Case study
Prof G, have you seen today’s news that MS progression may be due to genetic factors? What does it mean for someone like me who was recently diagnosed with secondary progressive MS? I am about to transition from dimethyl fumarate to siponimod. I am waiting for my genetic tests to return to see if I can take the drug. Does this mean siponimod won’t work in slowing down my MS?
Prof G’s opinion
The observation that subtle genetic variations in your genome are associated with MS severity is not surprising. You must remember that humans, like all forms of life, are biological machines that run off genetic algorithms encoded in our genomes. Everyone has subtle differences in these programmes due to changes or mutations across the genome when DNA is copied and mixed up when sperm and ova are formed. As you inherit half your DNA from each parent, more complexity and mixing are added. On top of the hardwired genetic code, epigenetic changes affect your biological algorithms by determining how they are augmented or suppressed.
This study (paper 1) shows that pwMS who inherit one or two, or possibly more variants, will likely get worse more quickly than people who don’t inherit these variants. The fact that longevity and many neurodegenerative diseases are affected by heritable traits put these findings into perspective.
These MS variants are called SNPs (single nucleotide polymorphisms) and represent a single change in the DNA code. It is not necessarily the SNPs themselves which are responsible for the more rapid worsening of MS, but the genes that are close to these variants in your DNA. These genes can be upstream or downstream of the identified SNPs.
Gene finding
What is interesting is what these investigators did next. They did a gene prioritisation analysis and identified four biologically plausible genes close to the two SNPs or loci. One of the genes I am very excited about is ZNF638, which encodes a protein that binds to DNA. The protein is from a family of proteins called DNA-binding zinc-finger proteins.
ZNF638 suppresses the reading of retroviral DNA in the genome via chromatin repressors involved in the epigenetic silencing of human endogenous retroviruses (HERVs). This would be like partitioning a computer’s hard drive and putting one partition in a secure vault that is password protected. This partition prevents any software program without the necessary permissions from reading data off that hard drive section. Why this is important is that HERVs and their expression are linked to MS and possibly smouldering MS. So pwMS who have a genetic variant that can suppress HERVs more efficiently may do better than people with a genetic variant that is less effective.
Importantly, ZNF638 is expressed in the brain, particularly in oligodendrocytes and their precursor cells, so this variant could be linked to myelination or remyelination biology. ZNF638 is implicated in general intelligence and cognitive ability, which may be relevant to its role in MS. ZNF638 is preferentially found in oligodendrocyte clusters predicted to be actively myelinating. And cell expression of ZNF638 is proportionally enriched in control brain tissue and chronic inactive MS lesions compared to other MS lesions. Therefore, ZNF638 is a good gene candidate. However, more work will need to be done to determine whether it is involved in MS prognosis.
Another potential gene is DYSF which encodes dysferlin, a transmembrane protein. In skeletal muscle, dysferlin participates in membrane repair and regeneration. Rarer pathogenic variants have been found to cause muscular dystrophies. Dysferlin is also found in oligodendrocytes and neurons, and dysferlin accumulates in plaques in Alzheimer's disease in proportion to severity. It has been proposed that dysferlin maintains neuronal or glial membranes, which could influence neuroaxonal survival or remyelination in pwMS.
The other possible variant or locus is found in the DNM3 gene, which encodes dynamin-3 and helps with synaptic vesicles' endocytosis. Expression of dynamin-3 is preferentially in the oligodendrocyte lineage and neurons. A closely-related protein called dynamin-2 is also involved in maintaining skeletal muscle membranes, which may point to a convergence of mechanisms with DYSF. This other variant is also within a gene PIGC that encodes phosphatidylinositol glycan C. Mutations in PIGC are associated with epilepsy and intellectual disability.
This work has opened up several new lines of research that will hopefully lead to new therapeutic targets to tackle the major unmet need in MS, smouldering-associated worsening (SAW) independent of focal inflammatory disease activity.
How strong is the link between these variants and worsening MS?
Both severity variants were shown to have a clinically meaningful association with time to EDSS 6.0 (needing a walking aid). People who carried two copies of the variant linked to ZNF638 and the dysferlin genes reached EDSS 6.0 (walking stick) 3.7 years earlier than those who did not inherit the variants. For the dynamin-3 and PIGC variants, this figure was for 3.3 years. The magnitude of these effects is relatively small and matches the treatment effect of low-efficacy DMTs such as interferon-beta on disability progression. Another comparator is vascular comorbidities. PwMS with one or more vascular comorbidities (hypertension, diabetes, smoking, hypercholesterolemia, dyslipidemia, or documented vascular disease) reach EDSS 6.0 approximately six years earlier than pwMS without vascular comorbidities. So the effect of these variants is relatively small.
