Key words antisense oligonucleotide - exon skipping - N-of-1 treatment - rare disease - therapy
Duchenne and Antisense-Mediated Exon Skipping
1
Duchenne and Antisense-Mediated Exon Skipping
Antisense oligonucleotides (ASOs) are short synthetic pieces of DNA that are chemically
modified to give them druglike properties.[1 ] ASOs can bind to gene transcripts in a sequence specific manner. They have been
employed to reduce the production of toxic proteins or to modulate the splicing process
to restore production of missing proteins. The latter is used in Duchenne muscular
dystrophy (DMD) to restore production of dystrophin.[2 ] DMD is a severe muscle-wasting disease resulting in loss of ambulation before the
age of 12, the need for assisted ventilation by the age of 20, and premature death
in the 2nd to 4th decade of life.[3 ] The disease is caused by mutations (mostly deletions of one or more exons) in the
DMD gene that disrupt the reading frame and thus prevent the production of functional
dystrophin. Splice modulating ASOs can induce exon skipping to restore the reading
frame, allowing the production of shorter, but partially functional dystrophin proteins,
such as those found in the later onset and less severely progressive Becker muscular
dystrophy.[2 ]
The idea for ASO-mediated exon skipping to restore the dystrophin transcript reading
frame was posed over 25 years ago and the approach was pioneered at different locations
around the world. Proof-of-concept studies showing ASO-mediated exon skipping and
reading frame restoration in patient-derived cell cultures and DMD mouse models were
published by groups in Australia, Japan, the Netherlands, and the UK.[4 ]
[5 ]
[6 ]
[7 ]
[8 ]
[9 ] The approach is mutation specific, as different exons have to be targeted based
on the size and location of the deletion.[10 ] However, ASOs inducing the skipping of all internal dystrophin exons were identified.[11 ]
As DMD affects most of the skeletal muscles, systemic treatment is required. It was
shown that ASO uptake by dystrophic muscles is more efficient than by healthy muscles,[12 ] which facilitated systemic treatment. Clinical trials in DMD patients were then
conducted in Japan and Europe and later also in the USA with ASOs targeting exon 44,
45, 51, and 53.[13 ]
[14 ]
[15 ]
[16 ] These revealed that systemic ASO treatment can result in exon skipping and dystrophin
restoration in skeletal muscles. However, efficiency is very low with dystrophin restoration
levels usually being less than 1% after a year of weekly intravenous ASO treatment.
Currently, 4 exon skipping ASOs are approved by the Food and Drug Administration of
the USA (FDA) targeting exon 45 (casimersen), 51 (eteplirsen), and 53 (golodirsen
and viltolarsen).[2 ] One exon skipping ASO is approved by the Japanese Ministry of Health, Labor and
Welfare (viltolarsen). The approvals were only based on dystrophin protein restoration
at low levels (0.4% for eteplirsen, ca. 1% for casimersen and golodirsen, and ca.
5–6% for viltolarsen). Clinical trials to assess whether the low amounts of dystrophin
are sufficient to slow down disease progression are currently ongoing.[2 ]
Opportunities for Treating Central Nervous System Diseases and Developing Individualized
ASOs for Central Nervous System Diseases
2
Opportunities for Treating Central Nervous System Diseases and Developing Individualized
ASOs for Central Nervous System Diseases
From over 2.5 decades of research into dystrophin exon skipping, we can conclude that
delivery of ASOs to skeletal muscles is challenging.[17 ] Delivery to the central nervous system, however, is relatively straightforward using
intrathecal injections.[18 ] After local delivery, ASOs are distributed throughout the central nervous system
and efficiently taken up by most residing cells.[19 ] The advantages of local delivery are that it requires a low dose, so systemic exposure
is low, and that the treatment frequency is low (3–6 times per year). This is exemplified
in nusinersen, an ASO for the treatment of spinal muscular atrophy that is approved
since 2016 in the USA[20 ] and currently marketed in many countries. Intrathecal treatment with nusinersen
in patients with the severe type 1 form of spinal muscular atrophy allows treated
patients to achieve milestones that are by definition not achieved by these individuals.
Furthermore, in a phase 3 clinical trial there was a significant reduction in death
or permanent ventilation for treated patients.[21 ]
As such, splice modulating ASOs offer great potential to treat brain diseases for
eligible genetic conditions.[22 ] The most notable eligible pathogenic variant would be the intronic ‘cryptic splicing
mutation’. This is a variant where a noncoding part of a gene is aberrantly included
into the messenger RNA transcript, thus abolishing protein production. ASOs can prevent
this aberrant inclusion thus restoring normal protein expression (Figure [1 ]).
