Selection of Well-Tolerated GalNAc-siRNAs by Screening for ... · of a subset of GalNAc-siRNA conjugates that fail in rodent toxicity studies at >100x PD dose • Careful selection

Selection of Well-Tolerated GalNAc-siRNAs

by Screening for RNAi-Mediated Off-Target

Effects in Rodent Toxicity Studies

Maja Janas, PhD, DABT

Senior Scientist, Investigative Toxicology

TIDES, May 2nd 2017

2

Agenda

• Nonclinical safety profile of GalNAc-siRNAs

• Mechanisms of GalNAc-siRNA hepatotoxicity in rodents

3

GalNAc-siRNA Conjugates

SC-Administered Platform for Targeted Delivery to Hepatocytes

siRNA Metabolic stability

Duration of effect

Intrinsic potency

Safety

CMC

Ligand Receptor affinity & specificity

Metabolic stability

Safety

CMC

Asialoglycoprotein Receptor (ASGPR)

Highly expressed in hepatocytes

High turnover (recycling time ~15 min)

Conserved across species

4

GalNAc-siRNA Conjugates

Nonclinical Considerations

• Highly specific distribution to liver via

ASGPR

• Limited uptake/distribution in other organs

• Liver to kidney ratio >30 at PD doses

• Short plasma half-life; long tissue half-life;

long PD effects

• Metabolized largely by intracellular exo-

and endo-nucleases

• Maximum liver drug concentration limits

relevant toxicological doses

◦ Loss of dose-proportionality, approaching

saturation of liver uptake via ASGPR

◦ Cannot achieve MTDs

– Excess circulating siRNA levels not taken up

by liver; excreted through kidneys

5

GalNAc-siRNA Conjugates

Platform-Wide Responses at Toxicological Doses

Rat

◦ Liver

– Hepatocellular vacuolation: increased number and size of normal rat hepatocellular

vacuoles, most contain lipid

– Increased single cell necrosis

– Increased mitosis and regeneration

◦ Kidney

– Basophilic granules in proximal tubular epithelium; represents drug accumulation

NHP

◦ Liver

– Basophilic granules in Kupffer cells and hepatocytes; represents drug accumulation

◦ Lymph nodes

– Vacuolated macrophages (with basophilic stippling); represents phagocytosis of drug

Primarily dose-dependent and the result of drug accumulation in tissue

6

In silico prediction & In vitro efficacy

In vitro screen for predicted off-targets

Rodent Knockdown

Rat Tox @ >100x PD dose

NHP Knockdown DC

Bad Actors (40%):

Show hepatotoxicity

Good Actors (60%):

No hepatotoxicity

Path Towards a Development Candidate (DC)

