Improved Anti-miRNA (AMOs) and Splice-Switching Oligonucleotides (SSOs), presented by Dr Mark Behlke, Chief Scientific Officer at Integrated DNA Technologies

1. Integrated DNA Technologies
Improved Anti-miRNA (AMOs) and Splice-
Switching Oligonucleotides (SSOs)
Mark Behlke MD, PhD Chief Scientific Officer
Biopolis, Singapore July 26, 2013

2. Inhibition of miRNAs by Antisense Oligonucleotides
RISC
RNA Induced
Silencing Complex
Target mRNA
Inhibit translation, mRNA cleavage …
AAAAA
Transfect AMO
miRNA
Steric blocking
Eventual Degradation?
2

3. Role of Chemical Modifications
1. Nuclease Stabilization
2. Increased binding affinity
a. Tighter binding  greater potency
b. Tighter binding  decreased specificity
3. Compatible with invasion of RISC?
4. Assist with delivery?
3

4. Chemical Modifications – Sugar alterations
4

5. Chemical Modifications – Phosphorothioate linkage
5

6. Newer AMO designs
Original “antagomir”
M*M*MMMMMMMMMMMMMMMMMM*M*M*M-Chol
M = 2’OMe
m = 2’MOE
F = 2’F
D = DNA
L = LNA
* = PS bond
Chemistries used in anti-miRNA Oligos (AMOs)
6
DLDDLDDLDDLDDLDDLDDLDD
D*D*L*D*D*L*D*D*L*D*D*L*D*D*L*D*D*L*D*D
L*D*L*D*D*L*L*D*D*L*D*L*D*L*L
m*m*F*F*F*F*F*F*F*F*F*F*F*F*F*F*F*F*F*m*m
M
MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM
M
MMMMMM MMMMMM
L*L*L*L*L*L*L*L*L

7. Luc/Luc Assay System
One perfect match miRNA binding site was cloned
Renilla
luciferase
Firefly
luciferase
miRNA
+ AMO
Translation
Light
miRNA
Renilla
Luciferase
miRNA
Binding
Site
Transfected into Cells
+
miRNA
Degradation
Renilla Luc
No Light
Cleavage
7

8. Rluc/Fluc assay in HeLa cells (express miR-21)
Direct Comparison of Different AMO Chemistries for miR-21 Knockdown
0
10
20
30
40
50
60
70
80
90
100
FoldincreaseinRL/FL
1nM
5nM
10nM
25nM
50nM
DNA 2'OMe 2'OMe
PS-ends
DNA/
LNA PS
2'OMe/
LNA
2'OMe/
LNA PS
2'F
LNAends
2'F
LNAends PS
HP+RC+HP
2'OMe
DNA/
LNA
2'OMe
PS
DNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
2’OMe ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe PS-ends ■•■•■•■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■•■•■•■
2’OMe PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
DNA/LNA ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
DNA/LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
2’OMe/LNA ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe/LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
2’F LNAends ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’F LNAends PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
HP+RC+HP 2’OMe ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
■ = DNA bases
■ = 2’OMe bases
■ = LNA bases
■ = 2’F bases
• = PS linkages
8

9. Unmodified DNA or 2’OMe oligos are rapidly degraded in
serum – need PS modification or hairpin
(Note: unmodified 2’OMe is stable in cell extracts, even though
it degrades in serum)
Stability in Serum
9

10. Interestingly, the DNA/LNA mixmers also require some PS
modification (at least on the ends)
(Note: unmodified are all stable in cell extracts)
Stability in Serum
10

11. Interestingly, the DNA/LNA mixmers also require some PS
modification (at least on the ends)
(Note: 2’F without PS rapidly degrades in cell extracts)
Stability in Serum
Degraded in
cell extracts
11

12. For More Information:
12

13. Insertion of “ZEN” between bases increases duplex stability
Temperature (oC)
20 30 40 50 60 70 80
%meltedduplex
0
20
40
60
80
100
Unmod DNA
Internal ZEN
5’-ATCGTTGCTA-3’
3’-TAGCAACGAT-5’
5’-ATCGTzTGCTA-3’
3’-TAGCA ACGAT-5’
vs.
+ ZEN
13

14. 2’OMe RNA is:
Natural
Safe / Nontoxic
Less expensive than LNAs or 2’F
Degraded by exonucleases in serum
Stable to endonucleases in cell extracts
The new napthyl-azo modification increases
Tm (PS decreases Tm), blocks exonuclease
action, and is compatible with RISC invasion
MzMMMMMMMMMMMMMMMMMMMMMzM
2’OMe with new napthyl-azo (“ZEN”) modifier between end bases
O P
O
O-
O
3'
N
O
N
N
5'
NO2
14

