Discovery of Risdiplam, a Selective Survival of Motor Neuron-2 (SMN2) Gene Splicing Modifier for the Treatment of Spinal Muscular Atrophy (SMA)

Drug Annotation

pubs.acs.org/jmc

Cite This: J. Med. Chem. XXXX, XXX, XXX −XXX

Discovery of Risdiplam, a Selective Survival of Motor Neuron‑2

(SMN2) Gene Splicing Modiﬁer for the Treatment of Spinal Muscular

Atrophy (SMA)

Hasane Ratni,*^,†Martin Ebeling,^†John Baird,^‡Stefanie Bendels,^†Johan Bylund,^†Karen S. Chen,^§

Nora Denk,^†Zhihua Feng,^∥Luke Green,^†Melanie Guerard,^†Philippe Jablonski,^†Bjoern Jacobsen,^†

Omar Khwaja,^†Heidemarie Kletzl,^†Chien-Ping Ko,^∥Stefan Kustermann,^†Anne Marquet,^†

Friedrich Metzger,^†Barbara Muelle_†r,^†Nikolai A. Naryshkin,^‡Sergey_†V. Paushkin,^§Emmanuel Pinard,^†

†

‡

,†

̀

Agnes Poirier, Michael Reutlinger, Marla Weetall, Andreas Zeller, Xin Zhao, and Lutz Mueller*

^†F. Hoﬀmann-La Roche Ltd., pRED, Pharma Research & Early Development, Roche Innovation Center Basel, Grenzacherstrasse

124, 4070 Basel, Switzerland

^‡PTC Therapeutics, Inc., 100 Corporate Court, South Plainﬁeld, New Jersey 07080, United States

^§SMA Foundation, 888 Seventh Avenue, Suite 400, New York, New York 10019, United States

^∥Section of Neurobiology, Department of Biological Sciences, University of Southern California, Los Angeles, California 90089,

United States

S

* Supporting Information

ABSTRACT: SMA is an inherited disease that leads to loss

of motor function and ambulation and a reduced life

expectancy. We have been working to develop orally

administrated, systemically distributed small molecules to

increase levels of functional SMN protein. Compound 2 was

the ﬁrst SMN2 splicing modiﬁer tested in clinical trials in

healthy volunteers and SMA patients. It was safe and well

tolerated and increased SMN protein levels up to 2-fold in

patients. Nevertheless, its development was stopped as a

precautionary measure because retinal toxicity was observed

in cynomolgus monkeys after chronic daily oral dosing (39

weeks) at exposures in excess of those investigated in patients. Herein, we describe the discovery of 1 (risdiplam, RG7916,

RO7034067) that focused on thorough pharmacology, DMPK and safety characterization and optimization. This compound is

undergoing pivotal clinical trials and is a promising medicine for the treatment of patients in all ages and stages with SMA.

The SMN2 gene, due to a translationally synonymous C- to T-

mutation in exon 7, encodes a pre-mRNA that undergoes

alternative splicing that excludes exon 7 from the majority of the

mRNA. This SMN2Δ7 mRNA produces a rapidly degrading

SMNΔ7 protein, whereas the reduced quantities of full length

SMN2 mRNA generate insuﬃcient levels of wild type full length

SMN protein. All SMA patients (with no functional SMN1

gene) have at least one SMN2 copy, and some have up to six

copies in a somatic cell.⁴Our strategy to treat SMA has focused

on the discovery and development of orally bioavailable and

brain penetrant small molecules that shift the outcome of SMN2

alternative splicing toward the production of full length SMN2

mRNA increasing functional SMN protein levels. The hallmark

of SMA is the progressive degeneration of α-motor neurons in

the brain stem and spinal cord that causes muscle atrophy and

disease-related complications that can impact survival. However,

INTRODUCTION

■

Spinal muscular atrophy (SMA)¹is an autosomal recessive

neuromuscular disorder characterized by the progressive loss of

spinal motor neurons leading to muscle weakness. It is the

leading genetic cause of mortality in infants and young children,

with an incidence of 1 in 11 000 live births.²SMA manifests in

various degrees of severity ranging from type 1 (patients never

achieve independent sitting) to type 2 (patients can sit but not

stand or walk) and type 3 (patients achieve walking ability), all of

which have in common clinical signs including hypotonia,

muscle weakness and atrophy, and impaired mobility.

SMA is caused by a homozygous deletion or mutation of the

survival of motor neuron 1 (SMN1) gene on chromosome 5q

(locus 5q13) which encodes survival motor neuron protein

(SMN), an essential protein for normal development and

functional homeostasis in all species,³expressed in both

neuronal and non-neuronal cells. Most humans carry a second,

very closely related gene SMN2 that also can produce SMN

protein; however, its level is mainly driven by the SMN1 gene.

Received: May 8, 2018

A

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

more and more data suggest that SMA is not only a disease of

motor neurons. SMA is increasingly described as a disease

aﬀecting tissues and cell types beyond the motor neuron. This

then strongly suggest that an eﬀective treatment may require

body-wide correction of SMN protein levels to achieve a

complete reversal or amelioration of the disease state. Hence we

are looking for small molecule increasing SMN protein in both

the CNS and peripheral organs and tissues upon oral

administration.

In recent years, a number of therapeutic options for SMA have

been assessed,⁵and among them the modiﬁcation of SMN2

alternative splicing appears highly promising. It can be achieved

either by antisense oligonucleotides (ASOs)⁶(nusinersen

approved by U.S. FDA in December 2016) that currently

require intrathecal administration or advantageously by small

molecules upon oral delivery leading to SMN protein level

increases in both the CNS and peripheral organs and tissues as

initially reported by us⁷and recently by others (5-(1H-pyrazol-

4-yl)-2-[6-[(2,2,6,6-tetramethyl-4-piperidyl)oxy]pyridazin-3-

yl]phenol (branaplam, LMI070), currently in phase 2 clinical

trials).⁸

In this paper, we describe the path from compound 2

(RG7800, RO6885247),⁹the ﬁrst molecule which entered

human clinical trials in SMA, to the discovery of compound 1

(risdiplam, RG7916, RO7034067, 7-(4,7-diazaspiro[2.5]octan-

7-yl)-2-(2,8-dimethylimidazo[1,2-b]pyridazin-6-yl)pyrido[1,2-

a]pyrimidin-4-one) currently undergoing pivotal clinical studies

in SMA patients (Figure 1). We utilized state-of-the-art

optimization and characterization of the pharmacology,

DMPK, and nonclinical safety.

length SMN2 mRNA increased, and the levels of SMN2Δ7

mRNA decreased, with no detectable change in SMN1 full

length mRNA levels. Those results supported the evaluation of

compound 2 in SMA patients to assess its safety and conﬁrmed

proof of mechanism measuring its eﬀect on SMN2 exon 7

splicing and on SMN protein levels in whole blood.¹⁰An

increase in full length SMN2 mRNA levels and up to a 2-fold

increase in SMN protein levels relative to those at baseline after

12 weeks of treatment were achieved. However, this clinical

study was put on hold as a precautionary measure due to safety

ﬁndings in cynomolgus monkeys (nonreversible histological

ﬁndings in the retina) in the long term chronic toxicity study (39

weeks) performed in parallel to the clinical trial. Exposure levels

of compound 2 in this long-term chronic preclinical toxicity

study were however considerably higher than those reached in

the clinical study. However, it seemed not feasible to explore the

full therapeutic potential of compound 2 while still maintaining a

safety margin to the animal toxicity ﬁndings.

Selectivity of 2. This proof of mechanism achieved in SMA

patients supported the subsequent identiﬁcation of a molecule

with improved properties. Eﬀorts focused on increasing further

the therapeutic index and improving the PK proﬁle.

One parameter evaluated was selectivity. We recently

reported on a transcriptional proﬁling analysis performed in

PNN1-46 SMA patient-derived ﬁbroblasts treated with various

compounds that all had an eﬀect on SMN2 splicing.¹¹Going

through the list of genes whose splicing was aﬀected by at least

one such compound (Figure 3b of our previously published

work;¹¹also see Figure 14), we identiﬁed several genes of

particular relevance for their established functional roles in cell

cycle regulation or cell death signaling. The functions of these

genes, such as FOXM1 and MADD, were consistent with

phenotypes seen in in vitro and animal toxicity studies in the

form of (stage-speciﬁc) cell cycle arrest, micronucleus induction,

and apoptosis formation as described later in this paper.

FOXM1, forkhead box protein M1, encodes a key cell cycle

regulator. FOXM1 protein is critically required for cell division

and is highly expressed in rapidly dividing cells, such as those

found in the gastrointestinal tract, male germ cells, skin, and

blood cell progenitors in the bone marrow. Knock-down of

FOXM1 or a change in its splice variants can lead to mitotic

arrest and apoptosis, depending on the stage of the cell cycle.

The transcriptionally inactive FOXM1a variant contains exon

A2 (full length), and the transcriptionally active FOXM1b/c

variants lack exon A2 (ΔA2). There is also a third, recently

described FOXM1ΔC isoform.¹²FOXM1ΔC is expressed in

several cancer cell lines, interferes with normal cell cycle

progression, induces chromosomal instability and aneuploidy.

Using RT-qPCR with speciﬁc primers for FOXM1a (full length)

and FOXM1b/c (ΔA2), the change in the alternative splicing

pattern of FOXM1 after treatment with compound 2 in SMA

patient-derived ﬁbroblasts was conﬁrmed with a FOXM1a EC₅₀

of 247 nM ( 60 nM) and FOXM1b/c EC₅₀352 nM ( 49 nM),

respectively. For comparison, the SMN2 splicing activity (FL

mRNA in HEK293H) of compound 2 was 23 nM. We also

evaluated the eﬀect of compound 2 on the expression level of

FOXM1 mRNA variants (FOXM1b/c) in human and

cynomolgus monkey induced pluripotent stem cells (iPSCs).

A similar potency was observed with an IC₅₀of 1.85 μM and 0.90

μM in human and cynomolgus monkey iPSCs, respectively.

These data conﬁrm similarities in responses of human and

cynomolgus monkey cells toward changes in splice variants of

FOXM1, which is particularly important as cynomolgus

Figure 1. Structures of compounds 1 and 2.

