Biological Efficacy and Toxicity of Diamidines in Myotonic Dystrophy Type 1 Models

Article abstract of DOI:10.1021/acs.jmedchem.5b00356

Myotonic dystrophy type 1 (DM1) is a disease characterized by errors in alternative splicing, or "mis-splicing". The causative agent of mis-splicing in DM1 is an inherited CTG repeat expansion located in the 3′ untranslated region of the DM protein kinase gene. When transcribed, CUG repeat expansion RNA sequesters muscleblind-like (MBNL) proteins, which constitute an important family of alternative splicing regulators. Sequestration of MBNL proteins results in the mis-splicing of its regulated transcripts. Previous work has demonstrated that pentamidine, a diamidine which is currently FDA-approved as an antiparasitic agent, was able to partially reverse mis-splicing in multiple DM1 models, albeit at toxic concentrations. In this study, we characterized a series of pentamidine analogues to determine their ability to reverse mis-splicing and their toxicity in vivo. Experiments in cell and mouse models demonstrated that compound 13, also known as furamidine, effectively reversed mis-splicing with equal efficacy and reduced toxicity compared to pentamidine.

Full text of DOI:10.1021/acs.jmedchem.5b00356

Subscriber access provided by UNIV OF MISSISSIPPI
Article
Biological Efficacy and Toxicity of Diamidines
in Myotonic Dystrophy Type 1 Models
Ruth B Siboni, Micah J Bodner, Muhammad Miza Khalifa, Aaron G Docter,
Jessica Y Choi, Masayuki Nakamori, Michael M. Haley, and J. Andrew Berglund
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00356 • Publication Date (Web): 23 Jun 2015
Downloaded from http://pubs.acs.org on July 9, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biological Efficacy and Toxicity of Diamidines in
Myotonic Dystrophy Type 1 Models
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ruth B. Sibonia, Micah J. Bodnerb, Muhammad M. Khalifab, Aaron G. Docterb, Jessica Y. Choib,
Masayuki Nakamoric, Michael M. Haleyb, J. Andrew Berglunda,b,*
a Institute of Molecular Biology, University of Oregon, Eugene, OR 97403
b Department of Chemistry and Biochemistry, University of Oregon, Eugene, OR 97403
c Department of Neurology, University of Osaka Graduate School of Medicine, Osaka, Japan
ABSTRACT. Myotonic Dystrophy type 1 (DM1) is a disease characterized by errors in
alternative splicing, or “misꢀsplicing”. The causative agent of misꢀsplicing in DM1 is an
inherited CTG repeat expansion located in the 3’ untranslated region of the DM protein kinase
gene. When transcribed, CUG repeat expansion RNA sequester MBNL proteins, which
constitute an important family of alternative splicing regulators. Sequestration of MBNL proteins
results in the misꢀsplicing of its regulated transcripts. Previous work has demonstrated that
pentamidine, a diamidine which is currently FDAꢀapproved as an antiꢀparasitic agent, was able
to partially reverse misꢀsplicing in multiple DM1 models, albeit at toxic concentrations. In this
study, we characterized a series of pentamidine analogues in order to determine their ability to
reverse misꢀsplicing and their toxicity in vivo. Experiments in cell and mouse models
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 35
1
2
3
4
demonstrated that compound 13, also known as furamidine, effectively reversed misꢀsplicing
5
6
with equal efficacy and reduced toxicity compared to pentamidine.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION.
Myotonic dystrophy type 1 (DM1) is an inherited disorder characterized by myotonia,
muscle wasting, insulin resistance, cardiomyopathy, and cognitive dysfunctions1,2,3. As a
trinucleotide repeat disorder, individuals with DM1 possess aberrant expansions of DNA repeats.
In particular, DM1 is caused by an expansion of CTG repeats in the 3’ untranslated region of the
DM protein kinase (DMPK) gene. These expansions can contain between 50 and thousands of
repeats 4,5,6. When transcribed into RNA, CUG expansions serve as binding sites for RNAꢀ
binding proteins, thereby aggregating and sequestering these proteins from their canonical
functions 7,8. One such class of proteins is the muscleblindꢀlike (MBNL) family of splicing
factors. Consistent with this model, in vivo fluorescent hybridization experiments of expanded
CUG repeats demonstrated that they form nuclear aggregates, or foci, with MBNL proteins9.
Members of the MBNL family, including MBNL1, MBNL2, and MBNL3, regulate the
alternative splicing of over 90 tissueꢀspecific transcripts, and are also involved in RNA
localization and processing events10,11. DM1 phenotypes are the result of misꢀspliced, MBNLꢀ
regulated events. In support of this, two of the transcripts misꢀspliced in DM1, pertaining to the
insulin receptor (INSR) and chloride channel (CLCN1) genes, have been directly correlated with
two key symptoms experienced by DM1 patients: insulin insensitivity and myotonia,
respectively12,13,14. While there are no current treatments for this disease, approaches under
development aim to reduce MBNL sequestration by CUG repeats and thereby reverse misꢀ
splicing and associated symptoms 15,16,17,18,19. One strategy to achieve this is through the
ACS Paragon Plus Environment

