16 Epigenetics and Epigenomics

16.1 Chapter Overview

This chapter explores the fascinating world of epigenetics—the study of heritable changes in gene function that occur without alterations to the DNA sequence itself. We’ll discover how cells with identical DNA can have vastly different functions, how environmental factors can influence gene expression across generations, and how epigenetic dysregulation contributes to disease. From DNA methylation to histone modifications, from chromatin remodeling to non-coding RNAs, we’ll examine the molecular mechanisms that allow organisms to adapt, develop, and respond to their environment.

16.2 Definition and Overview of Epigenetics

16.2.1 What Is Epigenetics?

Epigenetics studies changes in gene activity that DON’T change the DNA sequence itself.

The word “epigenetic” means:

Epi = above or on top of
Genetic = relating to genes
So: “Above genetics” or “On top of genes”

Think of it like:

DNA = the words in a book
Epigenetics = highlighting, bookmarks, and notes that change how you read the book
Same words, but different reading experience!

More formally, epigenetics refers to:

Heritable changes in gene expression or cellular phenotype
Changes that do not involve alterations to the underlying DNA sequence
Modifications that can be maintained through cell divisions
Reversible changes that respond to developmental and environmental cues

Historical context:

The term “epigenetics” was coined by Conrad Waddington in 1942 to describe “the interactions of genes with their environment that bring the phenotype into being.” Today, it encompasses all molecular mechanisms that regulate gene expression beyond the DNA sequence.

16.3 Difference between Genetics and Epigenetics

Understanding the distinction between genetics and epigenetics is crucial:

16.3.1 Genetics

What it is:

The study of genes and DNA sequences
Focuses on the genetic code itself (A, T, G, C)
Deals with heredity through DNA

Key characteristics:

Changes are typically permanent (mutations)
Follows Mendelian inheritance patterns
Sequence-based: altering the actual letters of genetic code
Relatively stable throughout life
Passed from parent to offspring through DNA replication

Example: A mutation in the BRCA1 gene sequence increases breast cancer risk—this change is permanent and inherited.

16.3.2 Epigenetics

What it is:

The study of changes in gene activity without changing DNA sequence
Focuses on chemical modifications to DNA and histones
Deals with gene expression regulation

Key characteristics:

Changes are potentially reversible
Can be influenced by environment, age, lifestyle, disease
Modification-based: adding or removing chemical groups
Dynamic and responsive to signals
Can be passed through cell divisions, sometimes across generations

Example: Identical twins have the same DNA, but different epigenetic marks lead to different disease susceptibility—these differences accumulate over time.

16.3.3 Key Comparisons

Feature	Genetics	Epigenetics
What changes?	DNA sequence	Gene expression
Permanence	Usually permanent	Reversible
Inheritance	Always inherited	Sometimes inherited
Environmental influence	Minimal	Significant
Therapeutic potential	Hard to change	Can be targeted by drugs
Stability	Very stable	Dynamic

Think of it this way:

Genetics = The hardware (fixed components)
Epigenetics = The software (changeable settings)

16.4 Epigenetic Inheritance and Reversibility

16.4.1 The Dynamic Nature of Epigenetic Marks

One of the most remarkable features of epigenetic modifications is their reversibility, which distinguishes them from genetic mutations.

Key properties:

Mitotic inheritance (cell division):
- Epigenetic marks are copied when cells divide
- This allows daughter cells to maintain the same identity as parent cells
- How liver cells stay liver cells through many divisions
- Involves maintenance mechanisms that “remember” marks
Meiotic inheritance (across generations):
- Most epigenetic marks are erased during reproduction
- Prevents inheritance of temporary environmental responses
- Ensures each generation starts relatively fresh
- Some marks escape erasure—this is transgenerational inheritance
Reversibility:
- Unlike DNA mutations, epigenetic changes can be reversed
- Environmental factors can add or remove marks
- Therapeutic interventions can modify epigenetic states
- This makes epigenetic changes both a risk and an opportunity

16.4.2 Types of Epigenetic Inheritance

Intergenerational inheritance:

Direct exposure affects the organism and its developing gametes
Example: Pregnant woman exposed to famine affects her child (F1) and the child’s developing eggs/sperm (F2)
Not “true” transgenerational—direct exposure occurred

Transgenerational inheritance:

Effects persist beyond F2 generation (F3 and beyond)
No direct exposure to the original environmental factor
More controversial in mammals
Well-documented in plants, worms (C. elegans), and some other organisms

Evidence in different organisms:

Plants: Strong evidence for transgenerational epigenetic inheritance
C. elegans: RNA-mediated inheritance can last many generations
Mammals: Most marks erased; rare examples of transgenerational effects
Humans: Limited but growing evidence

Reprogramming events:

Two major “epigenetic reset” events occur in mammals:

Zygotic reprogramming: Just after fertilization
Germline reprogramming: During germ cell (egg/sperm) development

These erasure events explain why most acquired characteristics are NOT inherited!

16.4.3 Why Epigenetics Matters

The Big Question: If all your cells have the same DNA, why are they different?

Brain cells and muscle cells have identical DNA
But they look and act completely different!
Answer: Epigenetics!

Epigenetic marks determine:

Which genes are turned ON in each cell type
Which genes are turned OFF
How cells remember their identity
How cells respond to the environment
How development proceeds in an orderly fashion
How organisms adapt to environmental changes
How experiences can affect gene expression
How diseases develop and might be treated

16.5 DNA Methylation

16.5.1 Adding Chemical Tags to DNA

DNA methylation = Adding a methyl group (CH₃) to DNA

Where it happens:

Usually on cytosine (C) bases
Specifically in “CpG” sequences (C next to G)
“p” stands for the phosphate between them

What it looks like:

Normal: Cytosine (C)
Methylated: 5-methylcytosine (5mC)
Same base, just with a tiny chemical decoration

Think of it like:

Putting a sticker on a letter
The letter is still there and readable
But the sticker changes how it’s treated

16.5.2 CpG Islands and Their Importance

CpG islands = Regions with lots of CpG sequences

Key facts:

“CpG” refers to cytosine-phosphate-guanine dinucleotides
Found near 60-70% of human gene promoters
Typically 500-2000 base pairs long
High GC content (>50%)
Usually UN-methylated in normal cells (genes are ready to be expressed)
If they GET methylated → gene is turned off

Why “islands”?

CpG dinucleotides are relatively rare in most of the genome (~1%)
They’re “depleted” because methylated CpGs mutate over evolutionary time (5mC → T)
CpG islands are protected regions where CpGs are preserved
These islands are like oases in a CpG-poor desert

Location and function:

Promoter CpG islands: Control gene transcription start sites
Intragenic islands: Found within gene bodies
Intergenic islands: Between genes, may regulate distant genes
Shore regions: Areas flanking CpG islands (up to 2kb away)—often show tissue-specific methylation patterns

Examples:

Cancer: Tumor suppressor genes get abnormally methylated → cancer
Development: Methylation patterns change as embryo develops
X-inactivation: Methylation silences one X chromosome in females
Aging: CpG islands gradually lose protection and become methylated

16.5.3 DNA Methyltransferases (DNMTs)

The enzymes that add methyl groups to DNA are called DNA methyltransferases (DNMTs).

DNMT1 - The “Maintenance” Methyltransferase:

Function: Copies methylation patterns during DNA replication
How it works: Recognizes hemi-methylated DNA (one strand methylated)
Role: Maintains existing patterns through cell divisions
Specificity: Prefers hemi-methylated substrates
Location: Associates with replication fork
Importance: Ensures daughter cells inherit same methylation pattern as parent

Think of DNMT1 as a photocopier—it copies existing marks onto new DNA strands!

DNMT3A and DNMT3B - The “De Novo” Methyltransferases:

Function: Establish NEW methylation patterns
When active: During development, germ cell formation
DNMT3A:
- Active in most tissues
- Important for somatic cell methylation
- Mutations cause growth disorders
DNMT3B:
- Particularly active in early development
- Critical for embryonic development
- Mutations cause ICF syndrome (Immunodeficiency, Centromeric instability, Facial anomalies)
Collaboration: Often work together with DNMT1

Think of DNMT3A/B as architects—they design new methylation patterns from scratch!

DNMT3L - The Helper:

Does NOT have methyltransferase activity itself
Stimulates DNMT3A and DNMT3B activity
Critical for establishing maternal imprints
Mutations cause infertility

Coordination of DNMTs:

DNA Replication:
Parent DNA: [methylated-C]==========[methylated-C]
               ↓                         ↓
           DNA copied
               ↓
New DNA:    [methylated-C]==========C
            C===========[methylated-C]
               ↓
           DNMT1 acts
               ↓
Daughter:   [methylated-C]===========[methylated-C]
            [methylated-C]===========[methylated-C]

16.5.4 Demethylation and TET Enzymes

For a long time, scientists thought DNA methylation was essentially permanent. Then the TET enzymes were discovered!

TET Enzymes (Ten-Eleven Translocation enzymes):

TET1, TET2, TET3 - Three family members in mammals
Function: Convert 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC)
Result: First step in active demethylation pathway

The Demethylation Pathway:

5mC (5-methylcytosine) ↓ TET enzymes
5hmC (5-hydroxymethylcytosine) ↓ TET enzymes
5fC (5-formylcytosine) ↓ TET enzymes
5caC (5-carboxylcytosine) ↓ TDG enzyme + BER pathway
C (unmodified cytosine)

Properties of oxidized methylcytosines:

5hmC: Not just an intermediate—also has its own regulatory functions
Enriched in brain tissue and embryonic stem cells
Can recruit specific protein readers
May represent a stable epigenetic mark

Two types of demethylation:

Active demethylation:
- TET-mediated oxidation pathway
- Can occur rapidly in response to signals
- Important in development and cell differentiation
- Critical for paternal genome demethylation in zygote
Passive demethylation:
- Occurs when DNMT1 is blocked or absent
- Methylation lost through DNA replication
- Slower process
- Important in some developmental contexts

TET enzyme importance:

Development: Essential for zygotic and germline reprogramming
Cancer: TET2 frequently mutated in blood cancers
Brain: High TET1 expression; important for neuronal function
Reprogramming: Critical for iPSC generation

16.5.5 Effects of DNA Methylation on Gene Expression

Promoter methylation → Gene Silencing:

How it works:

Physical blocking:
- Methyl groups prevent transcription factors from binding
- Like putting a lock on a door
Recruiting silencing complexes:
- Methylated DNA attracts methyl-binding domain (MBD) proteins
- MBD proteins recruit histone deacetylases (HDACs)
- HDACs remove activating marks from histones
- Creates compact, inactive chromatin
Creating heterochromatin:
- Methylated regions become densely packed
- Transcription machinery cannot access DNA
- Gene is stably silenced

Gene body methylation → Variable effects:

Surprisingly, methylation WITHIN gene bodies can: - Correlate with increased transcription - Prevent spurious transcription initiation - Regulate alternative splicing - Mark actively transcribed genes

Context-dependent effects:

Location	Methylation Level	Effect
Promoter CpG islands	High	Strong silencing
Gene bodies	High	Often associated with expression
Enhancers	High	Reduced activity
Repetitive elements	High	Silencing (protective)
Imprinted genes	Parent-specific	Monoallelic expression

16.5.6 Genomic Imprinting

Genomic imprinting = Parent-of-origin-specific gene expression

Key concept:

For most genes, both copies (maternal and paternal) are expressed
For imprinted genes, only ONE copy is expressed
Which copy is expressed depends on which parent it came from
The other copy is silenced by DNA methylation

How it works:

During germ cell formation, specific genes get methylated
This methylation is maintained after fertilization
Methylated allele is silenced
Unmethylated allele is expressed
Pattern is reset each generation in the germline

Why imprinting exists:

Parental conflict hypothesis: Mother and father have different evolutionary interests
Dosage regulation: Some genes need precise expression levels
Brain development: Many imprinted genes affect behavior and development

Imprinting disorders:

Prader-Willi syndrome: Loss of paternal 15q11-13 expression
Angelman syndrome: Loss of maternal 15q11-13 expression
Same chromosomal region, different diseases depending on parent of origin!

We’ll discuss specific examples in detail later in the “Genomic Imprinting” section.

16.5.7 X-Chromosome Inactivation (XCI)

The problem:

Females have TWO X chromosomes (XX)
Males have ONE X chromosome (XY)
How to balance X-linked gene expression?

The solution: X-Chromosome Inactivation (XCI)

One X chromosome in each female cell is randomly inactivated
Occurs early in development
Inactivated X becomes a Barr body (condensed, silent chromatin)
Same X remains inactive in all descendant cells

Role of DNA methylation in XCI:

Initiation: Guided by XIST lncRNA (more on this later)
Spreading: Silencing spreads along the chromosome
Maintenance: DNA methylation LOCKS in the silent state
Stability: Methylation ensures X remains inactive through divisions

Methylation patterns:

CpG islands on inactive X become heavily methylated
This stabilizes the inactive state
Active X remains unmethylated
Pattern is mitotically heritable

Clinical significance:

X-linked diseases can have variable severity in females
Depends on which X is inactivated in each cell
Random inactivation creates mosaic pattern

We’ll explore the molecular mechanisms in greater detail in the dedicated “X-Chromosome Inactivation” section.

16.6 Histone Modifications

16.6.1 Decorating the Histone Proteins

Remember histones? DNA wraps around them like thread on spools.

Histone modifications = Adding or removing chemical groups to histone proteins

Where modifications occur:

Histone “tails” sticking out from nucleosomes
Like ribbons on wrapped presents
These tails are accessible to modifying enzymes
Rich in lysine (K), arginine (R), serine (S), and threonine (T) residues

16.6.2 Types of Histone Modifications

1. Acetylation (adding acetyl groups):

Effect: Loosens chromatin, gene activation
Where: Lysine amino acids
Think of it as: Opening the curtains to let in light

2. Methylation (adding methyl groups):

Effect: Can activate OR repress genes (depends on location!)
Where: Lysine or arginine amino acids
Complex: Different methylation patterns mean different things

3. Phosphorylation (adding phosphate groups):

Effect: Various effects depending on location
Important for: Cell division, DNA repair

4. Ubiquitination (adding ubiquitin protein):

Effect: Can activate or repress genes
Also can: Mark histones for degradation

16.6.3 Histone Acetylation in Detail

Histone Acetyltransferases (HATs):

What they do: - Add acetyl groups (-COCH₃) to lysine residues - Neutralize positive charge on histones - Weaken histone-DNA interaction - Open up chromatin structure - Generally → Gene activation

Major HAT families:

GNAT family (Gcn5-related N-acetyltransferases):
- Gcn5, PCAF
- Often part of large complexes
- Acetylate H3 and H4
MYST family:
- MOZ, MOF, TIP60
- Important in development
- Some mutations cause cancer
p300/CBP family:
- Broad substrate specificity
- Act as transcriptional co-activators
- Integrate multiple signaling pathways

Histone Deacetylases (HDACs):

What they do: - Remove acetyl groups from lysines - Restore positive charge on histones - Strengthen histone-DNA binding - Condense chromatin - Generally → Gene repression

HDAC classes:

Class I HDACs (HDAC1, 2, 3, 8):
- Found mainly in nucleus
- Related to yeast Rpd3
- Broadly expressed
Class II HDACs (HDAC4, 5, 6, 7, 9, 10):
- Can shuttle between nucleus and cytoplasm
- Tissue-specific expression
- Regulated by cellular signals
Class III HDACs (Sirtuins: SIRT1-7):
- NAD⁺-dependent (different mechanism)
- Involved in aging and metabolism
- Connected to caloric restriction effects
Class IV HDACs (HDAC11):
- Shares features of Class I and II
- Least studied

Balance is key:

   HATs          HDACs
    ↓              ↓
Acetylation ←→ Deacetylation
    ↓              ↓
Open chromatin  Closed chromatin
    ↓              ↓
Active genes    Silent genes

Clinical relevance: - HDAC inhibitors are approved cancer drugs - Work by reactivating silenced tumor suppressor genes - Examples: Vorinostat, Romidepsin

16.6.4 Histone Methylation in Detail

Histone Methyltransferases (HMTs):

Complexity: - Can add 1, 2, or 3 methyl groups (mono-, di-, tri-methylation) - Different degrees of methylation have different meanings - Different lysines can be methylated (K4, K9, K27, K36, K79)

Major HMTs:

SET domain-containing HMTs:
- Largest family
- MLL, NSD, SET1/2, SUV39H, G9a, EZH2
Non-SET domain HMTs:
- DOT1L (methylates H3K79)
- Uses different mechanism

Key methylation marks:

Activating marks: - H3K4me3 (Histone 3, Lysine 4, tri-methylation): - Found at active promoters - Mark of active transcription - Added by MLL complexes

H3K36me3 (Histone 3, Lysine 36, tri-methylation):
- Found in gene bodies
- Mark of active transcription elongation
- Prevents cryptic transcription

Repressive marks: - H3K9me3 (Histone 3, Lysine 9, tri-methylation): - Found in heterochromatin - Added by SUV39H enzymes - Creates binding sites for HP1 protein - Maintains gene silencing

H3K27me3 (Histone 3, Lysine 27, tri-methylation):
- Added by PRC2 complex (contains EZH2)
- Important for development
- Silences developmental genes
- Maintains cell identity

H3K9 methylation pathway:

Unmethylated H3K9
      ↓ G9a/GLP (euchromatic regions)
H3K9me1/me2
      ↓ SUV39H (heterochromatic regions)
H3K9me3
      ↓ Recruits HP1
Heterochromatin formation
      ↓
Gene silencing

Histone Demethylases (KDMs):

Yes, histone methylation is reversible too!

