Chemical inhibitors make their RNA epigenetic mark - Nature

Credit: STORM Therapeutics

N e w s & A na ly s i s

Chemical inhibitors make

their RNA epigenetic mark

RNA epigenetic drug development enters a new era, as companies start revealing

targets and evidence of preclinical efficacy.

Megan Cully

Ever since Chuan He coined the term

¡®RNA epigenetics¡¯ in Nature Chemical

Biology in 2010 to describe the post-?

synthesis modifications that adorn RNA,

the University of Chicago chemist has

faced naysayers. From a nomenclatural

perspective, some critics pushed back

because these marks are not heritable,

while others argued that messenger

RNA (mRNA) was not truly a component

of genetics. From a biological perspective,

others questioned the relevance of this

field because neither the dynamic nature

of these marks nor their functional impact

was clear early on. Now, as increasing

evidence suggests that these marks play

a key role in disease from cancer to

infection, He and other RNA epigenetics

pioneers are making progress with

candidates that inhibit the enzymes that

regulate these marks.

At an inaugural conference on

RNA epigenetics in Cambridge, UK,

in September, STORM Therapeutics,

Accent Therapeutics and Gotham

Therapeutics all disclosed that they

have developed small-?molecule inhibitors

of the METTL3¨CMETTL14 complex,

which regulates epigenetic marks on

RNA (Table 1). Two other companies,

Twentyeight-?Seven Therapeutics and

EPICS Therapeutics, are also eyeing

RNA epigenetic enzymes. And others are

watching closely.

892 | December 2019 | volume 18

But questions also loom over this emerging

field, which is still trying to figure out how

and why around 170 different chemical

modifications impact RNA biology. New

chemical marks are still being discovered, with

one recent report showing that in addition to

the methyl and acetyl modifications that have

already been identified, mRNAs can also be

glycosylated. And on the drug development

front, there is debate about whether these

nearly ubiquitous marks can be altered in

specific tissues or diseases without incurring

too much toxicity.

For Gerhard M¨¹ller, CSO at Gotham

Therapeutics, the field is nevertheless ripe

for drug discovery work. ¡°It¡¯s really one

hundred per cent compatible with what

biopharma companies should be doing these

days,¡± he says. ¡°These companies should

not work on an overly developed, validated

target for which there are already three, four

or five drugs in clinical development.¡± The

niche for biotech companies is in translating

academic findings into proof-?of-concept

compounds, he argues, after which it is up to

big pharma to carry out larger development

programmes.

Robert Copeland, Accent¡¯s president

and CSO, and former CSO at the chromatin

epigenetics firm Epizyme, sees things

similarly. ¡°We¡¯re all in very good places

right now for early drug discovery. Where

companies will differentiate themselves

is really in how well they execute on the

medicinal chemistry and the preclinical and

clinical development. That¡¯s the hard part.¡±

Meddling with m6A

Despite the nomenclatural pushback,

Stacy Horner, a virologist from Duke

University Medical Center who studies

mRNA modifications, sees value in the

terminology that the field has adopted.

¡°I like the term RNA epigenetics,¡± she

says. The parallels between the well-?

studied marks on DNA and histones and

those on RNA provide a framework for

understanding and explaining the biology

of RNA modifications, she adds.

For DNA, RNA and histones alike,

¡®writers¡¯ add post-?synthesis modifications

that alter the structure of the molecule

to either recruit or repel ¡®readers¡¯ (Fig. 1).

Readers interpret those marks to alter

transcription, in the case of DNA and

histones, or to modulate translation and

degradation, in the case of RNA. And

¡®erasers¡¯ remove the modifications,

restoring the unaltered function of the

source material.

For histone biology, drug developers

have already made inroads against all three

components of the writer-?reader-eraser

paradigm. Four inhibitors of histone

deacetylase (HDAC), an eraser, are approved

for a subset of haematological malignancies.

Epizyme has submitted its inhibitor of the

histone methyltransferase EZH2, a writer,

for FDA approval, with a PDUFA date in

January 2020. And amongst the histone

readers, more than 20 BET inhibitors are

in clinical trials.

The race is now on to do the same with

RNAs, with a focus on mRNAs because

they are substrates for protein translation.