Page 3 of 35
Journal of Medicinal Chemistry
1
2
3
4
administration of small molecules that can bind nucleic acids and ultimately prevent formation of
5
6
CUG: MBNL aggregates 20,21
.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: Chemical structure of pentamidine.
Recent work from our lab has found that the small molecule pentamidine and the related
molecule heptamidine were able to rescue misꢀspliced events and reduce CUG RNA levels in
both cell and mouse models of DM1 15,22. Pentamidine, a diamidine composed of two phenyl
amidine groups joined by a fiveꢀcarbon methylene linker (Figure 1), is currently FDAꢀapproved
for the treatment of various parasitic infections, including pneumocystic pneumonia,
leishmaniasis, and trypanosomiasis23. While pentamidine’s biomolecular target within parasitic
organisms remains unclear, previous work has demonstrated that it is capable of binding the
minor groove of ATꢀrich DNA with high affinity 24. This is supported by the fact that
pentamidine has a number of important structural features which promote DNA interaction: the
amidines provide minor groove recognition by forming hydrogen bonds with base pair acceptors,
the overall positive charge of the molecule helps to offset the negatively charged DNA
backbone, and the flat arenes connected by a conformationally flexible linker promote
incorporation into, and stacking within, the walls of the narrow minor groove of ATꢀrich DNA
25,26. However, given our previous observation that pentamidine also modestly interacts with
CTG DNA in vitro and reduces CUG RNA levels in vivo 22, it is possible that ATꢀrich DNA is
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 35
1
2
3
4
5
6
7
8
9
not the only (or even the primary) target of diamidines within the DM1 framework. Our
observations are consistent with a model in which pentamidine inhibits transcription of CTG
DNA although other possible mechanisms (including the promoting of CUG RNA degradation)
remain to be explored.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While extensive SAR and analogue studies have been performed in order to optimize the
structural features of diamidines against eukaryotic parasites 25, 27ꢀ29, there is sparse information
concerning which physiochemical properties of pentamidine are necessary to rescue misꢀsplicing
in a DM1 model. An exploration of pentamidine derivatives varying in methylene linker length
(3ꢀ9 carbons) demonstrated that increasing linker length correlated with increasing efficacy in a
cell model, but also resulted in decreased solubility and increased toxicity; ultimately, seven–
carbon derivative, heptamidine, demonstrated the highest efficacy while remaining water soluble
22. However, heptamidine also demonstrated high toxicity in a mouse model, as some mice were
unable to tolerate a standard, oneꢀweek daily regimen. Modifications to other regions of
pentamidine or heptamidine (i.e., the phenyl amidine groups) have not been previously explored
in DM1 models. In this study, we present the results of an SAR study based on pentamidine and
heptamidine. Analogues were synthesized containing various substitutions that alter the size,
degrees of freedom, hydrophobicity and number of hydrogen bond donors of the parent
compounds. The analogues’ ability to ameliorate splicing defects in cellular and transgenic
mouse models of DM1, as well as their toxicity relative to the parent compound, were assessed.
RESULTS AND DISCUSSION.
SAR study identifies key structural elements of pentamidine required for splicing rescue
ACS Paragon Plus Environment

Page 5 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Diamidines such as pentamidine share three key structural features that are critical for
their antiꢀparasitic activity: a semiꢀcurved shape (or the ability to adopt this shape), two
positively charged terminal amidine groups with hydrogen bonding ability, and a relatively flat
conformation with some flexibility in the methylene linker 26. We hypothesized that the
structural requirements for therapeutic activity as an antiꢀparasitic may be similarly important for
misꢀsplicing rescue in a DM1 cell model, given our model that pentamidine may function
through transcription inhibition of CTG repeat DNA and thus may largely be dependent on its
DNAꢀbinding properties in order to reverse misꢀsplicing. To this end, we created a series of
structures (2a-c, Figure 2a) that were derived from pentamidine but lacked the second amidine
group (2a), both the second amidine group and the second ether linkage (2b), or an entire
phenoxyꢀamidine group (2c). The compounds were also distinct from pentamidine in that they
had shorter methylene linkers: compound 2a possessed a threeꢀcarbon methylene linker, 2b a
fourꢀcarbon methylene linker, and 2c an unꢀsubstituted butyl chain.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: SAR study of pentamidine derivatives.
(a) General synthetic schematic for pentamidine derivatives 2a-2c.
(b) Schematic depicting the effect of CUG repeat RNA on the splicing of INSR. Percentages
listed indicate percentage inclusion/exclusion of alternatively spliced exon 11, as
quantified in a minigene transfection system.
ACS Paragon Plus Environment

Page 7 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
(c) RTꢀPCR data of INSR splicing in cells incubated with increasing concentrations of
compounds 2a-2c. Results are depicted as average percentage of exon 11 inclusion +
standard deviation of three independent trials.
7
8
9
(d) Jitter plot representation of INSR splicing in cells treated with compounds 2a-2c. Each
point is one experiment and the line represents the average of all experiments for that
condition (at least three for each concentration). Gray area denotes range between typical
splicing and DM1ꢀinduced misꢀsplicing. Concentrations (in µM) are listed below.
Asterisks indicate a significant difference between the treatment concentration and misꢀ
splicing without compound.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The HeLa cell DM1 splicing assay is a commonly employed system used to assess the
efficacy of potential treatments for this disease in vivo. While HeLa cells do not endogenously
possess CTG expansions in their genome, overexpression of 960 interrupted CUG repeats
(CUG960) in vivo recapitulates DM1 molecular phenotypes 7,22,30. When these cells are coꢀ
transfected with MBNLꢀregulated minigenes, the resulting splicing variations can be analyzed in
the presence or absence of potential compounds. One wellꢀcharacterized minigene, INSR,
corresponds to exons 10ꢀ12 (and the intervening introns) of the human insulin receptor gene
(Figure 2b). The INSR minigene demonstrates splicing defects upon coꢀtransfection with
CUG96015. Within the INSR minigene, MBNL promotes the inclusion of alternatively spliced
exon 11. Exon 11 is included in 76 ± 6% of INSR minigene transcripts when CUG repeats were
absent. Upon coꢀtransfection of CUG960, this value decreased to 49 ± 3%. All compounds were
initially screened with this assay to determine if they could induce reversal of exon 11 inclusion
levels back to those exhibited in cells lacking the transfected CUG repeats.
Interestingly, our splicing results demonstrated removal of one amidine group (2a) did
not ablate the ability to rescue INSR splicing (Figures 2c and 2d), as compound 2a demonstrated
doseꢀdependent rescue of misꢀsplicing through 60 µM in the HeLa cell system. The calculated
concentration necessary to observe 50% splicing rescue (EC50) for this compound was 47 ± 5
µM, which is greater than the value calculated for pentamidine (31 ± 2 µM). However, when
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 35
1
2
compared to the EC50 value of propamidine22 (42 ± 19 µM), which similarly possesses a fourꢀ
carbon linker, there is no significant difference. We were unable to observe full splicing rescue
with 2a before extensive cell death occurred. In contrast, removal of the second phenolic linkage
in conjunction with the amidine group (2b) resulted in strongly reduced ability to rescue misꢀ
splicing, as little rescue was observed by 20 µM (Figures 2c and 2d). Due to the toxicity
observed with compound 2b before any substantial rescue was observed, we were unable to
calculate a proper EC50 value, and therefore no definitive conclusions can be made about the
importance of the ether linkage for splicing rescue. Finally, compound 2c, which consisted only
of a single phenyl amidine group with a butoxy chain, demonstrated significantly reduced
potential for misꢀsplicing rescue with an EC50 value of 237 ± 33 µM (Figures 2c and 2d). Taken
together, these results suggest that the presence of only one phenoxyamidine is sufficient in
pentamidine derivatives to induce a partial rescue of the INSR splicing event, as long as the other
phenoxy group is maintained. This is surprising considering amidines provide Hꢀbond donating –
NH groups that interact with Hꢀbond acceptors (such as those present in the minor groove of ATꢀ
rich sequences), and therefore both groups are presumed to play a crucial role in the hydrogen
bonds that stabilize pentamidine’s interaction with ATꢀrich DNA 25.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Heptamidine derivatives vary in splicing rescue efficacy
Heptamidine, a derivative of pentamidine with a seven carbon methylene linker, rescued
INSR misꢀsplicing with an EC50 value of 9 ± 1 µM in the HeLa cell system while still retaining
its water solubility 22. It was similarly able to rescue or partially rescue two endogenous splicing
events in a transgenic DM1 mouse model. Nonetheless, heptamidine was observed to be fairly
ACS Paragon Plus Environment