Two families:

LSD family (Lysine-Specific Demethylases):
- LSD1/KDM1A, LSD2/KDM1B
- Remove mono- and di-methyl groups
- FAD-dependent enzymes
- Cannot remove tri-methyl groups
JmjC domain-containing demethylases:
- Many family members (KDM2-KDM8)
- Can remove all methylation states (me1, me2, me3)
- Iron and α-ketoglutarate-dependent
- More versatile than LSDs

Example - LSD1: - Removes H3K4me1/me2 (activating marks) → repression - Can also remove H3K9me1/me2 (repressive marks) → activation - Context-dependent function - Targeted by drugs for cancer treatment

16.6.5 Other Important Histone Modifications

Phosphorylation:

Key marks: - H3S10ph (Histone 3, Serine 10, phosphorylation): - Added during mitosis - Chromosome condensation - Also involved in gene activation

H2AXS139ph (γH2AX):
- Mark of DNA damage
- Recruits DNA repair machinery
- Spreads megabases from break site

Function: - Rapid and reversible - Kinases add phosphate groups - Phosphatases remove them - Often cross-talks with other modifications

Ubiquitination:

H2AK119ub1 (H2A, Lysine 119, monoubiquitination): - Added by PRC1 complex - Gene repression - Works with H3K27me3 - Part of Polycomb silencing system

H2BK120ub1 (H2B, Lysine 120, monoubiquitination): - Gene activation! - Required for H3K4 and H3K79 methylation - Trans-histone regulation - Important for transcription elongation

Sumoylation:

Addition of SUMO proteins (Small Ubiquitin-like MOdifiers)
Generally associated with transcriptional repression
Less well understood than other modifications
Can compete with acetylation at same lysines

ADP-ribosylation:

Addition of ADP-ribose groups
Catalyzed by PARP enzymes
Important for DNA damage response
Can modify core histones and H1

16.6.6 The Histone Code Hypothesis

The “histone code” hypothesis:

Different combinations of modifications send different signals
Like a complex language written on histones
Specific proteins “read” these marks and respond accordingly

The code concept:

Histone tail:  [K4]---[K9]---[K14]---[K18]---[K27]
Modifications:  me3    ac     ac      ac      me3
Meaning:      "Bivalent promoter - poised for activation"

vs.

Histone tail:  [K4]---[K9]---[K14]---[K18]---[K27]
Modifications:  me1    me3    -       -       -
Meaning:      "Heterochromatin - permanently silenced"

Reader proteins:

Different protein domains recognize specific modifications:

Bromodomains:
- Recognize acetylated lysines
- Found in HATs, chromatin remodelers, transcription factors
- Example: BRD4 reads H3K27ac and H4K5ac
Chromodomains:
- Recognize methylated lysines
- Found in HP1 (reads H3K9me3)
- Found in Polycomb proteins
PHD fingers:
- Recognize H3K4me3 and H3K4me0
- Found in many chromatin proteins
- Can distinguish methylation states
Tudor domains:
- Recognize methylated arginines and lysines
- Found in multiple protein families
MBT domains:
- Recognize mono- and di-methylated lysines
- Found in Polycomb proteins

Combinatorial complexity:

Over 100 different histone modifications identified
Multiple modifications can occur on same histone tail
Modifications can influence each other (cross-talk)
Creates enormous information-coding capacity
Context matters: same mark, different location = different meaning

Examples of histone codes:

Code	Meaning	Location
H3K4me3 + H3K27me3	Bivalent/poised	ES cell genes
H3K4me3 + H3K9ac + H3K27ac	Active transcription	Active promoters
H3K9me3 + H4K20me3	Constitutive heterochromatin	Repetitive DNA
H3K27me3 + H2AK119ub1	Polycomb silencing	Developmental genes
H3K36me3	Transcription elongation	Gene bodies

Think of it like:

Different emoji combinations convey different meanings
Cells “read” histone modifications to know what to do
😊 + 👍 = approval, but 😊 + 😏 = sarcasm!
Similarly, H3K4me3 + H3K27ac = activate, but H3K4me3 + H3K27me3 = pause

Evidence and limitations:

Supporting evidence: - Specific modifications correlate with specific chromatin states - Reader proteins bind modifications and recruit effector proteins - Removing modifications changes gene expression predictably

Limitations and debates: - Not all modification combinations have been tested - Same modifications can have different effects in different contexts - Some argue it’s more about “histone language” than strict “code” - Redundancy exists—different marks can have similar effects

16.7 Chromatin Structure and Remodeling

16.7.1 Understanding Chromatin States

Chromatin is not uniform throughout the nucleus—it exists in different states of compaction that directly affect gene expression.

16.7.2 Euchromatin vs Heterochromatin

Euchromatin (“true chromatin”):

Characteristics: - Loosely packed chromatin - Light-staining under microscope - Transcriptionally active (genes can be expressed) - Rich in genes - Associated with activating histone marks (H3K4me3, H3K9ac, H3K27ac) - Typically unmethylated or hypomethylated DNA - Accessible to transcription factors

Location: - Interior of nucleus - Gene-rich chromosomal regions - Active during interphase

Think of euchromatin as open library stacks—books (genes) are accessible and can be read.

Heterochromatin (“different chromatin”):

Characteristics: - Tightly packed chromatin - Dark-staining under microscope - Transcriptionally silent (genes cannot be expressed) - Gene-poor - Associated with repressive marks (H3K9me3, H3K27me3, H4K20me3) - Highly methylated DNA - Inaccessible to transcription factors

Two types:

Constitutive heterochromatin:
- Always condensed, all cell types
- Found at centromeres and telomeres
- Contains repetitive DNA
- Structural role
- Examples: Centromeric repeats, satellite DNA
- Never transcribed (except specialized RNAs)
Facultative heterochromatin:
- Can switch between hetero- and euchromatin
- Cell type-specific and developmentally regulated
- Contains genes that CAN be expressed in some contexts
- Examples: Inactive X chromosome, imprinted genes, tissue-specific silent genes
- Reversible condensation

Think of heterochromatin as locked library storage—books exist but can’t be accessed.

Comparison table:

Feature	Euchromatin	Heterochromatin
Compaction	Loose	Tight
Staining	Light	Dark
Transcription	Active	Silent
DNA methylation	Low	High
Histone marks	H3K4me3, H3ac	H3K9me3, H4K20me3
Replication timing	Early S phase	Late S phase
Gene density	High	Low
Accessibility	High	Low
Location	Nuclear interior	Nuclear periphery, nucleolus

Transitions between states:

During development and differentiation:

Euchromatin ←→ Facultative Heterochromatin
     ↑                    ↑
  (Active)           (Silenced)
     ↑                    ↑
  Gene ON              Gene OFF

Constitutive heterochromatin remains stable:

Constitutive Heterochromatin (永远 always silenced)

16.7.3 Nucleosome Positioning

Nucleosome = DNA wrapped around histone octamer

Positioning matters:

Where nucleosomes sit affects gene expression
How tightly DNA wraps affects accessibility
Spacing between nucleosomes matters

Key regions:

Nucleosome-Free Regions (NFRs):
- Found at promoters and enhancers
- Typically 150-200 bp
- Allow transcription factor binding
- Mark of active regulatory elements
+1 nucleosome:
- First nucleosome downstream of transcription start site (TSS)
- Often contains H3.3 variant
- Precisely positioned
- Marks active genes
-1 nucleosome:
- Nucleosome just upstream of TSS
- Flanks promoter NFR
- Often has H2A.Z variant

Determinants of positioning:

DNA sequence:
- Some sequences favor nucleosome formation
- AT-rich sequences are more flexible
- Poly(dA:dT) tracts resist nucleosomes
- Creates intrinsic positioning preferences
Trans-acting factors:
- Transcription factors
- Chromatin remodelers (see below)
- Pioneer factors that can bind nucleosomal DNA
Histone variants:
- H2A.Z, H3.3, H2A.Bbd alter stability
- Affect positioning and dynamics
Boundary elements:
- Insulator proteins (like CTCF)
- Create barriers to nucleosome spreading

Dynamic nucleosome landscape:

Inactive gene:
[Nuc][Nuc][Nuc][Nuc][Nuc][Promoter buried][Nuc][Nuc]

↓ Activation signal

Active gene:
[Nuc][Nuc]------[NFR]------[+1]--[Nuc]--[Nuc]--[Nuc]
                 ↑          ↑
             TF binding   TSS

16.7.4 ATP-Dependent Chromatin Remodeling Complexes

Chromatin remodeling = Moving, removing, or changing nucleosomes

Chromatin remodeling complexes are molecular machines that:

Slide nucleosomes along DNA
Eject nucleosomes completely
Replace histones with variants
Loosen or tighten chromatin
Use ATP energy to do so

Effects:

Make DNA accessible (gene ON)
Make DNA inaccessible (gene OFF)
Respond to cellular signals

Think of chromatin remodeling like:

Rearranging furniture to access different parts of a room
Opening or closing window blinds
Making space for activities

16.7.5 Major Remodeling Complex Families

All use ATP as energy source, but have different mechanisms and outcomes.

1. SWI/SNF Family (Switch/Sucrose Non-Fermentable)

Members: BAF (Brahma-Associated Factors), PBAF complexes in mammals

Mechanism: - Slides and ejects nucleosomes - Creates nucleosome-free regions - Uses DNA translocation

Functions: - Gene activation (primarily) - Opens chromatin at promoters and enhancers - Allows transcription factor access - Important for development

Composition: - Core ATPase: BRG1 or BRM - Multiple accessory subunits (BAFs) - Subunit composition varies by cell type - Modular architecture allows functional diversity

Clinical relevance: - Frequently mutated in cancer (~20% of tumors) - SMARCA4 (BRG1) and SMARCB1 (SNF5) are tumor suppressors - Mutations in autism spectrum disorders - Targets for cancer therapy

Example function:

Closed promoter:
[Nuc][Nuc][Nuc] → No transcription

↓ SWI/SNF recruited

Open promoter:
[Nuc]---[NFR]---[Nuc] → Transcription possible

2. ISWI Family (Imitation SWI)

Members: Multiple complexes including NURF, ACF, CHRAC

Mechanism: - Primarily slides nucleosomes - Spaces nucleosomes evenly - Fine-tunes nucleosome positioning - Creates regular arrays

Functions: - Chromatin assembly and maintenance - Nucleosome spacing - Both activation and repression - Replication-coupled chromatin assembly

Features: - Recognizes unmodified histone tails - Inhibited by acetylation - Important for chromatin structure

Example function:

Disordered nucleosomes:
[Nuc]--[Nuc]-------[Nuc]--[Nuc]

↓ ISWI activity

Regularly spaced:
[Nuc]----[Nuc]----[Nuc]----[Nuc]

3. CHD Family (Chromodomain Helicase DNA-binding)

Members: CHD1-9, with diverse functions

Mechanism: - Slides and ejects nucleosomes - Contains chromodomains (read histone methylation) - Links histone modifications to remodeling

Functions: - CHD1: Active transcription, binds H3K4me3 - CHD3/4 (NuRD complex): Gene repression, associates with HDACs - CHD7: Development, mutations cause CHARGE syndrome - CHD8: Brain development, autism risk gene

Unique features: - Chromodomains allow “reading” histone code - Can be recruited to specific chromatin states - Different family members have opposite effects

Example - NuRD complex:

Active gene:
H3K4me3, acetylated histones → Transcription

↓ CHD3/4 + HDACs recruited

Repressed gene:
Deacetylated, repositioned nucleosomes → Silencing

4. INO80 Family

Members: INO80, SWR1 (SRCAP in mammals)

Mechanism: - Exchanges histone variants - Removes and replaces histones - Specialized functions

Functions:

INO80 complex: - DNA damage response - Replication fork stability - Removes H2A.Z - Transcription regulation

SWR1/SRCAP complex: - Deposits H2A.Z histone variant - Marks regulatory regions - Important for gene regulation

Unique features: - Large complexes (>1 MDa) - Multiple subunits - Specialized for histone variant exchange

H2A.Z exchange:

Standard nucleosome:
[H2A-H2B dimer] in nucleosome

↓ SWR1 complex

H2A.Z-containing nucleosome:
[H2A.Z-H2B dimer] in nucleosome
→ Less stable, more dynamic
→ Mark of regulatory regions

16.7.6 Coordination of Remodeling Activities

Remodelers work together:

Sequential action:
- SWI/SNF opens chromatin
- ISWI maintains spacing
- CHD fine-tunes for transcription
Antagonistic functions:
- Some open, some close
- Balance determines chromatin state
Signal integration:
- Recruited by transcription factors
- Read histone modifications
- Respond to cellular needs

Example—Gene activation:

1. Signal received
   ↓
2. Transcription factors bind
   ↓
3. HATs recruited → Acetylation
   ↓
4. SWI/SNF recruited → Nucleosome ejection
   ↓
5. Pol II recruited → Transcription begins
   ↓
6. CHD1 recruited by H3K4me3 → Maintains open state
   ↓
7. ISWI organizes nucleosomes in gene body

Dysregulation in disease:

Cancer: SWI/SNF mutations very common
Developmental disorders: CHD7, CHD8 mutations
Neurodevelopmental: Multiple remodeler genes
Immunodeficiency: Some SWI/SNF mutations

Therapeutic targeting:

Synthetic lethality approaches in cancer
BRG1/BRM ATPase inhibitors in development
Combination with other epigenetic drugs

16.8 Non-Coding RNA-Mediated Regulation

16.8.1 RNA Molecules That Control Genes

Non-coding RNAs (ncRNAs) play crucial roles in epigenetic regulation! Unlike mRNA that codes for proteins, these RNAs have regulatory functions.

Key concept: ncRNAs guide, recruit, and regulate epigenetic machinery to specific genomic locations.