Although researchers first identified

mRNA modifications in the 1970s, the

consequences of those modifications were

not clear early on. ¡°The real data that

got the field moving was the claim that

some of these modifications seem to be

dynamic and that they regulate the fate of

RNA,¡± says Oliver Rausch, CSO of STORM

Therapeutics. For example, in 2011 He

demonstrated that mRNA is the substrate

of FTO, a disease-?associated RNA ¡®eraser¡¯.

And a 2012 Nature paper showed that a

subset of sites are methylated in response to

growth stimuli. Coupled with the previous

realization that N6?methyladenosine (m6A)

likely accounts for the majority of internal

We¡¯re all in very good places

right now for early drug

discovery

nrd

N e w s & A na ly s i s

methylation marks on mRNA and is

altered in diverse processes including cell

differentiation and viral infection, the field

took off.

¡°It¡¯s not all about m6A, but the m6A story

is really what lifted the field above the parapet

and made everybody look up,¡± says Keith

Blundy, STORM¡¯s CEO.

It¡¯s not all about m6A, but the

m6A story is really what lifted

the field above the parapet

The research community has pinpointed

the METTL3¨CMETTL14 methyltransferase

complex as likely responsible for laying

down the m6A on most mRNAs. In 2016,

Richard Gregory, a biochemist from

Boston Children¡¯s Hospital and founder of

Twentyeight-?Seven Therapeutics, showed that

METTL3 is upregulated in certain cancers

and selectively promotes the translation of a

subset of mRNAs, many of which are known

oncogenes. And in a landmark Nature paper

in 2017, Tony Kouzarides¡¯ lab demonstrated

that by disrupting METTL3 or METTL16

using CRISPR they could prevent the growth

of acute myeloid leukaemia (AML) cells both

in vitro and in vivo.

With this evidence pointing to RNA

methylation as a potential driver of

AML oncogenesis, drug hunters saw an

opening. And METTL3 had a handful of

properties that make it an ideal target: it is

an enzyme, it is dynamically regulated and

it contains an S-?adenosyl-l-?methionine

(SAM)-binding pocket that has been

successfully targeted in histone epigenetic

enzymes.

All three of the lead companies in this

space ¡ª STORM Therapeutics, founded

by Kouzarides and Eric Miska, Accent

Therapeutics, founded by He, Copeland

and Howard Chang, and Gotham

Therapeutics, founded by Samie Jaffrey

from Weill Cornell and Gotham¡¯s

management team ¡ª are now working

on METTL3¨CMETTL14.

STORM found its METTL3 inhibitor

using high-?throughput screening

(HTS), biophysical methods and mass

spectrometry, it reported at the recent RNA

epigenetics conference in Cambridge. To

find molecules, it first screened a chemical

library using a biochemical assay of m6A

methylation, and then they culled the hits

using biophysical and structural approaches

Nature Reviews | DRug DIsCovERy

ahead of us, but that STORM has basically

validated our approach.¡±

The flip side of this coin is that class-?

specific failures will cast a shadow over the

whole field. When Incyte¡¯s IDO1 inhibitor

failed in the clinic, for example, New Link

Genetics ¡ª whose main asset was an IDO1

inhibitor ¡ª suffered a severe blow.

But STORM, Accent and Gotham also

have other targets on the go. STORM has

identified a group of 20 or so targets of

interest, mostly methyltransferases, via

CRISPR screens in multiple cancer cell

types. It originally feared that developing

inhibitors with specificity for a particular

methyltransferase would be difficult

because of the structural similarities in

their SAM-?binding pockets, where most

of the METTL3 inhibitors bind, as well

as in their substrate-?binding pockets.

However, ¡°one of the surprising findings

is that it is not as difficult as we thought to

make specific methyltransferase inhibitors,¡±

says Rausch, so it plans to pursue other

targets in this class.

Gotham is working on a reader and a

writer, both of undisclosed identity. Accent

has presented work on inhibitors of the RNA

¡®editor¡¯ ADAR1, an adenosine deaminase.

Korro Bio, who came out of stealth mode

in early October, is also developing

ADAR-?directed therapies. Although RNA

editors fall outside the reader-?writer-eraser

trio, this focus shows that enthusiasm for

other types of RNA-?modifying strategies

is growing.

EPICS Therapeutics is working hard

to catch up to STORM, Accent and

Gotham. ¡°We hope now we¡¯re in the game,¡±

says Francois Fuks, a biologist from the

University of Brussels who founded EPICS

together with Jean Combalbert. Although

EPICS is not disclosing its targets at the

moment, ¡°there are some fairly obvious ones,

that most of us will and should go after,¡±

says Fuks.

to assess binding with METTL3. At the

conference, STORM showed that in a

mouse model of AML, oral dosing of its

lead compound reduced both splenomegaly

and the number of circulating monocytes.