Page 9 of 35
Journal of Medicinal Chemistry
1
2
3
4
toxic in both the cell and mouse models. Because of these previously reported results, we opted
5
6
to synthesize two heptamidine derivatives (4 and 6, Figure 3a) to determine if we could increase
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3: Heptamidine analogues partially rescue mis-splicing of minigene reporters.
(a) General synthetic schematic for heptamidine derivatives 4 and 6.
(b) RTꢀPCR data of INSR splicing in cells incubated with increasing concentrations of
compounds 4 and 6. Results are depicted as average percentage of exon 11 inclusion +
standard deviation of three independent trials.
(c) Jitter plot representation of INSR splicing in cells treated with compounds 4 and 6.
Asterisks indicate a significant difference between the treatment concentration and misꢀ
splicing without compound.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 35
1
2
3
4
5
6
7
8
9
or maintain efficacy and decrease toxicity. In compound 4, we placed the amidine groups in the
metaꢀ position relative to the phenoxy linkage, rather than in the paraꢀ position. In compound 6, a
methoxy group was added in the orthoꢀ position relative to the phenoxy linkage. The five carbon
version of both of these compounds were previously synthesized and tested in a rat model for
P.carinii pneumonia, where they varied in their toxicity in comparison to pentamidine 31.
In the DM1 HeLa cell model, compound 4 was able to produce almost complete doseꢀ
dependent misꢀsplicing rescue at 25 µM, but was also very toxic over the range tested (10 – 25
µM, Figures 3b and 3c). The EC50 for this compound was 18 ± 7 µM, suggesting that compounds
with amidine groups in a metaꢀposition have half the efficacy of heptamidine. Compound 6
demonstrated significantly reduced potential for misꢀsplicing rescue with an EC50 value of 97 ±
17 µM, more than tenfold the original EC50 value calculated for heptamidine (Figures 3b and 3c).
Furthermore, compound 6 demonstrated significant toxicity by 100 µM, and therefore full rescue
was not achieved with this compound. As we were unable to significantly reduce toxicity within
the range of rescue for either heptamethyleneꢀlinked compound, we chose to focus on creating
more effective five carbon linker compounds that were less toxic.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N-substituted amidines do not reduce toxicity/increase efficacy in vivo
We next examined the previously reported strategy of substituting the amidine groups of
the diamidine compounds, thereby producing Nꢀsubstituted amidines 32, 33, 34. The aim of creating
Nꢀsubstituted compounds was to increase the size of the molecule in order to reduce the
concentration required to achieve full splicing rescue. Furthermore, we chose substitutions that
increased hydrogen bonding ability or enhanced hydrophobic interactions, some of which have
been previously examined in the context of antiꢀparasitic activity 35, 36. We first synthesized a
ACS Paragon Plus Environment

Page 11 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
precursor molecule, which contained an unꢀsubstituted propyl chain (8, Figure 4a), to determine
if the addition of a carbon chain somewhere other than within the central methylene domain
would increase the efficacy of splicing rescue. Compound 8 was able to reach full splicing rescue
by 150 µM, and was not toxic over the range of rescue (Figure 4b). While the EC50 value of the
compound was 70 ± 12 µM, the lack of toxicity exhibited with this compound was a promising
result that prompted us to examine whether substituting the propyl chain would increase efficacy
and maintain low toxicity. We created two series of compounds with either a phenylpropyl or
N,Nꢀdimethylaminopropyl substituent, with each series consisting of 3ꢀcarbon, 5ꢀcarbon, and 7ꢀ
carbon variants for both substitution types (9aꢀ9c and 10aꢀ10c, Figure 4c).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The phenylꢀsubstituted series (9aꢀ9c) demonstrated increased efficacy compared to
pentamidine and similar efficacy compared to heptamidine, with compounds 9b and 9c
producing EC50 values of 10.5 ± 5.4 and 10.4 ± 2.9 µM, respectively. However, all compounds
in this series were too toxic within our cell assay to observe full rescue (Figure 4d). Conversely,
the N,Nꢀdimethylamino series (10aꢀ10c) demonstrated decreased efficacy, with compounds 10a
through 10c demonstrating EC50 values in the 200ꢀ300 µM range; all compounds were also
observed to be too toxic to reach full rescue, although they were considerably less toxic then
their phenylꢀsubstituted counterparts (Figure 4e). As previously observed, both series
demonstrated a trend towards increased efficacy with increasing length of the central linker
region.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: N-substituted pentamidine derivatives demonstrate increased toxicity in cellulo.
(a) General synthetic schematic for precursor to aminopropylꢀsubstituted compounds.
(b) Jitter plot representation of INSR splicing in cells treated with compound 8. Below,
representative RTꢀPCR gel is demonstrated, with average percentage values (+ standard
deviation) listed per each condition. Asterisks indicate a significant difference between
the treatment concentration and misꢀsplicing without compound.
(c) General synthetic schematic for phenylꢀsubstituted (9a-9c) and N’Nꢀ
dimethylaminopropylꢀsubstituted (10a-10c) compounds.
ACS Paragon Plus Environment