16.8.2 MicroRNAs (miRNAs)

What they are:

Small RNA molecules (~22 nucleotides long)
Single-stranded
Evolutionarily conserved
Over 2,000 miRNAs in humans

How they’re made:

Transcription:
- RNA Pol II transcribes pri-miRNA (primary miRNA)
- Can be from intergenic regions or introns
Nuclear processing:
- Drosha enzyme cuts pri-miRNA → pre-miRNA (~70 nt hairpin)
- DGCR8 protein helps Drosha recognize correct sites
Export:
- Exportin-5 transports pre-miRNA to cytoplasm
Cytoplasmic processing:
- Dicer enzyme cuts pre-miRNA → mature miRNA duplex
- One strand selected as guide strand
RISC assembly:
- Guide strand loaded into RISC (RNA-Induced Silencing Complex)
- Argonaute (AGO) protein is catalytic component

What they do:

Bind to complementary sequences in mRNA (usually 3’ UTR)
Perfect match: mRNA cleavage (rare in animals)
Imperfect match: Translation repression (common)
Both: mRNA degradation

Result: Reduced protein production without changing DNA

Epigenetic aspects:

They regulate epigenetic enzymes:
- miR-29 family targets DNMTs → DNA demethylation
- miR-101 targets EZH2 → reduced H3K27me3
- miR-449a targets HDAC1 → increased acetylation
They’re regulated epigenetically:
- miRNA genes can be DNA methylated
- Silencing miRNAs can cause cancer
Heritable effects:
- miRNA expression patterns maintained through division
- Can be passed to daughter cells
- Some evidence for transgenerational inheritance

Examples:

let-7: Developmental timing, tumor suppressor
miR-21: Oncogenic, upregulated in many cancers
miR-34: p53-regulated, tumor suppressor
miR-155: Immune function

Clinical significance:

Biomarkers for cancer diagnosis
Therapeutic targets
miRNA mimics and inhibitors in development

16.8.3 Long Non-Coding RNAs (lncRNAs)

What they are:

RNA molecules >200 nucleotides
Do NOT code for proteins
Over 50,000 lncRNAs in humans
Often less conserved than protein-coding genes
Cell type and developmental stage-specific expression

How they work:

lncRNAs are incredibly versatile—they can:

Recruit chromatin-modifying complexes
Scaffold proteins into complexes
Decoy proteins away from DNA
Guide proteins to specific genomic locations
Organize nuclear domains

Major mechanisms:

1. Recruiters/Guides: - Bind to chromatin-modifying enzymes - Direct them to specific genomic loci - Can use complementary base-pairing with DNA/RNA

2. Scaffolds: - Bring multiple proteins together - Organize functional complexes - Multiple protein-binding domains

3. Decoys: - Sequester proteins away from their targets - Act as molecular sponges - Competitive inhibition

4. Structural organizers: - Form nuclear bodies - Organize chromatin domains - Create specialized compartments

16.8.4 Famous Examples of lncRNAs

XIST (X-Inactive Specific Transcript):

Function: Silences one X chromosome in females

Mechanism: 1. Expressed only from chromosome to be inactivated 2. Coats the X chromosome in cis (same chromosome) 3. Recruits PRC2 complex → H3K27me3 4. Recruits PRC1 complex → H2AK119ub1 5. Chromosome becomes heterochromatic 6. Most genes silenced

Size: ~17 kb

Importance: X-chromosome dosage compensation

Regulation: Controlled by TSIX (antisense lncRNA)

We’ll explore this in detail in the “X-Chromosome Inactivation” section!

HOTAIR (HOX Transcript Antisense RNA):

Function: Silences HOX genes

Mechanism: - Transcribed from HOXC locus - Acts in trans (affects distant HOX loci) - 5’ end binds PRC2 (deposits H3K27me3) - 3’ end binds LSD1/CoREST (removes H3K4me2) - Coordinates multiple repressive marks

Clinical significance: - Overexpressed in many cancers - Promotes metastasis - Prognostic biomarker

AIR (Antisense Igf2r RNA):

Function: Silences imprinted genes

Mechanism: - Antisense to Igf2r gene - Silences Igf2r and neighboring genes in cis - Important for genomic imprinting

MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1):

Function: Regulates alternative splicing

Location: Nuclear speckles

Mechanism: - Modulates splicing factor localization - Affects gene expression programs

H19:

Function: Imprinted lncRNA, multiple roles

Mechanism: - Regulates Igf2 expression - Processed into miR-675 - Acts as decoy for some proteins

Properties: - Maternally expressed - Paternally imprinted (silenced) - Important in development and growth

16.8.5 Small Interfering RNAs (siRNAs)

What they are:

Small RNA molecules (~21-23 nucleotides)
Double-stranded (initially)
Can be exogenous or endogenous

Origin:

Exogenous siRNAs:
- From viruses, transposons
- Cellular defense mechanism
- Cut by Dicer from long dsRNA
Endogenous siRNAs (endo-siRNAs):
- Produced from cellular sources
- Transposon-derived
- Pseudogene-derived
- Convergent transcription

How they work:

Long double-stranded RNA (dsRNA) enters cell or is produced
Dicer enzyme cuts dsRNA into ~21 bp fragments
One strand loaded into RISC complex
Guide strand directs RISC to complementary mRNA
Perfect complementarity → mRNA cleavage
Target destroyed

Epigenetic roles:

In many organisms (but not mammals), siRNAs can:

Direct DNA methylation to matching sequences
Induce histone modifications
Establish heterochromatin

RNA-directed DNA Methylation (RdDM) in plants:

siRNA production
    ↓
siRNA guides AGO protein to DNA
    ↓
Recruits DNA methyltransferases
    ↓
DNA methylation at matching sequence
    ↓
Gene silencing

Therapeutic uses:

First FDA-approved siRNA drug: Patisiran (2018)
Treats hereditary transthyretin amyloidosis
Silences TTR gene in liver
More siRNA drugs in development

Differences from miRNAs:

Feature	miRNA	siRNA
Origin	Endogenous	Endo- or exogenous
Structure	Hairpin precursor	Long dsRNA
Complementarity	Partial (usually)	Perfect
Effect	Translation repression	mRNA cleavage
Targets	Multiple mRNAs	Specific mRNA
Evolution	Conserved sequences	Variable

16.8.6 piRNAs (PIWI-Interacting RNAs)

What they are:

Largest class of small ncRNAs (~24-31 nucleotides)
Specific to animal germlines
Millions of unique sequences

Where found:

Germ cells (sperm and egg precursors)
Some somatic cells in certain organisms
Not in mammals: Only in germline

Function:

Primary role: Silence transposable elements (TEs)

Protect genome integrity in germline
Prevent transposon mobilization
Maintain genome stability across generations

How they work:

piRNA biogenesis:
- Transcribed from piRNA clusters
- Processed by PIWI proteins (not Dicer!)
- Ping-pong cycle amplifies piRNAs
Ping-pong cycle:

Sense transposon RNA
    ↓
Cleaved by piRNA-PIWI complex
    ↓
Antisense piRNA generated
    ↓
Guides another PIWI to sense RNA
    ↓
Cycle continues, amplifying piRNAs

Silencing:
- Post-transcriptional: Cleave transposon mRNA
- Transcriptional: Direct heterochromatin formation
- In some organisms, direct DNA methylation

PIWI proteins:

PIWI family of Argonaute proteins
PIWIL1-4 in mammals (also called MIWI, MILI, etc.)
Essential for spermatogenesis
Mutations cause male infertility

Epigenetic aspects:

Direct DNA methylation (in mammals):
- piRNAs guide DNA methylation to transposons
- Establishes germline methylation patterns
- Inherited in offspring
Heterochromatin formation:
- Recruit H3K9me3 machinery
- Silence transposon-rich regions
Germline genome defense:
- Protect germline from selfish genetic elements
- Maintain genome stability across generations

Transgenerational inheritance:

piRNA-directed silencing can persist for generations
Particularly well-studied in C. elegans
Provides heritable “memory” of transposon exposure

Clinical relevance:

Male infertility when piRNA pathway defective
Some cancer cells reactivate piRNA pathway
Potential therapeutic targets

16.8.7 Comparison of Small Regulatory RNAs

Feature	miRNA	siRNA	piRNA
Size	~22 nt	~21 nt	24-31 nt
Processing	Drosha, Dicer	Dicer	PIWI, no Dicer
Partner protein	Argonaute	Argonaute	PIWI
Source	Hairpin RNA	dsRNA	piRNA clusters
Targets	mRNAs	mRNAs, TEs	Transposons
Mechanism	Translation block	Cleavage	Cleavage + chromatin
Location	Ubiquitous	Various	Germline
Conservation	High	Variable	Low
Number in humans	~2,000	Variable	Millions

16.8.8 ncRNA Coordination

Regulatory networks:

ncRNAs don’t work in isolation:

lncRNAs can be processed into miRNAs:
- H19 → miR-675
- MIR100HG → miR-100, let-7, miR-125
miRNAs regulate lncRNA expression:
- Feedback loops
- Fine-tuning
lncRNAs can sponge miRNAs:
- Competing endogenous RNAs (ceRNAs)
- Sequester miRNAs away from targets
- Example: MALAT1 sponges several miRNAs
ncRNAs regulate epigenetic enzymes:
- Creating feedback loops
- Self-reinforcing gene expression states

Example network:

lncRNA recruits PRC2
    ↓
H3K27me3 deposited
    ↓
Gene silenced
    ↓
Gene encoded miRNA not produced
    ↓
Target of that miRNA is expressed
    ↓
Cellular phenotype changes

16.9 Epigenetic Control of Gene Expression

16.9.1 How Epigenetic Mechanisms Regulate Transcription

Epigenetic mechanisms don’t work in isolation—they coordinate to control when, where, and how much genes are expressed.

16.9.2 Promoter Silencing and Activation

Promoters = DNA sequences where transcription begins

Active promoter characteristics:

Open chromatin:
- Nucleosome-free region (NFR) at TSS
- Accessible to transcription factors
- DNase/ATAC-seq hypersensitive
Activating histone marks:
- H3K4me3 (strong signal)
- H3K9ac, H3K27ac
- H3K4me1 (at some promoters)
Unmethylated DNA:
- CpG island typically unmethylated
- Allows transcription factor binding
Bound transcription factors:
- Can recruit HATs
- Can recruit chromatin remodelers
- Initiate transcription

Silenced promoter characteristics:

Closed chromatin:
- Nucleosomes positioned over TSS
- Inaccessible to transcription factors
- Compact structure
Repressive histone marks:
- H3K27me3 (facultative silencing)
- H3K9me3 (constitutive silencing)
- Deacetylated histones
Methylated DNA:
- CpG island methylated
- Blocks transcription factor binding
- Recruits MBD proteins and repressors
Repressive complexes:
- PRC2/PRC1 (Polycomb)
- HP1 proteins
- HDACs

Transition between states:

SILENT PROMOTER:
Methylated DNA, H3K27me3, positioned nucleosomes
                ↓
           Activation signal
                ↓
Step 1: Pioneer factors bind (can access nucleosomal DNA)
Step 2: Recruit TET enzymes → DNA demethylation
Step 3: Recruit HATs → Histone acetylation
Step 4: Recruit chromatin remodelers → Nucleosome eviction
Step 5: H3K4 methyltransferases → H3K4me3 deposition
Step 6: Transcription factors bind
Step 7: RNA Pol II recruited
                ↓
ACTIVE PROMOTER:
Unmethylated DNA, H3K4me3, H3K27ac, open chromatin, active transcription

16.9.3 Enhancers, Insulators, and Boundary Elements

Enhancers:

What they are: - Regulatory DNA sequences - Activate transcription from a distance - Can be 10 kb or more from target gene - Can be upstream, downstream, or in introns

Epigenetic signatures of active enhancers:

H3K4me1 (mono-methylation):
- Marks both active and poised enhancers
- Distinguish from promoters (which have H3K4me3)
H3K27ac (acetylation):
- Marks ACTIVE enhancers specifically
- Poised enhancers lack this mark
Open chromatin:
- DNase/ATAC-seq accessible
- Nucleosome-depleted regions
Bound transcription factors:
- Cell-type specific
- Often clustered (“enhancer clusters”)
Enhancer RNAs (eRNAs):
- Transcribed from active enhancers
- Mark of enhancer activity

Enhancer states:

State	H3K4me1	H3K27ac	Activity
Active	+	+	Actively enhancing
Poised	+	-	Ready to activate
Primed	+	-	Developmental potential
Inactive	-	-	Not functional

Super-enhancers: - Clusters of enhancers - Very high levels of H3K27ac - Drive cell identity genes - Enriched for disease-associated variants - Often dysregulated in cancer

Enhancer-promoter interactions:

Physical contact through DNA looping:

Enhancer ----[looping]---- Promoter
  ↑                           ↑
H3K27ac                    H3K4me3
TF bound                   Pol II

      Mediator complex
      Cohesin complex

Insulators (Boundary elements):

What they are: - DNA sequences that block interactions - Create boundaries between chromatin domains - Prevent inappropriate enhancer-promoter contacts

CTCF - Master insulator protein:

Functions: 1. Enhancer blocking: - Prevents enhancers from activating wrong genes - Creates regulatory boundaries

Barrier function:
- Blocks spread of heterochromatin
- Maintains euchromatin/heterochromatin boundaries
Chromatin organization:
- Anchors chromatin loops
- Organizes TADs (Topologically Associated Domains)
Imprinting control:
- Regulates parent-specific expression
- Critical for imprinting centers

CTCF mechanism:

Enhancer [----CTCF----] Gene A    Gene B

Without CTCF:
Enhancer activates both Gene A and Gene B

With CTCF:
Enhancer ----| CTCF barrier
             ↓
Only Gene A activated
Gene B insulated

Epigenetic regulation of CTCF:

DNA methylation blocks CTCF binding
- Methylation at CTCF sites → loss of insulation
- Important for imprinting control
- Used at some imprinted loci
Example - H19/Igf2 locus:
- CTCF binds unmethylated ICR
- Blocks Igf2 enhancer access
- Methylation of ICR blocks CTCF
- Allows Igf2 activation

16.9.4 Role of Transcription Factors and Cofactors

Transcription factors (TFs):

Pioneer factors: - Can bind to nucleosomal DNA - First factors to access closed chromatin - Recruit chromatin remodelers - Examples: FoxA, GATA, PU.1, Oct4

Sequence-specific TFs: - Recognize specific DNA motifs - Recruit coactivators or corepressors - Integrate multiple signals

Epigenetic TF interactions:

TFs recruit epigenetic modifiers:
- HATs/HDACs
- Methyltransferases
- Chromatin remodelers
Epigenetic state affects TF binding:
- Open chromatin → easier binding
- DNA methylation → blocks binding
- Some TFs require specific histone marks

Positive feedback loops:

TF binds → Recruits HAT → Acetylation →
→ More TFs bind → More acetylation →
→ Stable active state

Coactivators:

p300/CBP: - HAT activity - Platform for multiple factors - Integrates signals from many TFs - Often rate-limiting for activation

Mediator complex: - Large complex (>30 subunits) - Bridge between enhancers and promoters - Interacts with RNA Pol II - Required for most transcription

TIP60/NuA4: - HAT complex - Activates transcription - DNA damage response

Corepressors:

NCoR/SMRT: - Recruit HDACs - Mediate transcriptional repression - Nuclear receptor corepressors

NuRD complex: - Contains CHD3/4 (remodeler) + HDAC1/2 - Both remodeling and deacetylation - Gene silencing

CtBP (C-terminal Binding Protein): - Recruits HDACs and HMTs - Mediates repression by many TFs

16.9.5 Polycomb and Trithorax Systems

Two ancient opposing systems that maintain cell identity:

Polycomb Group (PcG) Proteins - The Silencers:

PRC2 (Polycomb Repressive Complex 2): - Catalytic subunit: EZH2 (or EZH1) - Activity: Deposits H3K27me3 - Function: Initiates silencing - Targets: Developmental genes

PRC1 (Polycomb Repressive Complex 1): - Activity: Ubiquitinates H2A (H2AK119ub1) - Function: Maintains/reinforces silencing - Can: Compact chromatin - Recruitment: By H3K27me3 or DNA binding

Polycomb mechanism:

Development signal
    ↓
PRC2 recruited to gene
    ↓
H3K27me3 deposited
    ↓
PRC1 recruited
    ↓
H2AK119ub1 added
    ↓
Chromatin compaction
    ↓
Gene stably silenced
    ↓
Maintained through cell divisions

Trithorax Group (TrxG) Proteins - The Activators:

MLL complexes (Mixed Lineage Leukemia): - Activity: Deposit H3K4me3 - Function: Maintain active state - Oppose: Polycomb silencing

Function: - Keep developmental genes ON - Maintain differentiated state - Remember activation decisions

Balance determines fate:

Stem cells:
Many genes have BOTH H3K4me3 AND H3K27me3
→ "Bivalent" state
→ Poised for activation OR silencing

Differentiation:
Cell type A:
Gene X: H3K4me3 only → Active
Gene Y: H3K27me3 only → Silent

Cell type B:
Gene X: H3K27me3 only → Silent
Gene Y: H3K4me3 only → Active

Clinical significance: - EZH2 overexpressed in many cancers - MLL translocations cause leukemia - Therapeutic targets

16.9.6 Techniques in Epigenomics

1. Bisulfite Sequencing:

Detects DNA methylation
Treats DNA with bisulfite (changes unmethylated C to U)
Methylated C stays as C
Sequence and compare!

