Sex-specific DNA methylation in adult skeletal muscle

This study reveals that while spatial gene expression along the proximal-distal axis of adult tibialis anterior muscle is not driven by DNA methylation, sex-specific epigenetic programs characterized by widespread male hypermethylation serve as the primary determinant of methylation variation and associated transcriptional differences.

The Gist

The Big Picture: A Muscle Map vs. A Genetic Blueprint

Imagine your leg muscle (specifically the Tibialis Anterior, or TA) as a long, busy highway.

• The "Traffic" (Gene Expression): Scientists already knew that the "traffic" on this highway changes depending on where you are. The part of the muscle near the knee (proximal) has different cars and speed limits than the part near the ankle (distal). This is called spatial regionalization.
• The "Blueprint" (DNA Methylation): DNA methylation is like the construction blueprint or the lock-and-key system that tells the cells which genes to turn on or off.

The Big Question: The researchers wanted to know: Does the change in traffic (gene expression) happen because the blueprint (methylation) changes along the highway?

The Discovery: The Blueprint Stays the Same, But the Gender Changes Everything

The team took a "snapshot" of the muscle from the knee to the ankle and looked at two things:

1. What the cells were saying (RNA/Transcriptomics).
2. What the blueprint looked like (DNA Methylation).

Here is what they found, broken down simply:

1. The "Location" Myth (The Highway Analogy)

They expected that the blueprint would look different at the knee compared to the ankle, just like the traffic does.

• The Result: The blueprint was identical from top to bottom.
• The Analogy: Imagine a long train. The passengers in the front car are eating pizza, and the passengers in the back car are eating sushi. You'd expect the menu (the blueprint) to be different in each car. But when they checked the menu, it was the exact same in every car!
• Conclusion: The differences in how the muscle works near the knee vs. the ankle are not caused by changes in the DNA methylation blueprint. Something else (like immediate signals or protein activity) is driving those differences.

2. The "Gender" Reality (The Male vs. Female Analogy)

While the location didn't matter, gender was the superstar of the study.

• The Result: The blueprints for Male muscles looked completely different from the blueprints for Female muscles.
• The Analogy: Imagine two identical houses (one for a man, one for a woman). Even if they are built on the same street, the interior design is totally different.
  • Male Muscles: The blueprint had more "locks" (hypermethylation) on certain genes. This corresponds to muscles that are built for sprinting (fast, explosive, glycolytic).
  • Female Muscles: The blueprint had fewer locks on different genes. This corresponds to muscles built for marathons (endurance, slow, oxidative).
• Conclusion: In adult muscles, your sex is the biggest factor determining your epigenetic "blueprint," far more than where the muscle is located on your body.

3. The "Foreman" Theory (The Managers)

The researchers also looked for why the male and female blueprints were so different. They found three "Foremen" (genes) that were working overtime in the male muscles:

• Setd7, Gsk3a, and Bmyc.
• The Analogy: Think of these as construction managers. In the male construction site, these managers are shouting orders to lock down certain parts of the blueprint, ensuring the muscle is built for speed and power. In the female site, these managers are quieter, allowing for a different type of construction (endurance).

Why Does This Matter?

1. Stop Blaming the Location: If you are studying muscle differences between the top and bottom of a leg, don't look for DNA methylation changes there. They aren't the cause.
2. Start Respecting Gender: If you are studying muscle health, disease, or how drugs work, you must separate men and women. Their muscle "blueprints" are fundamentally different. Treating them as the same group is like trying to fix a Ferrari and a minivan with the exact same manual.
3. The Future: This tells us that while our DNA is the same, the "software" running on it (epigenetics) is heavily influenced by our sex, shaping how our muscles perform and age.

Summary in One Sentence

While the "traffic" in our leg muscles changes from knee to ankle, the "DNA blueprint" stays the same; however, the blueprint looks completely different depending on whether you are male or female, with men having a "locked-down" blueprint for speed and women having a blueprint for endurance.

Technical Summary

1. Problem Statement

Skeletal muscle exhibits functional heterogeneity, including spatial regionalization along the proximal-distal axis (e.g., differences in metabolic profiles and fiber types in the tibialis anterior, or TA) and significant sex-specific differences in physiology and metabolism. While previous transcriptomic studies (specifically Chapter 2 of the authors' work) have mapped robust spatial gene expression patterns and confirmed sex-specific transcriptional programs, the epigenetic mechanisms driving these phenomena remain unclear.

The central research question is whether DNA methylation underlies the spatial regionalization of gene expression in adult skeletal muscle or if it primarily reflects broader regulatory features, such as sex differences. Specifically, the authors hypothesized that proximal-distal differences in gene expression might be mirrored by corresponding variations in DNA methylation patterns.

2. Methodology

The study employed a multi-omic approach integrating spatial transcriptomics with genome-wide DNA methylation profiling on matched tissue sections from adult mouse TA muscles.

• Sample Preparation:

  • TA muscles from 4–6-month-old mice (3 females, 1 male for methylation; 2 females, 2 males for transcriptomics) were cryosectioned perpendicular to the proximal-distal axis (70 µm slices).
  • Odd-numbered sections were used for spatial transcriptomics (TOMOseq).
  • Even-numbered sections were used for DNA extraction and methylation profiling (MeDseq).
  • Spatial identity (proximal-distal vs. central) was assigned based on the transcriptomic annotation of the interleaved sections.