15. miR-21 2’OMe AMOs with internal ZEN insertion
2’OMe ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe 3PSends ■•■•■•■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■•■•■•■
2’OMe PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
HP+RC+HP 2’OMe ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe 3’-Z ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■z■
2’OMe 5’-Z ■z■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe 2-Z ■z■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■z■
2’OMe 2-Z-PS ■z■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■z■
2’OMe 3-Z ■z■ ■ ■ ■ ■ ■ ■ ■ ■ ■z■ ■ ■ ■ ■ ■ ■ ■ ■ ■z■
DNA LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
2’OMe LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
■ = DNA bases
■ = 2’OMe bases
■ = LNA bases
• = PS linkages
z = ZEN mod
Tm 72.7 72.2 68.1 76.3 71.4 74.5 87.3
DTm +0.5 -- -4.1 +4.1 -0.8 +2.3 +15.1
15

16. Small DTm can result in large DKa at 37oC
16
DNA t c a a c a t c a g t c t g a t a a g c t a 56.3 -16.4 -18.7 1.5 x1013
2′OMe U C A A C A U C A G U C U G A U A A G C U A 72.7 - -26.9 9.4 x1018
2′OMe 3PSends U*C*A*A C A U C A G U C U G A U A A G*C*U*A 72.2 -0.5 -26.4 4.3 x1018
2′OMe 5′inZEN,3′ZEN UzC A A C A U C A G U C U G A U A A G C U Az 76.3 3.6 -30.6 3.8 x1021
Ka(37°C)
(mol/L)-1Name miR-21 AMO Sequences (5′ to 3′) Tm (°C) ΔTm (°C)
ΔG o
37
(kcal/mol)
A 4.1oC increase in Tm between the 2’OMe-PSends AMO
and the ZEN-AMO results in an 880-fold increase in the
binding affinity (Ka) at 37oC

17. Importance of binding affinity for AMO potency
17
• It is generally accepted that high binding affinity improves potency for all steric
blocking antisense applications (AMO, SSO, mRNA …)
• miRNAs reside in RISC (complexed to protein) and can be stable for weeks. It is
critical to be able to invade RISC and inactivate these miRNAs
• miRNAs start as dsRNAs and get reduced to ssRNA form in RISC – thus RISC has
machinery that renders the miRNA duplex single-stranded : the AMO must
overcome these natural pathways so it does not get treated like a passenger strand
• Nuclease “slicer” function in Ago2
• Helicase “unwinding” of duplexes in Ago1, Ago2, Ago3, Ago4
• Thus the AMOs need to be nuclease resistant (cannot be cut by Ago2)
• Thus the AMOs need high enough binding affinity to overcome helicase activity
• After you reach the “threshold Tm” where helicase can no longer unwind the AMO from
the miRNA guide strand, then increases in binding affinity mostly serve to make cross-
reactivity for mismatches worse

18. miR-21 AMO length walk
18
For the miR-21 AMO with ZEN-2’OMe chemistry, the binding affinity threshold
to escape helicase unwinding must lie between the 14mer & 15mer

19. Specificity comparison of AMO chemistries
19
Mutant Type
Wildtype
MUT 1
MUT 2
MUT 3
a
Mutations are notated as blue nucleotides enclosed in red boxes
U C A A C A U C A G U C U G A U A A G C U A
U C A A C A U C A G U C A G A U A A G C U A
U C A A C C U C A G U C A G A U A A G C U A
U C A A C C U C A G U C A G A U A A C C U A
miR-21 AMO Sequences (5′ to 3′)
a
ZEN-2’OMe DNA/LNA-PS 2’OMe/LNA-PS2’OMe-PSends
“Antagomir”

20. ZEN is non-toxic, whereas PS mod and LNA mod show toxicity
2’OMe 2-Z ■Z■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■Z■
2’OMe 3PS-ends ■•■•■•■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■•■•■•■
DNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
DNA/LNA PO ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
DNA/LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
2’OMe/LNA PO ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■
2’OMe/LNA PS ■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■•■
■ = DNA bases
■ = 2’OMe bases
■ = LNA bases
Z = ZEN mod
• = PS linkages
20
0
20
40
60
80
100
120
%ViableCells
Cytotoxicity Data of "NC1" AMOs transfected
into HeLa Cells for 24hrs with INTERFERin™
50nM
100nM
2'OMe
2-Z
Reagent
Only
2'OMe
3PSends
DNA/
LNA PS
PS DNA/
LNA PO
2'OMe
LNA PS
2'OMe/
LNA PO
Stauro-
sporine
Pos Cont

21. 21
30 nM
100 nM
10 nM
DNA/LNA-PS 2’OMe-ZEN 2’OMe-3PSends 2’OMe2’OMe/LNA-PS

22. Potency differences do not relate to transfection efficiency
22
Different AMOs were
transfected at 30 nM.
AMO transfection efficiency
was assessed using ISH
(in situ hybridization).
Cells were fixed and
hybridized with a Dig-probe
and detected with an
A647-anti-Dig antibody

