Wave Life Sciences : Potent, Durable mRNA Knockdown in Extrahepatic Tissues Using siRNAs With Novel Phosphoryl Guanidine Backbone Variants

Chemical Design of Oligonucleotides That Support Targeted

RNA Editing in the CNS of Non-Human Primates

Chikdu Shivalila, Genliang Lu, Christopher Acker, Ian Harding, Jigar Desai, Jack Godfrey, Naoki Iwamoto, Pachamuthu Kandasamy, Hui Yu, Tomomi Kawamoto, Jake Metterville, Megan Cannon, Erin Purcell-

Estabrook, Alyse Faraone, Prashant Monian, Stephany Standley, Jesse Turner, Ryan Yordanoff, Fangjun Liu, Hailin Yang, Michael Byrne, and Chandra Vargeese

Wave Life Sciences, Cambridge, MA, USA

SUMMARY

RESULTS

Figure 3. Impact of AIMer optimization on editing

Figure 4. AIMers facilitate RNA editing in the CNS

Figure 5. ADAR editing for MECP2R168X protein restoration

• Leveraging PRISM™, our discovery and drug development platform,

we developed relatively short oligonucleotides that elicit A-to-I RNA

editing with high efficiency using endogenous ADAR enzymes, which

we call AIMers (Figure 1A).1-3

• Using N-Acetylgalactosamine(GalNAc)-modified AIMers with a

stereopure chimeric phosphodiester (PO)/ phosphorothioate (PS)/

phosphoryl guanidine (PN) backbone pattern, we previously

demonstrated up to 50% editing in nonhuman primate (NHP) liver.2

• Here, we have further optimized AIMer design by identifying base,

sugar, and backbone modification patterns that improve editing across

target and nearest neighbor sequences.

Figure 2. AIMer base, sugar and backbone modifications enhance editing efficiency across nearest neighbor combinations in cells

A	Approach to Improving Editing Efficiency
		Orphan site		Edit region
AIMer	3′		5′	N-3-uridine
NCN
HN
Target RNA
5′	XAX	3′	O N O
		Nearest Neighbors
			Sugar

1

Optimize orphan site base

5′

3′

Cytosine

N3U

efficiency in vitro

A		Cell-free System					Ugp2 Editing
	100				Primary Mouse Hepatocytes
										****
			50				****			****
							****
			40
	80				****
					****
		Editing	30		****
		20		****
Editing	60						*
%	10		ns
	0	6	24 48 72 96	6	24 48 72 96	6	24 48 72 96	6	24 48 72 96
%			Hours
40		Stereorandom PS	Stereorandom PN		Stereopure PS	Stereopure PN
						AIMer Abundance

A				In vitro Editing in CNS Cell Lines
	100		ACTB				100		UGP2
Editing	80					Editing	80
60					60
% ACTB	40					% UGP2	40
	20						20
	0						0
	0.01	0.1	1	10	100		0.01	0.1	1	10	100
		Concentration (μM)				Concentration (μM)
					iNeurons	iAstrocytes

• The Problem

Top 8 Hotspot Rett Syndrome (RTT) Mutations

10					T158M R168X				Nonsense
(%)Frequency		R106W	R133C	R255X R270X	R294X R306C
	Missense
8
6
4
2
	N-term		MBD	ID		TRD			C-term
1	78		162	207		310	487

Methyl-CpG Binding Protein 2 (MECP2)

The Approach

	Amino Acid		Arg			Glu			Glu
Normal	mRNA	C	G	G	C	G	A	G	A	G
DNA	C	G	G	C	G	A	G	A	G
	G	C	C	G	C	T C	T	C
										MECP2
RTT	Amino Acid		Arg			STOP
mRNA	C	G	G	U	G	A	G	A	G
(nonsense)
	DNA	C	G	G	T	G	A	G	A	G
	G	C	C	A	C	T C	T	C
	AIMer									MECP2R168X
RNA-	Amino Acid		Arg			Trp			Glu
mRNA	C	G	G	U	G	I	G	A	G
edited
	C	G	G	T	G	A	G	A	G
	DNA
	G	C	C	A	C	T C	T	C
										MECP2R168W

MECP2 Coregulatory

Protein Binding

293T Cells, Transfected

	MECP2
Mock	WT R168W
90	FLAG
70
	MECP2

NCoR1

260

90

70TBLR1

IgG Heavy chain

50HDAC3

• Here, we show that AIMer RNA base editing technology is applicable in

the central nervous system (CNS). AIMers support editing of

housekeeping RNA in neurons and astrocytes in vitro, and

unconjugated AIMers broadly direct durable RNA editing across the

CNS of mice and NHPs.