Page 13 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(d & e) Jitter plot representation of INSR splicing in cells treated with compounds 9a-9c (d)
and 10a-10c (e). Below, representative RTꢀPCR gels are demonstrated, with average
percentage values (+ standard deviation) listed per each condition. Asterisks indicate a
significant difference between the treatment concentration and misꢀsplicing without
compound.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While the observed toxicity of these compounds was qualitatively informative, we wanted to
quantitate toxicities and compare compounds based on these values alone. To quantify relative
toxicities, we devised a flow cytometryꢀbased cell viability assay using propidium iodide, a
membraneꢀimpermeant dye which binds exposed DNA in dead cells. Both phenylpropyl and
N,Nꢀdimethylaminopropyl series, as well as pentamidine and heptamidine, were tested in the
DM1 cell model using the same cell culture conditions as for the splicing assays. The
cytotoxicity of each compound was assessed by calculating the concentration necessary to
observe 50% cell death, or IC50 value (Table 1). Higher IC50 values indicate lower toxicity.
These results were then used to calculate EC50:IC50 ratios, which provided us with the criteria for
identifying compounds that are both effective and less toxic than pentamidine. For example, the
EC50:IC50 of pentamidine was calculated as 0.48 (value attained by dividing 31 µM /64 µM),
which demonstrates that twice as much pentamidine is necessary to see 50% cell death as is
needed to see 50% INSR splicing rescue. A value above 0.48 indicates that the compound had a
smaller window to rescue splicing before causing cell death. Therefore, desirable compounds
would have an EC50:IC50 value lower than 0.48. However, smaller EC50:IC50 ratios could also
result from structures that were less effective in splicing rescue compared to pentamidine but
were less toxic (for example, an EC50 of 200 µM and IC50 of 500 µM would yield a ratio of 0.4),
and so both EC50 and IC50 values were considered individually as well.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 35
1
2
3
4
EC50
IC50
EC50:IC50
5
6
7
8
9
31 + 2
9 + 1
64 + 12
15 + 0.46
30 + 4.9
0.48
0.60
0.47
0.82
0.20
3.08
0.95
2.08
0.66
0.95
1.18
0.93
0.15
N/A
Pentamidine
Heptamidine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14 + 5.2
75 + 11
4
6
91 + 8.4
70 + 12
355 + 45.1
13 + 2.9
8
40* + 9.3
10.5* + 5.4
10.4* + 2.9
366* + 14.9
393* + 94.2
256* 68.2
125* 54.4
31 + 4
9a
9b
9c
11 + 1.6
5 + 0.26
551 + 35.4
414 + 31.2
217 + 16.2
134 + 55.8
210 + 19.3
N/A
10a
10b
10c
11
12
13
30 +4
Table 1: Efficacy (EC50) and cytotoxicity (IC50) values for diamidine derivatives of
pentamidine and heptamidine. Asterisk indicated that EC50 values are mathematically
predicted, but where never reached experimentally due to toxicity. All values reported as +/ꢀ
standard deviation.
Since we were only interested in derivatives with higher or equal efficacy to pentamidine, a more
efficient alternative to pentamidine has to meet two criteria: a lower EC50 value than
pentamidine’s calculated value of 31 µM as well as an EC50:IC50 value lower than 0.48. All
ACS Paragon Plus Environment