Single cell necrosis and/or

hepatocellular degeneration with

↑LFT 2x upper limit of normal

7

Potential Causes of Hepatotoxicity in Rodent Toxicity Studies

1. Non-RNAi drug

effects e.g. protein binding

2. Competition for Ago

binding with miRNAs

3’-UTR

Off-target binding Partial sequence match

3. RNAi-mediated

off-target activity

RISC loading

mRNA

RISC

On-target binding Full sequence match

mRNA cleavage

Desired

on-target activity

Illustrative design

8

5’ Modifications on the Antisense Strand Can Block

RISC Loading With No Impact on Liver Exposure

RISC loading

mRNA

Sense strand removal

Target RNA cleavage

Dose: 30 mg/kg

Regimen: q2d x 6

Necropsy: Day 15

- + - +

1

1 0

1 0 0

R a t L iv e r E x p o s u re

ug

AS

/ g

liv

er

s iR N A -1 s iR N A -2

R IS C lo a d in g b lo ck- + - +

0 .0

0 .2

0 .4

0 .6

R a t R IS C L o a d in g

ng

AS

/ g

liv

er

s iR N A -1 s iR N A -2

R IS C lo a d in g b lo ck

9

Blocking Antisense RISC Loading Without Altering

the 2’F/2’OMe/PS Content Mitigates Hepatotoxicity

*

* ^

#

- + RISC loading block

Dose: 30 mg/kg

Regimen: q2d x 6

Necropsy: Day 15

S a lin e - + - +

0

1 0 0

2 0 0

3 0 0

R a t A L T

U/L

R e fe re n c e ra n g e

s iR N A -1 s iR N A -2

R IS C lo a d in g b lo ck

10

ESC

Advanced ESC

Changing siRNA Chemical Modification Pattern

Does Not Reduce Liver Exposure or RISC Loading Dose: 100 mg/kg

Regimen: qw x 9

Necropsy: Day 58

Altered modification pattern for

improved metabolic stability

E S C Ad v a n c e d E S C

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

R a t R IS C L o a d in g

ng

AS

/ g

liv

er

E S C Ad v a n c e d E S C

1

1 0

1 0 0

1 0 0 0

R a t L iv e r E x p o s u re

ug

AS

/ g

liv

er

Illustrative design

11

Changing siRNA Chemical Modification Pattern

Does Not Mitigate Hepatotoxicity Dose: 100 mg/kg

Regimen: qw x 9

Necropsy: Day 58

S a lin e E S C Ad v a n c e d E S C

0

1 0 0

2 0 0

3 0 0

R a t A L T

U/L

R e fe re n c e ra n g e

ESC Advanced ESC

#

* # *

12

Potential Causes of Hepatotoxicity in Rodent Toxicity Studies

1. Non-RNAi drug

effects e.g. protein binding

2. Competition for Ago

binding with miRNAs

3’-UTR

Off-target binding Partial sequence match

3. RNAi-mediated

off-target activity

RISC loading

mRNA

RISC

On-target binding Full sequence match

mRNA cleavage

Desired

on-target activity

Illustrative design

13

Reversir™ Platform Achieves Tailored Control of RNAi Pharmacology

0%

20%

40%

60%

80%

100%

120%

0 3 6 9 12 15 18 21 24

Rel

ativ

e Ta

rget

Pro

tein

Lev

el

No Reversir

Time (Days)

3 mg/kg

GalNAc-siRNA

0.1 mg/kg

GalNAc-Reversir

Reversir

mRNA cleavage

mRNA target

ASGPR

Inactive RISC

Functional

RISC

14

Prevention

Treatment

Reversir

siRNA

Reversir

siRNA

( )