• AIMers with optimized chemistry support editing of a disease-relevant

transcript in neuronal cells. MECP2 AIMers direct RNA editing to

2	Optimize sugar and backbone			AIMer-S	AIMer-D
modification pattern outside 5′		3′
	Pattern	Pattern
	the edit region
B	Orphan Site Base	C	Sugar and Backbone

5′

AIMer

3′

5′

3′

20				Primary Mouse Hepatocytes
				6 hr			96 hr
0		AIMer Concentration (ng/ml)	1000	****	%AIMer Remaining	100	****
10-710-610-510-410-3	10-210-1100 101	800		80	**
AIMer Concentration (µM)	600	***	60
400	40
Stereorandom PS	Stereopure PS	200		20
Stereorandom PN	Stereopure PN
	0			0
		Stereorandom PS Stereorandom PN	Stereopure PS Stereopure PN

B

CNS Editing in vivo 1-weekPost-dosing

Human ADAR1-p110Mice

Nonhuman Primates (NHPs)

80

PBS

80

aCSF

AIMer

AIMer

Editing

60

Editing

60

40

40

Ugp2

Ugp2

B			iPSC-derived Cortical Neurons			Primary Cortical Neurons
			Human Patient-derived			Mecp2R168X KI mouse
		30	PBS or NTC AIMers	AIMer 1		100
				NTC AIMer
			MECP2 AIMers			Mecp2 AIMer
	Editing				Editing	80
	20			60
	% MECP2	10			% Mecp2	40
				20

convert the Rett Syndrome mutation MECP2R168X into the missense

R168W

Editing	75

Editing	75

B

Cell-free System

Stereopure PS

Ugp2 Editing

%	20			%	20

0						0
0	20	40	60	80	100	0.00001	0.0001	0.001	0.01	0.1	1	10

codon MECP2	in human and mouse neuronal cell models.
• A-to-I(G) editing of RNA-encoding Mecp2R168X restores expression of
full-length Mecp2R168W protein in neuronal cells, which correctly
colocalizes with heterochromatin. MECP2R168W protein also associates
with wild-type MECP2 binding partners, suggesting functionality.

mRNA	50
%Ugp2	25
	0

D

		mRNA	50
		%Ugp2	25
			0
C	N3U		AIMer-S	AIMer-D
		Nearest Neighbors

	100		Stereopure PN	Primary Mouse Hepatocytes
							****		****		****
								****
			60				****			****
	80					****			****
		Editing	40			*				**
	60
Editing	%	20
	0
%
40		Hours	6	24 48 72 96	6	24 48 72 96	6	24 48 72 96	6	24 48 72 96

0

0

tex

us

Striatum

m

m

rd

tex

s

Callosum

m

m

Cord

u

u

r

p

e

llu

o

r

p

e

t

lC

t

ll

o

m

S

e

o

m

S

e

C

oca

ain

b

a

C

oca

in

b

Spinal

e

e

r

n

a

r

pp

r

e

p

pp

Br

e

B

C

S

Corpus

C

Hi

Hi

C

Durable CNS Editing in vivo

Human ADAR1-p110 Mice

PBS

AIMer

Peak

30%

>40%

25%

>40%

50%

>65%

editing

Cortex

Hippocampus

Striatum

Brain Stem

Cerebellum

Spinal Cord

Ranked by Potency from NTC AIMer

C		Mecp2R168X	WT
	Ladder	AIMer	1	PBS	PBS
Mouse	90					Mecp2
70
Cortical					HDAC2
Neurons	50					PBS
	Primary Cortical Neurons

Dose (µM)

Primary Cortical Neurons

Mecp2R168X KI mouse

DAPI

TUJ1

Mecp2

INTRODUCTION

• PRISM™ generates stereopure oligonucleotides with controlled sequence,

5′ A

Orphan 3′ A

Base:

C

5′ A 5′ A 5′ A	5′ C	5′ C	5′ G	5′ G	5′ G	5′ U	5′ U	5′ U
3′ C 3′ G 3′ U	3′ C	3′ G	3′ A	3′ G	3′ U	3′ A	3′ C	3′ U

AIMer-S

AIMer-S

AIMer-D

	AIMer Abundance
20	Stereopure PS Primary Mouse Hepatocytes
	Stereopure PN

6 hr

96 hr

80

60

80

60

80

60

80

60

*

*

80

60

*

*

* **

80

60

*

*

*

*

*	*
*
	*

*

	Mecp2R168X KI mouse
DAPI	Mecp2	Merge

chemistry, and stereochemistry (Figure 1A).1

• PRISM™ can be applied to optimize AIMer design for editing efficiency,

target sequence, and target tissue.

