Aseem Ansari started out as a chemist for a good reason. “What do biology students do for a living other than memorize taxonomy? I wasn’t interested,” he said. But a chance encounter put a fork in his road. “I ran into a very inspiring professor who dragged me out of the cafeteria and took me to this honors course.” 

Ansari attended a course on bringing chemistry and biology together in the mid-1980s at St. Xavier’s College. The first day he attended, it addressed how to synthesize DNA. Ansari was shocked enough that it knocked him off his chemistry path. 

Today, Ansari is the chair of chemical biology and therapeutics at St. Jude Children’s Research Hospital where he works on a new kind of genetic engineering that doesn’t edit genes at all. Instead, he focuses on editing transcription. By targeting transcription factors and creating small synthetic transcription factors, Ansari hopes to finetune gene expression and treat diseases that arise from complex genetic causes that can’t be treated with simple genetic changes. 

How did you incorporate biology into your chemistry background?

In that lecture, we talked about all sorts of things, including bacteriophages. A bacteriophage is just protein that can be crystallized from bacterial cultures. But if it has a genome, it can come alive. That caught my attention. 

We read Har Gobind Khorana’s work from the 1960s on synthesizing nucleotides and Mark Ptashne’s book, A Genetic Switch, on the early understanding of gene regulation. That introduced the idea that gene regulation may be the biggest control point for how genomes are used. From there, all sorts of interesting questions come up. Everybody gets a genome; that’s interesting. How does a genome, a chemical collection of nucleotides, give rise to a sentient being?   

Eventually, I made my way from India to Northwestern University where I went back to my chemistry roots. I heard about how a single atom of mercury can trigger a whole genetic program. I began studying how once it enters cells, mercury drives a transcriptional program. Mercury binds a transcription factor called the mercury resistance protein, or MerR. An atom of mercury reconfigures MerR, allowing it recruit RNA polymerase and turn on a new transcriptional cascade. 

What can transcriptional editing do that CRISPR cannot?

CRISPR editing and RNA interference can do wonderful things. When a researcher delivers genetic material, the cell responds in lots of interesting ways. In places that are accessible, where off-target effects are limited, CRISPR is the way to go. 

This is tunable gene expression, like turning the volume on the radio up and down instead of just smashing it with a hammer. 
- Aseem Ansari, St. Jude Children’s Research Hospital

It frequently gets more complicated than that, however. In cases where we try to toggle gene expression or modulate a genetic defect that’s difficult to target, we’re left with a challenge. Gene editing just isn’t there yet. Even next generation innovations like CRISPR interference aren’t mature technologies yet. We have to deliver a large protein complex; we have to do it repeatedly and consistently; and we have to do it in tissues that may not be easily accessible. In those cases, CRISPR is not the solution. There isn’t one hammer that serves as a tool for all jobs. 

The molecules we’re developing will ideally have the properties of drug-like small molecules. They will distribute better throughout the body and penetrate in a way that CRISPR doesn’t. Since they are more like drugs than CRISPR is, we can also develop them like drugs to make them orally bioavailable and easily administered. It can be just like a statin a patient takes to lower cholesterol.

Statins bind to 3-hydroxy-3-methylglutaryl coenzyme A reductase, which inhibits a key step in cholesterol biosynthesis without completely blocking it (1). Achieving a precise goal like this with gene editing would be difficult and potentially dangerous in many scenarios. A simpler approach is transcriptional editing, which like a statin, modulates genes and proteins instead of turning them on and off. It’s like turning the volume on the radio up and down instead of just smashing it with a hammer.

What types of disease might be treated with transcriptional editing?

Huntington’s disease, for example, is caused by increasing numbers of CAG repeats in the Huntingtin gene (HTT) (2). It sounds like a prime case for CRISPR: edit the HTT gene to give it fewer repeats, and the disease is cured. But CRISPR is difficult to deliver specifically to one location in the body. Instead, it spreads everywhere. It is extremely difficult to edit repetitive sequences like this with CRISPR, especially where some repeats are necessary but too many create a problem. 

On the other hand, there are genetic conditions that arise from multiple events. They are not caused by one or two genetic mutations. What if there were 150 different regulatory sites that all contributed to a disease? What can be done in a case like that? 

What is your research focus?