BACKGROUND

■

Characterization of Compound 2. A ﬁrst-in-human study

to assess compound 2 for safety, tolerability, pharmacokinetics

(PK), and pharmacodynamics (PD) in healthy subjects was

performed. Single oral doses were administered to healthy male

subjects in a single-ascending-dose, placebo controlled, double

blind study. SMN1 full length, SMN2 full length, and SMNΔ7

(generated predominantly by the SMN2 gene) mRNA levels

were quantiﬁed in whole blood samples using a newly developed

multiplex RT-qPCR assay. Compound 2 was safe and well

tolerated in this study at all doses tested, with the highest dose

selected being determined by animal no adverse eﬀect data and

the ability to generate reliable pharmacodynamic eﬀects (RNA

splicing data in blood cells as biomarker). The plasma exposure

increased in a slightly more than dose proportional manner. The

peak plasma concentration of compound 2 was reached at 5−8 h

after dose with a terminal half-life of 120 h. Due to long half-life,

subjects showed exposure to compound 2 in plasma for more

than 2 weeks. A dose- and exposure-dependent eﬀect on SMN2

exon 7 splicing in whole blood was observed: the levels of full

B

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 2. Structures of compounds 3, 4, and 5 with a more potent imidazopyridazine moiety.

monkeys were used for toxicity characterization of compound 1

and suﬃcient margins in terms of safety were established in vivo

to study the full therapeutic potential of compound 1 in patients.

Nonclinical Pharmaceutical Evaluation and Safety of

2. Other pharmaceutical and safety-related properties were also

evaluated. The eﬀect of compound 2 on the hERG channel was

assessed at 37 °C giving an IC₅₀and IC₂₀of 1.8 μM and 0.5 μM,

respectively. A telemetry study was performed in cynomolgus

monkeys in which a mild QTC interval prolongation of at most

10 ms was observed only at the highest dose tested. The

exposure at the highest dose in cynomolgus monkeys reached

the IC₂₀, and hence in vivo results were in line with the IC₂₀of

the hERG channel interaction in vitro and thus expected.

Clinical investigations and management of such eﬀects in the

clinic were put in place.

In an in vitro phototoxicity assessment of compound 2 using

the standard neutral red uptake assay upon UVA irradiation in

3T3 ﬁbroblasts an IC₅₀of 330 nM was measured (IC₅₀> 10 μM

in the dark). The phototoxic potential was subsequently

conﬁrmed in rats in vivo. Clinical safety measures were put in

place for the trials with compound 2.

As part of the testing for genotoxic activity, compound 2 was

found to be negative in the bacterial reverse mutation assay.

However, increased frequencies of micronucleated cells were

observed both in vitro in mammalian cells (mouse and human

cells) and in vivo in rats upon oral dosing. The increased

frequency of micronucleated cells can be expected based on the

changes in FOXM1 splicing observed in the presence of

compound 2. We also observed changes in male germ cell in

vivo in rats and cynomolgus monkeys after at least seven once-

daily oral doses of compound 2. Similar observations with

compound 1 were made, for which the mechanism is further

discuss herein after.

Microscopic examination in cynomolgus monkey tissue

revealed stage- and cell-speciﬁc germ cell degeneration in the

testes in males. Staging analysis of the germ cell degeneration

was performed, and it turned out that the arrest occurred

primarily in the pachytene stage of meiosis I and associated with

decreased sperm count, increased abnormal sperms, and

reduced testes size. In line with this observation, testis ﬁndings

fully reversed or showed ongoing recovery (normalization of

sperm count, sperm morphology and testis size, absence of germ

cell degeneration) following an extended period of observation

after treatment cessation. The master cell cycle regulating factor

FOXM1 (also termed HFH11) is known to be highly expressed

in spermatocytes but not spermatogonia of the testis.¹³Thus,

the eﬀect of treatment of cynomolgus monkeys with compound

2 on testes is stage-speciﬁc and reversible. Similar ﬁndings were

observed for compound 1.

Consistent with the large volume of distribution of compound

2 (V_ss= 29 L/kg in rat) and its basic charge, histopathological

observations of epithelial vacuolation and foamy macrophages

were observed suggestive of phospholipidosis in several tissues

in the rat. The eﬀect was less pronounced than with other

compounds in the same chemistry series. Overall, a large volume

of distribution appeared to be the key driver of phospolipidosis.

Thus, reduction of V_sswas part of the further optimization

process to arrive at compound 1.

Possible improvement of the PK proﬁle of compound 2

initially consisted of shortening the half-life as well as reducing

its volume of distribution (V_ss). We have reported a terminal

half-life of around 120 h in healthy adult humans. This is

consistent with our observation in rats and cynomolgus

monkeys with a respective half-life of 19 h and 42h after oral

dosing and a V_ssof around 20 to 30 L/kg for both species.

COMPOUND DESIGN AND OPTIMIZATION

RESULTING IN DISCOVERY OF 1

■

As a continuation of our eﬀorts to discover a SMN2 splicing

modiﬁer drug candidate with an improved safety, pharmacoki-

netic and pharmacodynamic proﬁle suitable for the chronic

treatment of SMA patients, we evaluated a novel class of

benzamide derivatives which has been recently reported.¹⁴

While this class initially held some promises, such as absence of

phototoxicity, diﬀerent chemotype and ﬁrst prototypes

displaying good in vivo eﬃcacy, this series was ultimately

terminated. The in vivo eﬃcacy could not be further improved,

with the best derivatives being less potent than compound 2 and

oﬀering no advantage in terms of selectivity. We therefore kept

our focus on the pyridopyrimidinone series, from which our ﬁrst

clinical candidate was identiﬁed.

To increase the therapeutic window versus nontarget related

potential side eﬀect (e.g., hERG channel interaction, phospho-

lipidosis, phototoxicity), we focused on changing key elements

in the molecules to reduce basicity (bpKa: basic pK_a), volume of

distribution, and UV absorption and at the same to improve the

potency so as to reduce the required eﬃcacious dose/exposure.

First, we evaluated several novel fused heteroaromatic in place of

the pyrazolopyrazine moiety as right-hand side. Gratifyingly, the

use of an imidazopyridazine fragment led to a potency increase

of roughly 10-fold providing us with a novel and promising

subclass (Figure 2, compounds 4 and 5 as compared to

C

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 3. Structural overlay (A) and gas phase QM torsional angle scan statistics for the indicated bond (B) for compound 3 in gray and compound 4 in

orange.

Figure 4. The potency of compounds in the pyridopyrimidinone series in modifying the splicing of SMN2 exon 7 and FOXM1 exon A2 is correlated.

compound 3). Compound 3 contains the same pyrazolopyr-

azine moiety as compound 2.⁹

Next, to establish a structure−activity relationship (SAR) for

FOXM1 splicing in order to further increase selectivity in favor

of SMN2 splicing, we tested an already existing pool of 210

derivatives from the pyridopyrimidinone series for their activity

in FOXM1 splicing assay. A clear correlation (Figure 4, red line,

regression analysis) was observed between the in vitro splicing

activity of SMN2 (EC_1.5×FL mRNA in HEK293H) and FOXM1

(IC₅₀ΔA2 in ﬁbroblast). Compounds were generally 10-fold

more potent against SMN2 than FOXM1. However, com-

pound-mediated modulation of SMN2 and FOXM1 alternative

splicing appear mechanistically related so that substantially

increasing the in vitro selectivity for SMN2 over FOXM1

appeared not to be feasible.

The novel imidazopyridazine fragment conserved the

conformational preorganization and planarity of the right-hand

side which has been shown previously to be crucial to obtain

potent compounds (Figure 3). SAR analysis indicates that the

pyrazine nitrogen provides an important molecular interaction

and is critical for a high in vitro potency. The change of the

hydrogen-bond binding energy in imidazopyridazine fragment

contained in compounds 4 and 5 relative to pyrazolopyrazine

from compound 3 (−8.06 kcal/mol vs −7.16 kcal/mol

respectively estimated by QM) could contribute to the observed

increase in potency.

D

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

We therefore adapted our optimization strategy and focused

on an increase of the in vivo selectivity between SMN2 and

FOXM1 splicing as an improvement versus compound 2. We

found that compound 2 can be metabolized in vitro to form N-

dealkylated metabolite 6 across all species tested (human, rat,

dog, minipig, cynomolgus monkey, mouse, and rabbit; Figure

5). Upon oral administration in rodents and cynomolgus

diversity of amine fragments is tolerated on the left as long as it

contains a basic amine with a pK_aabove 6.5 (to maintain in vitro

potency). To deﬁne the optimal amine moiety, we considered

several additional molecular properties descriptors. One of them

was the lipophilicity of the ﬁnal molecule. We have previously

shown that lipophilicity in this series was a strong predictor of

the P-gp eﬄux. Therefore, we continued designing molecules

with a calculated lipophilicity between 1.6 and 3.0 to achieve

respectively a human P-gp ER below 2.5 and to prevent a high

clearance and nonspeciﬁc protein binding.¹⁶

The hERG aﬃnity and phospholipidosis (seen with

compound 2) can generally be reduced by the lowering of the

basicity of an amine present in the molecule. In addition, we had

determined that volume of distribution of compounds in this

series (e.g., compound 2 V_ssis around 20−30 L/kg in

cynomolgus monkeys and rodents) is correlated with the

basicity of the molecule (Table 1, Figure 7). We had to strike an

optimal balance between basicity and lipophilicity while

maintaining potency. We focused on selecting a left-hand-side

basic amine moiety with the lowest bpKa (but still above 6.5).

Taking into consideration those criteria, we prepared a virtual

library of ﬁnal compounds containing amines, with a calculated

log D between 1.6 and 3 and a calculated basic pK_abetween 6.5

and 8 (the lower the better) and not subject to N-dealkylation

(Figure 8). The amine fragments were selected by a matched

pair analysis from in house molecules.

This exercise led us to a list of roughly 100 virtual compounds.

After a visual inspection of the chemical structure to remove any

amine fragments with unwanted features (reactive or unstable

functional groups for example), we identiﬁed 40 compounds of

interest which were prepared and characterized. Compound 1

came out on top due to a combination of very high potency,

strongly reduced basicity, no phototoxicity risk, and the lack of

active metabolites (Figure 9).

Figure 5. In vitro and in vivo metabolism of 2 leading to 6.

monkeys, we observed the formation of this metabolite 6 with

plasma concentrations reaching up to 9% of those for parent

compound 2.

Metabolite 6 is 10-fold more potent than parent compound 2

in vitro in SMA patient-derived ﬁbroblasts in both SMN2

(EC_1.5×, SMN FL = 2 nM) and FOXM1 (EC₅₀, FOXM1, ΔA2 =

23 nM) splicing assays. The higher potency of the metabolite

was also observed on FOXM1 mRNA expression levels in human

and cynomolgus monkey iPSCs with an IC₅₀of 119 nM and 103

nM, respectively. However, metabolite 6 was found to be a

strong P-gp substrate with an eﬄux ratio (ER, a measure of P-gp-

mediated eﬄux) of 18.7 in human and 20.0 in mouse cells.

Therefore, not surprisingly, metabolite 6 did not elicit SMN

protein increase in brain but only in muscles of adult C/C-allele

mice, a mouse model of mild SMA¹⁵when dosed orally once

daily for 10 days, despite the high drug levels in plasma.

Therefore, being peripherally restricted, metabolite 6 can

contribute to potential side eﬀects mediated by the induced

FOXM1 isoforms in peripheral organs and tissues but not

contribute to the CNS activity in term of SMN protein increase.