23. Final design rules
MzMMMMMMMMMMMMMMMMMMMMz (N-1)-length
MzMMMMMMMMMMMMMMMMMMMMzM Full-length
Either during transfection (in serum) or exposure to cytoplasmic
nucleases, the first AMO gets degraded to the second AMO. Positioning the
ZEN at the 3’-end removes most of the Nearest Neighbor effects and avoids
any degradation. Making the final AMO be 1 base shorter than the miRNA
target can also slightly increase potency.
23

24. For more information:
24
Molecular Therapy Nucleic Acids, 2013

25. Insulin Regulation in Pancreatic Islets
Eran Hornstein
Weizmann Institute
• Investigated a possible role for miRNAs in the
regulation of insulin secretion
• Dicer1 conditional KO using tamoxifen inducible Cre
recombinase with rat insulin promoter
• Examine changes in insulin levels and glucose
regulation with miRNA production disrupted
• Identified SOX6 and Bhlhe22 as negative regulators of
insulin secretion, with miRNAs 22, 24, 148, and 182
regulating expression of these genes, thereby indirectly
regulating insulin secretion
25

26. Knockout of Dicer leads to reduction in miRNA levels
Studied in isolated islets; b-cells comprise ~50-60% of cell mass
26

27. Dicer mutants have elevated glucose, decreased insulin
27

28. Up-regulation of transcription repressors in Dicer mutants
Sox6 and Bhlhe22 are repressors of insulin transcription.
Increased levels of repressors lower insulin and raise glucose levels.
28

29. miRNA regulation of Sox6 and Bhlhe22?
• miRNA expression in pancreatic islets
was examined using microarrays
• Potential binding sites for several
highly expressed miRNAs were found
in the 3’-UTRs of Sox6 and Bhlhe22
• AMOs were synthesized to specifically
suppress these miRNAs to investigate
the role of these species in normal
cells (not Dicer mutants)
29

30. AMOs increase SOX6 and Bhlhe22 levels, lowering glucose
30

31. miRNA regulatory pathway for Insulin secretion
31

32. For more information:
32

33. Regulation of Cystic Fibrosis (CFTR)
33
• Investigated role of miRNAs in regulation of CFTR
expression and found a major role for miR-138
• miR-138 does not directly regulate CFTR but instead
acts as a suppressor of SIN3A, which is a suppressor of
CFTR transcription
• miR-138 AMO increases SIN3A levels which lowers
CFTR levels
• miR-138 mimic lowers SIN3A levels which raises CFTR
levels (same effect is seen using anti-SIN3A siRNA)
• Not just transcription/translation effect – salvages D508
mutants!
Paul McCray
University of Iowa

34. Effect of miR-138 AMO vs. miR-mimic on WT CFTR
34

35. SIN3A knockdown rescues CFTR D508 expression on cell surface
35
The D508 mutant retains Cl- channel activity, but is degraded in the EPR
and never reaches the cell surface; SIN3A knockdown not only increases
CFTR expression, it alters processing and permits the semi-functional
mutant CFTR protein to reach the cell surface.

36. miR-138 mimic restores Cl- conductance in CF airway cells
36
A new target for CF therapy?

37. For more information, see:
37

38. Use of the ZEN modification: Splice Switching Oligos (SSOs)
Another use for this kind of antisense technology: SSOs
• Steric blocking mechanism of action
• 2’OMe RNA, LNA, PMO, PNA
• Bypass stop codon or other errors present by causing
splicing to shift and deleting the affected mutation. Many
diseases exist which could be treated by this
mechanism, including DMD, SMA, b-Thalassemia, and many
more
38

39. Duchenne Muscular Dystrophy (DMD)
Genetics – An X-linked recessive disorder
affecting approx 1 in 3500 boys
Cause - An absence of dystrophin, a protein
that helps keep muscle cells intact.
Onset - Early childhood - about 2 to 6
years.
Symptoms - Generalized weakness and
muscle wasting first affecting the muscles
of the hips, pelvic area, thighs and
shoulders. Loss of
ambulation, development of respiratory
problems (diaphragm), cardiomyopathy
and death in 20’s/30’s.
39

40. Collaboration with the Wood lab to study SSO in “mdx” mouse
C57BL/10ScSn-Dmdmdx/J mouse
• Stop codon in Exon 23
• Develops DMD phenotype
• Salvage with exon skipping SSOs = skip exon 23 and
you get an in-frame semi-functional dystrophin
protein
• PCR assay:
• Full length = 1kb product
• Skip exon 23 = 700 bp product
• Skip exons 22+23 = 550 bp product
Samir EL Andaloussi
Suzan Hammond
Graham McClorey
Matthew Wood
40