2. ChIP-seq (Chromatin Immunoprecipitation + Sequencing):

Detects histone modifications
Uses antibodies to pull down modified histones
Sequence the attached DNA
Map where modifications are genome-wide

3. ATAC-seq (Assay for Transposase-Accessible Chromatin):

Finds open (accessible) chromatin regions
Open chromatin = active genes
Closed chromatin = silenced genes

16.10 Differences in Prokaryotic vs. Eukaryotic Epigenomes

16.10.1 Prokaryotes: Limited Epigenetics

Prokaryotes have:

DNA methylation (but different purposes)
No histones (so no histone modifications)
Simpler regulation

Prokaryotic methylation:

Protects DNA from restriction enzymes
Helps DNA replication timing
Some gene regulation

16.10.2 Eukaryotes: Complex Epigenetics

Eukaryotes have:

Extensive DNA methylation
Many histone modifications
Chromatin remodeling
Non-coding RNAs
Complex epigenetic inheritance

Why the difference?

Eukaryotes need more sophisticated regulation
Multicellular organisms need cell type identity
Complex development requires epigenetic memory

16.11 Epigenetics and Environment

16.11.1 Nature AND Nurture

Epigenetics is the bridge between genes and environment!

Environmental factors that affect epigenetics:

Diet: Nutrients affect methylation patterns
Stress: Changes chromatin structure
Toxins: Can alter epigenetic marks
Exercise: Modifies histones in muscle
Social interaction: Affects brain epigenetics

Famous examples:

1. Dutch Hunger Winter (1944-1945):

Pregnant women experienced famine
Their children had epigenetic changes
These changes persisted into adulthood
Even affected grandchildren!

2. Agouti Mouse Experiment:

Mother mice fed different diets
Offspring had different coat colors
Same genes, different epigenetic marks!
Diet changed epigenetics, which changed appearance

3. Bee Development:

Queen bees and worker bees have identical DNA
Different diets (royal jelly vs. regular food)
Different epigenetic patterns
Completely different bodies and behaviors!

16.12 Epigenetic Inheritance

16.12.1 Passing On More Than Just DNA

Can epigenetic marks be inherited? Sometimes, YES!

16.12.2 Types of Epigenetic Inheritance

1. Mitotic inheritance (cell division):

Epigenetic marks copied when cells divide
How liver cells stay liver cells
Normal and common!

2. Transgenerational inheritance (across generations):

Some epigenetic marks passed from parent to offspring
More controversial
Happens in plants and worms
Rare in mammals

Reset mechanisms:

Most epigenetic marks are erased during reproduction
Prevents inheriting temporary changes
Some marks escape erasure!

16.13 Epigenetics and Disease

16.13.1 When Epigenetics Goes Wrong

Cancer:

Abnormal DNA methylation
Tumor suppressor genes get silenced
Oncogenes get activated
Epigenetic drugs can help!

Developmental Disorders:

Imprinting disorders (Prader-Willi, Angelman syndromes)
Rett syndrome (affects histone modifications)
Fragile X syndrome (abnormal methylation)

Aging:

Epigenetic marks change with age
“Epigenetic clock” can predict biological age
Some age-related changes are reversible!

Neurological Disorders:

Alzheimer’s disease
Depression (responds to environment via epigenetics)
Addiction (epigenetic changes in brain)

16.14 Epigenetic Therapies

16.14.1 Reversing Epigenetic Changes

Unlike DNA mutations, epigenetic changes are potentially reversible!

Epigenetic drugs:

DNMT inhibitors: Block DNA methylation
HDAC inhibitors: Affect histone acetylation
Already used to treat some cancers!

Future therapies:

Editing epigenetic marks precisely
Reversing age-related changes
Treating complex diseases

16.15 Epigenetic Reprogramming

16.15.1 Resetting the Epigenetic Landscape

Epigenetic reprogramming = Widespread erasure and re-establishment of epigenetic marks

This is essential for development and can be induced for research/therapeutic purposes.

16.15.2 Germline Reprogramming

When it happens: During formation of sperm and eggs (gametogenesis)

Why it’s necessary: - Erase somatic cell epigenetic patterns - Reset for totipotency - Establish sex-specific imprints - Prepare for next generation

Timeline in mammals:

1. Primordial Germ Cell (PGC) Specification (E6.5-E7.25 in mice): - Cells destined to become gametes specified - Still retain somatic epigenetic marks

2. Migration Phase (E7.5-E10.5): - PGCs migrate to developing gonads - Begin losing DNA methylation

3. Major Reprogramming (E10.5-E13.5): - Massive DNA demethylation: - Global loss of DNA methylation - Down to ~10% of normal levels - Most CpG islands demethylated - Imprints erased - Histone modification changes: - Loss of repressive marks - H3K9me2 reduction - H3K27me3 preserved at some regions - Chromatin remodeling: - More open chromatin state - Increased accessibility

4. Re-methylation (sex-specific):

In males (prospermatogonia): - Begins at E14.5 - Continues post-natally - Establishes paternal imprints - Methylation of repetitive elements

In females (oocytes): - Begins at birth (in humans, in utero) - Continues during oocyte growth - Establishes maternal imprints - Different pattern from sperm

Mechanisms of germline demethylation:

TET-mediated oxidation:
- TET1 and TET2 highly expressed
- Convert 5mC → 5hmC → 5fC → 5caC
- Base excision repair removes modified bases
Passive dilution:
- PGC proliferation without maintenance methylation
- Dilutes existing marks through replication
Active demethylation:
- TDG-mediated base excision repair
- Removes oxidized methylcytosines

Importance: - Prevents transgenerational inheritance of acquired marks (mostly) - Resets cellular potential - Allows each generation to start fresh - Some marks escape erasure (transgenerational epigenetic inheritance)

Marks that escape reprogramming: - Some transposable element methylation - Some imprinted genes (depends on locus) - Basis for transgenerational inheritance - More common in plants and invertebrates

16.15.3 Zygotic Reprogramming (During Embryogenesis)

When it happens: Shortly after fertilization

Why it’s necessary: - Transition from gamete to embryo - Establish totipotency - Prepare for development - Different from germline reprogramming in several ways

Asymmetric reprogramming:

Paternal genome (from sperm): - Rapid, active demethylation - Occurs within hours of fertilization - TET3-mediated oxidation - 5mC → 5hmC → 5fC → 5caC - Protamines replaced with histones - Most methylation lost by pronuclear stage

Maternal genome (from egg): - Slower, passive demethylation - Lost during DNA replication - DNMT1 excluded from nucleus (some embryos) - Gradual loss over cleavage divisions - Some methylation persists longer

Timeline (in mammals):

Fertilization
    ↓
Hours 0-6: Paternal active demethylation (TET3)
    ↓
Day 1-4: Maternal passive demethylation
    ↓
Day 3-4: Morula stage, minimal methylation (~10%)
    ↓
Day 4-6: Blastocyst, begin re-methylation
    ↓
Day 6+: Implantation, lineage-specific methylation
    ↓
Gastrulation: Establish somatic methylation patterns

De novo methylation in embryo:

Begins at blastocyst stage
DNMT3A and DNMT3B active
DNMT3L helps in some regions
Establishes embryonic patterns
Lineage-specific differences emerge

Regions protected from reprogramming:

Imprinted genes:
- Maintain parent-of-origin-specific methylation
- Protected from both germline and zygotic reprogramming
- Essential for normal development
Some repetitive elements:
- IAPs (Intracisternal A Particles) in mice
- Remain methylated
- Genomic defense mechanism
Some CpG islands:
- Housekeeping genes
- Essential genes

Histone modifications during reprogramming:

Sperm chromatin: Mostly protamines, some histones
Rapid histone incorporation after fertilization
H3K4me3 and H3K27me3 re-established
Bivalent domains form in ES cells

16.15.4 Induced Pluripotency (iPSCs)

What it is: Reprogramming somatic cells back to stem cell state

Discovery: Shinya Yamanaka, 2006 (Nobel Prize 2012)

The Yamanaka Factors - Four transcription factors: 1. Oct4 (Octamer-binding transcription factor 4) 2. Sox2 (Sex determining region Y-box 2) 3. Klf4 (Kruppel-like factor 4) 4. c-Myc (Cellular Myc)

How it works:

Initial phase (Days 0-3): - Forced expression of OSKM factors - Changes in cell morphology - Metabolic shift - MET (Mesenchymal-Epithelial Transition)

Intermediate phase (Days 3-9): - Stochastic gene expression changes - Some cells begin reprogramming - Many cells fail or senesce - Low efficiency (~0.1-1%)

Maturation phase (Days 9-15+): - Endogenous pluripotency genes activated - Oct4/Sox2/Nanog expression from endogenous loci - Exogenous factors no longer needed - iPSC colonies form

Epigenetic changes during reprogramming:

DNA demethylation:
- Promoters of pluripotency genes demethylated
- TET enzymes activated
- Gradual global changes
Histone modifications:
- Loss of differentiation-associated marks
- Gain of ES cell marks
- H3K4me3 at pluripotency genes
- Bivalent domains re-established
Chromatin remodeling:
- More open chromatin
- Nucleosome repositioning
- Changes in TAD structure
X-reactivation (in female cells):
- Inactive X can reactivate
- XIST silenced
- Both X chromosomes active (like ES cells)

Barriers to reprogramming:

Epigenetic memory:
- Somatic methylation patterns
- Silenced pluripotency genes
- Hard to erase
Senescence:
- p53/p21 pathway activation
- Cell cycle arrest
- Prevents transformation
Apoptosis:
- Stress response
- Many cells die

Enhancing reprogramming:

Vitamin C (enhances TET activity)
HDAC inhibitors (Valproic acid)
Demethylating agents (limited use)
Replace c-Myc with safer factors
Small molecules instead of factors

Applications:

Disease modeling:
- Generate patient-specific cells
- Study disease mechanisms
- Drug screening
Regenerative medicine:
- Personalized cell therapies
- Avoid immune rejection
- Generate replacement tissues
Basic research:
- Study cell fate decisions
- Understand pluripotency
- Model development

Challenges:

Incomplete reprogramming
Epigenetic memory persists
Genomic instability risk
Tumor formation risk
Efficiency still low
Quality control needed

Comparison of reprogramming types:

Feature	Germline	Zygotic	iPSC
Trigger	Developmental program	Fertilization	Yamanaka factors
Speed	Weeks	Days	Weeks
Completeness	Very high	Very high	Variable
Methylation change	Extensive	Extensive	Moderate
Histone changes	Extensive	Extensive	Moderate
Efficiency	100%	100%	<1%
Natural	Yes	Yes	No

16.16 Genomic Imprinting

16.16.1 Parent-of-Origin Specific Gene Expression

Genomic imprinting = Genes expressed from only one parental allele

Key concept: - Most genes: Both copies (maternal + paternal) expressed - Imprinted genes: Only ONE copy expressed - Which copy expressed depends on parent of origin - Established by epigenetic marks (primarily DNA methylation)

Statistics: - ~150 known imprinted genes in mice - ~100 known imprinted genes in humans - Often found in clusters - Evolutionarily conserved in mammals - Rare in other organisms (except some plants, insects)

16.16.2 Mechanism of Imprinting

Imprinting Control Regions (ICRs):

What they are: - Regulatory DNA sequences - Control imprinted gene clusters - Differentially methylated between parents - Act as master switches

How they work:

During gametogenesis:

Germline reprogramming
    ↓
Erasure of imprints
    ↓
Sex-specific de novo methylation
    ↓
Maternal ICRs: methylated in egg OR
Paternal ICRs: methylated in sperm
    ↓
Imprint established

After fertilization:

Embryo has:
- Maternal chromosome (ICR pattern from egg)
- Paternal chromosome (ICR pattern from sperm)
    ↓
ICR methylation maintained through development
    ↓
Differential expression based on parent of origin

Two main mechanisms:

1. Enhancer competition model:

Maternal chromosome (ICR unmethylated):
CTCF binds ICR → Blocks enhancer → Gene A silenced
                → Allows enhancer → Gene B expressed

Paternal chromosome (ICR methylated):
CTCF can't bind → Enhancer accesses Gene A → Gene A expressed
                → Gene B silenced

2. ncRNA-mediated silencing:

Maternal chromosome (ICR unmethylated):
lncRNA transcribed from ICR
    ↓
lncRNA spreads in cis
    ↓
Recruits repressive complexes
    ↓
Neighboring genes silenced

16.16.3 Examples of Imprinted Genes

IGF2 (Insulin-like Growth Factor 2):

Expression: Paternally expressed (maternal copy silenced)

Function: Promotes fetal growth

Mechanism: - H19/IGF2 locus on chromosome 11 - Shared enhancers - ICR between H19 and IGF2

Maternal allele:

ICR unmethylated
    ↓
CTCF binds ICR
    ↓
Insulator blocks IGF2 enhancer access
    ↓
IGF2 silenced, H19 expressed

Paternal allele:

ICR methylated
    ↓
CTCF cannot bind
    ↓
Enhancers access IGF2
    ↓
IGF2 expressed, H19 silenced

Clinical significance: - Beckwith-Wiedemann Syndrome: Loss of imprinting → IGF2 overexpression - Associated with childhood cancers - Overgrowth, organomegaly

H19:

Expression: Maternally expressed (paternal copy silenced)

Function: - lncRNA (doesn’t code for protein) - Tumor suppressor properties - Processed into miR-675 - Growth regulator

Relationship to IGF2: - Reciprocally imprinted with IGF2 - Share regulatory elements - Opposite expression patterns

UBE3A (Ubiquitin Protein Ligase E3A):

Expression: Brain-specific maternal expression

Unique feature: - In most tissues: Biallelic (both copies expressed) - In neurons: Monoallelic (maternal only) - Paternal silencing by SNHG14 antisense lncRNA

Function: - E3 ubiquitin ligase - Targets proteins for degradation - Important for synaptic function

Clinical significance:

Angelman Syndrome (maternal loss): - Deletion of maternal 15q11-13 - OR maternal UBE3A mutation - OR paternal uniparental disomy (both chr 15 from father) - Symptoms: - Severe intellectual disability - Speech impairment - Happy demeanor, frequent laughter - Ataxia, seizures

Prader-Willi Syndrome (paternal loss): - Deletion of paternal 15q11-13 - OR paternal mutations - OR maternal uniparental disomy - Different genes affected (SNRPN and others) - Symptoms: - Intellectual disability (milder) - Hyperphagia (excessive eating) - Obesity - Hypogonadism - Short stature

Same region, opposite syndromes!