Blocking RISC-Loaded Antisense Strand with Reversir

Does Not Reduce Liver Exposure or RISC Loading

3-10 mg/kg

30 mg/kg

3-10 mg/kg

30 mg/kg

1

1 0

1 0 0

1 0 0 0

R a t L iv e r E x p o s u re

ug

A

S /

g l

ive

r

- Targeting Ctr - Targeting Ctr Targeting Ctr RVR

siRNA1 2 3

1

1 0

1 0 0

1 0 0 0

R a t R IS C L o a d in g

ng

A

S /

g l

ive

r

- Targeting Ctr - Targeting Ctr Targeting Ctr RVR

siRNA1 2 3

15

Reversing Activity of the RISC-Loaded Antisense

Strand Mitigates Hepatotoxicity

Pre

ve

nti

on

+ Ctr RVR + Targeting RVR

# *

#

+

*

* #

* *

*

^

Tre

atm

en

t

0

5 0

1 0 0

1 5 0

R a t G L D H

U/L

R e fe re n c e ra n g e

- Targeting - Targeting Ctr RVR

siRNA- 1

0

1 0

2 0

3 0

4 0

R a t G L D H

U/L

R e fe re n c e ra n g e

- Targeting - Targeting Ctr RVR

siRNA- 2

0

1 0

2 0

3 0

4 0

R a t G L D H

U/L

R e fe re n c e ra n g e

- Targeting - Targeting Ctr RVR

siRNA- 3

16

Swapping Seed Regions Does Not Reduce Liver

Exposure or RISC Loading Dose: 30 mg/kg

Regimen: q2d x 6

Necropsy: Day 15

1

1 0

1 0 0

R a t L iv e r E x p o s u re

ug

AS

/ g

liv

er

Toxic Non-toxic Non-toxic Toxic Seed

Toxic Non-toxic Toxic Non-toxic Non-seed

Parent Seed swap

0

5 0

1 0 0

1 5 0

2 0 0

R a t R IS C L o a d in g

ug

AS

/ g

liv

er

Toxic Non-toxic Non-toxic Toxic Seed

Toxic Non-toxic Toxic Non-toxic Non-seed

Parent Seed swap

Illustrative design

17

Swapping Seed Regions Mitigates Hepatotoxicity

* #

*

#

* *

To

xic

Toxic Non-toxic

No

n-t

ox

ic

SEED

NO

N-S

EE

D

Dose: 30 mg/kg

Regimen: q2d x 6

Necropsy: Day 15

S a lin e b a d C 5 g o o d C 5g o o d s e e db a d s e e d

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

R a t A L T

U/L

R e fe re n c e ra n g e

Toxic Non-toxic Non-toxic Toxic Seed

Toxic Non-toxic Toxic Non-toxic Non-seed

Parent Seed swap

18

GalNAc-siRNA Conjugates Can Cause Seed-Mediated Off-Targets

at High Doses

RNAseq in Rat Hepatocytes (24 hrs, 10 nM)

siRNA-2 siRNA-1

Downregulated genes Upregulated genes

Antisense seed

enrichment

Sense seed

enrichment

Antisense seed

enrichment

Sense seed

enrichment

siRNA-1 p < 2.2e-16 p = 0.05286 p = 0.667 0.2174

siRNA-2 p < 2.2e-16 p = 0.9679 p = 0.6345 p = 0.3364

65

199

160

190

19

RNAi Activity is Driving Gene Dysregulation In Vivo at High Doses RNAseq in Rat Liver (24 hrs, 50 mg/kg)

siR

NA

-1

siR

NA

-2

- RISC loading block

16

44

8

1

-50%

-98%

196

311

7

18

-96%

-94%

+

RISC

Loading

Block

Downregulated genes Upregulated genes

Antisense seed

enrichment

Sense seed

enrichment

Antisense seed

enrichment

Sense seed

enrichment

siRNA-1 - p = 1.25e-03 p = 0.3521 p = 1 p = 0.7103

+ p = 1 (NA) p = 1 (NA) p = 1 p = 0.5548

siRNA-2 - p = 2.06e-10 p = 0.1843 p = 1 p = 0.9872

+ p = 1 p = 1 p = 1 p = 1

20

Conclusions

• Antisense strand-driven RNAi-mediated hybridization-based off-target

effects, not chemical modifications, are a major driver of hepatotoxicity

of a subset of GalNAc-siRNA conjugates that fail in rodent toxicity

studies at >100x PD dose

• Careful selection of "good actor" siRNAs is an important part of

identifying optimal siRNA molecules for clinical development

3’-UTR

Off-target binding Partial sequence match

RNAi-mediated off-

target activity

21

Nonclinical Safety and Bioanalysis

Carole Harbison

Yongfeng Jiang

Onur Yilmaz

Michael Placke

Natalie Keirstead

Kirk Brown

Brenda Carito

Samantha Chigas

Sean Dennin

Kristin Fong

Paul Gedman

Toni Hayes

Emma Henchy

Lauren Moran

Elena Ooms

Rachel Peters

Kristina Perry

Kellie Sawyer

Catrina Wong

Research

Adam Castoreno

Rubina Parmar

Mark Schlegel

Svetlana Shulga-Morskaya

Huilei Xu

Ivan Zlatev

Martin Maier

Vasant Jadhav

Terence Cawley

Klaus Charisse

Anna Bisbe

Rajeev Kallanthottathil

Ryan Malone

Mano Manoharan

Jonathan O'Shea

Nate Taneja

Christopher Theile

Jeff Rollins

Stacy Seide

Thank You!