The authors state that at least 13% of the so-called variance in long-term MS severity can be attributed to common and low-frequency SNPs, which explains some of the variability in MS outcomes. Importantly, the genes linked to this heritability are enriched or expressed in the brain and spinal cord, which differs from the immune signals for MS susceptibility. The divergence of the two MS-associated heritability signals is congruent with our understanding of MS; the cause of MS is linked to inflammation and immune cells and disability or severity to the resilience of the nervous system.
What do these findings mean for people with MS?
At the moment, not much. In the future, these variants may be incorporated into prognostic risk scores. However, as the effects of these putative genes are downstream of damage, if you prevent the damage from occurring in the first place with early effective therapy, then having them or not may make no difference.
I suspect a flurry of research examining these variants' functional effects in cell cultures and humanised knock-in animal models. MRI and PET imagers will dichotomise their patients into groups to see if these variants are associated with accelerated brain volume loss, slowly expanding lesions, paramagnetic rims lesions, TSPO-hot lesions on PET, etc. The soluble biomarker brigade will want to see if the variants are linked to raised neurofilament levels and other damage biomarkers (GFAP, …). People working on the eye in MS will explore whether these variants predict retinal nerve fibre loss on OCT and whether carriers have accelerated retinal nerve fibre loss over time. Researchers with large datasets will look to see if these variants are associated with SAW or PIRA (progression independent of relapse activity).
Pharmaceutical companies with neuroscience programmes will undoubtedly put teams of people onto the genes and pathways identified by this study to see if they can find druggable targets to augment or inhibit them. A quick screen of the drug-gene-interaction database only identified DYSF as potentially druggable as it is a transporter.
To develop drugs to target dysferlin and other potential targets in the pathways involved with these candidate genes, we will need cell culture and/or animal models to screen and test drugs. If these pathways prove ‘druggable’, candidate molecules must be optimised. For a molecule to become a candidate drug for clinical development requires the molecules to be able to be taken orally, have a relatively long half-life in the body, be able to cross the blood-brain barrier, not be associated with any liver or cardiac toxicity, to ideally not interact with other drugs, and to clear standard toxicology testing. It is often stated that for every 10,000 potential drug candidates, only one molecule clears these hurdles.
Pharma will then need to do phase 1 first in human studies, which has two phases. A single-ascending dose (SAD) phase, which is then followed by a multiple-ascending dose (MAD) phase. The SAD and MAD phases are usually done in healthy male volunteers. Once the drug(s) are safe in healthy volunteers, they may go into people with MS as part of a phase 1b (safety) study or directly into a phase 2 dose-finding and proof of concept trial. There needs to be some objective readout using one or more of the biomarkers mentioned above to get proof of biology. Once the company is confident of getting a potential result, they typically launch two or more phase 3 registration trials. Before starting the studies, the trial designs must be discussed with the regulatory authorities.
The point of telling you all this is that from the time a potential treatment target is identified, for example, dysferlin, to a licensed product takes about 15 years, with many potential pitfalls on the way. This time lag is why I don’t think these discoveries will make much difference to pwMS in the short term.
Does this mean siponimod won’t work in slowing down my MS?
When I searched the drug-gene-interaction database, none of the identified genes (DNM3, PIGC, DYSF, ZNF638) appeared to interact with siponimod. The only genes that came up for siponimod were S1PR1, S1PR5 and CYP2C9, which we already know about. S1PR1 and S1PR5 are the genes siponimod binds to, and CYP2C9 is the metabolising enzyme in the liver, which has two variants, i.e. a slow and fast metabolising variant. This is the genetic test you are waiting for; if you have two copies of the slow-metabolising variant, you will be unable to take siponimod; if you have one copy of each, you will be prescribed low-dose siponimod and if you have two copies of the fast-metabolising variants you will be prescribed high-dose siponimod.
However, it may be possible to analyse the trial cohorts who received siponimod or placebo to see if these disease outcome variants predict treatment response to siponimod.