SNRPN/SNURF:

Expression: Paternally expressed

Function: - SNRPN: Splicing factor - Also produces SNHG14 lncRNA - SNHG14 silences paternal UBE3A in neurons

Clinical relevance: Prader-Willi Syndrome

16.16.4 Evolutionary Theories of Imprinting

Parental Conflict Hypothesis (David Haig):

Theory: - Maternal and paternal genes have conflicting interests - Related to resource allocation

Paternal genes: - Favor increased resource extraction from mother - Promote fetal growth - Tend to be growth-promoting when expressed

Maternal genes: - Must balance resources among all offspring - Limit any single fetus’s demands - Tend to be growth-limiting when expressed

Examples: - IGF2 (paternal): Growth promoter - IGF2R (maternal): Degrades IGF2 - Opposing effects on fetal growth

Evidence: - Paternally expressed genes often promote growth - Maternally expressed genes often restrict growth - Important in placental mammals

16.16.5 Imprinting Disorders

Disorder	Genetic Cause	Chromosome	Key Features
Prader-Willi	Loss of paternal 15q11-13	15	Hyperphagia, obesity, ID
Angelman	Loss of maternal UBE3A	15	Severe ID, happy demeanor
Beckwith-Wiedemann	Loss of maternal 11p15	11	Overgrowth, cancer risk
Silver-Russell	Loss of paternal 11p15	11	Growth restriction
Transient neonatal diabetes	Loss of maternal 6q24	6	Temporary diabetes

16.17 X-Chromosome Inactivation (XCI)

16.17.1 Dosage Compensation in Mammals

The problem: - Females: XX (two X chromosomes) - Males: XY (one X chromosome) - X chromosome has ~1000 genes - Without compensation: Females would have 2X gene expression

The solution: X-Chromosome Inactivation (XCI)

Inactivate one X chromosome in each female cell
Occurs early in development
Random choice of which X to silence
Maintains choice through all cell divisions
Creates mosaic pattern in females

16.17.2 The XIST/TSIX Regulatory System

XIST (X-Inactive Specific Transcript):

Type: Long non-coding RNA (~17 kb in humans)

Function: Master regulator of XCI

Mechanism:

Counting phase:
- Cells “count” how many X chromosomes present
- Prepare for XCI if more than one X
Choice phase:
- ONE X chosen to remain active
- OTHER X(s) will be inactivated
- Random in most cells
- Some cells show skewing
Initiation:
- XIST transcribed from Xi (inactive X)
- Accumulates in cis (on same chromosome)
- Coats the entire X chromosome
- Visible as “cloud” in nucleus
Spreading:
- XIST spreads along chromosome
- ~150 copies of XIST coat Xi
- Recruits silencing complexes
Maintenance:
- DNA methylation locks in silence
- Histone modifications
- Late replication
- Nuclear periphery localization

Protein partners of XIST:

PRC2 complex → Deposits H3K27me3
PRC1 complex → Adds H2AK119ub1
SMCHD1 → DNA methylation, chromatin compaction
HDACs → Histone deacetylation
MacroH2A → Histone variant enrichment

TSIX (XIST antisense):

Type: lncRNA antisense to XIST

Function: Represses XIST

Mechanism: - Transcribed from Xa (active X) - Antisense orientation to XIST - Blocks XIST expression - Ensures only one X makes XIST

Regulatory loop:

Before XCI:
Both X chromosomes: TSIX expressed, XIST repressed

XCI initiation:
Future Xi: TSIX silenced, XIST activated
Future Xa: TSIX maintained, XIST repressed

Result:
Xi: No TSIX, high XIST → Inactivation
Xa: High TSIX, no XIST → Remains active

16.17.3 Stages of XCI

1. Initiation (Early embryo, day 5-6):

Counting and choice
XIST upregulation on Xi
Chromosome-wide transcriptional silencing begins

2. Spreading (Day 6-7):

XIST coating entire chromosome
Histone modifications accumulate:
- H3K27me3 enrichment
- H3K9me3 at some regions
- Loss of H3K4me2/me3
- Histone deacetylation
- MacroH2A incorporation

3. Establishment (Day 7-8):

Transcriptional silencing complete
CpG island methylation begins
Most genes silenced
Some “escape” genes remain active

4. Maintenance (Throughout life):

DNA methylation locks in silence
XIST may be dispensable
Chromatin state maintained
Late replication
Xi visible as Barr body

Chromatin features of Xi:

Feature	Active X (Xa)	Inactive X (Xi)
XIST coating	No	Yes
H3K27me3	Low	High
H3K9me3	Low	Moderate
H4 acetylation	High	Low
DNA methylation	Low	High
MacroH2A	Normal	Enriched
Replication	Early S	Late S
Location	Nuclear interior	Periphery
Structure	Euchromatic	Heterochromatic

16.17.4 Random vs Imprinted X Inactivation

Imprinted XCI (Extraembryonic tissues):

Occurs very early (4-cell stage)
Always silences paternal X
Non-random
In placenta and yolk sac
Evolutionary significance debated

Random XCI (Embryo proper):

Occurs at blastocyst stage
Random choice of which X
~50:50 ratio (but can skew)
In all somatic cells
Creates mosaic

Skewed XCI:

When inactivation isn’t 50:50:

Causes: - Random chance (especially with few cells) - Selection advantage/disadvantage - X-linked mutations favoring one X - Age (skewing increases with age)

Consequences: - Variable disease severity in carriers - Some females with X-linked diseases symptomatic - Example: Some female DMD carriers have muscle weakness

16.17.5 Escape from XCI

Not all genes are silenced:

~15% of X-linked genes “escape” XCI in humans
Expressed from both X chromosomes
Variable between individuals
May contribute to sex differences

Examples of escape genes: - XIST itself - KDM6A (H3K27me3 demethylase) - KDM5C (H3K4 demethylase) - SHOX (short stature gene) - Many pseudoautosomal region genes

16.17.6 Clinical Significance of XCI

1. X-linked disease carriers: - Usually protected by having one normal X - Severity depends on XCI pattern - Skewed XCI can lead to symptoms

2. Turner Syndrome (45,X): - Only one X chromosome - No XCI needed - But escape genes affected - Short stature, ovarian dysgenesis

3. Klinefelter Syndrome (47,XXY): - One X inactivated (like normal females) - But still have phenotypic effects - Possibly due to escape genes

4. Female mosaicism: - Every female is mosaic for X-linked genes - Can see this in: - Anhidrotic ectodermal dysplasia (patchy sweating) - Ornithine transcarbamylase deficiency - Some calico cats!

5. Cancer: - XCI maintenance lost in some cancers - Xi can reactivate - XIST deletions in some tumors

16.18 Epigenetics in Development and Differentiation

16.18.1 Lineage Commitment

Developmental progression:

Zygote (Totipotent)
    ↓
Blastocyst
    ↓ Lineage decisions
├─ Inner Cell Mass (Pluripotent) → Embryo
│   └─ Epiblast → Three germ layers
│       ├─ Ectoderm → Nervous system, skin
│       ├─ Mesoderm → Muscle, blood, bone
│       └─ Endoderm → Gut, lungs, liver
└─ Trophectoderm → Placenta

Epigenetic changes during differentiation:

1. Progressive restriction of potency:

Pluripotent stem cells: - Open chromatin - Bivalent promoters (H3K4me3 + H3K27me3) - Low DNA methylation at regulatory elements - High plasticity

Multipotent progenitors: - Some lineage genes silenced (H3K27me3) - Other lineage genes activated (H3K4me3) - Increased DNA methylation at silenced loci - Reduced plasticity

Differentiated cells: - Extensive DNA methylation - Clear H3K4me3 or H3K27me3 (not both) - Stable chromatin states - Committed identity

2. Bivalent domains in ES cells:

What they are: - Promoters with BOTH H3K4me3 AND H3K27me3 - “Poised” for activation or silencing - Found at developmental transcription factors - Unique to stem cells

Function: - Keep genes repressed but ready - Maintained by PRC2 (H3K27me3) and MLL (H3K4me3) - Allow rapid activation when needed - Resolution during differentiation

Resolution during differentiation:

ES cell: Gene X has H3K4me3 + H3K27me3 (bivalent)
                    ↓
            Differentiation signal
                    ↓
        ┌───────────┴───────────┐
    Cell A                  Cell B
H3K4me3 only            H3K27me3 only
Gene ON                 Gene OFF

3. Pioneer factors:

Definition: Transcription factors that: - Bind to closed chromatin - Can access nucleosomal DNA - Initiate opening of chromatin - Recruit other factors

Examples: - FoxA: Liver development - GATA: Blood cell development - PU.1: Macrophage differentiation - Oct4, Sox2, Nanog: Pluripotency - MyoD: Muscle differentiation

Mechanism:

Closed chromatin
    ↓
Pioneer factor binds
    ↓
Recruits chromatin remodelers
    ↓
Recruits HATs
    ↓
Open chromatin
    ↓
Other TFs can bind
    ↓
Gene activation
    ↓
Lineage commitment

16.18.2 Tissue-Specific Gene Expression

How cells maintain identity:

DNA methylation patterns: - Liver-specific genes: Unmethylated in liver, methylated in brain - Neuron-specific genes: Unmethylated in neurons, methylated in liver - Maintained through divisions - Stable and heritable

Enhancer landscapes: - Different cell types activate different enhancers - Marked by H3K27ac in active tissues - Cell-type-specific transcription factors - Super-enhancers at identity genes

Example - Blood cell differentiation:

Hematopoietic Stem Cell
        ↓
    ┌───┴───┐
Myeloid    Lymphoid
    ↓          ↓
  ┌─┴─┐      ┌─┴─┐
Granulocyte Monocyte  B cell  T cell

Epigenetic changes: - PU.1 high: Macrophage fate - GATA1 high: Erythrocyte fate - Antagonistic interactions - Mutual exclusion by chromatin changes - Progressive methylation of alternate fates

16.18.3 Stem Cell Plasticity

Types of stem cells:

1. Embryonic Stem Cells (ESCs): - Pluripotent - Can become any cell type - Bivalent chromatin domains - Open chromatin overall - Low DNA methylation

2. Adult Stem Cells: - Multipotent (limited fates) - Tissue-specific (hematopoietic, neural, etc.) - More restricted chromatin - Higher DNA methylation - Tissue-specific enhancers active

3. Induced Pluripotent Stem Cells (iPSCs): - Reprogrammed somatic cells - Similar to ESCs (but not identical) - Some epigenetic memory retained - Variable quality

Maintenance of stemness:

Pluripotency network: - Oct4, Sox2, Nanog transcription factors - Auto-regulatory loops - Repress differentiation genes - Maintain open chromatin

Epigenetic requirements: - Active DNA demethylation (TET1/2) - Polycomb silencing of differentiation genes - Open chromatin (chromatin remodelers) - Specific histone marks

Niche signals: - LIF/STAT3 (mouse ES cells) - FGF/ERK and TGFβ (human ES cells) - Maintain pluripotency network - Prevent spontaneous differentiation

16.19 Environmental Epigenetics

16.19.1 Nutrition, Toxins, and Stress Effects

Epigenetics is the molecular bridge between environment and gene expression!

16.19.2 Nutritional Effects

Methyl donors and one-carbon metabolism:

Key nutrients: - Folate (Vitamin B9) - Vitamin B12 - Choline - Methionine - Betaine

Biochemical pathway:

Dietary methyl donors
    ↓
S-Adenosyl Methionine (SAM)
    ↓ (methyl donor for DNMTs)
DNA methylation
    ↓
S-Adenosyl Homocysteine (SAH)

Effects of deficiency: - Reduced SAM availability - Global hypomethylation - CpG island hypermethylation (paradoxically) - Neural tube defects - Cancer risk increased

Agouti Mouse Experiment (Classic example):

Background: - Avy allele: Agouti gene with retrotransposon - Methylation of retrotransposon → normal brown coat - Unmethylated → yellow coat, obesity, diabetes

Experiment: - Pregnant mice fed methyl-donor-rich diet - OR control diet

Results: - Supplemented diet → More brown offspring - Control diet → More yellow offspring - Same genes, different epigenetic marks! - Diet changed epigenetics, changed phenotype

Implications: - Maternal nutrition affects offspring epigenome - Effects can persist into adulthood - Demonstrates environmental epigenetic effects

Genistein (soy compound): - Also shifts Agouti phenotype - DNMT inhibitor - Shows dietary compounds can affect epigenome

16.19.3 Toxic Exposures

Endocrine disruptors:

Bisphenol A (BPA): - Plastic component - Estrogen mimic - Effects: - Alters DNA methylation patterns - Affects developmental genes - Transgenerational effects in rodents - Linked to obesity, reproductive issues

Diethylstilbestrol (DES): - Synthetic estrogen (used 1940s-1970s) - Given to pregnant women - Epigenetic effects: - Altered methylation in offspring - Increased cancer risk (daughters) - Effects in grandchildren too - Classic transgenerational example

Heavy metals:

Arsenic: - Interferes with SAM metabolism - Alters DNA methylation globally - Histone modification changes - Cancer risk

Lead: - Changes DNA methylation - Affects neurodevelopment - Persistent effects

Cadmium: - Mimics zinc, interferes with enzymes - Epigenetic dysregulation - Found in cigarette smoke

Air pollution: - Particulate matter exposure - Changes DNA methylation - Effects on cardiovascular genes - Respiratory disease risk

16.19.4 Stress Effects

Psychological stress:

Hypothalamic-Pituitary-Adrenal (HPA) axis: - Stress → Cortisol release - Affects gene expression - Can cause epigenetic changes - Long-lasting effects

Early life stress:

Maternal care in rats (Michael Meaney’s work):

High maternal care: - Increased licking/grooming - Offspring: Less stressed as adults - Glucocorticoid receptor gene LESS methylated - Higher receptor expression - Better stress response

Low maternal care: - Less licking/grooming - Offspring: More stressed as adults - Glucocorticoid receptor gene MORE methylated - Lower receptor expression - Poor stress response

Reversibility: - Cross-fostering reverses effect - HDAC inhibitors reverse methylation - Shows epigenetic, not genetic

Human studies:

Childhood adversity: - Abuse, neglect alter epigenome - Changes in stress response genes - Increased psychiatric disorder risk - Changes detectable in blood cells

PTSD (Post-Traumatic Stress Disorder): - Altered methylation patterns - Especially in stress-related genes - May predispose or result from PTSD - Potential biomarkers

16.19.5 Transgenerational Epigenetic Inheritance

Definitions:

Intergenerational (F1, F2): - Direct exposure occurred - F0: Pregnant mother exposed - F1: Developing fetus exposed - F2: Developing germ cells in F1 exposed - Not “true” transgenerational

Transgenerational (F3+): - No direct exposure - Effects in F3 generation - True epigenetic inheritance - Controversial in mammals

Evidence in different organisms:

Plants: - Well-documented - Stress-induced changes persist - Paramutation phenomena - RNA-based inheritance

C. elegans: - Strong evidence - RNAi effects last generations - piRNA-based inheritance - Can last >100 generations

Mammals: - Limited clear examples - Most marks erased in germline - Some evidence for: - Metabolic effects - Stress responses - Toxin exposure effects - Mechanisms unclear

Dutch Hunger Winter (1944-1945):

Background: - Severe famine in Netherlands - Well-documented population - Pregnant women exposed

Findings in offspring (F1): - Exposed in early pregnancy: Increased obesity, cardiovascular disease - Exposed in late pregnancy: Glucose intolerance - DNA methylation changes detected decades later - Specific genes affected (IGF2, others)

F2 generation: - Some effects in grandchildren - Lower birth weights - Increased disease risk - Debated if true transgenerational (F1 gametes exposed)

Överkalix Study (Sweden):

Design: - Historical records of food availability - Multi-generational health data

Findings: - Paternal grandfather’s food abundance - Affected grandson’s longevity - Sex-specific effects - Suggested epigenetic inheritance

Interpretation: - Controversial - Confounding factors possible - Mechanism unknown - Needs replication

16.19.6 Epigenetic Epidemiology

Epigenome-Wide Association Studies (EWAS):

Similar to GWAS, but: - Measure DNA methylation (usually) - Associate with disease/trait - Identify epigenetic risk factors

Challenges: - Tissue specificity (blood vs target tissue) - Causation vs correlation - Environmental confounders - Temporal dynamics

Successes: - Smoking: Robust methylation signatures - Aging: Epigenetic clock markers - Some diseases: Consistent findings

Biobanks: - Large-scale epigenetic profiling - UK Biobank, others - Longitudinal data - Gene-environment interactions

16.20 Epigenetic Crosstalk

16.20.1 Interaction Between DNA Methylation, Histone Marks, and ncRNAs

Epigenetic mechanisms don’t work in isolation—they communicate and reinforce each other.

16.20.2 DNA Methylation ↔︎ Histone Modifications

1. DNA methylation recruits histone modifiers:

MBD proteins (Methyl-CpG Binding Domain): - Bind methylated DNA - Recruit HDACs - Remove activating histone marks - Example: MeCP2, MBD1-4

Pathway:

DNA methylation
    ↓
MBD proteins bind
    ↓
Recruit NuRD complex (HDAC + chromatin remodeler)
    ↓
Histone deacetylation
    ↓
Chromatin compaction
    ↓
Gene silencing reinforced

2. Histone modifications guide DNA methylation:

H3K9me3 → DNA methylation: - DNMT3A/B recognize H3K9me3 - UHRF1 binds H3K9me3 - Recruits DNMT1 - Establishes/maintains DNA methylation

H3K4me3 → Prevents DNA methylation: - DNMT3A/L inhibited by H3K4me3 - Protects CpG islands - Keeps promoters unmethylated

H3K36me3 → Gene body methylation: - DNMT3B recruited to H3K36me3 - Explains high methylation in gene bodies - Different from promoter methylation

Feedback loops:

Scenario 1 - Silencing:
DNA methylation → MBD → HDAC → H3K9 deacetylation
       ↑                              ↓
       ←─────── SUV39H ←──── H3K9me3 ←┘
       
Self-reinforcing silencing

Scenario 2 - Activation:
Unmethylated DNA → TET enzymes → 5hmC
       ↑                              ↓
       ←─────── Active marks ←──── Open chromatin
       
Self-reinforcing activation

16.20.3 Histone Modifications ↔︎ ncRNAs

1. ncRNAs recruit histone modifiers:

lncRNAs as guides: - HOTAIR recruits PRC2 (H3K27me3) and LSD1 (removes H3K4me) - XIST recruits PRC2 and PRC1 - AIR recruits G9a (H3K9me2) - Provides specificity to enzyme complexes

2. Histone marks affect ncRNA expression:

miRNA genes: - Can be silenced by DNA methylation - Regulated by histone modifications - Example: miR-34 regulated by p53 and epigenetics

lncRNA loci: - Often have bivalent marks in ES cells - Developmentally regulated - Epigenetic control of their expression

16.20.4 DNA Methylation ↔︎ ncRNAs

1. ncRNAs direct DNA methylation (in some organisms):

In plants (RdDM pathway): - siRNAs guide DNA methylation - Target complementary sequences - Direct de novo methylation

In mammals (limited): - piRNAs in germline - May guide DNA methylation to transposons - Mechanism less clear than plants

2. DNA methylation affects ncRNA expression:

miRNA silencing: - miRNA gene promoters can be methylated - Turns off tumor suppressor miRNAs - Cancer mechanism

lncRNA regulation: - Many lncRNA loci controlled by methylation - Tissue-specific patterns

16.20.5 Three-Way Interactions

Polycomb silencing:

lncRNA (e.g., XIST)
    ↓
Recruits PRC2
    ↓
H3K27me3 deposited
    ↓
Recruits PRC1
    ↓
H2AK119ub1 added
    ↓
Recruits DNMTs
    ↓
DNA methylation
    ↓
Stable, maintained silencing

All three mechanisms collaborate!

