Epigenetic Editors: CRISPR's Geneless Gene Control Revolution!

Genome editing is revolutionizing biology. But since when have scientists stopped there?

Today, we go a layer higher with epigenetic editing! Let’s jump into it.

No Breaks for DNA

CRISPR Comes for the Epigenome

Genome editing has reshaped biology and medicine.

Genome editing tools help us understand how our genes work and fight diseases, with the first personalized therapy saving a baby’s life just last year!

But scientists are never satisfied.

Traditional genome editing relies on DNA breaks. Powerful, yes, but also potentially harmful. And researchers started asking: can we expand the potential of genetic medicine, without cutting the DNA?

Epigenetic editing promises a solution.

Epigenetics studies how reversible modifications, like DNA methylation or changes in DNA-associated histones, affect the way your genes work, without altering the DNA sequence itself.

These modifications matter in everything from cancer to neurodegenerative diseases. Being able to write or erase them at will would be amazing!

Epigenetic editors do just that. These tools repress or activate gene expression by rewriting epigenetic modifications. No harmful DNA breaks, no permanent sequence change!

These editors have two parts:

• Effector domain: Transcriptional activators or repressors, these proteins write or erase epigenetic information.

• Programmable DNA-binding proteins: The targeting system that guides the effector domain to the chosen sequence.

Scientists have tried the whole spectrum of DNA-targeting proteins, but dCas9 (deactivated Cas9) is emerging as the standard. It’s easy to program, versatile, and safe. No catalytic activity → no accidental DNA breaks.

But is it perfect? Not quite.

CRISPR-dCas9: Powerful, and Bulky

All CRISPR-Cas systems share two traits: they’re powerful, and they’re bulky.

CRISPR-based epigenetic editors like CRISPRoff (dCas9 fused to epigenetic editors) can silence multiple genes at once. But getting these giant complexes into cells is a problem!

Most “classical” delivery systems don’t work well for them:

• Adeno-associated viruses (AAV): These viruses can package target DNA, but they are too small for these big systems.

• mRNA electroporation: This commonly used method has high cytotoxicity, reducing the efficiency of transformation.

• Lipid nanoparticle-mRNA complexes: They require cell type-specific formulation, which is still a challenge.

So, DNA or RNA delivery seems like a no-go!

Delivering the Complete Package with RNPs

The alternative is directly delivering the ribonucleoprotein (RNP) complexes.

In this way, the dCas9 protein and the sgRNA are already assembled by the time they arrive in the cell. This has some advantages:

• No transgenic or viral expression, with no risk of viral DNA integration!

• The RNPs stay for a bit, and then they degrade: this transient delivery decreases off-target editing.

And the best vehicle for RNPs? Virus-like particles.

Virus-like particles, or VLPs for short, are protein structures similar to viruses, but without the genetic material. No genetic materials means no infection and safe nanoparticles!

RENDER: Engineered VLPs Platform

And here comes today’s paper!

The authors developed RENDER (Robust ENveloped Delivery of Epigenome-editor Ribonucleoproteins), an engineered VLP platform that delivers big epigenome editors (as RNPs) into human cells.

RENDER enables both transient and durable epigenetic silencing and activation of endogenous genes, using different epigenome editors. All without DNA breaks or viral integration!

RENDER packages full epigenome editors inside VLPs adapted from murine leukemia virus. In practice, the VLPs are protein nanostructures enveloping the RNPs!

Initial Demonstration and Optimization

The researchers started with silencing a GFP reporter in human cells. A single dose of VLPs packaging an epigenetic repressor yielded up to 75% silencing of GFP expression!

And the silencing duration could be modulated. VLPs loaded with a CRISPRi system showed a transient effect that reverted in a week or so, while the ones loaded with a CRISPRoff system silenced genes for weeks!

To improve the efficiency, the researchers optimized RENDER:

• Increasing the fraction of CRISPRoff plasmids during packaging improved the RNP payload and brought more effective gene silencing. This was called RENDER-CRISPRoff v2.

• Adding extra sgRNA plasmids during packaging enhanced the performance of v2 at lower doses, yielding v3. The combined optimizations produced a ~4.4-fold potency increase relative to v1!

And a single v3 dose maintained GFP cells in 92% of cells for 49 days post-treatment!

Expanding to Different Cell Types

Many delivery systems only work sometimes.

Let’s say you develop it for neurons: the same delivery system will probably not work in T cells. This creates big headaches for researchers and a huge bottleneck for translation.