Hence, part of our revised strategy to improve the therapeutic

window in vivo was to avoid formation of peripherally restricted

active metabolites. Thereby, the peripheral side eﬀects induced

by FOXM1 splicing could be minimized.

Preclinical Characterization of 1. Compound 1 was

subjected to a full characterization. It displayed an improved and

very high in vitro potency on SMN2 splicing, while the basicity

has been drastically reduced by nearly 4 log units (Table 2). The

lipophilicity was maintained to 2.5; hence compound 1 was not a

human P-gp substrate. Fraction unbound in plasma was around

10% and similar across species.

Compound 1 was active in vitro in SMA patient-derived

ﬁbroblasts and in motor neurons generated from induced

pluripotent stem cells (iPSCs) derived from SMA type 1 patient

ﬁbroblasts, promoting the inclusion of exon 7, to generate full-

length (FL) mRNA. SMN protein levels were also increased

(Figure 10).

We evaluated two approaches to suppress the N-dealkylation:

(1) making the N-alkyl residue a part of a ring or (2) moving the

alkyl residue to an adjacent location (Figure 6).

At this stage, the central core (the pyridopyrimidinone

moiety) and the right-hand-side fragment (a novel imidazo-

pyridazine moiety) had been ﬁxed. We knew that a very large

Figure 6. Two approaches to suppress the formation of N-dealkylated active metabolites.

E

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

marked basicity reduction of the compound 1 (mpKa of 6.8) as

initially hypothesized. Furthermore, the half-life of the

compound was also markedly reduced being in a more favorable

range of 5−6 h upon oral administration.

Table 1. Volume of Distribution Measured in Rat and pKa

Preclinical Eﬃcacy of 1. We then evaluated compound 1 in

two mouse models of SMA. Adult C/C-allele mice with mild

SMA phenotype were treated for 10 days with compound 1

given once a day orally at three diﬀerent doses (1, 3, and 10 mg/

kg). The level of SMN protein was assessed in brain and in

quadriceps muscle (Table 4). In the Δ7 mouse model of severe

SMA,¹⁷compound 1 was administered by intraperitoneal (ip)

injection once daily starting on postnatal day (PND) 3 and

continued through PND9. The level of SMN protein was

assessed in brain and in quadriceps muscle as well (Table 5). A

maximum SMN protein increase was already reached at 1 mg/kg

(free AUC_0−24hextrapolated from 3 mg/kg exposure = 73 ng.h/

ml), whereas a dose of 3 mg/kg (free AUC_0−24h= 118 ng.h/ml)

for compound 2 led to a lower SMN protein increase. These data

demonstrate that 1 potently increases SMN protein in both

brain and muscle tissues of transgenic mouse models of SMA.

The impact on the lifespan and body weight of Δ7 SMA mice

treated with compound 1 was also assessed daily (Figure 12).

During the observation period, only two heterozygous

littermates died. In contrast, all vehicle-treated mice died before

PND21 with a median survival time (MST) of 10.5 days.

Compound 1 treatment prolonged animal survival at all doses,

with a minor but signiﬁcant eﬀect already at the lowest dose

tested (MST of 26 days), whereas for the other dose groups

minimal diﬀerences from heterozygous littermates were

observed. Body weight gain followed a similar trend. These

data indicate that treatment with compound 1 prevents the

manifestation of the SMA phenotype in the severely aﬀected Δ7

SMA mice when dosing is started at PND3.

Finally, we evaluated the eﬀect of compound 1 on

neuromuscular pathology in the Δ7 SMA mouse model.

Animals were treated from PND3 to PND14 by ip

administration of vehicle or compound 1 at doses of 0.1, 0.3,

and 1 mg kg⁻¹day⁻¹. Spinal cord and muscle tissues were

processed for histological assessment. Relative to heterozygous

littermates, SMAΔ7 mice showed a signiﬁcant loss of vesicular

glutamate transporter 1 (vGlut1)-positive proprioceptive inputs

onto motor neuron, loss of motor neurons, neuromuscular

junction (NMJ) denervation in the longissimus muscle, and

muscle atrophy. Upon treatment with compound 1, we observed

a signiﬁcant and dose-dependent increase of number of vGlut1

inputs, the number of motor neurons, the percentage of fully

innervated NMJs and increased muscle size in the extensor

digitorum longus (EDL) muscles relative to those in vehicle-

treated SMNΔ7 mice (Figure 13).

Nonclinical Safety of 1. Compound 1 was also subjected to

preclinical safety assessment. The eﬀect on the hERG channel

was assessed at 37 °C. No eﬀect on the hERG K⁺current was

observed (IC₂₀> 5 μM). This marked improvement compared

to compound 2 can be attributed to the strongly reduced basicity

of the molecule. A telemetry study was performed in

cynomolgus monkey which conﬁrmed the absence of QTC

interval prolongation as well as absence of any other

cardiovascular ﬁndings.

In vitro metabolism studies using human liver microsomes

and human hepatocytes identiﬁed N-hydroxylated derivative 18

(Figure 11) as the main metabolite (although formed at a low

level of 3.8% and 1.7%, respectively). Compound 18 was

assessed for its potency on both SMN2 and FOXM1 splicing

assays and was inactive as anticipated (due to removal of

basicity).

This was conﬁrmed by assessing the eﬀect of compound 1 and

its metabolite 18 on the expression level of FOXM1 mRNA

variants (FOXM1b/c) in human and cynomolgus monkey

induced pluripotent stem cells (iPSCs). While 1 modiﬁed

FOXM1 splicing with IC₅₀, ΔA2 of 113 nM and 155 nM

respectively, its metabolite 18 did not show any eﬀect. This is in

sharp contrast to the compound 2 and its metabolite 6 and met

our design criteria.

A favorable drug metabolism and pharmacokinetic (DMPK)

proﬁle in the rat and cynomolgus monkey was observed for 1

upon intravenous and oral administration (Table 3). Note-

worthy, the volume of distribution 3.1 L/kg (rat) and 2.0 L/kg

(cynomolgus monkey) is markedly reduced compared to our

ﬁrst compound 2. This outcome is perfectly in line with the

In an in vitro phototoxicity assessment using the standard

neutral red uptake in 3T3 ﬁbroblasts assay, the compound did

not show any eﬀect up to the highest solubility-limited

concentration of 9 μM (∼3600 ng/mL), i.e., well above any

expected therapeutic free concentration. This improvement is

F

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 7. Volume of distribution (V_ss) measured in rat versus calculated basicity (cpKa).

Table 2. In Vitro Potency and Physicochemical

Characterization

SMN2 FOXM1

SMN

splicing splicing protein

f_u[%]

h/m

EC_1.5×

[nM]

EC₅₀

[nM]

EC_1.5×

[nM] log D

h P-gp

ER

p

compd

mpKa

1

2

4

23

67

247

29

87

2.5

2.3

2.2

2.0

6.8

10.9

11/10

13/2

Figure 8. Virtual library to deﬁne compounds with optimal

combination of basicity, lipophilicity, and potency.

observations made in the repeat dose animal toxicity studies

with compounds 1 and 2. The selected animal species for these

studies have been rats and cynomolgus monkeys based on their

well-known characterization in toxicity studies for small

molecules while acknowledging at the same time that SMN2

gene is unique to humans and absent in the commonly used

animal species in toxicity testing. Hence, toxicities seen in

animals with compounds 1 and 2 were either characterized as

unrelated to its action on pre-mRNA splicing (oﬀ-target) or as

related to the compound interactions with the splicing

machinery but aﬀecting splicing targets other than SMN2

(termed secondary targets). Yet, similar to their quick onset of

action on the SMN2 target in vitro and in transgenic mice, the

results of engagement of secondary splice targets in SMN2

nonresponder animals became visible soon after start of repeat

dose toxicity studies and did not change much in severity with

chronic dosing. Findings of toxicological signiﬁcance for

compound 1 (as with compound 2) were observed mainly in

organs with rapid cell turnover in mice, rats, and cynomolgus

monkeys. Looking into the range of secondary splice targets

most likely due to the use of the novel right-hand-side fragment

(the imidazopyridazine).

No evidence of phospholipidosis has been observed in vivo in

nonclinical safety studies which again can be attributed to the

reduced basicity (4 units lower), reduced calculated amphiphi-

licity, reduced volume of distribution (from 20 to 2 L/kg), and

an improved potency (leading to lower eﬃcacious AUC).

Speciﬁcally, compound 1 did not display any evidence for

phospholipidosis in any tissues, at the doses tested, when

administered chronically with daily oral gavage for up to 26

weeks to rats or up to 39 weeks to cynomolgus monkeys up to

the highest doses tested.

Role of Selectivity in Safety. In a comprehensive strategy

to proﬁle fully compounds 1 and 2 for their eﬀects on pre-

mRNA splicing in general, the safety assessment strategy

included a genome-wide splice site analysis, which identiﬁed a

potential of compounds 1 and 2 to engage splice sites beyond

the SMN2 target. Results of this analysis and speciﬁc functions

of gene splice variants associated with it were compared with the

Figure 9. Optimization from 2 leading to the discovery of 1.

G

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 10. Compound 1 tested in (A) SMN2 splicing in SMA type I ﬁbroblasts, (B) SMN protein in SMA type I ﬁbroblasts, and (C) SMN protein in

SMA type I motor neurons. FL = full length, and Δ7 = mRNA lacking exon 7.

Figure 11. In vitro metabolism of 1.

a

Table 3. In Vivo Single-Dose Pharmacokinetic (SDPK) Proﬁle for Compounds 1 and 2

rat

cynomolgus monkey

compd

Cl^a(mL min⁻¹kg⁻¹

)

V_ss^a(L/kg)

T_1/2^b(h)

F^b(%)

Cl^c(mL min⁻¹kg⁻¹

)

V_ss^c(L/kg)

T_1/2^d(h)

F^d(%)

1

2

8.9

25

3.1

29

6.4

19

∼100

5.7

5

2.0

20

5.4

42

43

52

a

For compound 1: (a) iv, 1.9 mg/kg; (b) po, 5.5 mg/kg; (c) iv, 0.1 mg/kg; (d) po, 0.5 mg/kg. For compound 2: (a) iv, 2 mg/kg; (b) po, 5 mg/kg;

(c) iv, 0.3 mg/kg; (d) po, 1.3 mg/kg.

eﬀect of this compound on genes such as FOXM1 and/or

MADD, i.e., genes the functions of which are associated with cell

cycle and apoptosis processes and for which splice variant

changes result in changes in function. The ﬁndings were

reversible or partially reversible in nature (depending on the

chosen reversibility period and cell turnover in the respective

organs) and included the following:

Table 4. Pharmacokinetics and SMN Induction in Adult C/

C-Allele Mouse Model

a

total plasma

AUC_0−24h

% SMN

increase in

brain

% SMN

increase in

quadriceps

dose

compd (mg/kg po)

(μg·h/mL)

vehicle

0

1

3

0

na

9.0

0

17

0

b

1

b

• micronucleation (MN) induction in vitro in mouse cell

lines and in rat bone marrow erythroblasts and decreased

cellularity in bone marrow;

10

100

49

a

b

Plasma AUC determined in a satellite group. Not assessed.