Page 15 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds except 8 had EC50:IC50 values above 0.48 due to their low IC50 values. While
structure 8, demonstrated a promising EC50:IC50 values of 0.20, it was half as effective as
pentamidine in the splicing assay (EC50 value of 70 ± 12 µM). Since none of the phenyl or
substituted compounds (9a-9c, 10a-10c) met these criteria, we concluded these substitutions are
not more effective than pentamidine at improving splicing defects in vivo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds 12 and 13 effectively rescue mis-splicing of two minigenes
The final set of compounds analyzed consisted of three conformationally restricted
molecules, planar compounds 11ꢀ13 (Figure 5a). Both linearity and planarity have been
previously implicated in promoting binding of these molecules to the minor groove of DNA 37, 38
Minor groove binding is best mediated when the curvature of the inner atoms of a molecule are
around 20 Å, but recent studies demonstrated that linear compounds can also bind the minor
groove of DNA through waterꢀmediated interactions 25,39. Compounds 11 and 12 are essentially
linear, while 13 has a concave shape, considered best suited for binding in the minor groove of
DNA 37.Planarity, meanwhile, is commonly associated with intercalation; however, CD spectra
of 12 and 13 demonstrate that neither compound intercalates with DNA39. Furthermore,
compound 11, a 2,7ꢀbisamidine carbazole, has been previously demonstrated to bind both GC
and AT DNA 40,41. Despite this fact, 11 was the least effective in the DM1 splicing assay (EC50
value of 125 µM) and was also the most toxic with an IC50 value of 134 ± 56 µM (Figure 5b). In
contrast, both 12 and 13 demonstrated similar EC50 values to pentamidine (31 ± 4 and 30 ± 4
µM, respectively). Compound 12 was also significantly less toxic than 11 with an IC50 value of
210 ± 19.3 µM, making it the molecule with the lowest EC50:IC50 ratio. Interestingly, at the
highest concentration tested (150 µM, not shown), 12 increased exon 11 inclusion levels above
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 35
1
2
3
4
5
6
7
8
9
that of wild type. While a reproducible IC50 value could not be attained for compound 13
because the fluorescence of the furan group produced high background in our flow cytometry
assay, this molecule was also observed to be nonꢀtoxic over the range of rescue. High levels of
cell death were not observed with this compound in the 20ꢀ80 µM range, unlike pentamidine.
Based on these results, it was clear that both 12 and 13 were the most effective compounds and
have a larger window of activity compared to pentamidine and heptamidine.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DM1 pathophysiology is the result of numerous misꢀsplicing events across skeletal
muscle, striated muscle, and neuronal tissues 2, 42. Therefore, small molecule approaches which
aim to reduce DM1 pathology must target multiple MBNL1ꢀdependent misꢀsplicing events to be
successful. In light of this fact, we tested compounds 12 and 13 for their ability to rescue splicing
of another transiently transfected minigene, cardiac troponin T (TNNT2). MBNL proteins are
negative regulators of exon 5 in TNNT2, thereby facilitating the exclusion of this exon in the
final mRNA product (Figure 5c). Typical inclusion levels of TNNT2 exon 5 within the HeLa cell
system were observed at 69 ± 4%, and this value increased to 87 ± 3% in the presence of CUG960
(Figure 5d). Compound 12 was able to achieve full misꢀsplicing reversal by 120 µM, while 13
achieved full misꢀsplicing reversal by 80 µM. The calculated EC50 values for TNNT2 rescue by
12 and 13 were 77 ± 6 and 42 ± 7 µM, respectively. The difference observed in EC50 values
between INSR and TNNT2 systems has been previously observed with other diamidine
derivatives 22 and may be due to the fact that MBNL targets differ in the amount of MBNL they
require in order to be properly regulated 43. These results suggest that planarity, but not linearity,
of diamidine compounds may be an important component for splicing rescue.
ACS Paragon Plus Environment

Page 17 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 35
1
2
3
4
5
Figure 5: Two conformationally-constrained diamidines (12 and 13) demonstrate decreased
toxicity in cellulo.
(a) General synthetic schematic for conformationallyꢀconstrained compounds 11-13.
(b) Jitter plot representation of INSR splicing in cells treated with compounds 11-13. Below,
representative RTꢀPCR gels are demonstrated, with average percentage values (+
standard deviation) listed per each condition. Asterisks indicate a significant difference
between the treatment concentration and misꢀsplicing without compound.
(c) Schematic depicting the effect of CUG repeat RNA on the splicing of TNNT2.
Percentages shown indicate percentage inclusion/exclusion of alternatively spliced exon
11, as quantified in a minigene transfection system.
(d) Jitter plot representation of TNNT2 splicing in cells treated with compounds 12 and 13.
Below, representative RTꢀPCR gels are demonstrated, with average percentage values (+
standard deviation) listed per each condition. Asterisks indicate a significant difference
between the treatment concentration and misꢀsplicing without compound.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While reversal of misꢀsplicing is an important measure of the efficacy of small molecules
within the DM1 framework, it is also important to determine if these molecules alter splicing in
the absence of CUG repeats. We therefore performed control splicing experiments with both
compounds 12 and 13 in the absence of transfected CUG960 (Figure S1). With the INSR
minigene, inclusion levels of exon 11 vary between ~70% and ~80 %. This range of exon 11
inclusion is likely due to a combination of experimental nuances, including variability in cell
passage number, cell plating, and transfection efficiency. Compound 12 increased inclusion of
exon 11 to 90 + 2 % with a 60 µM dosage (Figure S1a). In contrast, compound 13 treatment
didn’t modulate exon 11 inclusion splicing above or below the expected range, even at an 80 µM
dosage (81+ 4% exon inclusion observed, Figure S1b). Similar trends were observed for the
control experiment with TNNT2 minigene splicing. Compound 12 again demonstrated a doseꢀ
dependent reduction of exon 5 inclusion in the absence of CUG repeats, while compound 13 did
not alter TNNT2 splicing over the tested range (Figure S1cꢀd). These results indicate compound
12 modulates splicing through mechanisms beyond targeting the CUG repeats. Due to the nonꢀ
specific results with compound 12, we focused the remainder of our studies on compound 13.
ACS Paragon Plus Environment

Page 19 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
Compound 13 reduces nuclear foci in vivo
7
8
9
A distinguishing characteristic of DM1 cells is that CUG RNA aggregate with MBNL
proteins to form nuclear foci 9. To directly visualize the effect of compound 13 on MBNL
sequestration, CUG foci formation and nuclear MBNL1 localization were examined in the HeLa
cell system in the presence or absence of these compounds. AntiꢀMBNL1 antibody MB1a (4A8,
MDA Monoclonal Antibody Resource) was used to probe for MBNL1 localization, and a CAG
probe was employed to bind CUG RNA 44. Untreated HeLa cells not transfected with CUG960
demonstrated diffuse localization of MBNL1 throughout the nucleus (Figure 6, row 1). Upon
transfection of the CUG960 plasmid, characteristic RNA foci were observed (Figure 6, row 2).
Complete coꢀlocalization of MBNL1 was not observed, most likely a result of varying CAGꢀ
probe accessibility to CUG RNA once in aggregate form.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Upon addition of 80 µM of compound 13, we observed a decrease in the number of foci
per cell when compared to the transfected controls (Figure 6, rows 4 & 5). This result is
comparable to that observed for 50 µM pentamidine (Figure 6, row 3). While most cells treated
with 13 still possessed modest numbers of nuclear foci, there was an 89% reduction with
compound treatment (n = 121, p < 0.05), which was statistically significant. Nuclear MBNL
levels in treated cells were observed to be more diffuse and evenly spread compared to untreated
cells, despite the persistence of some foci. These results indicate that compound 13 is able to
disrupt formation of the CUG: MBNL complex.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6: Compound 13 reduces nuclear aggregation of MBNL1.
Compound 13 reduces ribonuclear foci and relieve MBNL1 sequestration. Top row:
untreated HeLa cells, with MBNL1 localizing diffusely throughout the nucleus. Second
row: Transfection of CUG960 plasmid results in the formation of ribonuclear foci.
Addition of 50 µM pentamidine (third row) or 80 µM compound 13 (fourth row) results
in significant reduction of ribonuclear foci.
Compound 13 rescues mis-splicing in a DM1 mouse model
The HSALR mouse model expresses 220 CUG repeats under the skeletal promoter . In
47
this model, two wellꢀcharacterized endogenous preꢀmRNAs were analyzed: Clcn1, which has
been linked to myotonia when misꢀsplicing occurs, and Atp2a1 (also known as Serca), which
encodes a sarcoplasmic calcium ATPase and exhibits a broad splicing differential in DM114, 48
.
ACS Paragon Plus Environment