Activating network:

TF binding
    ↓
Recruits TET enzymes (DNA demethylation)
    ↓
Recruits HATs (histone acetylation)
    ↓
Chromatin remodeling
    ↓
H3K4me3 deposition
    ↓
Blocks DNA methylation
    ↓
Transcription (including eRNAs, miRNAs)
    ↓
Positive feedback

16.20.6 Crosstalk in Different Contexts

Development: - Coordinated changes in all marks - Sequential recruitment - Stable cell identity

Cancer: - Dysregulation of crosstalk - Hypermethylation + loss of H3K4me3 - Abnormal ncRNA expression - Vicious cycles

Aging: - Loss of coordination - Stochastic changes - “Epigenetic drift” - Accumulated errors

16.20.7 Therapeutic Implications

Combination therapies: - DNMT inhibitors + HDAC inhibitors - Synergistic effects - FDA-approved combinations - Target multiple mechanisms

Understanding crosstalk helps: - Predict drug effects - Design better therapies - Avoid resistance mechanisms - Target feedback loops

16.21 Epigenetic Mechanisms in Disease

Epigenetic dysregulation is a hallmark of many diseases, particularly cancer. Unlike genetic mutations, epigenetic changes are potentially reversible, making them attractive therapeutic targets.

16.21.1 Cancer

Epigenetic alterations in cancer are as important as genetic mutations!

Three major epigenetic changes:

1. CpG island hypermethylation:

Tumor suppressor silencing: - VHL (von Hippel-Lindau): Renal cell carcinoma - BRCA1: Breast/ovarian cancer - MLH1: Colorectal cancer (Lynch syndrome) - CDKN2A (p16): Many cancers - MGMT: Glioblastoma

Mechanism:

Normal cell:
CpG island unmethylated → Tumor suppressor ON

Cancer cell:
CpG island methylated → Tumor suppressor OFF
→ Loss of growth control
→ Cell proliferation

CIMP (CpG Island Methylator Phenotype): - Widespread CpG island methylation - Subset of colorectal, gastric, gliomas - Distinct molecular subtype - Different prognosis and treatment response

2. Global hypomethylation:

Effects: - Chromosomal instability - Activation of transposons - Oncogene activation - Loss of imprinting

Examples: - LOI (Loss of Imprinting) at IGF2 - Hypomethylation of repetitive elements - Reactivation of silenced genes

3. Histone modification alterations:

Common changes: - Loss of H4K20me3 (genome stability mark) - Loss of H4K16ac (active mark) - Aberrant H3K27me3 patterns - EZH2 overexpression (many cancers)

Mutations in epigenetic regulators:

Frequently mutated: - DNMT3A: Acute myeloid leukemia (AML) - TET2: Myeloid malignancies - IDH1/2: Gliomas, AML (produce oncometabolite) - EZH2: Lymphomas, solid tumors - KMT2A (MLL): Leukemias (translocations) - SWI/SNF components: 20% of cancers - ARID1A: Ovarian, endometrial cancers

Cancer-specific patterns:

Cancer Type	Key Epigenetic Change	Target Gene/Pathway
Colorectal	MLH1 methylation	DNA mismatch repair
Glioblastoma	MGMT methylation	DNA repair
Breast	BRCA1 methylation	DNA repair
Leukemia	MLL translocations	Histone methylation
Lymphoma	EZH2 mutations	H3K27 methylation
Renal	VHL methylation	HIF pathway

16.21.2 Neurological Disorders

Rett Syndrome:

Cause: Mutations in MECP2 (MethylCpG-binding Protein 2)

Function of MeCP2: - Binds methylated DNA - Recruits transcriptional repressors - Regulates thousands of neuronal genes - Critical for brain development and function

Symptoms: - Normal development until 6-18 months - Then regression: loss of speech, hand skills - Autism-like features - Seizures, breathing problems - Almost exclusively affects females (X-linked)

Epigenetic mechanism: - Loss of gene repression - Inappropriate gene expression in neurons - Synaptic dysfunction

Fragile X Syndrome:

Cause: CGG repeat expansion in FMR1 gene

Epigenetic silencing: - Repeats trigger DNA methylation - FMR1 promoter methylated - Gene silenced - No FMRP protein produced

Symptoms: - Intellectual disability - Autism spectrum features - Physical features (long face, large ears) - Most common inherited intellectual disability

Mechanism: - Abnormal methylation spreads from repeats - Heterochromatin formation - Complete gene silencing - Loss of translational regulation

Alzheimer’s Disease:

Epigenetic changes: - Global DNA hypomethylation - Region-specific hypermethylation - Histone deacetylation - Changes in genes for synaptic function

Therapeutic potential: - HDAC inhibitors improve memory in models - Targeting epigenetic changes may help - Reversibility is promising

Huntington’s Disease:

Epigenetic effects: - Mutant huntingtin affects HATs - Reduced histone acetylation - Transcriptional dysregulation - HDAC inhibitors show promise in models

16.21.3 Metabolic and Cardiovascular Diseases

Type 2 Diabetes:

Epigenetic factors: - Methylation changes in insulin signaling genes - Pancreatic β-cell dysfunction - Skeletal muscle insulin resistance - Influenced by maternal nutrition

Fetal programming: - Maternal diabetes affects offspring epigenome - Increased diabetes risk in offspring - Transgenerational effects possible

Cardiovascular Disease:

Atherosclerosis: - Endothelial dysfunction - Inflammatory gene activation - Smooth muscle proliferation - Epigenetic changes in all cell types

Heart failure: - Altered gene expression patterns - Fetal gene program reactivation - Epigenetic regulation of cardiac remodeling

Hypertension: - Prenatal influences - Epigenetic programming of blood pressure regulation

Obesity:

Epigenetic contribution: - Adipocyte differentiation - Metabolism regulation - Inflammatory state - Influenced by diet and exercise

Developmental origins: - Maternal obesity affects offspring - Epigenetic changes in metabolic genes - Increased obesity risk in children

16.21.4 Aging and Epigenetic Drift

Epigenetic changes with age:

Global patterns: - Overall DNA hypomethylation - CpG island hypermethylation - Loss of heterochromatin - Changes in histone marks

Epigenetic clock: - Specific CpG sites change predictably with age - Can estimate biological age - Developed by Steve Horvath and others - Different clocks for different purposes

Consequences: - Increased transcriptional noise - Loss of cellular identity - Senescence - Increased cancer risk - Reduced regenerative capacity

Interventions: - Caloric restriction affects epigenome - Exercise modifies epigenetic marks - Potential to slow epigenetic aging - Reprogramming partially reverses age

16.22 Epigenetic Biomarkers

16.22.1 Diagnostic Methylation Markers

Clinical utility: Epigenetic marks can indicate disease presence, type, or outcome.

MGMT methylation (Glioblastoma):

What it is: - MGMT = O6-Methylguanine-DNA Methyltransferase - DNA repair enzyme

Clinical significance: - Methylated MGMT → Gene silenced - Tumors can’t repair alkylation damage - Better response to temozolomide (chemotherapy) - Better prognosis - Routine clinical test!

Testing: - Bisulfite sequencing - Methylation-specific PCR - Guides treatment decisions

BRCA1 methylation (Breast/Ovarian Cancer):

Significance: - Silences BRCA1 without mutation - Creates “BRCAness” phenotype - Sensitive to PARP inhibitors - May respond to platinum therapy

Potential as biomarker: - Identify patients for targeted therapy - Complement genetic testing - More common than mutations

Septin 9 methylation (Colorectal Cancer):

Application: - Blood-based screening test - Detects methylated SEPT9 DNA - FDA-approved (Epi proColon) - Non-invasive alternative to colonoscopy

Performance: - ~70-90% sensitivity for CRC - ~80-95% specificity - Detects some advanced adenomas

Cell-free DNA methylation:

Liquid biopsies: - Detect tumor DNA in blood - Methylation patterns indicate: - Presence of cancer - Tissue of origin - Tumor burden - Treatment response

Advantages: - Non-invasive - Can monitor over time - Detect minimal residual disease - Earlier than imaging

Examples: - Galleri test (multi-cancer detection) - GRAIL’s technology - Guardant Health products

16.22.2 Prognostic and Therapeutic Indicators

Prognostic methylation patterns:

AML (Acute Myeloid Leukemia): - DNMT3A mutations → Worse prognosis - IDH1/2 mutations → Altered methylation, better with targeted therapy - TET2 mutations → Affect outcome

Neuroblastoma: - Methylation classification - Predicts outcome - Guides intensity of treatment

Medulloblastoma: - DNA methylation profiling - Identifies subgroups - Different prognoses and treatments

Predictive biomarkers:

Predict response to specific therapies:

MGMT (as above): Temozolomide response

MLH1 methylation: - Colorectal cancer - Microsatellite instability - May respond to immunotherapy

PITX2 methylation: - Breast cancer - Predicts benefit from tamoxifen

DNA methylation signatures:

Smoking signature: - Robust, persistent - Lung cancer risk - cg05575921 (AHRR gene) - Reverses slowly after quitting

Alcohol signature: - Associated with consumption - Cancer and disease risk

Exercise signature: - Changes with physical activity - Metabolic benefits

16.23 Epigenetic Therapies

16.23.1 DNMT Inhibitors

Mechanism: Block DNA methyltransferases → Reduce DNA methylation → Reactivate silenced genes

5-Azacytidine (Vidaza) and Decitabine (Dacogen):

Mechanism: - Nucleoside analogs - Incorporated into DNA during replication - Form covalent complex with DNMTs - Trap and degrade DNMTs - Result: DNA demethylation

Approved uses: - Myelodysplastic syndromes (MDS) - Acute myeloid leukemia (AML) - Chronic myelomonocytic leukemia (CMML)

Effects: - Reactivate tumor suppressors - Increase immune recognition - Differentiation therapy - Improve survival

Challenges: - Non-specific (genome-wide effects) - Require cell division (S-phase) - Transient effects - Need ongoing treatment

16.23.2 HDAC Inhibitors

Mechanism: Block histone deacetylases → Increase histone acetylation → Open chromatin → Gene activation

FDA-Approved HDAC Inhibitors:

Vorinostat (SAHA, Zolinza): - Approved for cutaneous T-cell lymphoma (CTCL) - Pan-HDAC inhibitor - Oral administration

Romidepsin (Istodax): - Approved for CTCL and peripheral T-cell lymphoma - Class I HDAC inhibitor - IV administration

Belinostat (Beleodaq): - Approved for peripheral T-cell lymphoma - Pan-HDAC inhibitor

Panobinostat (Farydak): - Approved for multiple myeloma - Pan-HDAC inhibitor - Used with bortezomib and dexamethasone

Mechanisms of action: - Increase acetylation → gene expression changes - Induce apoptosis - Cell cycle arrest - Differentiation - Immune effects - DNA damage sensitization

Combination therapies:

DNMT + HDAC inhibitors: - Synergistic effects - Used together in MDS/AML - More effective than either alone - Target complementary mechanisms

Rationale: - DNA demethylation + histone acetylation - Maximize gene reactivation - Overcome resistance

16.23.3 BET Inhibitors

Target: BET (Bromodomain and Extra-Terminal) proteins

BET proteins: - BRD2, BRD3, BRD4, BRDT - Read acetylated histones (via bromodomains) - Recruit transcriptional machinery - Especially active at super-enhancers

BET inhibitors (e.g., JQ1, I-BET, OTX015):

Mechanism: - Compete with acetylated lysines - Displace BET proteins from chromatin - Disrupt super-enhancer function - Silence oncogene expression

Targets: - MYC oncogene (many cancers) - Other super-enhancer-driven genes - Particularly effective in certain leukemias

Status: - Multiple clinical trials - Promising preclinical data - Not yet FDA-approved - Challenges with resistance

16.23.4 EZH2 Inhibitors

Target: EZH2 (Enhancer of Zeste Homolog 2)

Function of EZH2: - Catalytic component of PRC2 - Deposits H3K27me3 - Silences tumor suppressors - Often overexpressed/mutated in cancer

EZH2 inhibitors (Tazemetostat):

Mechanism: - Block EZH2 methyltransferase activity - Reduce H3K27me3 - Reactivate silenced genes

FDA approval: - Epithelioid sarcoma (2020) - Follicular lymphoma with EZH2 mutation (2020)

Clinical applications: - EZH2-mutant cancers - EZH2-overexpressing cancers - Being tested in various tumors

16.23.5 LSD1 Inhibitors

Target: LSD1 (Lysine-Specific Demethylase 1)

Function of LSD1: - Removes H3K4me1/me2 (activating marks) - Can also remove H3K9me1/me2 - Important in cancer and differentiation

Inhibitors (tranylcypromine, ORY-1001, others):

Applications: - Acute myeloid leukemia - Small cell lung cancer - Promotes differentiation - Clinical trials ongoing

16.23.6 Limitations and Future Directions

Current challenges:

Specificity: - Current drugs affect genome-wide - Hard to target specific genes - Off-target effects

Resistance: - Cells adapt - Redundant pathways - Requires combination therapy

Delivery: - Some drugs have poor bioavailability - Toxicity concerns - Need better formulations

Patient selection: - Who will respond? - Biomarker-driven approaches needed - Precision epigenetic medicine

Future strategies:

More selective inhibitors
Combination therapies
Epigenome editing (next section!)
Personalized approaches
Early intervention/prevention

16.24 Epigenome Editing and CRISPR Applications

16.24.1 Targeted Epigenetic Modification

Unlike genome editing (which changes DNA sequence), epigenome editing changes epigenetic marks at specific loci without altering sequence.

16.24.2 dCas9-Based Epigenome Editors

dCas9 (dead Cas9): - Cas9 with inactivated nuclease domains - Can still bind DNA when guided by sgRNA - Doesn’t cut—just sits there - Platform for recruiting effectors

General strategy:

sgRNA designs → Target locus
    ↓
dCas9 binds to target
    ↓
Fused epigenetic effector acts locally
    ↓
Epigenetic change at specific site

16.24.3 dCas9-TET1 Systems

For DNA demethylation:

Design: - dCas9 fused to TET1 catalytic domain - Or recruit via adapter proteins (SunTag, etc.)