• histopathological changes in gastrointestinal tract epi-

thelia (increased apoptosis/single cell necrosis) and

lamina propria (vacuolation) and exocrine pancreas

epithelia (single cell necrosis) in mice, rats, and/or

cynomolgus monkeys;

• parakeratosis/hyperplasia/degeneration of the skin,

tongue, and larynx epithelia with associated inﬂammation

in cynomolgus monkeys;

Table 5. Pharmacokinetics and SMN Induction in Neonatal

Δ7 Mouse Model

a

total plasma

AUC_0−24h

% SMN

increase in

brain

% SMN

increase in

quadriceps

dose

compd (mg/kg ip)

(μg·h/mL)

vehicle

0

na

0

28

90

206

202

0

32

64

210

241

b

1

0.1

0.3

1

b

• degeneration of germ cells in the testis of cynomolgus

b

monkeys and rats.

3

2.2

Most importantly for clinical evaluation of safety and eﬃcacy of

compound 1, the in vitro and animal toxicity studies with

compound 1 yielded clearly no observed adverse eﬀect levels

(NOAELs) for all of the ﬁndings listed above. Consistent with

the hypothesis of a transient eﬀect on splice variants, adverse

eﬀects were reversible upon cessation of treatment. Consistent

with the approach to improve potency, adverse eﬀects were only

a

Plasma AUC determined in a satellite group of neonatal (P10) wild-

b

type mice. Not assessed.

identiﬁed with the genome-wide splice-site analysis, these

phenotypes of toxicity observed in the toxicity studies above

the no observed adverse eﬀect level (NOAEL) are in line with an

H

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

analysis for its relatively homogeneous cell population). Figure

14 shows the eﬀects of compounds 1, 2, and 6 and branaplam on

the splicing of the primary target SMN2 and on the most

pronounced secondary splice targets (of either compound, note

the diﬀerences in speciﬁcity). This is based on in vitro

experiments with SMA patient-derived ﬁbroblasts at concen-

trations active on SMN2 splicing; thus diﬀerent concentrations

of the respective compounds were used. Most prominent

secondary splice targets in these experiments were STRN3, FOX

M1, MADD, APLP2, and SLC25A17 mRNAs.

In summary, compounds 1 and 2 do indeed show good

selectivity toward the SMN2 target, with most of the secondary

splice target eﬀects seen with very weak or even undetectable

levels at the concentrations with full pharmacodynamics eﬀect

on the SMN2 target.

The splicing events described cover a range of biological

eﬀects.

For SMN2, exon 7 inclusion is promoted, leading to the

desired increased production of a functional full length SMN

protein. In the animal experiments in vitro or in vivo, no changes

in SMN splicing could be observed with compound 1 or 2 due to

the lack of SMN2 gene and divergent 5′ splice site in Smn1.

For STRN3, targeting of a donor site in all species (mice, rats,

cynomolgus monkeys, and humans) leads to the production of

an at best rare, maybe nonphysiological protein variant that

includes only one of two exons that usually come together.

However, the event will “only” introduce an additional 37 amino

acids into the protein and may not abolish the protein

completely. It remains unclear whether there could be any

safety issue stemming from changes in STRN3 splice variants as

there is little evidence from the published literature which would

be compatible with the phenotypes seen in toxicity studies.

For FOXM1, the increased use of two diﬀerent 5′ splice sites

in the same exon can play a role. The annotated donor splice site

can lead to the formation of a dominant-negative form of the

protein, whereas use of a cryptic internal donor site results in the

production of the C terminus-truncated FOXM1ΔC isoform

Figure 12. Survival and body weight gain of Δ7 mice treated with

compound 1. Animals were treated once daily with vehicle or

compound at doses of 0.1, 0.3, 1, and 3 mg kg⁻¹day⁻¹from PND3

to PND23 by ip injection and thereafter by oral gavage at doses of 0.3, 1,

3, and 10 mg kg⁻¹day⁻¹.

which has been observed in several cancer cell lines¹²

.

FOXM1ΔCis a dominant negative inhibitor of wild type

FOXM1 function as a transcriptional factor and has been

shown to interfere with normal cell cycle progression, induce

chromosome instability and aneuploidy. Changes in FOXM1

with formation of a dominant negative exon inclusion form,

which inhibits cell cycle progression, were seen in cynomolgus

monkey cells and tissues. In the rat Foxm1 gene the region

encoding exon A2 is absent, and therefore, these splice variants

cannot form but an overall reduction of FoxM1 expression was

observed when rodent cells used for the in vitro micronucleus

test were analyzed for eﬀects on Fox M1. On the basis of the

known involvement of FOXM1 in cell cycle progression,

changes in its splice variants may be the most important factor

for understanding of the toxicities observed in particular in

cynomolgus monkeys with compounds 1 and 2 and other SMN2

splice modiﬁers.

For APLP2 and MADD, rare but physiologically relevant

exons are included at an increased rate, modulating the protein

function in ways that are not fully understood. However, clear

changes in the splice variants of MADD were also seen in cells

and tissues from rats and monkeys treated with compounds 1

and 2. One of the known splice variants of MADD, IG20, was

upregulated in tissues or rats and cynomolgus monkeys, in which

apoptosis was seen. IG20 is described as a proapoptotic splice

variant of MADD.

observed at exposures signiﬁcantly above exposures predicted to

be associated with the desired degree of modiﬁcation of SMN2

pre-mRNA splicing and SMN protein increase in SMA patients.

Thus, the overall clear NOAEL for adverse eﬀects seen in

animals enabled the identiﬁcation of eﬃcacious exposures for

compound 1 which could be safely tested in healthy volunteers

and in SMA patients while still maintaining a full pharmacody-

namic response. The ﬁndings also identiﬁed with reliability

safety parameters that can be monitored in human clinical trials

with compound 1 in the unexpected case of a higher sensitivity

of humans as compared to animals used for toxicological

evaluation. The currently available clinical evidence for safety

and pharmacodynamic response of compound 1 supports this

strategy.

To determine possible mechanisms behind these animal

toxicities and their relevance to humans, compounds 1 and 2 and

a series of earlier SMN2 splice modiﬁers were tested for their

speciﬁcity of interaction with the splicing machinery.¹¹These

included primary and secondary splice target investigations in

(1) patient cells, (2) induced pluripotent stem cells from human

and cynomolgus monkey, (3) cells of tissues of rats and

cynomolgus monkeys aﬀected with the observed toxicities

(compared with the reference organ spleen, which was selected

based on absence of major toxicities but easily amenable to

I

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 13. Δ7 SMA mice treated with 1 from PND3 to PND14. Neuromuscular connectivity and muscle atrophy were assessed by

immunohistochemistry. (A) vGlut1-positive proprioceptive inputs in L3−5 spinal cord. (B) Ventral motor neurons in L3−5 spinal cord. (C) NMJ

innervation onto longissimus muscle. (D) EDL muscle cross-sectional area. ∗ = p < 0.05; ∗∗ = p < 0.01; ∗∗∗ = p < 0.001 vs vehicle-treated mice.

Finally, for SLC25A17, inclusion of exon 3 was observed and

in a small fraction of products a downstream cryptic 5′ splice site

was also activated. According to the current genome assignment,

exon 3 encodes an alternative translation start codon predicted

to generate SLC25A17 protein with an alternative N-terminus.

To elucidate the mechanism of micronucleus (MN)

formation and/or exclude direct DNA reactivity, a number of

in vitro and in vivo studies were performed using known

methodologies.²⁰No induction of γH2AX, a biomarker for DNA

double-stranded breaks, was observed at concentrations

compatible with cell survival in L5178Y mouse lymphoma

cells or TK6 human lymphoblastoid cells. The proportion of

apoptotic cells was increased at cytotoxic concentrations, but

cell cycle perturbations or relevant responses in other in vitro

end points like phosphorylation of histone H3 or polyploidy

were not observed. Acute treatment of rats over 3 days with up to

25 mg kg⁻¹d⁻¹in a combined MN/comet study did not induce

primary DNA-damage as measured with the comet assay in liver

while a clear, dose-dependent increase of the frequency of

micronucleated erythrocytes was observed. A conclusive

mechanism for the induction of micronuclei by compound 1

could not be demonstrated, but the evidence described above

strongly points toward an indirect mode of action (e.g., related

to secondary splice target interaction) and excludes direct DNA-

reactivity. Therefore, a linear dose−response relationship must

not be assumed but rather a sublinear, “threshold-type” curve

should be used to derive appropriate safety margins. On the basis

of the most recent recommendations of the International

Workshop on Genotoxicity Testing (IWGT),²¹the benchmark-

dose (BMD) concept was applied to determine the point-of-

departure (PoD) with the software PROAST²²using Hill

modeling. The variability of solvent control samples (i.e., the

“background noise”) of the micronucleus assays suggests a

critical eﬀect size (CES) of approximately 50% to calculate the

benchmark-dose lower conﬁdence limit (BMDL, equivalent to

the critical eﬀect dose lower conﬁdence limit CEDL) of

exposure to compound 1 for micronucleus induction.²³Studies

performed with young (QD 13 week treatment duration) and

adult (both acute 2 days and QD 4 week treatment duration)

rats were used to derive C_maxand AUC_0−24hvalues for compound

1 that would not lead to relevant increases of micronucleated

erythrocyte frequencies. Graphical representations of the

modeling are shown Figure 15.

Other Types of Toxicities Observed in Animals

Treated with Compounds 1 and 2. In addition to the

above-mentioned ﬁndings in animal toxicity studies, histological

ﬁndings were present in the retina of cynomolgus monkeys.

These were noted with chronic daily oral treatment for both

compounds 1 and 2. Retinal degeneration was observed during

and at the end of daily oral dosing for 39 weeks. During the in-life

J

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 14. Splicing eﬀects on target and selected oﬀ-target transcripts. The diagrams show numbers of reads mapped to six human genomic loci in

various conditions. Each peak corresponds to an exon present in mature mRNA. Note that exons and introns are not drawn to scale, and lengths of loci

diﬀer substantially. The diagrams on the left present complete overviews of the loci, and a detailed view of the critical region (aﬀected by at least one of

the compounds tested) is shown enlarged on the right. Black: vehicle-treated control samples (human patient ﬁbroblasts). Red: 131 nM (light) and

655 nM (dark) doses, respectively, of compound 2. Blue: 121 nM (light) and 605 nM (dark) doses, respectively, of compound 1. For the detailed views

on the right, also added are compound 6, the potent metabolite of compound 2 (at 95 nM, in magenta), and branaplam, at 24 nM (cyan). (A) SMN2.