• MeDseq (Methylation-dependent sequencing):

  • DNA was digested with LpnPI, a methylation-dependent restriction enzyme that cleaves DNA 16 bp upstream and downstream of methylated CpG sites.
  • This generates short fragments (~32 bp) containing a centrally positioned CpG.
  • Libraries were prepared, size-selected, and sequenced (Illumina HiSeq 2500).
  • Data Processing: Reads were mapped to the GRCm38 genome. Read counts at LpnPI sites served as a proxy for methylation levels (presence of reads = methylated; absence = unmethylated). Data was aggregated into non-overlapping 1,000 bp genomic bins.

• Statistical Analysis:

  • Spatial Analysis: Compared methylation levels between "proximal-distal" and "central" sections for genes identified as regionally expressed.
  • Linear Modeling: Applied a linear regression model with interaction effects: Methylation ~ Sex + Spatial Annotation + (Sex × Annotation). This was performed on 1,000 bp bins across TSS regions, gene bodies, and regulatory elements to quantify the relative contribution of each factor.
  • Differential Analysis: Performed t-tests to identify differentially methylated regions (DMRs) between sexes and differentially expressed genes (DEGs) between sexes.
  • Integration: Correlated methylation changes with gene expression changes and identified upregulated epigenetic regulators (e.g., Setd7, Gsk3a, Bmyc).

3. Key Results

A. Lack of Spatial Methylation Variation

• Finding: Despite robust spatial differences in gene expression (transcriptome), DNA methylation patterns were largely uniform along the proximal-distal axis.
• Evidence:
  • Pearson correlation analysis showed highly similar methylation levels between central and proximal-distal sections for both TSS regions and gene bodies.
  • No discernible inverse trend (for TSS) or concordant trend (for gene bodies) was observed between methylation and regional gene expression.
  • Conclusion: DNA methylation does not drive the spatial regionalization of gene expression in adult TA muscle.

B. Sex as the Primary Determinant of Methylation

• Finding: Sex is the dominant source of DNA methylation variation, far exceeding spatial location or sex-by-location interactions.
• Evidence:
  • Linear Model: In 1,000 bp bins, Sex was significant in thousands of regions (e.g., 8,873 bins in TSS, 105,770 in gene bodies), whereas spatial annotation was significant in only single-digit or low-hundreds of bins.
  • Global Trend: Male muscles exhibited widespread hypermethylation compared to females across TSS, gene bodies, and regulatory regions.
  • Correlation: Male sections showed higher inter-sample correlation than female sections, likely due to higher methylation levels improving signal-to-noise ratios in the MeDseq assay.

C. Correlation with Fiber-Type and Metabolic Profiles

• Finding: Sex-specific methylation patterns align with known physiological differences in fiber-type composition.
• Evidence:
  • Males: Enriched in fast-glycolytic fibers. Showed hypermethylation at the TSS of oxidative markers (e.g., Myl3) but hypermethylation in the gene body of glycolytic markers (e.g., Pkm).
  • Females: Enriched in fast-oxidative/slow-oxidative fibers. Showed higher methylation at the TSS of glycolytic markers (e.g., Aldoa).
  • Transcriptomics: Confirmed sex-biased expression of fiber-type markers (e.g., upregulation of Myh2/Myh1 in females; Pkm and Kdm5d in males).

D. Mechanistic Links via Epigenetic Regulators

• Finding: Several transcriptional regulators known to influence chromatin state were upregulated in males, potentially driving the observed hypermethylation.
• Key Genes:
  • Setd7 (Histone-lysine methyltransferase): Upregulated in males; linked to sarcomere organization and fast-twitch fiber specification.
  • Gsk3a: Upregulated in males; known to regulate global methylation levels at imprinted loci.
  • Bmyc: Upregulated in males; a regulator of Myc, which recruits DNA methyltransferases.
• Implication: These genes suggest a coordinated axis where sex-specific transcriptional programs reinforce male-biased methylation landscapes.

4. Key Contributions

1. Decoupling of Spatial and Epigenetic Regulation: The study definitively demonstrates that spatial gene expression gradients in adult skeletal muscle are transcriptionally driven rather than established by static CpG methylation differences. This challenges the assumption that spatial heterogeneity always requires a corresponding epigenetic map.
2. Quantification of Sex Bias: Provides a genome-wide map confirming that sex is the primary driver of methylation variation in adult muscle, with males showing global hypermethylation.
3. Multi-omic Integration: Successfully linked spatial transcriptomics with methylomics to show that while spatial patterns are not epigenetically encoded, sex-specific epigenetic programs are tightly coupled with sex-specific metabolic and contractile phenotypes.
4. Identification of Drivers: Proposed specific molecular candidates (Setd7, Gsk3a, Bmyc) that may mechanistically link sex-biased gene expression to the observed epigenetic state.

5. Significance and Implications

• Biological Insight: The findings suggest that adult muscle spatial specialization relies on dynamic regulatory mechanisms (e.g., histone modifications, 3D chromatin architecture, or local innervation) rather than stable DNA methylation patterns.
• Experimental Design: The study underscores the critical necessity of stratifying by sex in multi-omic studies of skeletal muscle. Failure to account for sex could obscure biological signals, as sex differences in methylation are orders of magnitude larger than spatial differences.
• Future Directions: The results open avenues for investigating how sex hormones interact with chromatin modifiers to shape muscle identity and how these epigenetic landscapes respond to interventions like exercise or aging. The authors note limitations in sample size and coverage (MeDseq does not cover all CpGs), suggesting future work should utilize single-nucleus multi-omics to refine these observations.