad dreams or can't sleep, trying to "turn off" the dream machine. But the study suggests the relationship is complex: the genes that make you prone to nightmares might also be the same ones making you prone to using weed to fix it.

5. The "Self-Medication" Loop

The paper suggests a cycle:

1. You have a genetic makeup that makes it hard to sleep or makes you a night owl.
2. You feel tired or anxious.
3. You use cannabis to help you sleep or relax (self-medication).
4. Over time, this might change your brain chemistry, making sleep even harder without it, or causing those intense "rebound" dreams when you stop.

The Bottom Line

This study is a big step forward because it moves beyond just asking people "Do you smoke?" and looks at the biological blueprint underneath.

It tells us that cannabis use and sleep disorders are cousins in the genetic family. They share the same "address" in our DNA. This means that as cannabis becomes more common (both for fun and for medicine), we need to be very careful about how it affects our sleep cycles, because our brains are already wired to link the two together.

In short: If your DNA makes you a night owl or prone to sleep trouble, your brain might also be wired to reach for cannabis. Understanding this link helps doctors and scientists figure out better ways to treat both sleep issues and cannabis use in the future.

Technical Summary

1. Problem Statement

Cannabis is the most widely used psychoactive drug globally, with usage rising in both recreational and medicinal contexts (e.g., pain management, sleep aid). While individuals often self-medicate with cannabis for sleep disorders, the biological mechanisms connecting cannabis use to sleep regulation remain poorly understood. Previous genetic studies on cannabis use were limited in sample size and scope, failing to systematically characterize the shared genetic architecture between cannabis use liability and sleep traits (such as insomnia, circadian rhythm, and sleep duration). This study aims to bridge that gap by performing a large-scale meta-analysis to identify genetic loci associated with cannabis use and elucidate their overlap with sleep and circadian phenotypes.

2. Methodology

Cohorts and Data Sources:

• UK Biobank (UKB): ~500,000 participants. Cannabis use was defined via self-report (Field 29104).
• All of Us Research Program: ~400,000 participants. Cannabis use defined via lifestyle survey (Concept Code 1585636).
• Sample Size: The meta-analysis included 152,807 cases (cannabis users) and 220,272 controls (non-users), focusing primarily on individuals of European ancestry to ensure homogeneity in the GWAS.

Genetic Analysis Pipeline:

• Genotyping & Imputation: UKB data utilized Affymetrix arrays imputed via the TOPMed panel. All of Us utilized Illumina Global Diversity Array-8 and whole-genome sequencing data.
• GWAS: Performed using REGENIE v4.0 (logistic regression) with covariates including age, sex, genotyping array (UKB), and 10 principal components (PCs) for ancestry.
• Meta-analysis: Fixed-effects inverse variance weighted meta-analysis using METAL.
• Conditional Analysis: GCTA-COJO was used to identify independent association signals (39 independent loci) by accounting for linkage disequilibrium (LD).
• Functional Annotation & Enrichment:
  • Pathway Analysis: GProfiler for gene set enrichment.
  • Tissue Enrichment: Stratified LD Score Regression (S-LDSC) and FUMA to test enrichment across tissues and cell types.
  • Colocalization: PLACO (for pleiotropic loci) and SuSiE-COLOC (fine-mapping and colocalization) to determine if cannabis use and sleep traits share causal variants.
  • Mendelian Randomization (MR): TwoSampleMR and SMR Portal to infer causal relationships between insomnia and cannabis use.
• Phenotypic Validation: Fitbit data from All of Us (n=90,787) was used to correlate cannabis use with objective sleep duration and physical activity patterns.