Similarly, patient-?derived xenografts grew

more slowly when the mice were treated

with its METTL3 inhibitor.

STORM is now investigating how its

candidate fares in other tumour types,

including solid tumours, and is aiming to

have a clinical lead in phase I trials in 2021.

Accent Therapeutics has also identified

METTL3 inhibitors using a structure-?guided

approach, and it hopes to have a compound

ready for phase I trials for 2021. It plans to

start in AML, but is also eyeing a defined

group of solid tumours.

Gotham Therapeutics, the third

company with a METTL3 inhibitor in

preclinical development, tried an HTS

campaign, a fragment-?based approach and

a DNA-?encoded library screen in parallel

to find hits. The most promising chemical

matter emerged from fragment screens

and subsequent fragment evolution. It is

aiming for a 2022 clinical trial with its

METTL3 inhibitor.

Frenemies, unite

Given how young the field is, there is

plenty of room for friendly competition

between these firms. ¡°We all come

from different directions,¡± says M¨¹ller.

¡°We all want to approach those problems

with slightly different reference angles,

and we will prosecute those projects in

different ways.¡± He sees room for multiple

METLL3¨CMETTL14 compounds from

different developers.

Any successes of these programmes will

likely boost the others. STORM, as the only

company with proof-?of-concept animal

studies in the bag, is already giving hope

to the others. M¨¹ller plans to ¡°bring home

the message not that STORM is a year or so

Table 1 | Companies targeting RNA epigenetics

Company

Named targets

Likely lead

indication

STORM Therapeutics

METTL3, other methyl transferases AMLa

Accent Therapeutics

METTL3, ADAR1

Estimated phase I

trial start date

2021a

AMLa, NSCLCa 2021a, 2022b

Gotham Therapeutics METTL3, undisclosed ¡®reader¡¯,

undisclosed ¡®eraser¡¯

AMLa

2021a

EPICS Therapeutics

Undisclosed RNA modifying

enzymes

Cancer

ND

Twentyeight-?Seven

Therapeutics

Undisclosed RNA modifying

enzymes

Cancer

ND

Korro Bio

ADAR1

ND

ND

AML , acute myeloid leukaemia. For METTL3 inhibitor. For ADAR1 inhibitor. ND, no data available.

a

b

volume 18 | December 2019 | 893

N e w s & A na ly s i s

Nucleus

FTO or

ALKBH5

m6A writer

complex

Ribosome

A

DC1

Export

AAA(n)

m7G

mA

6

Pol II

DNA

Other

m6A

readers?

AAA(n)

snRNA

m6Am

Am

Transcription compartment

Writing

DF1

DF2

DF3

mRN A

METTL3

METTL14

Cytoplasm

FTO

Splicing

Erasing

Degradation

Translation

Reading

Fig. 1 | The life cycle of an m6A mRNA. Methylation of mRNA occurs cotranscriptionally in the

nucleus; erasers and readers are present in the nucleus and cytoplasm. Some erasers, such as FTO,

may also act on non-?mRNA species. DC1, YTHDC1; DF1, YTHDF1; METTL3, methyltransferase-?like

protein 3. Modified from Nat. Rev. Mol. Cell Biol. 20, 608¨C624 (2019), Springer Nature Limited.

Other marks, other RNAs

Two other potential targets that researchers

brought up repeatedly at the conference were

the eraser FTO and the family of RNA readers

called YTHDF.

From a pathobiological perspective,

the RNA demethylase FTO offers a

particularly strong rationale, points out

Copeland. A variant of FTO was the first

obesity susceptibility locus identified by a

genome-?wide association study, as reported

in Science in 2007. Subsequent work showed

that inactivation of FTO protects mice from

obesity. Obesity is an increasingly prevalent

disease with few existing therapies, so FTO

ticks the ¡®unmet medical need¡¯ box there too,

even if the clinical trial pathway for such

candidates remains fraught.

FTO also has a potential role in cancer, but

this seems to be dependent on the tissue type.

FTO is a proposed oncogene in AML, but a

potential tumour suppressor in breast cancer,

as Fuks presented at the Cambridge meeting.