Figure 1. Introduction to PRISM™, PN chemistry, and AIMers

N3U

C

N3U

AIMer-D

0	10-310-210-1100	101	1000		****	%AIMer Remaining	80	***	ns
10-710-610-510-4	800
AIMer Concentration (µM)	AIMer Concentration (ng/ml)			60
600	****		40
		400
Stereopure PS	Stereopure PN
200			20
AIMer-S	AIMer-S
	0				0
AIMer-D	AIMer-D
		AIMer-S	AIMer-D			AIMer-S	AIMer-D

% Ugp2 Editing

40

20

*

40

*

*

*

*

20

*	*
*
*

*

*

* *

*

*

40

20

*	*	*
	*
*
*		*
	*

40

20

*	*
*
	*
	* *
	*	40
	*
		20

* *

*

* *

*

*

*

40

20

**

PBS

NTC

AIMer 1

AIMer 1

25

75

25

75

25

75

25

75

25

75

25

75

25

75

25

75

25

75

25

75

25

75

25

75

%Ugp2 mRNA Editing

(A, B) Left: Cell-free editing assays: Lysates from 293T cells transfected with human ADAR-p110 (48h) were incubated with either UGP2-targeting AIMer (A) or ACTB-targeting

0

0

0

0

0

0

A

O

O B

O

O B

O

O

B

5′

B

(Rp)

O

O

R

O

O

R

N O

O

R

3′

2′

O

P O

O

B

S P

O

O

B

N P

O

O

B

X

R

N

5′

B

B Base

O

R

O

R

O

R

R

2'-Ribose

PO

PS

PN

3′

R

2′

X

Stereochemistry and

backbone modification

Phosphodiester

Phosphorothioate

Phosphoryl Guanidine

B

ADAR Editing of RNA

A	AIMer		I(G)
		ADAR	Edited
RNA		RNA
Oligonucleotides can direct A-to-I(G)	ADAR1 is ubiquitously	Cellular reservoir of ADAR
RNA editing by recruiting	expressed across tissues,	capacity supports directed editing
endogenous ADAR enzymes3	including liver and CNS	in addition to homeostatic function

(A) Schematic of approach to improving editing efficiency through AIMer backbone, sugar, and base chemistry. (B, C, D) Primary mouse hepatocytes from human ADAR1-p110 hemizygous mice were treated with 3 μM AIMers (unconjugated), directed toward the Ugp2 mRNA, with variable edit region sequence, chemistry pattern (AIMer-S or AIMer-D), and orphan base (C or N3U) for 72 hours. Ugp2 RNA editing was quantified by Sanger sequencing. (B) Lines connect complexes (represented by circles) with identical 5′- and 3′- nearest neighbors and chemistry format. (C) Lines connect complexes (represented by circles) with identical 5′- and 3′-nearest neighbors and orphan base. Stats: mean of n=3; error bars represent SEM.

• AIMers with orphan site N3U supported higher mean percent RNA editing than AIMers with orphan site C for all nearest neighbor combinations tested, although the magnitude of increase varies (Figure 2B).
• The AIMer-D pattern conferred a higher mean percent RNA editing compared to the AIMer-S patternfor most sequences tested (Figure 2C).
• The impacts of orphan site N3U base modification and the AIMer-D pattern appear largely additive (Figure 2D).
• AIMers with orphan site N3U and the AIMer-D pattern support highly efficient editing for many nearest neighbor combinations in primary mouse hepatocytes.

AIMer (B) at the concentration indicated for 1h, then RNA was extracted from lysates and RNA editing was quantified by Sanger sequencing. Stats: n=3 per dose, per condition; mean ± SEM shown. Right: Primary murine hepatocytes were treated gymnotically with 3 μM Ugp2-targeting AIMers for 6 hours. Cells were refreshed with maintenance media and collected at the indicated time point. RNA editing was quantified by Sanger sequencing. AIMer concentration was quantified by hybridization ELISA 6 hr or 96 hr after the start of the pulse. Stats: A two-way ANOVA was used to calculate statistical significance; * p<0.05,

• p<0.01, *** p<0.001, **** p<0.0001, ns significant.

• Incorporating stereopure PN linkages in AIMers enhances maximum editing compared to either stereopure PS or stereorandom PN in both cell-free and hepatocyte RNA editing assays (Figure 3A).
• The AIMer-D pattern further enhances the editing efficiency benefits of incorporating stereopure PN linkages in AIMers (Figure 3B).
• The AIMer-D pattern does not appear to enhance editing in cell-free systems but does lead to an increased cellular concentration of AIMers immediately after treatment (Figure 3B).
• Collectively, incorporation of stereopure PN linkages and the AIMer-D pattern improve AIMer-mediated RNA editing efficiency. This impact may occur through multiple mechanisms, including enhancing enzyme activity and AIMer uptake.