Function: - TET1 converts 5mC → 5hmC → demethylation - Targeted to specific CpGs - Can reactivate methylation-silenced genes

Applications demonstrated: - Reactivate BRCA1 in cancer cells - Demethylate specific CpG islands - Reverse pathological methylation - Research tool for causality studies

Examples:

Cancer cell: BRCA1 promoter methylated → Gene OFF
    ↓
dCas9-TET1 targeted to BRCA1
    ↓
Local demethylation
    ↓
BRCA1 reactivated → Gene ON

16.24.4 dCas9-DNMT3A Systems

For DNA methylation:

Design: - dCas9 fused to DNMT3A catalytic domain - Target hypomethylated oncogenes

Function: - Add methylation to specific sites - Silence target genes - Establish stable repression

Applications: - Silence oncogenes - Model disease methylation patterns - Study function of methylation - Potential therapeutic approach

16.24.5 Targeted Histone Modification

dCas9-p300 (Histone acetyltransferase):

Function: - Acetylate histones at target locus - Activate gene expression - Open chromatin

Applications: - Activate tumor suppressors - Rescue disease mutations - Enhance endogenous expression

Results: - Robust activation (>100-fold) - Specific to target gene - Durable effects

dCas9-LSD1 (Histone demethylase):

Function: - Remove H3K4me2 at targets - Repress gene expression

dCas9-EZH2 or dCas9-KRAB:

Function: - Deposit H3K27me3 (EZH2) - Or recruit heterochromatin (KRAB) - Silence genes

CRISPRa and CRISPRi:

CRISPRa (Activation): - dCas9-VP64 or dCas9-VPR - Strong transcriptional activators - Epigenetic remodeling follows

CRISPRi (Interference): - dCas9-KRAB - Deposits H3K9me3 - Stable repression - “Epi-silencing”

16.24.6 Multiplexing and Screening

Multiple targets simultaneously: - Array of sgRNAs - Edit multiple loci - Create complex patterns - Model disease states

Epigenome-wide screens: - Perturb chromatin regulators - Identify functional elements - Map enhancers - Understand gene regulation

16.24.7 Challenges and Future Directions

Current limitations:

Specificity: - Off-target binding possible - Need highly specific sgRNAs - Validation required

Efficiency: - Variable efficacy at different loci - Some sites hard to edit - Chromatin accessibility matters

Durability: - Some changes temporary - Need stable modification - Maintenance mechanisms

Delivery: - Get tools into cells - In vivo delivery challenging - Size of constructs

Future improvements:

Better editors: - More specific Cas variants - Smaller size (Cas12, Cas14) - Improved catalytic domains

Delivery methods: - AAV vectors - Nanoparticles - Exosomes - Direct protein delivery

Clinical translation: - Ex vivo editing (like CAR-T) - In vivo targeting - Disease-specific applications - Combination with genome editing

Therapeutic potential:

Cancer: - Reactivate tumor suppressors - Silence oncogenes - Reverse abnormal methylation

Genetic diseases: - Compensate for mutations - Activate compensatory genes - Silence toxic repeat expansions (like Huntington’s)

Regenerative medicine: - Enhance reprogramming - Direct cell fate conversion - Improve differentiation protocols

Research applications:

Dissect cause vs. effect of epigenetic changes
Map functional elements
Understand gene regulation
Validate therapeutic targets
Model diseases

16.25 Epigenomics Techniques

16.25.1 Studying Epigenetic Marks Genome-Wide

Epigenomics = Studying all epigenetic marks across the entire genome

Like proteomics studies all proteins, epigenomics studies all epigenetic modifications!

Modern techniques allow genome-wide mapping of: - DNA methylation - Histone modifications - Chromatin accessibility - 3D chromosome organization - Non-coding RNA expression

16.25.2 Bisulfite Sequencing

Purpose: Map DNA methylation at single-base resolution

Principle: Bisulfite converts unmethylated C to U, methylated C stays as C

Chemistry:

Unmethylated Cytosine + Bisulfite → Uracil → reads as T
Methylated Cytosine + Bisulfite → Cytosine → reads as C

Methods:

1. Whole Genome Bisulfite Sequencing (WGBS):

Protocol: 1. Extract genomic DNA 2. Treat with sodium bisulfite 3. PCR amplify (U → T) 4. High-throughput sequencing 5. Map reads to genome 6. Calculate methylation %

Advantages: - Complete coverage - Single-base resolution - Quantitative - Unbiased

Disadvantages: - Expensive (high coverage needed) - Requires significant DNA input - Complex data analysis - Can’t distinguish 5mC from 5hmC

2. Reduced Representation Bisulfite Sequencing (RRBS):

Protocol: - Digest DNA with MspI (cuts at CCGG) - Enriches for CpG-rich regions - Size selection - Bisulfite treatment - Sequencing

Advantages: - More cost-effective - Covers most CpG islands - Less DNA required - Good for precious samples

Disadvantages: - Incomplete coverage - Biased toward CpG islands - Misses intergenic regions

3. Targeted Bisulfite Sequencing:

Focus on specific regions
PCR amplification of targets
Deep coverage of regions of interest
Very cost-effective for limited targets

5hmC detection:

oxidative bisulfite sequencing (oxBS-seq)
TET-assisted bisulfite sequencing (TAB-seq)
Can distinguish 5hmC from 5mC
More complex protocols

16.25.3 ChIP-Based Methods

Chromatin Immunoprecipitation (ChIP): Core technique for studying histone modifications and protein-DNA interactions

Basic ChIP protocol:

Crosslinking:
- Formaldehyde fixes proteins to DNA
- Captures transient interactions
- Preserves in vivo state
Chromatin fragmentation:
- Sonication or enzymatic digestion
- 200-500 bp fragments
- Nucleosome-sized pieces
Immunoprecipitation:
- Add specific antibody
- Antibody binds target (histone mark, TF, etc.)
- Pull down with beads
- Wash away non-specific binding
DNA purification:
- Reverse crosslinks
- Purify DNA
- DNA that was bound by target protein
Analysis:
- qPCR for specific regions (ChIP-qPCR)
- OR sequencing for genome-wide (ChIP-seq)

ChIP-seq (ChIP + sequencing):

Workflow: - Standard ChIP protocol - Library preparation - High-throughput sequencing - Map reads to genome - Call “peaks” (enriched regions)

Applications: - Map histone modifications genome-wide - Find transcription factor binding sites - Identify regulatory elements - Compare between conditions/cell types

Advantages: - Genome-wide - Quantitative - High resolution (depends on fragment size) - Well-established

Challenges: - Antibody quality critical - Requires many cells (millions) - Some epitopes hard to ChIP - Analysis complexity

Variations:

ChIP-exo: - Adds exonuclease digestion - Near base-pair resolution - Very precise binding sites

Mint-ChIP: - Multiplexed barcoding - Many samples simultaneously - Cost-effective

16.25.4 CUT&RUN and CUT&Tag

Improvements over ChIP: Less input, better signal-to-noise, no crosslinking artifacts

CUT&RUN (Cleavage Under Targets & Release Using Nuclease):

Protocol: 1. Bind cells to beads 2. Permeabilize cells (keep nuclei intact) 3. Add antibody against target 4. Add protein A-MNase fusion 5. Activate MNase with calcium 6. MNase cuts DNA around antibody binding 7. Released fragments diffuse out 8. Sequence released DNA

Advantages: - Low cell number (1,000-100,000 cells) - Low background - No crosslinking needed - Fast protocol - Better signal-to-noise than ChIP

CUT&Tag (Cleavage Under Targets & Tagmentation):

Protocol: - Similar to CUT&RUN - But uses Tn5 transposase - Tn5 cuts AND adds adapters - Streamlined library prep

Advantages: - Even lower cell number (100-1,000 cells) - Faster library prep - Very low background - Excellent for rare cell types

Applications: - Same as ChIP (histone marks, TFs) - Better for limited samples - Single-cell versions emerging

16.25.5 Chromatin Accessibility

Principle: Open chromatin (active regulatory regions) is accessible to enzymes

ATAC-seq (Assay for Transposase-Accessible Chromatin):

Protocol: 1. Isolate nuclei 2. Add Tn5 transposase 3. Tn5 inserts into accessible DNA 4. Simultaneously fragments and tags 5. PCR amplify 6. Sequence

What it reveals: - Open chromatin regions - Nucleosome positioning - Transcription factor footprints - Active regulatory elements

Advantages: - Very fast (hours, not days) - Low cell number (500-50,000 cells) - Simple protocol - No antibodies needed

Applications: - Map enhancers and promoters - Identify active regulatory regions - Compare cell types/conditions - TF footprinting

DNase-seq (DNase I hypersensitive sites sequencing):

Protocol: - Treat nuclei with DNase I - DNase cuts accessible regions - Sequence cut fragments - Map hypersensitive sites

History: - Older method than ATAC-seq - ENCODE project used extensively - Now largely replaced by ATAC-seq - But large existing datasets

MNase-seq: - Uses micrococcal nuclease - Cuts between nucleosomes - Maps nucleosome positions - Reveals phasing, spacing

16.25.6 Chromosome Conformation Capture

Purpose: Understand 3D chromosome organization—which DNA regions physically interact

3C (Chromosome Conformation Capture):

Protocol: 1. Crosslink proteins to DNA (formaldehyde) 2. Digest with restriction enzyme 3. Ligate nearby fragments (captures interactions) 4. Reverse crosslinks 5. PCR to detect specific interactions

Limitation: Only one-to-one interactions, specific loci

4C (Circular 3C): - One-to-all interactions - Pick one “viewpoint” - Find all interactions with that locus

5C (3C-Carbon Copy): - Many-to-many within a region - Uses primers for many loci - More throughput than 3C

Hi-C:

Protocol: - Similar to 3C - But biotin mark at junctions - Pull down with streptavidin - Pair-end sequencing - All-to-all interactions genome-wide!

What it reveals: - Topologically Associated Domains (TADs) - A/B compartments (active/inactive) - Chromatin loops - Enhancer-promoter interactions - 3D genome structure

Applications: - Map genome organization - Identify regulatory interactions - Understand structural variants - Disease mechanisms

Micro-C and Micro-C XL: - Use MNase instead of restriction enzymes - Higher resolution - Nucleosome-level interactions

ChIA-PET (Chromatin Interaction Analysis by Paired-End Tag): - Combines ChIP + Hi-C - Protein-mediated interactions - Example: Map all loops involving a specific TF

16.25.7 RNA-seq for Epigenetic Studies

RNA-seq: Sequence all RNA to measure gene expression

Standard RNA-seq: - Measures mRNA levels - Shows which genes are ON/OFF - Consequence of epigenetic state

Small RNA-seq: - Specifically sequences small RNAs - miRNAs, siRNAs, piRNAs - 18-40 nucleotide size selection - Different library prep

Applications for epigenetics: - Profile miRNA expression - Discover novel small RNAs - Study RNA-mediated silencing

GRO-seq / PRO-seq (Global Run-On sequencing): - Measures active transcription - Nascent RNA - More direct measure than RNA-seq - Detects enhancer RNAs (eRNAs)

16.25.8 Single-Cell Epigenomics

Why single-cell?: - Cell-to-cell heterogeneity - Rare cell populations - Developmental trajectories - Tumor heterogeneity

Technologies:

scATAC-seq: - Single-cell chromatin accessibility - 100s-1000s of cells - Identify cell types by accessibility - Trace lineages

scChIP-seq: - Challenging (low material) - Limited success - Being improved

scCUT&Tag: - Better for single cells - Low background - Works well

scBS-seq / scWGBS: - Single-cell methylation - Sparse coverage - Can still identify patterns

sc-Multi-omics: - Combine multiple measurements - Example: scNMT-seq (methylation + transcriptome + accessibility) - Comprehensive cell state

16.25.9 Spatial Epigenomics

Emerging field: Map epigenetic marks while preserving tissue spatial information

Technologies: - Spatial ATAC-seq - Spatial Cut&Tag - In situ sequencing methods

Applications: - Understand tissue organization - Cell-cell interactions - Disease microenvironments

16.26 Epigenetic Databases and Bioinformatics

16.26.1 Major Epigenomics Resources

ENCODE (Encyclopedia of DNA Elements):

What it is: - International consortium - Map functional elements in human genome - Comprehensive epigenomic data

Data types: - ChIP-seq for many histone marks, TFs - DNA methylation (WGBS, RRBS) - Chromatin accessibility (DNase-seq, ATAC-seq) - RNA-seq - Hi-C - Multiple cell types and tissues

Access: - www.encodeproject.org - Freely available - Searchable database - Track hubs for genome browsers

Roadmap Epigenomics:

Focus: Human tissue and cell type reference

Coverage: - 127 reference epigenomes - Primary cells and tissues - Histone modifications - DNA methylation - RNA-seq

Applications: - Reference for “normal” epigenomes - Identify regulatory elements - Understand disease variants - Tissue-specific patterns

Access: - www.roadmapepigenomics.org - Integrated with ENCODE

IHEC (International Human Epigenome Consortium):

Global collaboration
1000 reference epigenomes planned
Multiple countries contributing
Standardized protocols

GEO (Gene Expression Omnibus):

What it is: - NIH database - Archive for functional genomics data - Includes epigenomics datasets

Contents: - ChIP-seq, ATAC-seq, RNA-seq, methylation - Published and unpublished data - Searchable by technique, organism, etc.

Access: - www.ncbi.nlm.nih.gov/geo - Free download - GEOquery R package

ArrayExpress (EBI): - European equivalent of GEO - Similar data types - www.ebi.ac.uk/arrayexpress

TCGA (The Cancer Genome Atlas):

For cancer epigenomics: - DNA methylation (450K, EPIC arrays) - RNA-seq - Clinical data - Many cancer types

4D Nucleome: - Chromosome structure - Hi-C and other 3D data - Nuclear organization - www.4dnucleome.org

Blueprint Epigenome: - Blood cell epigenomes - Hematopoietic system focus - Reference for blood disorders

16.26.2 Visualization Tools

UCSC Genome Browser:

Features: - Visualize genomic data - Multiple tracks simultaneously - Compare samples - Custom track upload

For epigenomics: - Display ChIP-seq peaks - Show methylation levels - Overlay multiple marks - Connect to ENCODE data

Access: genome.ucsc.edu

IGV (Integrative Genomics Viewer):

Features: - Desktop application - Fast visualization - Multiple file formats - Good for detailed examination

Use for: - View BAM/BigWig files - Check specific loci - Compare conditions - Quality control

Download: software.broadinstitute.org/software/igv

WashU Epigenome Browser: - Excellent for Hi-C data - Virtual 4C - Long-range interactions - epigenomegateway.wustl.edu

HiGlass: - Specialized for Hi-C - Fast, interactive - Compare multiple Hi-C datasets - higlass.io

16.26.3 Analysis Tools and Pipelines

For ChIP-seq: - MACS2: Peak calling - DiffBind: Differential binding - ChIPseeker: Annotation - deepTools: Visualization

For ATAC-seq: - Same as ChIP-seq tools - Plus footprinting tools - HINT, Wellington for TF footprints

For BS-seq: - Bismark: Alignment and methylation calling - methylKit: Differential methylation - bsseq: Statistical analysis - dmrseq: Find DMRs

For Hi-C: - Juicer: Processing Hi-C data - HiC-Pro: Complete pipeline - cooltools: Analysis and visualization - HiCExplorer: TAD calling, interactions

Machine Learning: - Predict enhancers from sequence - Classify chromatin states - Impute missing marks - ChromHMM, Segway: State segmentation

16.27 Epigenetics and Evolution

16.27.1 Role in Adaptation and Phenotypic Plasticity

Epigenetics as evolutionary mechanism:

Unlike genetic mutations, epigenetic changes: - Arise more rapidly - Are often reversible - Can respond to environment - May facilitate adaptation

Phenotypic plasticity:

Definition: Same genotype produces different phenotypes based on environment

Epigenetic basis: - Environmental signals → epigenetic changes - Rapid phenotypic adjustments - Without genetic change - Reversible if environment changes

Examples:

Honeybees: - Queen vs. worker determined by diet - Same genome, different methylation - Royal jelly alters epigenome - Produces dramatically different phenotypes

Locusts: - Solitary vs. gregarious forms - Crowding induces epigenetic changes - Color, behavior, morphology change - Serotonin-mediated epigenetic switch

Plants: - Same plant produces different leaves - Sun vs. shade leaves - Underwater vs. aerial leaves - Methylation differences

16.27.2 Lamarckian Inheritance Revisited

Lamarck’s hypothesis (1809): - Organisms pass acquired characteristics to offspring - “Use and disuse” - Giraffes stretch necks → offspring have longer necks

Modern view: - Mostly wrong for genetic inheritance - BUT: Epigenetic inheritance shows Lamarckian features - Environment can affect heritable marks (sometimes) - Transgenerational epigenetic inheritance IS real (in some cases)

Evidence:

C. elegans starvation: - Starvation alters small RNA pathways - Effects last 3+ generations - Metabolic adaptations inherited - piRNA-mediated

Plant stress memory: - Stress induces methylation changes - Some changes inherited - Offspring more stress-resistant - Adaptive advantage

Human examples (debated): - Dutch Hunger Winter effects - Överkalix study - Mechanisms unclear in mammals