41. H2k (myoblast) cell culture
• H2k cells were grown at 33°C in 10% CO2 atmosphere using DMEM media
supplemented with 20% FBS and 0.5% chick embryo serum; grown on gelatinised
plates.
• After 24 h, cells were moved to 37°C in 5% CO2 and media is replaced with
differentiation media (DMEM with 5% horse serum). Myotubes form within 3-5
days. These are very difficult to transfect compared to the undifferentiated
myoblasts.
• Cells were transfected with LF2000 at the indicated concentrations or naked SSOs
are added at 2-4 µM concentration and cells were incubated for 48-96 h in
optiMEM.
41

42. PS linkage is important for function (not just nuclease stability)
80 40 20 10 80 40 20 10 U
PS20 EndPS
nM
Myoblasts, lipid transfection, studied at 48h
SSO-EndPS
mG*mG*mC*mC mA mA mA mC mC mU mC mG mG mC mU mU mA*mC*mC*mU
SSO-PS
mG*mG*mC*mC*mA*mA*mA*mC*mC*mU*mC*mG*mG*mC*mU*mU*mA*mC*mC*mU 42

43. Unlike AMOs, PS linkage is important for function in ZEN SSOs
Myoblasts, lipid transfection, studied at 48h
SSO-ZEN
mGzmG mC mC mA mA mA mC mC mU mC mG mG mC mU mU mA mC mCzmU
SSO-ZEN-PS
mGzmG mC*mC*mA*mA*mA*mC*mC*mU*mC*mG*mG*mC*mU*mU*mA*mC mCzmU
80 40 20 10 80 40 20 10 U
ZEN-PO ZEN-PS
nM
43

44. Pilot study: direct intramuscular injection in “mdx” mice
2’OMePS ZENPS
SSO-PS
mG*mG*mC*mC*mA*mA*mA*mC*mC*mU*mC*mG*mG*mC*mU*mU*mA*mC*mC*mU
SSO-ZEN-PS
mGzmG mC*mC*mA*mA*mA*mC*mC*mU*mC*mG*mG*mC*mU*mU*mA*mC mCzmU
• IM injection of 30 µg SSO
• Harvest muscle at day 14
• Immunohistochemical
visualization of dystrophin
protein
44

45. Functional testing in “mdx” mice
• Plan: 50 mg/kg IV weekly for 8 weeks
• Reality: 50 mg/kg IV 2x weekly for 4 weeks
1 week off
50 mg/kg IP 2x weekly for 4 weeks
Study phenotype for a week, then collect tissue
• 4 cohorts, WT untreated, “mdx” untreated, “mdx” 2’OMe-PS, “mdx” ZEN
• Study animals for functional activity
The monitoring system provides an assessment of the motor activity and behavior of
the mice, measuring both anxiety-related behavior and locomotor behavior associated
with muscle strength. IR beams of light pass through the cage. When the mouse
crosses a beam, the light is broken and this is recorded in the software, Digiscan. The
system measures 22 forms of activity including rearing, active time, static time as well
as distance travelled.
• Examine muscles for dystrophin protein and mRNA splice forms
45

46. Functional testing in “mdx” mice
• Expt was not optimal – think that
the IV phase worked but IP did not
• No splice-shifted mRNA was
detected (short half life)
• Dystrophin protein was present (long
half life)
• Repeating with a 4 week IV regimen
46

47.  New ZEN (napthyl-azo) modifier inserted between the terminal
bases of a steric-blocking antisense oligo improves nuclease stability
and increases binding affinity
 Particularly useful when used with 2’OMe RNA; anti-miRNA (AMO;
PO form) and splice switching (SSO; PS form) applications; more?
 The designs show excellent mismatch specificity, similar to low
potency unmodified 2’OMe AMOs, yet achieve the high level of
potency normally associated with use of more toxic DNA/LNA-PS
mixmers
 Both ZEN-AMOs and ZEN-SSOs are being tested in mice now
 Available now “off catalog” – just call IDT Tech Support. Full product
line in catalog will be available later this year.
Summary
47

48. Integrated DNA Technologies
Kim Lennox
Scott Rose
Richard Owczarzy
Yong You
Mike Marvin
Anton Holets
Jess Alexander
Joseph Walder
Mark Behlke
Weizmann Institute
Tal Melkman-Zehavi
Sharon Kredo-Russo
Amitai Mandelbaum
Eran Hornstein
Thanks to all the scientists whose work was discussed today!
University of Iowa
Shyam Ramachandram
Michael Welch
Paul McCray
Bev Davidson
48
University of Oxford
Samir EL Andaloussi
Suzan Hammond
Graham McClorey
Matthew Wood