Figure 1 Cryptic splicing variants cause the inclusion of part of an intron (cryptic exon)
into the messenger RNA, thus abolishing protein production. Antisense oligonucleotides
(ASOs) can target cryptic exons, thus preventing their inclusion into the messenger
RNA during the splicing process and restoring production of a missing protein.
However, these cryptic splicing variants are incredibly rare, often even identified
in single individuals. Therefore, there is no commercial incentive to develop these
ASOs for eligible patients. Tim Yu (Boston Children Hospital) showed it is possible
to develop an individualized ASO for a patient with a cryptic splicing variant with
Batten disease, a severe disease characterized by blindness, frequent epileptic seizures,
and early onset dementia. This ASO was delivered intrathecally and significantly reduced
the number and duration of the seizures. The ASO was called milasen after the patient,
Mila.[23 ] This group later developed an ASO for an ataxia telangiectasia patient with a cryptic
splicing variant and further showed that probably ca. 15% of disease causing variants
in ataxia telangiectasia might be amenable to ASO-induced splicing modulation.[24 ]
Collaborative Spirit to Develop Individualized Treatments Globally
3
Collaborative Spirit to Develop Individualized Treatments Globally
The milasen story inspired many academics with expertise in oligonucleotide therapy
to embark on similar efforts to develop ASOs for very eligible candidates within an
academic setting. In the Netherlands, the Leiden University Medical Center (LUMC),
Radboudumc Nijmegen, and Erasmus Medical Center Rotterdam established the Dutch Center
for RNA Therapeutics.[25 ] This is a collaborative effort of institutes with expertise in ASO development and
treatment. The goal is to develop individualized ASOs for eligible patients with genetic
brain or eye diseases and to provide them to patients without a profit.[26 ] Similar initiatives were set up, such as the N-lorem foundation.[27 ] Furthermore, umbrella initiatives were started to align efforts in Europe (1 Mutation
1 Medicine, 1M1M)[28 ] and globally (N-of-1 collaborative, N1C).[29 ] Recently, also a taskforce on N-of-1 therapy development in general (so not ASO
specific) was launched by the International Rare Disease Research Consortium (IRDIRC).[30 ]
These initiatives are crucial as individualized treatment development is a pioneering
effort, where the traditional drug development pathway does not apply. Most obviously,
the development time needs to be much quicker than the >10 years it generally takes
from target validation to approval, as the individual for whom individualized treatment
is an option, generally cannot wait 10 years. Either the disease by then will have
progressed so much that no treatment benefit is to be expected, or the patient will
have passed away. Still, it is pertinent that the individualized treatments show efficacy
and safety in preclinical studies before treatment of a patient in a clinical setting
is initiated. Another difference is that for individualized therapies, placebo-controlled
clinical trials cannot take place. Instead, the treatment will be provided as an experimental
treatment in a patient care setting. However, monitoring benefit and side effects
will be crucial, so collecting natural history data before the treatment, while the
individualized treatment is in preclinical development, will be imperative to allow
comparison of trajectories before and after treatment on top of comparison of the
individual trajectory of the treated patient with the natural history. Here, outcome
measures and start and stop criteria will have to be discussed by the clinician and
the patient/family.
A roadmap covering all the steps involved in individualized treatment development,
from assessing patient eligibility to clinical implementation is being constructed
by the N-of-1 IRDIRC Taskforce.[31 ] Furthermore, the DCRT has produced educational papers in collaboration with N1C
on which ASO modality apply to which genetic disease,[22 ] including also considerations and caveats, as well as guidance on preclinical assessment
of efficiency for exon skipping ASOs.[32 ]
[33 ] The 1M1M network has produced a guidance document on eligibility criteria from a
genetic, clinical, and ethical perspective.[34 ] Notably, in Europe individualized treatment can be done in a named patient setting
and does not need approval from the European Medicines Agency (EMA), while in the
USA an investigational new drug (IND) application has to be done to FDA even for individualized
treatments. While regulatory approval is not required in Europe, 1M1M and DCRT has
approached EMA for advice and have shared these experiences.[33 ]
4
Global Implementation
Current efforts are ongoing to develop individualized ASOs. In the USA, Tim Yu’s group
has started treatment of 6 patients with 4 different ASOs, 1 in collaboration with
N-lorem and Children’s hospital in Colorado. N-lorem has initiated treatment of 5
patients, 2 are pending institutional review boards (IRB) approval, and 2 INDs are
pending (presented at N-lorem colloquium Oct 12, 2023). In Europe, ASOs are in preclinical
development. One individual with AT is being treated with atipeksen. Notably, after
cross referencing ASO amenable AT variants the PIs involved in the N-of-1 collaborative,
it became clear that an AT patient in Germany carried the same variant as the individual
already treated by Tim Yu with atipeksen. The patient moved to Boston to initiate
treatment there and is currently being treated and monitored in Germany. Additional
individuals with the same pathogenic variant have been identified in Turkey.