All compounds show very comparable on-target eﬀects. For compounds 2, 6, and 1, increasing the doses to higher values did not have any marked

additional eﬀect on the on-target activity, indicating a saturation of the on-target eﬀect. The 3′-most part of the exon is higher in controls and repressed

by all the treatments, consistent with a switch between the spliced and read-through forms. (B) STRN3. Next to SMN2, striatin 3 splicing is most

consistently aﬀected by all compounds tested, probably due to a very similar splicing mechanism.¹¹For this gene, two variants are well described that

either include or exclude two adjacent exons, exons 8 and 9 (highlighted). In the control samples (black), both exons are present at similarly low levels,

indicating that they are skipped in the majority of transcripts. As shown in the right panel, all the tested compounds have a rather speciﬁc splicing eﬀect

on only the ﬁrst of the two, exon 8, to produce a splice variant that was described earlier but is not well characterized and that does not appear to play

any major role under physiological conditions. Compounds 2 and 1 at lower doses have minor eﬀects on STRN3 exon8 (left) that however increase

dose dependently. In comparison, exon 9 (right) is aﬀected to a much lower degree. Both compound 6 and branaplam have eﬀects comparable to the

higher doses of compounds 2 and 1 (and about twice as high as for the lower doses of compounds 2 and 1), indicating some selectivity for both

compounds toward the on-target, SMN2. Exon 8 has a length of 111 bp, divisible by 3, so inclusion of this exon alone does not interrupt the STRN3

open reading frame (exon 9 is 141 bp long, again an “in-frame” exon). (C) FOXM1 is a key cell cycle regulator and, as such, potentially critical for oﬀ-

target eﬀects. The aﬀected splicing event, at ﬁrst glance, is the inclusion or exclusion of exon A2 (also referred to as exon VIIa or exon 9) that disrupts a

protein domain of FOXM1, leading to a dominant-negative variant of the protein that retains the ability to bind DNA but lacks the transactivation

domain and therefore cannot promote DNA transcription.¹⁸This domain is usually not included, or only at very low levels, in controls (black line).

Compounds 2 and 1 at their lower concentrations have a very moderate eﬀect on inclusion of the complete exon A2. In addition, compounds 6 and

branaplam and compounds 2 and 1 at their higher doses appear to activate, to a much higher degree, an alternative 5′ splice site located inside exon A2

leading to the generation of the FOXM1ΔC isoform that has been previously described.¹²This isoform contains only the ﬁrst 34 nucleotides of exon A2

resulting in a frameshift and the generation of the truncated FOXM1ΔCprotein. (D) APLP2 is another pre-mRNA whose splicing was found to be

aﬀected by various compounds, albeit very weakly by the low doses of compounds 1 and 2. The aﬀected exon is 168 bp long (no change in the reading

frame) and is used in about one-quarter of the total transcripts in controls. Compounds 6 and branaplam increased the usage of this exon by more than

2-fold in this experiment, whereas compounds 2 and 1 have slightly lower eﬀects even at their higher doses. (E) MADD has a complex gene structure

with many exons and various splice variants described, most of which are not fully annotated or characterized. The aﬀected exon here is absent in

controls, as shown in the panel on the right (the neighboring exon to the right is included for reference in the detailed view). The black arrow points to

another facultative exon that is coexpressed with the boxed exon in all annotated variants of this gene. As for several examples above, compounds 2 and

1, at their lower concentrations, have only a very small eﬀect on this exon. Compound 6 at the dose used here and compounds 2 and 1 at their higher

doses show somewhat stronger eﬀects that are, however, still very weak. In contrast, branaplam leads to an almost complete inclusion of the exon into

all MADD transcripts. The second exon (arrow) does not appear to be aﬀected by any of the compounds and always remains very low or absent. (F)

K

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 14. continued

SLC25A17. Exon 3 comes in two forms of diﬀerent lengths. The shorter, canonical form of the exon (5′ part) is not included in controls

but is very weakly aﬀected by compounds 2 and 1 at their lower doses. At higher doses, and for branaplam, this exon is included in about

50% of all SLC25A17 transcripts. There is a very weak eﬀect of all compounds on an extended, noncanonical form of this exon. These

graphs were generated using the software Manananggal.¹⁹

phase, multifocal peripheral retina degeneration in the photo-

receptor layer and microcystic spaces in the inner retinal layers

were detected using optical coherence tomography (OCT).

This was associated with a depressed B-wave which was more

obvious in a scotopic (rod-driven) than a photopic (cone-

driven) light setting in the electroretinogram (ERG). Histolog-

ically, the most aﬀected layers were the outer nuclear layer

(ONL) and photoreceptor layer with additional vacuolation of

the inner nuclear layer (INL). Incidence and severity of the

ﬁndings were dose dependent, and the changes were located

mostly in the periphery. A no observed eﬀect level (NOEL) for

the retinal ﬁndings was established. In a 22 week recovery phase

after 39 weeks of treatment, retinal degeneration with peripheral

photoreceptor loss remained present. However, as observed

using OCT, microcystic macular degeneration decreased

signiﬁcantly and areas of white speckled hyperﬂuorescence

became less dense. In previously treated animals, ERG

amplitudes noticeably increased by the end of the recovery

phase showing normal ERGs in 3 out of 4 animals.

Histologically, retinal degeneration with photoreceptor loss

and retinal pigment epithelium (RPE) hypertrophy were still

present, but the vacuolation of the INL had reversed which can

explain the recovery from depressed ERG responses. Exper-

imental evidence suggested that the eﬀect on the retina is not

directly associated with eﬀects on tissue proliferation but related

to high melanin binding observed in vitro, to tissue retention in

the retina, and impairment of lysosomal/autophagosomal

function in retinal pigmented epithelial cells. For compounds

1 and 2, similar NOELs were established for these retinal

changes in chronic dosing studies in cynomolgus monkeys. As

retinal changes were seen with both compounds, they could

potentially represent a class eﬀect of this series of SMN2 splice

modiﬁers, which are independent of their action on the

spliceosome. This was followed up with investigation of both

compounds with chronic treatment in pigmented rats versus

albino rats. Similar to pronounced accumulation with repeated

dosing in melanin-containing retinal tissue in monkeys, we have

found pronounced accumulation of compounds 1 and 2 in these

retinal tissues in pigmented rats but not in albino rats. Yet,

chronic treatment of pigmented (and albino) rats for 26 weeks

with compounds 1 and 2 at systemic exposures in excess of those

in monkeys did not yield any evidence for retinal changes

detectable with OCT, ERG, and histopathology similar to those

observed in the monkey, while other types of toxicities such as to

the male germ cells were congruent between albino and

pigmented rats and monkeys. Hence, the retinal changes seen

with chronic treatment of compounds 1 and 2 appear to be

speciﬁc for the monkey, but human relevance cannot be

excluded. Since accumulation in melanin-containing tissue was

present in both pigmented rats and monkey, the toxicity is not

directly associated with the melanin-binding potential of

compounds 1 and 2. Consolidation of the mechanism of action

and judgment of whether this is a class eﬀect would require

conduct of further monkey studies with chronic treatment.

Overall, the translatability of compound-induced retinal changes

from cynomolgus monkeys to humans is presently unknown, but

OCT was established as a highly sensitive method to detect early

changes to the retina, which in our studies in cynomolgus

monkeys were present after 5−6 months of treatment.

Compound 1 was preferred for further clinical exploration due

to its higher potency for SMN2 exon 7 inclusion and hence a

potentially higher therapeutic window with respect to secondary

splice target engagement, absence of safety concerns regarding

phospholipidosis, phototoxicity, QT-prolongation, and an

improved therapeutic window for retinal changes in the

cynomolgus monkey. A full therapeutic eﬀect is expected at

exposures in patients not exceeding the no adverse eﬀect level

from animal toxicity studies.

CLINICAL EVALUATION

■

Compound 1, is currently under clinical evaluation in three

ongoing trials: a study in type 1 infants with SMA, a study in

types 2 and 3 SMA patients age 2−25 years, and a study in non-

naive SMA patients. These studies are aiming to assess safety,

tolerability, and eﬃcacy of 1 in the respective patient

populations. A single ascending dose study in healthy subjects

has been completed previously. The safety, pharmacokinetic,

and pharmacodynamic data in human obtained so far conﬁrm

the successful optimization of the compound as described in this

summary, when compared to data in human from the previous

compound 2.

CHEMISTRY

■

The discovery chemistry synthetic route for the preparation of

compound 1 is highlighted in Scheme 1.²⁴The pyridazine

derivative 19 underwent a nucleophilic substitution with

aqueous ammonia, giving a mixture of regioisomers which

were not separated at this stage. A subsequent N-alkylation of

with 1-bromo-2,2-dimethoxypropane followed by an intra-

molecular cyclization gave the imidazopyridazine 20 (both

regioisomers were separated), which was readily converted into

boronic ester 21. In a convergent manner, a condensation

between the 5-ﬂuoropyridin-2-amine 22 and dimethyl malonate,

followed by chlorination with POCl₃, aﬀorded 23. A Suzuki

cross coupling between 21 and 23 led to 24. Finally, a

regioselective aromatic nucleophilic substitution with 4,7-

diazaspiro[2,5]octane gave the ﬁnal derivative 1.

CONCLUSIONS

■

RNA splice modiﬁers are a new class of small molecule

therapeutics. In terms of speciﬁcity and safety they represent

profound challenges for medicinal chemistry. In-depth charac-

terization of compound 2, our ﬁrst SMN2 splicing modiﬁer

which entered human clinical trials, and understanding of its

safety and pharmacokinetics limitations mainly based on

nonclinical data have led to the rational discovery of 1 displaying

a superior proﬁle, which is currently being veriﬁed in clinical

trials. Additionally, our small molecule, unlike an ASO (which

required an intrathetical administration and have a limited tissue

distribution), is orally available and displayed a wide tissue

L

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Figure 15. (A) BMD of AUC_0−24h(left) and C_max(right) vs MN in acute (48 h) study with adult rats; (B) BMD of AUC_0−24h(left) and C_max(right) vs

MN in 4 week study in adult rats; (C) BMD of AUC_0−24h(left) and C_max(right) vs MN in 13 week study with juvenile rats. CES 50% (i.e., 1.5-fold

increase over concurrent controls) was used with PROAST software. CEDL and CEDU are the critical size lower and upper conﬁdence limits.

distribution. A partly novel optimization process was used,

clearly beyond standard screening. Some critical standard

parameters included potency both in vitro and in vivo, volume

of distribution, and general safety considerations like genotox-

icity, hERG channel, phototoxicity, and phospholipidosis.