Page 21 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In Clcn1, MBNL is responsible for promoting the exclusion of exon 7a. Wild type adult mice
demonstrated 7a included at 5 + 0.9 %, this value increased 8ꢀfold to 39 + 2 % in HSALR mice
(Figure 7a). Treatment with a 10 and 20 mg/kg regimen of compound 13 resulted in a doseꢀ
dependent reversal of exon 7a levels to 11 + 3 %, which was close to a full rescue.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7: Compound 13 rescues misꢀsplicing in a DM1 mouse model. a) Compound 13 was
able to partially rescue the chloride channel endogenous event by 20 mg/kg for 7 d. b)
Compound 13 was also able to partially rescue the sarcoplasmic calcium ATPase 1
endogenous event by 20 mg/kg for 7 d. No toxicity was observed over the concentration
range tested. Each symbol represents the splicing outcome for quadriceps muscle of a
single mouse. Gray area denotes range of splicing between wild type and DM misꢀ
splicing patterns. Asterisks denote events that were significantly different from HSALR
mice.
Similar rescue of splicing was observed for the Atp2a1 event, for which MBNL is a positive
regulator of exon 22 (Figure 7b). Wild type adult mice included exon 22 at 100 + 0.2 %, while
HSALR mice only included this exon at 28 + 2 %. Treatment with 10 and 20 mg/kg of compound
13 over 7 days caused exon 22 inclusion levels to shift to 64 + 4 % and 83 + 4 %, respectively.
While a full rescue was not observed for either splicing event, the rescue of misꢀsplicing was
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 35
1
2
3
4
5
6
7
8
9
significant at both dosages and for both events investigated. These data demonstrate that
compound 13 was able to reverse misꢀsplicing in DM1 mice with similar activity compared to
heptamidine and did so without the observed toxicity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS.
An appropriate assessment of small molecules as potential lead compounds for
development of therapeutics for DM1 requires two key considerations: the effectiveness of the
compound in biological systems as well as the toxicity of the compound. Previous work
demonstrated the efficacy of pentamidine and various pentamidine derivatives as potential lead
compounds for DM1 models 15,22, but cytotoxicity and lethality in transgenic mice had not been
thoroughly explored as secondary criteria for success. Here, we report an SAR study of
pentamidine and heptamidine derivatives in the context of DM1 molecular pathology. A
deconstructionist approach was used to determine the structural elements most critical for
splicing rescue of the INSR minigene in a DM1 HeLa cell model. Results from these studies
demonstrated that one amidine group was sufficient for splicing rescue; however, both phenoxy
groups were necessary to retain efficacy in the DM1 cell model.
After thorough examination of previous work on diamidines, we synthesized three series
of molecules varying in amidine substitution and central linker length, as well as planarity. We
uncovered two planar systems, compounds 12 and 13, which exhibited similar activity to
pentamidine but are both considerably less toxic in cellulo and one was less toxic in DM1
transgenic mice. Control splicing experiments did demonstrate that compound 12 (but not
compound 13) caused splicing changes in the absence of CUG repeats, suggesting that the
splicing rescue observed in the DM1 HeLa cell model may be a result of global effects on
ACS Paragon Plus Environment