Importantly, the FTO work is also pushing

the RNA epigenetics community to think

broadly beyond the m6A modification, and

beyond mRNA. In a 2019 Nature Chemical

Biology paper, Jaffrey¡¯s lab showed that FTO

works on a dimethylated form of adenosine

(N6,2¡ä?O?dimethyladenosine, or m6Am),

and that it may be removing these marks

from small nuclear RNAs (snRNAs) as well

as mRNAs. snRNAs modulate the splicing

of mRNAs, and Jaffrey¡¯s group showed that

inhibition of FTO increases the inclusion of

certain exons in mRNAs.

The YTHDF family, for its part, has

been associated with learning and memory,

cancer and antiviral immunity. Because

these RNA readers look structurally quite

similar to each other, some researchers have

voiced concern that it may be particularly

894 | December 2019 | volume 18

challenging to develop subtype-?selective

inhibitors for these targets. But that might

not be an issue, says Jaffrey, on the basis

of some of his unpublished work in AML

cells lacking all three YTHDF family

members. Pan-?YTHDF inhibitors might be

effective and not toxic, he speculates, and

tissue-?specific expression patterns for the

YTHDFs could provide a means of achieving

specificity if needed.

Twentyeight-?Seven Therapeutics is

meanwhile going after epigenetic modulation

of a different RNA: microRNA (miRNA),

another small regulatory RNA. Let7 is a

tumour-?suppressive miRNA, levels of which

are controlled by post-?synthesis uridylation.

The uridylation of this RNA results in its

degradation. Uridylation itself is catalysed

by terminal uridylyltransferases (TUTases).

So TUTases control overall levels of the

tumour-?suppressive miRNA and could

therefore be targets. ¡°LIN28 itself might

be a challenging target given that it¡¯s an

RNA-binding protein, so we¡¯re quite excited

that there¡¯s an enzymatic activity involved

in that pathway,¡± says Gregory.

Gregory sees RNA regulation as ¡°the next

frontier¡±, and the identification of regulatory

enzymes here could provide a new way to

think about challenging targets.

But there are questions and controversies

that could yet hold up all of this science.

Measuring global changes in RNA

methylation, especially for marks other than

m6A, is challenging. And even with m6A, the

best studied of the methylation marks, there

is still debate about how it regulates mRNAs

¡ª whether by altering their translation or by

driving their degradation.

Liquid-?liquid phase separation, which

produces transient membraneless organelles

comprised of proteins and RNA, further

adds to this complexity. RNAs with more

than one m6A mark bind to YTHDF family

members and phase separate in vitro and

in cells, thereby removing the mRNA from

the translational system. The therapeutic

implications of this finding, published in

Nature in July 2019, are not yet clear. But

they could range from the risk of off-?target

effects on membraneless organelle formation

to the possibility that phase-?separated

proteins will have reduced exposure to drugs

in the cytosol.

Whether adenosine methylation is

increased or decreased in cancer is similarly

complicated. The regulatory dynamics

of RNA methylation is likely to be context

dependent and could differ between

cancers.

¡°The understanding of the biology is still

in its infancy,¡± concedes Gregory.

RNA methylation controls numerous

different processes in a context-?specific

manner, and messing with this process could

also have unintended and toxic consequences.

The results from preclinical and clinical

studies of first-?generation compounds could

clarify the potential usefulness of targeting

various pathways, but until recently nobody

could properly explore the consequences of

small molecule intervention. Thankfully, the

compounds needed to ask those questions are

now in hand.

The understanding of the

biology is still in its infancy

Things are moving on the technical side,

too. Quantitative methods that can be used to

assess RNA methylation at the transcriptome

level have come online this year, for example.

Previous technologies tended to use either

antibodies or site-?specific cleavage followed

by radiolabelling to measure methylation at

a given site, but not on a transcriptome-?wide

scale. Two newly disclosed methods ¡ª

MAZTER-?seq and m6A-?REF-seq ¡ª rely on

an m6A-?sensitive endonuclease to drive RNA

digestion, and then digested and undigested

fragments that cover the same sequence

can be compared to quantify site-?specific

m6A methylation.

And this kind of progress is buoying

RNA epigenetic drug developers. ¡°It¡¯s still

early days, but it¡¯s not my first rodeo,¡± says

Copeland, who has brought 19 drugs into

the clinic during his career, including

four epigenetic drugs during his time

at Epizyme.

nrd

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download