3. Key Contributions

• Expanded Genetic Loci: Identified 39 independent genome-wide significant loci associated with cannabis use, significantly expanding the known genetic architecture from previous studies (which reported ~11 loci).
• Replication and Novelty: Successfully replicated the canonical CADM2 locus and discovered over 20 previously unreported loci.
• Sleep-Cannabis Genetic Link: Provided the first systematic evidence of a robust shared genetic architecture between cannabis use and sleep disorders, specifically clinical insomnia and circadian rhythms.
• Tissue Specificity: Confirmed that the genetic liability for cannabis use is heavily enriched in the Central Nervous System (CNS) and specific brain cell types, rather than peripheral tissues.
• Causal Inference: Utilized MR to suggest a bidirectional or causal link where insomnia liability increases the risk of cannabis use.

4. Key Results

A. Genetic Associations (GWAS):

• 39 Independent Loci: Identified via meta-analysis.
• Key Genes:
  • CADM2: Replicated; associated with risk-taking, sleep, and insomnia.
  • HTR1A: Serotonin receptor 1A; a target for REM sleep suppression.
  • SLC39A8: Associated with multiple sleep and neuropsychiatric traits; showed near-perfect colocalization with clinical insomnia.
  • NCAM1: Neural cell adhesion molecule; linked to memory formation.
  • RAI1: Associated with chronotype and circadian preference.
  • MAD1L1: Previously linked to sleep duration.
  • NPC1: Involved in lipid transport; potentially linking cannabis lipophilicity to lysosomal storage and sleep/wake circuitry.
  • FOXO6 & AKT3: Linked to memory consolidation defects.

B. Tissue and Pathway Enrichment:

• Brain Specificity: S-LDSC and FUMA analyses revealed that genetic heritability for cannabis use is exclusively enriched in the brain, particularly in the CNS and specific neuronal cell types.
• Pathways: Significant enrichment in "brain morphogenesis" and "FOXO-mediated transcription."

C. Genetic Correlations with Sleep Traits:

• Positive Correlations:
  • Clinical Insomnia ($r_g = 0.184$).
  • Insomnia medication usage ($r_g = 0.286$).
  • Nighttime physical activity.
• Negative Correlations:
  • Morningness chronotype ($r_g = -0.198$).
  • Ease of awakening.
  • Daytime physical activity.
• Objective Sleep Data: Cannabis users in the All of Us cohort had significantly longer objective sleep duration (7.10h vs. 7.006h in controls, $P = 1.10 \times 10^{-18}$).

D. Colocalization and Causality:

• Colocalization: Strong evidence of shared causal variants between cannabis use and clinical insomnia at SLC39A8 (Posterior Probability = 1) and CADM2/NCAM1 (PP.H4 > 0.7).
• Mendelian Randomization: Suggested a causal effect of insomnia liability on cannabis use ($P = 0.00566$) and vice versa ($P = 0.02$).
• Dreaming Phenotypes: Cannabis use was genetically associated with higher dream recall and nightmare frequency, potentially mediated by REM suppression and subsequent REM rebound upon cessation.

5. Significance and Implications

• Biological Mechanism: The study establishes a clear biological link between cannabis use and sleep regulation, mediated by neuronal genes involved in synaptic transmission (HTR1A, CADM2), circadian rhythms (RAI1), and memory (NCAM1, FOXO6).
• Self-Medication Hypothesis: The strong genetic correlation with insomnia and the MR results support the hypothesis that individuals with a genetic liability for sleep disorders may use cannabis as a form of self-medication, despite the risk of developing tolerance or disorder.
• Clinical Relevance: As cannabis legalization expands, understanding these shared genetic risks is crucial for clinicians. Patients with a genetic predisposition to insomnia or circadian misalignment may be at higher risk for cannabis use disorders.
• Future Directions: The findings highlight the need for further research into how specific cannabis strains interact with these genetic pathways, particularly regarding REM sleep suppression and the "REM rebound" effect that causes vivid nightmares.

In conclusion, this meta-analysis provides a comprehensive genetic map of cannabis use, demonstrating that it is a highly heritable trait deeply rooted in brain biology and inextricably linked to the genetic architecture of sleep and circadian rhythms.