This example highlights several aspects. First, that ASOs developed for what is thought
to be a unique patient, can turn out to apply to additional individuals. Furthermore,
due to different regulations in different jurisdictions, it is not always straightforward
to initiate treatment of a second patient with such an ASO. Especially in Europe,
routes for clinical implementation within a named patient setting vary per country
and sometimes even per region or hospital. Ideally, in the future treating patients
in an “N-of-1-at-a-time” fashion is streamlined better to facilitate quick implementation
of treating additional patients. Notably, treating a second patient with an ASO reduces
the uncertainty with regards to safety and efficacy, as some extrapolation can be
made from the first treated individual.
With the rapid development of N-of-1 ASOs, many patients will be the first to be treated
with a given ASO. While these are tested in vitro to confirm efficacy and in vivo
to confirm safety, there may be unexpected side effects. Valeriasen was developed
for a patient with a severe infantile onset epilepsy syndrome caused by variants in
KCNT1 .[35 ] This RNase H activating ASO was effective in reducing seizure frequency in a first
patient, Valeria, as well as a second patient with the same disease for whom treatment
was initiated later. However, it became apparent that the ASO also induced hydrocephalus,
leading to a decision pursue hospice for the first patient and placement of a shunt
for the second patient. Shunt placement relieved the hydrocephalus, but after pausing
the ASO treatment, the epileptic seizures returned in this individual. None of the
safety studies hinted at the toxicity. Note that safety studies are conducted in healthy
rats, and it is conceivable that an interaction between the ASOs in a pathological
brain of the patient induced the hydrocephalus. Indeed, it has also been reported
in several nusinersen-treated patients and increased intracranial pressure and enlarged
ventricles were reported in a clinical trial for Huntington disease patients treated
with tominersen.[36 ]
[37 ] Now that hydrocephalus has been reported for valeriasen in individuals with KCNT1-variant-induced
pathology, reverse translation experiments have been initiated to develop preclinical
safety models that can predict it, so that this information can be used to avoid this
toxicity for future ASOs.
A challenge for the N-of-1 treatment development is the lack of infrastructure at
several fronts. First, at diagnosis, it is generally not flagged that a pathogenic
variant might be treatable by ASOs, and some treatable variants such as cryptic splicing
variants, are not routinely screened for. Secondly, processes are needed to ensure
eligible patients are identified justly and ethically, based on eligibility. Thirdly,
processes for ASO development and efficacy and safety testing can be further optimized
and streamlined. Fourthly, to measure clinical benefit, a toolkit of patient-relevant
outcome measures is needed to allow individualized measurements of treatment effects.
Fifthly, ideally, catalogues of ASOs are shared – both those that are safe and effective,
and those that are not safe or not effective. Sixth, processes to facilitate and streamline
clinical implementation within a hospital setting are required. Currently, the administrative
burden and the clinical monitoring comes down mainly on the clinician who will treat
the patient. Within each center, getting approval to start treatment of the first
individual will be the most challenging, as many IRBs are not familiar with the N-of-1
treatment setting. Regardless, support for the clinicians for the time invested and
the administrative burden would lighten the workload. The N-of-1 collaborative is
working on providing guidance on development and implementation of this missing infrastructure.
Finally, a model to cover the costs of development of individualized ASOs is required.
N-lorem currently develops ASOs for free and provides them for free for life. However,
this does not cover the costs made by the hospitals and families to allow treatment
and management. In other areas, development is currently funded via crowd funding,
funding from the government or by individual institutes. While existing efforts is
sufficient to provide proof-of-concept for a few patients, there has yet to emerge
a clearly sustainable or scalable solution. The individualized ASO development requires
innovation and pioneering of the entire drug development and reimbursement model and
the IRDRIC N-of-1 taskforce will discuss possible solutions.
5
Concluding Remarks
There is a long way to go to make individualized ASOs and individualized treatments
a reality around the world, and likely the process will not be identical in different
regulatory jurisdictions and for different therapeutic modalities. It is clear this
can only be achieved through sharing knowledge and expertise, successes and failures,
and only through the combined effort of academics, patients, and regulators can we
make this a reality.