Beyond standard eﬀorts included a genome-wide splice site

and gene expression analysis to assess the selectivity, veriﬁcation

of this in induced pluripotent stem cells of man and monkey, and

analysis of several organs from toxicity studies in rodents and

monkeys. Eventually, the clear association between pharmacol-

ogy and toxicity phenotypes in vitro and in vivo led to an

elucidation of both the unique mechanism of action of our class

M

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

a

Scheme 1. Discovery Chemistry Synthetic Route of Compound 1

a

Conditions: (a) (i) aqueous ammonia, 110 °C, 48 h; (ii) 1-bromo-2,2-dimethoxypropane (1.2 equiv), PPTS (0.072 equiv), 2-propanol, 105 °C,

12 h, 21% over 2 steps; (b) 4,4,4′,4′,5,5,5′,5′-octamethyl-2,2′-bi(1,3,2-dioxaborolane) (1.0 equiv), KOAc (2.0 equiv), dioxane, 110 °C; (c) (i)

dimethyl malonate (5 equiv), 230 °C, 1.5 h; (ii) POCl₃, i-Pr₂NEt (1.0 equiv), 110 °C, 15 h, 50% over 2 steps; (d) 21 (1.2 equiv), Pd(PPh₃)₄(0.05

equiv), aqueous K₂CO₃(2M, 3.0 equiv), CH₃CN, 100 °C, 6.0 h, 60%; (e) 4,7-diazaspiro[2,5]octane dihydrochloride (3.0 equiv), i-Pr₂NEt (4.0

equiv), DMSO, 130 °C, 48 h, 18%.

ing to Scheme 1: Step A. Preparation of 20. (i) In a sealed ﬂask, 3,6-

dichloro-4-methylpyridazine (27.0 g, 161 mmol) was suspended in

aqueous ammonia (25%, 300 mL). The reaction mixture was heated at

110 °C for 48 h (turned into a solution after 1 h). After cooling to room

temperature, the reaction was poured on CH₂Cl₂, and the organic phase

was separated, dried over Na₂SO₄, concentrated under vacuum to give

22.4 g of 6-chloro-4-methylpyridazin-3-amine and 6-chloro-5-methyl-

pyridazin-3-amine as a mixture of regioisomers which were used

directly in the next step. (ii) The mixture of regioisomers was

suspended in 2-propanol (300 mL). 1-Bromo-2,2-dimethoxypropane

(36.0 g, 26.6 mL, 193 mmol) and PPTS (2.96 g, 11.6 mmol) were

added, and the resulting solution was heated at 105 °C overnight. The

volatiles were removed under vacuo, and the residue was taken up in

CH₂Cl₂and washed with NaHCO₃. The organic phases were dried over

Na₂SO₄, concentrated under vacuo and the crude light brown solid was

chromatographed (EtOAc/heptane 1/2 to 1/1) to give separately 6-

chloro-2,7-dimethyl-imidazo[1,2-b]pyridazine (5.9 g, 20%) and 6-

chloro-2,8-dimethylimidazo[1,2-b]pyridazine 20 (6.1 g, 21%) as a

white solid. ¹H NMR (600 MHz, DMSO-d₆) δ ppm 8.04 (d, J = 0.6 Hz,

1H), 7.20 (d, J = 1.1 Hz, 1H), 2.54 (d, J = 1.1 Hz, 3H), 2.38 (d, J = 0.7

Hz, 3H); ¹³C NMR (151 MHz, DMSO-d₆) δ ppm 145.0, 143.0, 138.1,

137.7, 117.0, 114.8, 15.9, 14.5. LC−HRMS: m/z = 182.0490 [(M +

H)⁺calcd for C₈H₉ClN₃, 182.0485; Diﬀ 0.5 mDa].

of compound and the toxicity phenotypes including genotoxicity

(micronucleus induction, apoptosis, and cell cycle control).

Compound 1, orally available with wide tissue distribution,

suitable half-life, predictable pharmacokinetics, and very

eﬃcient in brain penetration, is currently undergoing testing

in pivotal clinical trials in types 1, 2, and 3 SMA patients, with a

potential for a life transformational therapy. This work enabled

the selection of a drug candidate for splice modiﬁcation, which

can be clinically tested in patients of all ages with doses not

exceeding the no observed eﬀect levels established in extensive

animal studies including juvenile animals.

The work described in this paper demonstrated that it is

possible to design safe, selective, and eﬃcacious small molecule

splicing modiﬁers, with a potential widespread implications in

the research and development of several additional RNA-

targeting therapies.

EXPERIMENTAL SECTION

Nonclinical Animal Studies. All animal studies were performed

under IACUC approved protocols at AAALAC-certiﬁed animal

facilities.

■

Step B. Preparation of 21. A mixture of 6-chloro-2,8-dimethyl-

imidazo[1,2-b]pyridazine (0.9 g, 4.96 mmol), 4,4,4′,4′,5,5,5′,5′-

octamethyl-2,2′-bi(1,3,2-dioxaborolane) (1.26 g, 4.96 mmol), KOAc

(0.97 g, 9.91 mmol), and Pd(dppf)Cl₂·CH₂Cl₂(363 mg, 0.49 mmol) in

dioxane (50 mL) was degassed and heated under argon at 110 °C. After

15 h, the mixture was diluted with EtOAc, ﬁltered through Celite, and

concentrated under vacuum to give 2,8-dimethyl-6-(4,4,5,5-tetrameth-

yl-1,3,2-dioxaborolan-2-yl)imidazo[1,2-b]pyridazine 21 which was

used directly in the next step.

Compound Synthesis and Characterization. Chemistry.

Reactions were carried out under argon atmosphere. Unless otherwise

mentioned, all reagents and chemicals were obtained from commercial

suppliers and used without further puriﬁcation. All reactions were

followed by TLC (TLC plates F254, Merck) or LCMS (liquid

chromatography−mass spectrometry) analysis. The purity of ﬁnal

compounds as measured by HPLC was at least above 95%. Flash

column chromatography was carried out either using cartridges packed

with silica gel (Isolute Columns, Telos Flash Columns) or on glass

columns on silica gel 60 (32−60 mesh, 60 Å). LC high resolution

spectra were recorded with a Agilent LC system consisting of Agilent

1290 high pressure system, a CTC PAL autosampler, and a Agilent

6520 QTOF. The separation was achieved on a Zorbax Eclipse Plus

C18 1.7 μm, 2.1 mm × 50 mm column at 55 °C; A = 0.01% formic acid

in water; B = 0.01% formic acid in acetonitrile at ﬂow 1 mL/min with

gradient 0 min 5% B, 0.3 min 5% B, 4.5 min 99% B, 5 min 99% B. The

NMR spectra were measured on a Bruker 600 MHz machine in a 5 mm

TCI cryoprobe at 298 K. TMS was used for referencing for experiment

done in CDCl₃. The deuterated DMSO-d₆solvent signal was used as

reference with 2.50 ppm.

Step C. Preparation of 23. (i) A mixture of 2-amino-5-ﬂuoropyridine

22 (11.20 g, 0.10 mol) and dimethyl malonate (57.0 mL, 0.50 mol) was

heated at 230 °C for 1.5 h. After cooling to rt, the precipitate was

collected by ﬁltration and washed with CH₃CN to give 7-ﬂuoro-2-

hydroxy-4H-pyrido[1,2-a]pyrimidin-4-one as a light brown solid,

which was used directly in the next step. (ii) A mixture of crude 7-

ﬂuoro-2-hydroxy-4H-pyrido[1,2-a]pyrimidin-4-one in POC1₃(50

mL) and i-Pr₂NEt (13.3 mL, 77 mmol) was heated at 110 °C for 15

h. The volatiles were removed and the dark residue was treated with

ice−water, washed with water (3×), and dried to give a brown solid.

The crude brown solid was puriﬁed by column chromatography (SiO₂,

CH₂C1₂/MeOH 95:5) to give 2-chloro-7-ﬂuoropyrido[1,2-a]-

7-(4,7-Diazaspiro[2.5]octan-7-yl)-2-(2,8-dimethylimidazo-

[1,2-b]pyridazin-6-yl)pyrido[1,2-a]pyrimidin-4-one (1). Accord-

1

pyrimidin-4-one 23 (9.84 g, 50%) as a yellow solid. H NMR (600

N

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

MHz, DMSO-d₆) δ ppm: 8.99 (dd, J = 4.7, 2.9 Hz, 1H), 8.25 (ddd, J =

9.8, 7.1, 2.9 Hz, 1H), 7.85 (dd, J = 9.7, 5.3 Hz, 1H), 6.56 (s, 1H). ¹³C

NMR (151 MHz, DMSO-d₆) δ ppm 156.7 (td, J = 1.3, 0.5 Hz), 156.3

(d, J = 1.9 Hz), 154.4 (d, J = 243.6 Hz), 148.4 (d, J = 1.0 Hz), 131.4 (d, J

= 25.0 Hz), 127.7 (d, J = 8.0 Hz), 114.3 (d, J = 41.6 Hz), 100.9. LC−

HRMS: m/z = 199.0067 [(M + H)⁺calcd for C₈H₅ClFN₂O, 199.0074;

Diﬀ 0.7 mDa].

Molecular formula strings and some data (CSV)

AUTHOR INFORMATION

Corresponding Authors

*H.R.: e-mail, hasane.ratni@roche.com; phone, (+41) 61-688-

2748.

■

Step D. Preparation of 24. To a solution of 2-chloro-7-ﬂuoro-4H-

pyrido[1,2-a]pyrimidin-4-one 23 (750 mg, 3.78 mmol) in ACN (36

mL) were added 2,8-dimethyl-6-(4,4,5,5-tetramethyl-1,3,2-dioxabor-

olan-2-yl)imidazo[1,2-b]pyridazine 21 (1.24 g, 4.53 mmol), Pd(Ph₃P)₄

(218 mg, 0.189 mmol), and an aqueous solution of K₂CO₃(3.78 mL,

7.55 mmol). The mixture was degassed and heated under argon at 100

°C for 6 h. The reaction was cooled down to rt and ﬁltered. The

precipitate was washed with Et₂O and then water, dried under vacuo to

give 2-(2,8-dimethylimidazo[1,2-b]pyridazin-6-yl)-7-ﬂuoropyrido[1,2-

a]pyrimidin-4-one 24 (700 mg, 60%) as a light brown solid. ¹H NMR

(600 MHz, DMSO-d₆) δ ppm 9.03 (dd, J = 4.7, 2.9 Hz, 1 H), 8.21 (ddd,

J = 9.8, 7.1, 2.9 Hz, 1 H), 8.17 (s, 1 H), 7.98 (br s, 1 H), 7.98 (ddd, J =

9.8, 5.5, 0.6 Hz, 1 H), 7.18 (s, 1 H), 2.65 (s, 3 H), 2.44 (s, 3 H); ¹³C

NMR (151 MHz, DMSO-d₆) δ = 157.1, 156.4, 154.4 (d, J = 243.5 Hz),

149.2, 147.7, 143.7, 139.1, 135.7, 130.3 (d, J = 25.3 Hz), 128.7 (d, J =

8.0 Hz), 114.8, 114.3, 113.7 (d, J = 41.5 Hz), 98.2, 16.3, 14.6. LC−

HRMS: m/z = 310.1098 [(M + H)⁺calcd for C₁₆H₁₃FN₅O, 310.1104;

Diﬀ 0.6 mDa].