Reasoning by analogy
Please don’t get disheartened by this study's findings. Even if one or two of the identified genes are validated in other studies and contribute to a worse outcome, you may be able to nullify their effects with DMTs and/or lifestyle changes—an example of this is type 2 diabetes mellitus (T2DM). T2DM has a large genetic component; despite this, a new study (paper 2 below) shows moderate-to-vigorous-intensity physical activity is very effective in preventing disease onset. Yes, exercise trumps genetics. Similarly, I would not be surprised if the same thing happens in MS. This is why you need to adopt my marginal gains philosophy and try and optimise all the small things going into improving MS outcomes. This includes your brain, metabolic and social health.
I would be interested to hear if you understand this Newsletter and how you feel about potentially important research findings being discussed on this platform.
Paper 1 - genetics
Multiple sclerosis (MS) is an autoimmune disease of the central nervous system (CNS) that results in significant neurodegeneration in the majority of those affected and is a common cause of chronic neurological disability in young adults1,2. Here, to provide insight into the potential mechanisms involved in progression, we conducted a genome-wide association study of the age-related MS severity score in 12,584 cases and replicated our findings in a further 9,805 cases. We identified a significant association with rs10191329 in the DYSF–ZNF638 locus, the risk allele of which is associated with a shortening in the median time to requiring a walking aid of a median of 3.7 years in homozygous carriers and with increased brainstem and cortical pathology in brain tissue. We also identified suggestive association with rs149097173 in the DNM3–PIGC locus and significant heritability enrichment in CNS tissues. Mendelian randomization analyses suggested a potential protective role for higher educational attainment. In contrast to immune-driven susceptibility3, these findings suggest a key role for CNS resilience and potentially neurocognitive reserve in determining outcome in MS.
Paper 2 - exercise
Objective: Although 30 min/day of moderate-intensity physical activity is suggested for preventing type 2 diabetes (T2D), the current recommendations exclusively rely on self-reports and rarely consider the genetic risk. We examined the prospective dose-response relationships between total/intensity-specific physical activity and incident T2D accounting for and stratified by different levels of genetic risk.
Methods: This prospective cohort study was based on 59 325 participants in the UK Biobank (mean age=61.1 years in 2013-2015). Total/intensity-specific physical activity was collected using accelerometers and linked to national registries until 30 September 2021. We examined the shape of the dose-response association between physical activity and T2D incidence using restricted cubic splines adjusted for and stratified by a polygenic risk score (based on 424 selected single nucleotide polymorphisms) using Cox proportional hazards models.
Results: During a median follow-up of 6.8 years, there was a strong linear dose-response association between moderate-to-vigorous-intensity physical activity (MVPA) and incident T2D, even after adjusting for genetic risk. Compared with the least active participants, the HRs (95% CI) for higher levels of MVPA were: 0.63 (0.53 to 0.75) for 5.3-25.9 min/day, 0.41 (0.34 to 0.51) for 26.0-68.4 min/day and 0.26 (0.18 to 0.38) for >68.4 min/day. While no significant multiplicative interaction between physical activity measures and genetic risk was found, we found a significant additive interaction between MVPA and genetic risk score, suggesting larger absolute risk differences by MVPA levels among those with higher genetic risk.
Conclusion: Participation in physical activity, particularly MVPA, should be promoted especially in those with high genetic risk of T2D. There may be no minimal or maximal threshold for the benefits. This finding can inform future guidelines development and interventions to prevent T2D.
Subscriptions and donations
Paid subscriptions to MS-Selfie are being used to administer the Newsletter and associated MS-Selfie microsite, which is now open to all readers. At the request of several readers, I have now added the option of making a one-off donation. To keep this initiative open to all readers, I would appreciate it if those who can afford a subscription to subscribe. For active paying subscribers, thank you; your contribution is much appreciated. Because of the falloff in paying subscribers, I am considering returning to a paywall that will give paying subscribers six months of unlimited access to all newsletters. At the same time, free subscribers will have a wait to access the newsletters later, not as an email, but on the substack site.
General Disclaimer: Please note that the opinions expressed here are those of Professor Giovannoni and do not necessarily reflect the positions of Barts and The London School of Medicine and Dentistry nor Barts Health NHS Trust. The advice is intended as general and should not be interpreted as personal clinical advice. If you have problems, please tell your own healthcare professional, who will be able to help you.
CNS resilience in progression of MS