PBS	1	4	8 1216	PBS	1	4	8 1216	PBS	1	4	8 1216	PBS	1	4	8 1216	PBS	1	4	8 1216	PBS	1	4	8 1216
								Weeks

ACTB and UGP2 percent editing measured by Sanger sequencing. (A) iNeurons and iAstrocytes were treated gymnotically with ACTB or UGP2 AIMers for 5 days. (B) Left: human ADAR1-p110 mice were administered phosphate buffered saline (PBS) or 100 μg AIMer by intracerebroventricular (ICV) injection (n=5) on day 0 and necropsied on day 7. Right: Cynomolgus monkeys (NHPs) were administered 10 mg ACTB AIMer or artificial CSF (aCSF) by intrathecal administration (n=2) on day 0 and necropsied on day 7. (C) human ADAR1-p110 mice were administered 100 μg AIMer or PBS by ICV injection on day 0 and evaluated for Ugp2 editing across CNS tissues at 1, 4, 8, 12 and 16- weeks post dose. Stats: 2-way ANOVA with post-hoc comparison to PBS (n=5 per time point, per treatment) *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

• AIMers with the AIMer-S pattern, optimized for the CNS, support dose- dependent editing of ubiquitous housekeeping transcripts in multiple CNS cell lines in vitro (Figure 4A).
• In a human ADAR1-p110 transgenic mouse model, and in NHPs, AIMer- directed editing of housekeeping transcripts were observed across the CNS at one week post single dose (Figure 4B).
• In human ADAR1-p110 mice, AIMer-directed editing peaked at 4 weeks

and persisted 4 months post-single ICV injection (Figure 4C).

(A) Left: Graph adapted from (4) with updated data from RettBASE (data pulled 7 Aug 23). Middle: Proposed MECP2 restoration strategy via

AIMer-basedA-to-I RNA editing. Right: 293T cells transfected (72 h) with FLAG-MECP2WT or FLAG-MECP2R168W, immunoprecipitated with anti-FLAG magnetic beads. Western blot of immunoprecipitated eluates probed for FLAG and NCoR1/SMRT complex members. (B) MECP2 AIMers direct editing in neuronal cells. Percentage editing determined by next-generation sequencing (NGS). Left: Patient iPSC-derived cortical neurons (MECP2R168X) treated gymnotically with 10 μM AIMer (mean ± SEM; n=2 per AIMer). Right: Primary cortical neurons from Mecp2R168X knock-in (KI) mice (E18), treated gymnotically with AIMer for 5 days (mean ± SEM; n=3 per dose/condition). (C) AIMer-based editing of Mecp2R168X restores protein expression. Western blot of nuclear extracts from mouse primary cortical neurons (E18) treated gymnotically with 10 μM AIMer. Primary cortical neurons from Mecp2R168X KI mice (E18) treated with PBS or gymnotic AIMer (30 μM, left; or 1 μM, right) for 5 days. Immunofluorescence staining for nuclei (DAPI, blue) and Mecp2 (magenta) or neuronal marker Tuj1 (Green). Magnification 40X (left) or 10X (right). NTC, nontargeting control. HDAC2, Histone deacetylase 2. WT, wild type.

• We hypothesized that AIMers could be used to correct MECP2R168X, the most common nonsense mutation found in Rett Syndrome (RTT), by converting the premature stop codon to a Tryptophan (W) codon in MECP2 mRNA (Figure 5A).
• We show that exogenous, plasmid-expressed MECP2R168W protein associates with endogenous co-regulatory proteins NCoR1, TBLR1, and HDAC3, suggesting edited MECP2 may retain wild type MECP2 functionality (Figure 5A).
• MECP2 AIMer incorporating the AIMer-D format directs editing of MECP2R168X in human patient-derived cortical neurons and primary cortical neurons isolated from the Mecp2R168X KI mouse (Figure 5B).
• MECP2 AIMer incorporating the AIMer-D pattern restores Mecp2 protein expression and localization in primary cortical neurons isolated from the Mecp2R168X KI mouse (Figure 5C).

References: 1. Kandasamy, et al., 2022. Nuc Acids Res 50(10):5443-5466; 2. Monian et al., 2022 Nature Biotech 40(7):1093-1102. doi: 10.1038/s41587-022-01225-1; 3. Woolf, et al., 1995 Proc Natl Assoc Sci 92:8298-8302; 4. Krishnaraj, et al., Hum Mutat 2017;00:1-10.Acknowledgments: The authors are grateful to Nicole Neuman (Wave Life Sciences) and Eric Smith for editorial and graphical support, respectively. This work was funded by Wave Life Sciences. Patient fibroblast cells were kindly provided by Rett Syndrome Research Trust and Harvard Stem Cell Institute iPS Core Facility.

Presented at the 27th Annual Meeting of the American Society of Gene & Cell Therapy, May 7-11, 2024 - Baltimore, MD

Supported by Wave Life Sciences, Cambridge, MA, USA