Step E. Preparation of 1. In a sealed tube, 2-(2,8-dimethylimidazo-

[1,2-b]pyridazin-6-yl)-7-ﬂuoropyrido[1,2-a]pyrimidin-4-one 24 (50

mg, 0.162 mmol), i-Pr₂NEt (0.22 mL, 1.29 mmol), and 4,7-

diazaspiro[2.5]octane dihydrochloride (32 mg, 0.320 mmol) were

heated in DMSO (2 mL) at 130 °C for 48 h. The solvent was removed

under high vacuum. The residue was taken up in CH₂Cl₂and washed

with an aqueous saturated solution of NaHCO₃. The organic layer was

separated and dried over Na₂SO₄and concentrated under vacuo. The

crude was puriﬁed by column chromatography (SiO₂, CH₂Cl₂/MeOH

= 98/2 to 95/5) to aﬀord 7-(4,7-diazaspiro[2.5]octan-7-yl)-2-(2,8-

dimethylimidazo[1,2-b]pyridazin-6-yl)pyrido[1,2-a]pyrimidin-4-one 1

(12 mg, 18%) as a pale yellow solid. ¹H NMR (600 MHz,CDCl₃) δ ppm

8.45 (d, J = 2.4 Hz, 1H), 7.92 (d, J = 1.0 Hz, 1H), 7.73 (d, J = 9.6 Hz,

1H) 7.80 (s, 1H), 7.70 (dd, J = 9.7, 2.5 Hz, 1H), 7.38 (s, 1H), 3.31−

3.22 (m, 2H), 3.20−3.16 (m, 2H), 3.08 (s, 2H), 2.74 (d, J = 0.9 Hz, 3H)

2.55 (s, 3H), 1.68 (br s, 1H), 0.77−0.75 (m, 2H), 0.67−0.64 (m, 2 H);

¹³C NMR (151 MHz,CDCl₃) δ ppm 158.2, 156.3, 148.5, 147.2, 144.1,

142.2, 140.0, 135.6, 131.2, 126.7, 114.9, 114.7, 110.1, 99.3, 56.7, 49.9,

44.5, 36.5, 16.9, 15.0, 13.0. LC−HRMS: m/z = 402.2051 [(M + H)⁺

calcd for C₂₂H₂₄N₇O, 402.2042; Diﬀ 0.9 mDa].

*L.M.: e-mail, lutz.mueller@roche.com; phone, (+41) 61-688-

2733.

ORCID

Hasane Ratni: 0000-0002-8151-8295

Author Contributions

H.R. was the discovery chemistry project leader and designed

risdiplam. L.M. was the nonclinical safety leader.

Notes

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

We thank Thierry Meyer, Serge Burner, Virginie Brom, Anke

Kurt, and Heidi Schar for the synthesis of the compounds

described, Josef Schneider for the NMR spectroscopy, Christian

Bartelmus for the mass spectroscopy, Liudmila Polonchuk for

hERG assessment, Stephan Kirchner for MNT characterization,

Katja Heinig for BioAnalysis, Massimiliano Donzelli for plasma

protein binding, Mohammed Ullah for P-gp assay, and Claudia

Senn for in vivo PK studies in rodents and cynomolgus monkeys.

Beyond these, the authors of this paper speciﬁcally extend their

gratitude to the extended teams at the F. Hoﬀmann-La Roche

R&D Innovation Center Basel, who supported these research

and development activities in various functions with great

enthusiasm to generate a potentially eﬃcacious and safe

medicine for SMA patients. The research described in the

manuscript was funded by F. Hoﬀmann-La Roche AG, PTC

Therapeutics, and the SMA Foundation.

ABBREVIATIONS USED

■

AUC_0−24h, area under the curve from 0 to 24 h; CEDL/U,

critical eﬀect dose lower/upper conﬁdence limit; DMA, N,N-

dimethylacetamide; DMAP, N,N-dimethylpyridin-4-amine;

DMF, N,N-dimethylformamide; DMSO, dimethyl sulfoxide;

Dppp, bis(diphenylphosphino)propane; ip, intraperitoneal;

MN, micronucleus; NMR, nuclear magnetic resonance; P-gp,

P-glycoprotein; PPTS, pyridinium p-toluenesulfonate; rt, room

temperature; SMA, spinal muscular atrophy; SMN, survival of

motor neuron; TFA, triﬂuoroacetic acid; THF, tetrahydrofuran

SMN HTRF Assay. The level of SMN protein in lysates of

compound-treated cells was quantiﬁed as described previously.⁷

Lipophilicity (log D) Determination. The log D was measured as

previously described.⁹

Ab Initio Calculation. Density functional calculations were

performed using B3LYP/6-31G**. The Jaguar package (release

̈

2017-4, Schrodinger, LLC, New York, NY) was used for all

calculations.²⁵The initial conformation was assigned by OPLS force-

ﬁeld minimization. For calculating the torsion proﬁles, a relaxed

coordinate scan around the dihedral of interest was performed with a 5°

increment.

REFERENCES

■

(1) Spinal Muscular Atrophy: Disease Mechanism and Theraphy, 1st ed.;

Sumner, C. J., Paushkin, S., Ko, C. P., Eds.; Academic Press: Cambridge,

MA, 2016.

cpKa Calculation. The calculation of the pK_avalues were done

using MoKa 2.0,²⁶which was trained with an in-house proprietary pK_a

training set.

(2) (a) Pearn, J. Classification of Spinal Muscular Atrophies. Lancet

1980, 315, 919−922. (b) Crawford, T. O.; Pardo, C. A. The

Neurobiology of Childhood Spinal Muscular Atrophy. Neurobiol. Dis.

1996, 3, 97−110. (c) Sugarman, E. A.; Nagan, N.; Zhu, H.; Akmaev, V.

R.; Zhou, Z.; Rohlfs, E. M.; Flynn, K.; Hendrickson, B. C.; Scholl, T.;

Sirko-Osadsa, D. A.; Allitto, B. A. Pan-Ethnic Carrier Screening and

Prenatal Diagnosis for Spinal Muscular Atrophy: Clinical Laboratory

Analysis of >72,400 Specimens. Eur. J. Hum. Genet. 2012, 20, 27−32.

(3) (a) Fallini, C.; Bassell, G. J.; Rossoll, W. Spinal Muscular Atrophy:

The Role of SMN in Axonal mRNA Regulation. Brain Res. 2012, 1462,

81−92. (b) Schrank, B.; Gotz, R.; Gunnersen, J. M.; Ure, J. M.; Toyka,

K. V.; Smith, A. G.; Sendtner, M. Inactivation of the Survival Motor

Neuron Gene, a Candidate Gene for Human Spinal Muscular Atrophy,

Leads to Massive Cell Death in Early Mouse Embryos. Proc. Natl. Acad.

Neuromuscular Pathology. Neuromuscular pathology was

examined as described previously.⁷

ASSOCIATED CONTENT

■

S

* Supporting Information

The Supporting Information is available free of charge on the

ACS Publications website at DOI: 10.1021/acs.jmed-

chem.8b00741.

¹H NMR and ¹³C NMR of 1 (PDF)

O

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

Sci. U. S. A. 1997, 94 (18), 9920−9925. (c) Paushkin, S.; Gubitz, A. K.;

Massenet, S.; Dreyfuss, G. The SMN Complex, an Assemblyosome of

Ribonucleoproteins. Curr. Opin. Cell Biol. 2002, 14 (3), 305−312.

(4) Crawford, T. O.; Paushkin, S. V.; Kobayashi, D. T.; Forrest, S. J.;

Joyce, C. L.; Finkel, R. S.; Kaufmann, P.; Swoboda, K. J.; Tiziano, D.;

Lomastro, R.; Li, R. H.; Trachtenberg, F. L.; Plasterer, T.; Chen, K. S.

Evaluation of SMN Protein, Transcript, and Copy Number in the

Biomarkers for Spinal Muscular Atrophy (BforSMA) Clinical Study.

PLoS One 2012, 7, e33572.

(5) (a) Kaczmarek, A.; Schneider, S.; Wirth, B.; Riessland, M.

Investigational Therapies for the Treatment of Spinal Muscular

Atrophy. Expert Opin. Invest. Drugs 2015, 24, 867−881. (b) Scoto,

M.; Finkel, R. S.; Mercuri, E.; Muntoni, F. Therapeutic Approaches for

Spinal Muscular Atrophy (SMA). Gene Ther. 2017, 24 (9), 514−519.

(c) Groen, E. J. N.; Talbot, K.; Gillingwater, T. H. Advances in Therapy

for Spinal Muscular Atrophy: Promises and Challenges. Nat. Rev.

Neurol. 2018, 14 (4), 214−224.

(6) (a) Hua, Y.; Sahashi, K.; Rigo, F.; Hung, G.; Horev, G.; Bennett, C.

F.; Krainer, A. R. Peripheral SMN Restoration is Essential for Long-

Term Rescue of a Severe Spinal Muscular Atrophy Mouse Model.

Nature (London, U. K.) 2011, 478, 123−126. (b) Passini, M. A.; Bu, J.;

Richards, A. M.; Kinnecom, C.; Sardi, S. P.; Stanek, L. M.; Hua, Y.;

Rigo, F.; Matson, J.; Hung, G.; Kaye, E. M.; Shihabuddin, L. S.; Krainer,

A. R.; Bennett, C. F.; Cheng, S. H. Antisense Oligonucleotides

Delivered to the Mouse CNS Ameliorate Symptoms of Severe Spinal

Muscular Atrophy. Sci. Transl. Med. 2011, 3, 72ra18. (c) Sahashi, K.;

Ling, K. K. Y.; Hua, Y.; Wilkinson, J. E.; Nomakuchi, T.; Rigo, F.; Hung,

G.; Xu, D.; Jiang, Y.-P.; Lin, R. Z.; Ko, C.-P.; Bennett, C. F.; Krainer, A.

R. Pathological Impact of SMN2 mis-Splicing in Adult SMA Mice.

EMBO Mol. Med. 2013, 5, 1586−1601.

(12) Kim, Y. H.; Choi, M. H.; Kim, J.-H.; Lim, I. K.; Park, T. J. C-

Terminus-Deleted FoxM1 is Expressed in Cancer Cell Lines and

Induces Chromosome Instability. Carcinogenesis 2013, 34 (8), 1907−

1917.