Why limited in mammals?: - Extensive epigenetic reprogramming - Two waves of erasure (germline + zygotic) - Protects against inheriting damage - But some marks escape

16.27.3 Evolutionary Dynamics

Neutral drift: - Epigenetic marks can drift - Creates epigenetic diversity - Raw material for selection

Canalization: - Developmental buffering - Epigenetics maintains consistent development - Despite genetic or environmental variation

Evolutionary capacitance: - Epigenetic variation hidden - Released under stress - Source of rapid adaptation - HSP90 example in Drosophila

Facilitated variation: - Epigenetics enables phenotypic exploration - Without genetic commitment - “Try before you buy” - If beneficial, may become genetic

16.28 Epigenetics in Reproduction

16.28.1 Gametogenesis

Sperm epigenome:

Unique features: - Most histones replaced by protamines - Hypercondensed DNA - Some histones retained at: - Developmental genes - Imprinted loci - Regulatory elements

Retained histones: - Carry modifications (H3K4me3, H3K27me3) - May mark important genes - Potential information transfer

DNA methylation: - Established during spermatogenesis - Paternal imprints set - Maintained after fertilization (at imprints)

Small RNAs: - Abundant in sperm - piRNAs, miRNAs, tsRNAs - May carry information to egg - Role unclear but intriguing

Egg epigenome:

Characteristics: - Histones with maternal modifications - Maternal imprints established - Large cytoplasm with RNAs - TET3 protein for paternal demethylation

Maternal factors: - Proteins inherited from egg - Guide early development - Include epigenetic regulators - Critical for zygotic reprogramming

16.28.2 Imprinting Maintenance

Challenge: Maintain imprints through reprogramming

Protection mechanisms:

At imprinting centers: - Resist demethylation in PGCs - Protected by specific factors - Maintain parental identity - Critical for development

PGC7/STELLA: - Protein that protects methylation - Binds to H3K9me2 - Prevents TET-mediated demethylation - Important for imprints

ZFP57: - Zinc finger protein - Recognizes methylated ICRs - Recruits maintenance machinery - Mutations cause imprinting disorders

16.28.3 Embryo Development

Preimplantation: - Zygotic genome activation - First cell fate decisions - Establishing pluripotency - ICM vs. trophectoderm

Epigenetic control: - Different methylation patterns - Histone modifications guide fate - ncRNAs regulate development - Chromatin remodeling essential

Implantation and gastrulation: - Lineage commitment - Tissue-specific patterns established - Bivalent domains resolved - Irreversible differentiation begins

16.29 Epigenetics and the Microbiome

16.29.1 Microbial Metabolites Influencing Host Epigenome

Key concept: Gut microbes produce metabolites that enter circulation and affect host epigenetics

Short-Chain Fatty Acids (SCFAs):

Butyrate, propionate, acetate:

Production: - Bacterial fermentation of fiber - Produced in gut - Absorbed into bloodstream - Reach tissues

Mechanism: - HDAC inhibitors! - Increase histone acetylation - Alter gene expression - Especially in colon cells

Effects: - Anti-inflammatory - Enhance barrier function - Reduce cancer risk - Affect metabolism

Studies: - High-fiber diet → More SCFAs → Beneficial epigenetic changes - Germ-free mice have different histone acetylation - Restore with SCFA supplementation

Other microbial metabolites:

Folate: - Produced by gut bacteria - Methyl donor - Affects DNA methylation

Trimethylamine N-oxide (TMAO): - From dietary choline - Bacterial metabolism required - Linked to cardiovascular disease - May have epigenetic effects

Polyphenol metabolites: - Plant compounds metabolized by microbes - Can affect DNMTs, HDACs - Green tea catechins, resveratrol

Immune effects: - Microbiome affects immune cell epigenetics - Regulatory T cell differentiation - TET expression affected by metabolites - Impacts inflammation, tolerance

Disease implications:

Colon cancer: - Dysbiosis alters metabolite profile - Changes in colonic cell epigenetics - Reduced butyrate production - Loss of protective effects

Obesity and metabolism: - Microbiome affects host metabolism partly via epigenetics - SCFA effects on adipocytes - Changes in metabolic gene expression

IBD (Inflammatory Bowel Disease): - Altered microbiome and metabolites - Epigenetic changes in intestinal cells - Barrier dysfunction - Chronic inflammation

Therapeutic potential: - Prebiotics, probiotics - Fecal transplants - Microbiome modulation - Indirect epigenetic therapy

16.30 Epigenetics in Plants

16.30.1 Paramutation

Paramutation = One allele influences the epigenetic state of another allele

Classic example: Maize b1 locus (pigmentation)

Mechanism: - B-I allele (high pigment) - B’ allele (low pigment) - When together, B-I → B’ - Change is heritable! - Mediated by small RNAs and chromatin changes

Mechanism:

B-I/B' heterozygote
    ↓
siRNAs from B' locus
    ↓
Guide DNA methylation at B-I
    ↓
B-I becomes B'*
    ↓
Change maintained in offspring
    ↓
Even without B' allele!

Significance: - Trans-generational epigenetic change - Communication between alleles - Role of small RNAs - Examples in multiple organisms

16.30.2 Transposon Silencing

Challenge: Plants have many transposons

Epigenetic silencing: - DNA methylation - Histone modifications (H3K9me2) - Small RNA-directed

RdDM pathway (RNA-directed DNA Methylation):

Mechanism: 1. Transposon transcribed (Pol IV) 2. Produces dsRNA 3. Dicer processes to siRNAs 4. siRNAs guide AGO proteins 5. AGO recruits DNA methylation machinery 6. Transposon methylated and silenced

Importance: - Genome defense - Prevent transposon mobilization - Maintain genome stability - Heritable silencing

Transgenerational: - Silencing inherited - Can last many generations - Adaptive advantage

16.30.3 Vernalization

Vernalization = Cold treatment induces flowering

Epigenetic mechanism in Arabidopsis:

FLC gene (Flowering Locus C): - Repressor of flowering - Prevents flowering before winter

Before cold: - FLC high → No flowering

During cold: - FLC gradually silenced - PRC2 recruits to FLC - H3K27me3 accumulates - Takes weeks of cold

After cold: - FLC remains silenced - H3K27me3 maintained - Flowering can occur - Memory of winter!

Mechanism details:

Cold exposure
    ↓
COLDAIR lncRNA induced
    ↓
Recruits PRC2 to FLC
    ↓
H3K27me3 spreads
    ↓
FLC silenced
    ↓
Mitotically stable
    ↓
Flowering competence acquired

Reset: - Each generation must experience cold - Reprogramming in germline - FLC reactivated in seeds

Agricultural importance: - Many crops require vernalization - Winter wheat vs. spring wheat - Understanding allows manipulation - Breeding applications

16.30.4 Other Plant Epigenetic Phenomena

Hybrid vigor (heterosis): - Small RNAs from parents - Create new epialleles - Contribute to hybrid performance

Grafting effects: - Mobile small RNAs - Move between stock and scion - Can cause heritable changes - Epigenetic communication

Stress memory: - Plants “remember” stress - Faster/stronger response on re-exposure - Some changes heritable - Adaptive transgenerational plasticity

16.31 Epigenetic Clock and Aging

16.31.1 DNA Methylation-Based Aging Biomarkers

Epigenetic clock = Predictor of age based on DNA methylation

Key concept: Specific CpG sites change methylation predictably with age

16.31.2 Horvath Clock (2013)

Development: - Steve Horvath analyzed thousands of samples - Found 353 CpG sites that change with age - Works across tissues - Predicts chronological age (r=0.96!)

Features: - Pan-tissue (works in most cell types) - Accurate across ages - Can estimate age within 3-4 years - Ticks in post-mitotic cells (neurons)

Applications: - Estimate age of sample - Forensics - Biological age vs. chronological age

Biological age: - Epigenetic age ≠ chronological age - Accelerated aging associated with: - Obesity - Smoking - Stress - Disease - Decelerated aging with: - Healthy lifestyle - Caloric restriction (in models)

16.31.3 Hannum Clock

Different approach: - Blood-specific - 71 CpG sites - Also accurate - Captures different aspects

16.31.4 PhenoAge

Purpose: Predict mortality and healthspan

Design: - Trained on mortality data - 513 CpG sites - Better predicts outcomes than chronological age

Clinical utility: - Risk stratification - Intervention trials - Track biological age changes

16.31.5 GrimAge

Most recent and predictive: - Predicts mortality, healthspan - Includes smoking pack-years - Multiple protein levels estimated - Best performer for outcomes

16.31.6 Mechanisms of Epigenetic Aging

What causes clock changes?

Theories:

Epigenetic drift:
- Random accumulation of errors
- Loss of maintenance fidelity
- Stochastic changes
Programmed aging:
- Developmental program continues
- Regulated changes
- Adaptive (or maladaptive) response
Damage accumulation:
- DNA damage
- Oxidative stress
- Inflammatory signals
- Drive epigenetic changes

Consequences: - Loss of cell identity - Transcriptional noise - Dysfunction - Senescence - Disease risk

16.31.7 Interventions

Can we slow the clock?

Caloric restriction: - Slows epigenetic aging (in models) - Effects in humans being studied - Metabolic benefits

Exercise: - Associated with younger epigenetic age - Reduces accelerated aging - Muscle-specific effects

Reprogramming: - Partial reprogramming reverses age - Yamanaka factors (transient) - Rejuvenates cells - Without losing identity

Senolytic drugs: - Clear senescent cells - Improve healthspan - May affect epigenetic age

NAD+ boosters: - Sirtuin activators - May slow some aging marks - Ongoing research

Future: - Epigenetic age as clinical outcome - Personalized interventions - Track intervention efficacy - “Add years to life, life to years”

16.32 Systems and Computational Epigenetics

16.32.1 Integration of Multi-Omics

Multi-omics approach: Combine epigenomics with other -omics for comprehensive understanding

Data types integrated:

Genomics: - DNA sequence variants - Mutations - SNPs

Epigenomics: - DNA methylation - Histone modifications - Chromatin accessibility - 3D structure

Transcriptomics: - mRNA levels - ncRNA expression - Alternative splicing - Single-cell resolution

Proteomics: - Protein abundance - Post-translational modifications - Protein-protein interactions

Metabolomics: - Small molecules - Metabolic pathways - Metabolic state

Integration benefits: - Understand causality - Gene regulation logic - Disease mechanisms - Identify therapeutic targets

Example workflow:

Genomics: What CAN happen (potential)
    ↓
Epigenomics: What's ACCESSIBLE (possibility)
    ↓
Transcriptomics: What's TRANSCRIBED (activity)
    ↓
Proteomics: What's MADE (function)
    ↓
Metabolomics: What's HAPPENING (outcome)

16.32.2 Machine Learning Models for Epigenetic Prediction

Applications of ML/AI:

1. Predicting chromatin states:

ChromHMM, Segway: - Hidden Markov Models - Integrate multiple histone marks - Classify genome into states - Promoter, enhancer, quiescent, etc.

Deep learning: - Neural networks - Predict chromatin state from sequence - More flexible than HMMs

2. Predicting regulatory elements:

DeepSEA: - Predicts chromatin features from DNA sequence - Trained on ENCODE data - Identifies functional variants

Basenji, Enformer: - Predict gene expression from sequence - Incorporate epigenomic features - Very large models (transformers)

3. Imputation of missing marks:

ChromImpute: - Predict unmeasured epigenetic marks - Use correlation with other marks - Fill in incomplete datasets

Avocado: - Deep tensor factorization - Impute missing experiments - Pan-tissue, pan-mark

4. Variant effect prediction:

DeepSEA, Basset: - Predict effects of SNPs on regulatory elements - Identify functional variants - Prioritize GWAS hits

5. Single-cell analysis:

Dimensionality reduction: - Handle sparse scATAC-seq data - Identify cell types - Pseudotime trajectories

Integration: - Seurat, Signac - Combine scRNA + scATAC - Multi-modal analysis

6. Disease classification:

DNA methylation classifiers: - Brain tumor classification - Cancer type identification - Better than histology alone

Example - CNS tumors: - 450K methylation array - Classifier identifies >100 tumor types - Clinically implemented - Changes diagnosis in ~10% of cases

16.32.3 Network Modeling

Gene regulatory networks: - Transcription factors - Enhancers - Target genes - Epigenetic states

Tools: - SCENIC (scRNA-seq) - PECA (integrative) - CausalPath

Chromatin interaction networks: - Hi-C data - Enhancer-promoter links - Predict target genes of enhancers

16.32.4 Challenges and Future Directions

Data integration challenges: - Different platforms - Batch effects - Missing data - Causality inference

Interpretability: - Black-box models - Need biological insight - Explainable AI important

Future directions: - Foundation models (like GPT for genomes) - Better causal inference - Personalized predictions - Clinical decision support - Drug response prediction

16.33 Key Takeaways

Epigenetics = Changes in gene activity without changing DNA sequence
Main mechanisms:
- DNA methylation: Adding CH₃ to cytosines (usually silences genes)
- Histone modifications: Chemical tags on histones (activate or repress genes)
- Chromatin remodeling: Moving nucleosomes (changes accessibility)
- Non-coding RNAs: Guide epigenetic changes
Key players:
- DNMTs (DNMT1, DNMT3A/B): Add methyl groups
- TET enzymes: Remove methyl groups
- HATs/HDACs: Add/remove histone acetylation
- HMTs/KDMs: Add/remove histone methylation
- Chromatin remodelers (SWI/SNF, ISWI, CHD, INO80): Reorganize nucleosomes
- lncRNAs, miRNAs, siRNAs, piRNAs: Regulatory RNAs
Important concepts:
- CpG islands: Usually unmethylated, mark promoters
- Histone code: Modification combinations convey meaning
- Bivalent domains: Poised genes in stem cells (H3K4me3 + H3K27me3)
- Polycomb/Trithorax: Maintain silenced/active states
- Imprinting: Parent-of-origin-specific expression
- X-inactivation: Dosage compensation in females
Development and differentiation:
- Progressive restriction of potency
- Lineage-specific epigenetic patterns
- Maintained through cell divisions
Reprogramming:
- Germline: Reset between generations
- Zygotic: Reset after fertilization
- iPSCs: Artificial reprogramming to pluripotency
Environmental sensitivity:
- Diet (methyl donors, SCFAs)
- Toxins (BPA, heavy metals, smoke)
- Stress (early life experiences)
- Exercise and lifestyle
- Potentially transgenerational
Disease involvement:
- Cancer: Hypermethylation of tumor suppressors, global hypomethylation
- Neurological: Rett (MECP2), Fragile X (FMR1 methylation)
- Metabolic: Diabetes, obesity, cardiovascular
- Aging: Epigenetic drift, clock changes
Biomarkers:
- MGMT methylation (glioblastoma prognosis)
- Epigenetic clocks (biological age)
- Liquid biopsies (cancer detection)
- Disease classification
Therapies:
- DNMT inhibitors: 5-azacytidine, decitabine (FDA-approved for MDS/AML)
- HDAC inhibitors: Vorinostat, romidepsin (FDA-approved for lymphoma)
- EZH2 inhibitors: Tazemetostat (FDA-approved)
- Combination therapy: DNMT + HDAC inhibitors
- Epigenome editing: dCas9-based tools (experimental)
Techniques:
- BS-seq (WGBS, RRBS): DNA methylation mapping
- ChIP-seq: Histone modifications and TF binding
- CUT&RUN/CUT&Tag: Improved ChIP alternatives
- ATAC-seq: Chromatin accessibility
- Hi-C: 3D genome organization
- Single-cell methods: Cell-to-cell variation
Resources:
- ENCODE: Comprehensive functional elements
- Roadmap Epigenomics: Reference epigenomes
- UCSC Genome Browser, IGV: Visualization
- Machine learning: Prediction and integration
Evolution and adaptation:
- Phenotypic plasticity
- Potential Lamarckian inheritance (limited)
- Evolutionary capacitance
Special topics:
- Plants: Paramutation, vernalization, transposon silencing
- Microbiome: Metabolites affect host epigenome
- Reproduction: Germline reprogramming, imprint maintenance
- Aging: Epigenetic clocks, potential interventions

Sources: Information adapted from MedlinePlus Genetics, Nature Reviews, NCBI, current epigenetics research literature, ENCODE Project, Roadmap Epigenomics Project, and recent publications in epigenetics and epigenomics.