Page 23 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
splicing. In future studies, it would be interesting to examine the effects of compounds 12 and 13
on all endogenous splicing events through RNAꢀseq analyses of treated DM1 model cells and/or
transgenic DM1 mice. For the purposes of this study, however, the results from these control
experiments prompted us to focus on compound 13, the best compound from our screen with
demonstrated specificity for CUG repeat molecular pathology.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fluorescent microscopy further confirmed that 13 reduced pathological nuclear
aggregates. The reduction of nuclear aggregates is consistent with several different mechanisms.
Pentamidine has been previously demonstrated to reduce CUG RNA levels, suggesting that it
may inhibit transcription of CTG DNA or reduce the stability of the transcript 22. Another
possibility is that these molecules are able to bind directly to CUG RNA and therefore displace
MBNL proteins from nuclear aggregates, which is a commonly employed mechanism by many
other compounds in the study of DM1 20, 21, 44. Future studies will aim to distinguish if amidines
work through one or both of these mechanisms or through additional mechanisms.
EXPERIMENTAL METHODS.
General Methods for Synthesis. 1H and 13C NMR spectra were obtained on a Varian Mercury
300 MHz spectrometer (1H: 300.09) or Inova 500 MHz spectrometer (1H: 500.10 MHz, 13C:
125.75 MHz). Chemical shifts (δ) are expressed in ppm downfield from tetramethylsilane (TMS)
using nonꢀdeuterated solvent present in the bulk deuterated solvent (CDCl3: 1H 7.26 ppm, 13C
77.16 ppm; DMSOꢀd6: 1H 2.50 ppm, 13C 39.52 ppm). All NMR spectra were processed using
MestReNova NMR processing software. High resolution mass spectra were acquired using a
JEOL MSꢀRoute mass spectrometer. THF was distilled from sodium prior to use. The purity of
all compounds was >95% and was confirmed by HPLC analysis using an Agilent 1100 Series
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 35
1
2
3
4
5
6
7
8
9
chromatograph with Diode Array Detector (DAD) UV detection. The chromatographic method
utilized a Waters XTerra MS Cꢀ18 3.9 x 100 mm, 3.5 m column; UV detection wavelength =
240 nm; flow rate = 1 mL/min; gradient = 0ꢀ100% (0.1% TFA in acetonitrile) over 25 min with a
hold of 5 min at 100% (0.1% TFA in acetonitrile); the aqueous mobile phase contained 0.1%
TFA. All chemicals were of reagent grade and used as obtained from manufacturers. For
compound synthesis and characterization, see Supporting Information text.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Splicing Analysis in Cell Culture. Approximately 2.1 x 105 cells were plated in 6ꢀwell plates and
1 µg total plasmid (either all pcDNA3 for mock treatment or 500 ng pcDNA3 and 500 ng
CUG960 plasmid) was transfected 24–36 h later using Lipofectamine 2000 (Invitrogen). After
transfection, cells were incubated in OPTIꢀmem (GIBCO) for 6 h and then washed with 1x PBS
and placed in DMEM with GLUTAMAX (GIBCO) supplemented with 10% FBS (v/v) (GIBCO)
along with indicated concentrations of compounds. Cells were harvested 18 h after addition of
compound and RNA was isolated immediately using an RNeasy kit (QIAGEN) and frozen at –
80°C until use. RNA concentration was determined using a NanoDrop (Thermo).After DNase
treatment, RNA was reverse transcribed with SuperScript II and a plasmid specific reverse
primer. This cDNA was then subjected to PCR (22 rounds for INSR or 24 rounds for TNNT2).
Resulting PCR products were then run on a 6% (w/v) 19:1 native polyacrylamide gel containing
0.5x TB at 300 V for 90 min. The gel was then stained with 1x SYBR I dye (Applied
Biosystems) in 0.5xTB for 15 min. Quantification of bands was performed using the Alpha
Imager HP software from Alpha Innotech. EC50 values were determined by performing linear
regression analysis of the data using Excel (Microsoft), and determining the concentration which
would result in 50% splicing rescue, calculated as halfway between wild type and disease state
ACS Paragon Plus Environment

Page 25 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
levels of exon inclusion. Statistical significance of splicing change due to molecule treatment
was calculated by tꢀtest in excel. Tꢀtests compared splicing values of treatment to splicing values
in the disease control. Each data set was analyzed for equal/unequal variance using Fꢀtest in
excel in order to determine appropriate Tꢀtest.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The following primers were used. For the TNNT2 minigene, the RT primer was 5’ AGC ATT
TAG GTG ACA CTA TAG AAT AGG G. The forward primer for PCR was 5’ GTT CAC AAC
CAT CTA AAG CAA GAT G and the reverse primer was 5’ GTT GCA TGG CTG GTG CAG
G. For the INSR minigene, the RT primer was 5’ GCT GCA ATA AAC AAG TTC TGC. The
forward primer for PCR was 5’ CGA ATT CGA ATG CTG CTC CTG TCC AAA GAC AG,
and the reverse primer was 5’ TCG TGG GCA CGC TGG TCG AG. The TNNT2 minigene was
a gift from the lab of Thomas Cooper and the INSR minigene was a gift from Nicholas Webster.
Cell Viability Analysis. Cells were plated and transfected as in the splicing analysis protocol.
After an 18 h incubation with compound, both dead (floating) cells and live (adhered) cells were
carefully harvested into 2 mL, lightꢀsensitive tubes to avoid rupturing and lysis. Cells were
centrifuged at 13,000 rcf for 3 minutes and residual media was aspirated from the tubes, leaving
cell pellets. A 100 µL volume of 1x PBS was then added to the cell pellets and each was
vortexed on high to reꢀsuspend cells in buffer. Cell suspensions were then incubated at room
temperature for 15 minutes with 1 µg/mL final concentration of propidium iodide. Cell
suspensions were then analyzed using an Attune Acoustic Focusing Flowcytometer (Life
technologies). IC50 values were determined with KaleidaGraph (Synergy) software using the
equation
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 35
1
2
3
4
5
6
7
8
9
where m0 = small molecule concentration, m1= IC50, m2 = Hill coefficient, and m3= fraction of
MBNL1 bound without small molecule present. Errors were determined by calculating the
standard deviation of triplicate data.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fluorescent Microscopy. Imaging of foci was performed as previously described15. Briefly, HeLa
cells were plated in 6ꢀwell plates onto coverslips and transfected with 500 ng of CUG960 plasmid
for each experiment. After transfection, respective compounds (pentamidine, 12, or 13) were
added in the indicated concentrations and the cells were fixed 16 h later. Cells were fixed with
4% PFA and washed in 1× PBS, then permeabilized with 0.5% triton Xꢀ100, in 1× PBS. Cells
were prewashed with 30% formamide, 2× SSC. Cells were then probed overnight at 37 °C with a
Cy3 CAG10 probe (IDT, IA). The next day cells were washed 30% formamide, 2× SSC, and then
with 1× SSC. Cells were next washed twice in 1× PBS and then probed overnight with the MB1a
antiꢀMBNL1 antibody (1:5000 dilution, MDA Monoclonal Antibody Resource) in 1× PBS. Cells
were washed with 1× PBS and incubated with donkey antiꢀmouse Alexa 488 (1:500 dilution).
Cells were washed 2 times with 1× PBS and then mounted onto glass slides using hardꢀset
mounting media that contains DAPI (Vectashield). Cells were imaged on an Olympus Fluoview
FV1000 with a Bx61 scope. Nuclear foci for each cell were quantified using ImageJ. F test
analyses were performed to determine if the data sets possessed equal or unequal variances.
Appropriate Tꢀtest analyses were then performed using Excel (Microsoft).
Compound 13 treatments of mice. Homozygous genderꢀmatched, HSALR transgenic mice of 10–
14 wks age were treated with compound 13 (dissolved in 5% glucose solution) at the indicated
ACS Paragon Plus Environment