Friedrich’s ataxia is a disease my  team has done quite a lot of work on. It occurs because of a trinucleotide repeat expansion in an intron of the frataxin gene (FTX) between the first and second exons. In a healthy sequence, there are eight or nine repeats, usually with the sequence GAA, although they can go up to as high as 34 repeats and still be healthy (3). In patients with Friedrich’s ataxia, that intron balloons with 120 to as many as 1700 GAA repeats. The frataxin protein stops functioning, and there’s a huge loss of FTX mRNA in cells as well. Patients experience a progressive, terminal neurodegenerative condition.

As it progresses, the repeats continue to expand; this may be due to replication errors, subsequent genome repair errors, even some cell division errors. It’s a highly transcribed gene; it’s extremely active; and errors and repeats pile up in this intron.

How do you use transcriptional editing in a case like that?

The repeats block RNA polymerase transcription elongation. RNA polymerase gets through the repeated stops and terminates. It sits there for a while and eventually falls off. This blocks production of FTX mRNA or produces truncated transcripts that make nonfunctional, pathogenic proteins. This is a classic pause mechanism that’s seen in developmental genes across animals and multicellular organisms. 

These pauses are normal. Sometimes RNA polymerase encounters a DNA sequence that is physically difficult to traverse, like hiking across rocky terrain. Cells have machinery that can push RNA polymerase along and get it going again. In recent years, there has been a huge burst of papers that describe transcriptional pausing in other scenarios, particularly in cancer where people are trying to design small molecules that reinforce pausing in the hope of turning off critical genes and killing cancer cells (4).

That inspired us, but we took the opposite view. By hitting transcription machinery, we thought we could use those small molecules as hooks or mimics of epigenetic marks that recruit the RNA elongation machinery. In that way, we can increase expression of a gene or turn it down. Our challenge is how to deliver these hooks to a specific site in the genome as opposed to having them everywhere.

In the end, we used six different synthetic transcription factors. In the case of Friedrich’s ataxia, they recruit elongation machinery to a stuck RNA polymerase that would otherwise stall in the repeats. The exact mechanism differs from case to case just as with any other drug. In this disease, the medication needs to encourage the transcription machinery to transcribe because the disease occurs when repeat expansion blocks gene expression. In another case, a synthetic transcription factor could be designed to discourage, but not block altogether, gene expression if that is what’s needed.

Is there a guiding principle for picking future gene targets?

The guide is the underlying basic biology. These synthetic transcription factors work by induced proximity. That happens everywhere in the cell, like when a kinase comes in contact with another protein. That’s old hat for biologists, but it’s just coming to the chemistry community now.

In all of these cases, medicine needs to act at the DNA level. 
- Aseem Ansari, St. Jude Children’s Research Hospital

There are complex diseases like diabetes that do not have a single causative mutation, but instead arise from a problem with gene regulation. Unfortunately, gene regulation is a phenomenon with almost infinite complexity. In some cases, a treatment can target relevant transcription factors, but frequently these proteins do not have the deep binding pockets that drugs love to stick to. In other conditions, there are transcription factors that are missing altogether and can’t be targeted. Sometimes the genome itself is physically fragile and needs to be held together. A scientist could design a synthetic transcription factor that replaces a missing one or a synthetic transcription factor that stabilizes fragile DNA (5). In all of these cases, medicine needs to act at the DNA level. 

This interview has been condensed and edited for clarity.

References

1. Jiang, S.-Y. et al. Discovery of a potent HMG-CoA reductase degrader that eliminates statin-induced reductase accumulation and lowers cholesterol. Nat Commun  9, 5138 (2018).
2. Datson, N. A. et al. The expanded CAG repeat in the huntingtin gene as target for therapeutic RNA modulation throughout the HD mouse brain. PLoS ONE  12, e0171127 (2017).
3. Cossée, M. et al. Evolution of the Friedreich’s ataxia trinucleotide repeat expansion: Founder effect and premutations. Proc Natl Acad Sci US.A  94, 7452–7457 (1997).
4. Mitra, P. Transcription regulation of MYB: a potential and novel therapeutic target in cancer. Ann Transl Med  6, 443–443 (2018).
5. Finn, P. B. et al. Single position substitution of hairpin pyrrole-imidazole polyamides imparts distinct DNA-binding profiles across the human genome. PLoS ONE  15, e0243905 (2020).