The researchers tested RENDER on eight cell types, ranging from embryonic kidney to liver and glioblastoma. They focused on silencing the expression of 2 genes encoding membrane proteins.

RENDER-CRISPRoff v3 produced robust silencing in many lines, but the performance varied. This highlights the importance of considering the target regulatory context.

Also, some lines with lower efficiency responded to higher doses or enhancers! So, it can work with some more optimization.

Therapeutic Angle: T cells and Neurons

The researchers applied RENDER to two therapeutically relevant systems:

• Primary human T cells:
  The traditional way to deliver epigenetic editors to T cells is via electroporation. Can RENDER work better? The team delivered RENDER-CRISPRoff v3 to primary T cells and compared it to CRISPRoff mRNA electroporation. RENDER caused minimal viability loss, while electroporation caused 40% cell death. On the other hand, the silencing efficiency was much lower for RENDER (21%) than for electroporation (97%). Increasing the dose of RENDER moved the silencing to 44%, still lower, but better. So, there is probably a tradeoff, with RENDER being more gentle, but requiring more process optimization.

• iPSC-derived neurons and Tau protein:
  It’s pretty hard to do genome editing in neurons. This sucks because many neurodegenerative diseases could be treated this way, in theory. In particular, the accumulation of Tau protein is linked to many diseases, and the reduction of its levels improved the conditions of mouse models with Alzheimer’s. Targeting the gene responsible, RENDER reduced Tau protein by over 60% at day 8, and 50% at day 15 in iPSC-derived neurons! A substantial and durable reduction in a neuron type that is hard to edit.

To Conclude

A super interesting work!

I’m no expert in genone/epigenome editing, but I see some strengths in RENDER:

• Increased size for RNP delivery: VLPs delivered larger epigenome editors, hard to package in AAV or LNPs.

• Transient RNP exposure: This minimizes prolonged exposure and reduces off-target risk.

• Versatile applications: It worked across many cell lines, and it can multiplex targets!

There are also limitations:

• Efficiency depends on cell type

• Electroporation had better silencing efficiency, though with much higher toxicity. A tradeoff to consider!

There’s a lot of work to do before we see RENDER in the clinic! But this was a cool paper, and very well written!

Go and get all the details here!

If you made it this far, thank you! What do you think of epigenome editing? Can it solve some of the problems of genome editing? Have you worked on something similar? Reply and let me know!

More Room:

• AI, but with DNA: Yeah, DNA can do everything, I promise you. In this study, the authors show that DNA-based molecular circuits can autonomously perform supervised learning in vitro. They build a DNA neural network that learns pattern classification directly from molecular training examples, storing learned information in molecular concentrations rather than digital memory. The system successfully classifies multiple sets of 100-bit patterns using these learned molecular “weights.” This work demonstrates that molecular circuits can implement embedded learning, opening the door to smart molecular systems capable of decision-making in applications ranging from biomedicine to soft materials. 

• Living Bacteria as Biosensors: Since DNA can do everything, it’s important to detect it. Living bacteria can be engineered as DNA biosensors that detect and analyze foreign DNA in vitro or in vivo by exploiting natural DNA uptake and processing pathways. This review outlines design strategies for these bactosensors, including genetic toolkits such as CRISPR–Cas systems, and surveys applications in biomedicine, agriculture, and food and water safety. It discusses choices of bacterial chassis, DNA uptake mechanisms, signal transduction, and output formats (for example, fluorescence or drug release), and compares living bacterial sensors with in vitro DNA biosensors in terms of sensitivity, specificity, multiplexing, and practical deployment.

• MISO: From Japan to Cryo-EM. I’m a sucker for good names, and for cryo-EM. This paper has both! Single-particle cryo-EM can, in principle, determine protein structures from small amounts of material, but in practice requires much larger samples. In this study, the authors introduce a micro isolation (MISO) workflow that combines microfluidic protein purification with direct cryo-EM grid preparation. Using this approach, they solve structures of soluble and membrane proteins starting from less than 1 µg of protein and going from cells to grids within hours. The method dramatically reduces material requirements and enables structural analysis of proteins that were previously difficult or impossible to study.

• Share Plenty of Room with founders or builders

  I help biotech and deep tech companies transform complex technologies into engaging content that builds credibility with investors, partners, and potential hires. Let’s chat!

• Know someone who’d love this?
  Pass it on! Sharing is the easiest way to support the newsletter and spark new ideas in your circle.

• Got a tip, paper, or topic you want me to cover?
  I’d love to hear from you! Just reply to this email or reach out on social.