Page 27 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dose by daily intraperitoneal injection for 7 d. Control mice received 5% glucose injections.
Mice were sacrificed 1 d after the final injection and quadriceps muscle was obtained for splicing
analysis. RNA was isolated, reverse transcribed, and amplified by PCR, and analyzed on agarose
gels using a fluorimager.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT.
Supporting Information Available: Synthetic methods and spectroscopic characterization of
compounds used in this study are provided herein. This material is available free of charge via
the Internet at http://pubs.acs.org.
Corresponding Author: *To whom correspondence should be addressed.
Email: aberglun@uoregon.edu
Conflicts of Interest: The authors declare the following competing financial interest(s): The
University of Oregon and J. Andrew Berglund have patented pentamidine and analogues (US
8,436,049 B2) as possible compounds for therapeutic use in Myotonic Dystrophy.
Acknowledgements: This work was supported by grant AR0599833 from NIAMS/NIH and
Myotonic Dystrophy Foundation (fellowship to Micah Bodner). We thank T. Cooper for the gifts
of TNNT2 and DMPK960 minigenes and N. Webster for the INSR minigenes. HRMS were
obtained at the Biomolecular Mass Spectrometry Core of the Environmental Health Sciences
Core Center, Oregon State University, supported by grant #P30ꢀES00210, National Institute of
Environmental Health Sciences, National Institutes of Health. We also thank the MDA
Monoclonal Resource for providing the MBNL1 antibody.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 35
1
2
3
4
Abbreviations used:
5
6
Myotonic dystrophy type 1: DM1
7
8
9
DM protein kinase gene (human): DMPK
Muscleblindꢀlike 1: MBNL1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Insulin receptor gene (human): INSR
Cardiac troponin T gene (human): TNNT2
Chloride channel (human): CLCN1
Plasmid containing 960 CUG repeats: CUG960
Chloride channel (mouse): Clcn1
Sarcosplasmic ATPase Atp2a1 (mouse): Atp2a1
Half maximal concentration (splicing rescue): EC50
Half maximal concentration (cell toxicity): IC50
References
(1) Harper P.S. Major problems in neurology: myotonic dystrophy; WB Saunders: London,
2001.
(2) Ranum L. P. and Cooper, T. A. RNAꢀmediated neuromuscular disorders. Annu. Rev.
Neurosci. 2006, 29, 259–277.
(3) Cho D. H., and Tapscott S. J. Myotonic dystrophy: emerging mechanisms for DM1 and
DM2. Biochim. Biophys. Acta, Mol. Basis Dis. 2007, 1772, 195–204.
(4) Harley H.G.; Brook J.D.; Rundle S.A.; Crow S.; Reardon W.; Buckler A.J.; Harper P.S.;
Housman D.E.; Shaw D.J. Expansion of an unstable DNA region and phenotypic variation in
myotonic dystrophy. Nature 1992, 355(6360), 545–546.
ACS Paragon Plus Environment

Page 29 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(5) Mahadevan M.; Tsilfidis C.; Sabourin L.; Shutler G.; Amemiya C.; Jansen G.; Neville C.;
Narang M.; Marcelo J.; O’Hoy K.; Leblond S.; EarleꢀMacdonald J.; de Jong PJ.; Wieringa
B.; Korneluk RG. Myotonic dystrophy mutation: an unstable CTG repeat in the 3’
untranslated region of the gene. Science 1992, 255, 1253–1255.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(6) O'Rourke J. R. and Swanson M. S. Mechanisms of RNAꢀmediated disease. J. Biol. Chem.
2009, 284, 7419–7423.
(7) Ho T.H.; Savkur R.S.; Poulos M.G.; Mancini M.A.; Swanson M.S.; Cooper T.A.
Colocalization of muscleblind with RNA foci is separable from misꢀregulation of alternative
splicing in myotonic dystrophy. J. Cell Sci. 2005, 118(13), 2923–2933.
(8) Osborne, R.J.; Thornton, C.A. RNA dominant diseases. Hum. Mol. Genet. 2006, 15, 162ꢀ69.
(9) Fardaei M.; Rogers M.T.; Thorpe H.M.; Larkin K.; Hamshere M.G.; Harper P.S.; Brook J.D.
Three proteins, MBNL, MBLL and MBXL, coꢀlocalize in vivo with nuclear foci of
expandedꢀrepeat transcripts in DM1 and DM2 cells. Hum. Mol. Genet. 2002, 11, 805–814.
(10) Wang E.T.; Cody N.A.; Jog S.; Biancolella M.; Wang T.T.; Treacy D.J.; Luo S.; Schroth
G.P.; Housman D.E.; Reddy S.; Lécuyer E.; Burge C.B. Transcriptomeꢀwide regulation of
preꢀmRNA splicing and mRNA localization by muscleblind proteins. Cell 2012, 15, 710–
724.
(11)
Konieczny P.; StepniakꢀKonieczna E.; Sobczak K. MBNL proteins and their target
RNAs, interaction and splicing regulation. Nucleic Acids Res. 2015, 1;42(17), 10873ꢀ87.
(12)
Kanadia R.N.; Johnstone K.A.; Mankodi A.; Lungu C.; Thornton C.A.; Esson D.;
Timmers A.M.; Hauswirth W.W.; Swanson M.S. A muscleblind knockout model for
myotonic dystrophy. Science 2003, 302, 1978–1980.
(13)
Philips, A. V.; Timchenko, L. T.; Cooper, T. A. Disruption of splicing regulated by a
ACS Paragon Plus Environment