(13) (a) Ye, H.; Kelly, T. F.; Samadani, U.; Lim, L.; Rubio, S.;

Overdier, D. G.; Roebuck, K. A.; Costa, R. H. Hepatocyte Nuclear

Factor 3/Fork Head Homolog 11 is Expressed in Proliferating

Epithelial and Mesenchymal Cells of Embryonic and Adult Tissues.

Mol. Cell. Biol. 1997, 17 (3), 1626−1641. (b) Chaudhary, J.; Mosher,

R.; Kim, G.; Skinner, M. K. Role of Winged Helix Transcription Factor

(WIN) in the Regulation of Sertoli Cell Differentiated Functions: WIN

Acts as an Early Event Gene for Follicle-Stimulating Hormone.

Endocrinology 2000, 141 (8), 2758−2766.

(14) Pinard, E.; Green, L.; Reutlinger, M.; Weetall, M.; Naryshkin, N.

A.; Baird, J.; Chen, K. S.; Paushkin, S. V.; Metzger, F.; Ratni, H.

Discovery of a Novel Class of Survival Motor Neuron 2 Splicing

Modifiers for the Treatment of Spinal Muscular Atrophy. J. Med. Chem.

2017, 60 (10), 4444−4457.

(15) Osborne, M.; Gomez, D.; Feng, Z.; McEwen, C.; Beltran, J.;

Cirillo, K.; El-Khodor, B.; Lin, M.-Y.; Li, Y.; Knowlton, W. M.;

McKemy, D. D.; Bogdanik, L.; Butts-Dehm, K.; Martens, K.; Davis, C.;

Doty, R.; Wardwell, K.; Ghavami, A.; Kobayashi, D.; Ko, C.-P.;

Ramboz, S.; Lutz, C. Characterization of Behavioral and Neuro-

muscular Junction Phenotypes in a Novel Allelic Series of SMA Mouse

Models. Hum. Mol. Genet. 2012, 21 (20), 4431−4447.

(16) Poirier, A.; Cascais, A.-C.; Bader, U.; Portmann, R.; Brun, M.-E.;

Walter, I.; Hillebrecht, A.; Ullah, M.; Funk, C. Calibration of In Vitro

Multidrug Resistance Protein 1 Substrate and Inhibition Assays as a

Basis to Support the Prediction of Clinically Relevant Interactions In.

Drug Metab. Dispos. 2014, 42 (9), 1411−1422.

(7) Naryshkin, N. A.; Weetall, M.; Dakka, A.; Narasimhan, J.; Zhao,

X.; Feng, Z.; Ling, K. K. Y.; Karp, G. M.; Qi, H.; Woll, M. G.; Chen, G.;

Zhang, N.; Gabbeta, V.; Vazirani, P.; Bhattacharyya, A.; Furia, B.;

Risher, N.; Sheedy, J.; Kong, R.; Ma, J.; Turpoff, A.; Lee, C.-S.; Zhang,

X.; Moon, Y.-C.; Trifillis, P.; Welch, E. M.; Colacino, J. M.; Babiak, J.;

Almstead, N. G.; Peltz, S. W.; Eng, L. A.; Chen, K. S.; Mull, J. L.; Lynes,

M. S.; Rubin, L. L.; Fontoura, P.; Santarelli, L.; Haehnke, D.; McCarthy,

K. D.; Schmucki, R.; Ebeling, M.; Sivaramakrishnan, M.; Ko, C.-P.;

Paushkin, S. V.; Ratni, H.; Gerlach, I.; Ghosh, A.; Metzger, F. SMN2

Splicing Modifiers Improve Motor Function and Longevity in Mice

with Spinal Muscular Atrophy. Science (Washington, DC, U. S.) 2014,

345, 688−693.

(8) Palacino, J.; Swalley, S. E.; Song, C.; Cheung, A. K.; Shu, L.; Zhang,

X.; Van Hoosear, M.; Shin, Y.; Chin, D. N.; Keller, C. G.; Beibel, M.;

Renaud, N. A.; Smith, T. M.; Salcius, M.; Shi, X.; Hild, M.; Servais, R.;

Jain, M.; Deng, L.; Bullock, C.; McLellan, M.; Schuierer, S.; Murphy, L.;

Blommers, M. J. J.; Blaustein, C.; Berenshteyn, F.; Lacoste, A.; Thomas,

J. R.; Roma, G.; Michaud, G. A.; Tseng, B. S.; Porter, J. A.; Myer, V. E.;

Tallarico, J. A.; Hamann, L. G.; Curtis, D.; Fishman, M. C.; Dietrich, W.

F.; Dales, N. A.; Sivasankaran, R. SMN2 Splice Modulators Enhance

U1-pre-mRNA Association and Rescue SMA mice. Nat. Chem. Biol.

2015, 11, 511−517.

(9) Ratni, H.; Karp, G. M.; Weetall, M.; Naryshkin, N. A.; Paushkin, S.

V.; Chen, K. S.; McCarthy, K. D.; Qi, H.; Turpoff, A.; Woll, M. G.;

Zhang, X.; Zhang, N.; Yang, T.; Dakka, A.; Vazirani, P.; Zhao, X.;

Pinard, E.; Green, L.; David-Pierson, P.; Tuerck, D.; Poirier, A.; Muster,

W.; Kirchner, S.; Mueller, L.; Gerlach, I.; Metzger, F. Specific

Correction of Alternative Survival Motor Neuron 2 Splicing by Small

Molecules: Discovery of a Potential Novel Medicine to Treat Spinal

Muscular Atrophy. J. Med. Chem. 2016, 59 (13), 6086−6100.

(10) Moonﬁsh clinical study, NCT02240355. www.clinicaltrials.gov.

(11) Sivaramakrishnan, M.; McCarthy, K. D.; Huber, S.; Meier, S.;

Augustin, A.; Heckel, T.; Meistermann, H.; Hug, M. N.; Birrer, P.;

Schmucki, R.; Berntenis, N.; Giroud, N.; Golling, S.; Tzouros, M.;

Banfai, B.; Duran-Pacheco, G.; Lamerz, J.; Luebbers, T.; Ratni, H.;

Ebeling, M.; Metzger, F.; Liu, Y. H.; Krainer, A. R.; Campagne, S.;

Moursy, A.; Khawaja, S.; Clery, A.; Allain, F. H. T.; Paushkin, S. Binding

to SMN2 pre-mRNA-Protein Complex Elicits Specificity for Small

Molecule Splicing Modifiers. Nat. Commun. 2017, 8 (1), 1476.

(17) Le, T. T.; Pham, L. T.; Butchbach, M. E. R.; Zhang, H. L.;

Monani, U. R.; Coovert, D. D.; Gavrilina, T. O.; Xing, L.; Bassell, G. J.;

Burghes, A. H. M. SMNΔ7, the Major Product of the Centromeric

Survival Motor Neuron (SMN2) Gene, Extends Survival in Mice with

Spinal Muscular Atrophy and Associates with Full-Length SMN. Hum.

Mol. Genet. 2005, 14 (6), 845−857.

(18) Korver, W.; Roose, J.; Clevers, H. The Winged-Helix

Transcription Factor Trident is Expressed in Cycling Cells. Nucleic

Acids Res. 1997, 25 (9), 1715−1719.

(19) Barann, M.; Zimmer, R.; Birzele, F. Manananggal - a Novel

Viewer for Alternative Splicing Events. BMC Bioinf. 2017, 18, 120.

(20) (a) Torous, D. K.; Hall, N. E.; Murante, F. G.; Gleason, S. E.;

Tometsko, C. R.; Dertinger, S. D. Comparative Scoring of Micro-

nucleated Reticulocytes in Rat Peripheral Blood by Flow Cytometry

and Microscopy. Toxicol. Sci. 2003, 74 (2), 309−314. (b) Bryce, S. M.;

Bernacki, D. T.; Bemis, J. C.; Spellman, R. A.; Engel, M. E.; Schuler, M.;

Lorge, E.; Heikkinen, P. T.; Hemmann, U.; Thybaud, V.; Wilde, S.;

Queisser, N.; Sutter, A.; Zeller, A.; Guerard, M.; Kirkland, D.;

Dertinger, S. D. Interlaboratory Evaluation of a Multiplexed High

Information Content In Vitro Genotoxicity Assay. Environ. Mol.

Mutagen. 2017, 58 (3), 146−161.

(21) MacGregor, J. T.; Frotschl, R.; White, P. A.; Crump, K. S.;

Eastmond, D. A.; Fukushima, S.; Guerard, M.; Hayashi, M.; Soeteman-

Hernandez, L. G.; Johnson, G. E.; Kasamatsu, T.; Levy, D. D.; Morita,

T.; Muller, L.; Schoeny, R.; Schuler, M. J.; Thybaud, V. IWGT Report

on Quantitative Approaches to Genotoxicity Risk Assessment II. Use of

Point-of-Departure (PoD) Metrics in Defining Acceptable Exposure

Limits and Assessing Human Risk. Mutat. Res., Genet. Toxicol. Environ.

Mutagen. 2015, 783, 66−78.

(22) Software PROAST, version 38.9. www.rivm.nl/proast (accessed

2016).

(23) Zeller, A.; Duran-Pacheco, G.; Guerard, M. An Appraisal of

Critical Effect Sizes for the Benchmark Dose Approach to Assess Dose-

Response Relationships in Genetic Toxicology. Arch. Toxicol. 2017, 91

(12), 3799−3807.

(24) Ratni, H.; Green, L.; Naryshkin, N. A.; Weetall, M. L. 2-

(Imidazo[1,2-b]pyridazin-6-yl)pyrido[1,2a]pyrimidin-4-one Deriva-

tives as SMN Modulators and Their Preparation and Use for the

Treatment of Spinal Muscular Atrophy. WO2015173181 A1, 2015.

P

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Drug Annotation

(25) Bochevarov, A. D.; Harder, E.; Hughes, T. F.; Greenwood, J. R.;

Braden, D. A.; Philipp, D. M.; Rinaldo, D.; Halls, M. D.; Zhang, J.;

Friesner, R. A. Jaguar: A High-Performance Quantum Chemistry

Software Program with Strengths in Life and Materials Sciences. Int. J.

Quantum Chem. 2013, 113 (18), 2110−2142.

(26) Milletti, F.; Storchi, L.; Sforna, G.; Cruciani, G. New and Original

pKa Prediction Method Using Grid Molecular Interaction Fields. J.

Chem. Inf. Model. 2007, 47 (6), 2172−2181.

Q

DOI: 10.1021/acs.jmedchem.8b00741

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Discovery of Risdiplam, a Selective Survival of Motor Neuron-2 (SMN2) Gene Splicing Modifier for the Treatment of Spinal Muscular Atrophy (SMA)

DOI: 10.1021/acs.jmedchem.8b00741

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.8b00741

Full text of DOI:10.1021/acs.jmedchem.8b00741

Products guided by the article

R&D Labs maybe for 1449598-85-9

Relevant to this article

Hot Product