Question
Prof G, could you do an article on what is going on with tolebrutinib in the States? Having a hard time processing the FDA rejection letter. Not sure what this means for the field.
NOTE: General Substack newsletters and the microsite are free; only Q&A sessions are restricted to paying subscribers. I can’t run and maintain the MS-Selfie microsite, so I must pay people to assist me. If people want to ask medical questions unrelated to the Newsletters or Podcasts, they either need to become paying subscribers or email (ms-selfie@giovannoni.net) to request a complimentary subscription.
Prof G’s response
You can download the FDA response here.
This is the informing Genzyme Corporation (Sanofi) that its application for the tolebrutinib cannot be approved in its current form. The primary obstacle to authorisation is a significant risk of severe liver injury, which the FDA believes outweighs the potential clinical benefits.
Despite Sanofi’s attempts to manage this toxicity through intensive monitoring, the FDA identified several “life-threatening cases” during clinical trials. Additionally, the FDA noted substantial uncertainty regarding the drug’s effectiveness across multiple sclerosis subpopulations, particularly in those with inactive disease. To move forward, Sanofi must provide more robust safety data and potentially identify a specific patient group or groups where the therapeutic advantages clearly justify the high risks.
The most critical factor in the FDA’s decision was the risk of severe drug-induced liver injury (DILI). The FDA characterised this risk as “substantial and unusually high” for drug development programmes in general, and specifically for MS therapies. In the Phase 3 development programme, there were 6 cases meeting Hy’s Law criteria (indicators of severe hepatotoxicity) out of approximately 2,700 subjects, including one patient who required a liver transplant and subsequently died.
The FDA noted that even a single Hy’s Law case is a signal of high hepatotoxicity, and most drugs withdrawn from the market for this reason have rates of severe injury far lower than what was observed with tolebrutinib. While DILI is a known class effect of BTK inhibitors, tolebrutinib’s risk of fatal DILI appears to be “among the highest in the class” and exceeds the risk observed with all other approved MS therapies.
The FDA concluded that the proposed Risk Evaluation and Mitigation Strategy (REMS) was insufficient. Although Sanofi implemented weekly liver monitoring, severe liver injury cases continued to occur even under this strict protocol. Consequently, the Agency determined that the benefit-risk assessment must assume severe and potentially fatal DILI cases would occur in a postmarketing setting regardless of monitoring requirements.
The FDA identified significant issues with the efficacy data supporting the proposed indication for non-relapsing secondary progressive MS (nrSPMS). The study population was heterogeneous, containing both “active” and “non-active” SPMS patients, which complicated the assessment of benefit. The treatment effect was largely driven by a small subgroup (13% of enrolled subjects) with baseline gadolinium-enhancing lesions, indicating active SPMS. While the drug showed efficacy here, there are already approved therapies for active SPMS that do not carry the same magnitude of liver risk. Therefore, the benefits in this group were not anticipated to outweigh the severe DILI risk.
For patients without baseline enhancing lesions—the population with the greatest unmet need—the treatment effect was significantly smaller and uncertain. The FDA stated that therapies with greater risks require a “greater magnitude and certainty of benefit” to support approval. The potential benefit in this population was deemed inadequate to justify the severe liver risks.
Their analysis of the data showed that the treatment effect was substantially diminished in patients who had previously tried two or more prior MS therapies. The FDA rejected Sanofi’s claim regarding “slowing disability accumulation independent of relapse activity”. The Agency cited several reasons for this rejection. The concept lacks widely accepted criteria and relies on post hoc analyses. There is limited understanding of whether BTK inhibitors address the underlying pathophysiology of progression. They also highlighted a separate study in primary progressive MS that has recently failed to demonstrate a benefit on its primary endpoint of confirmed disability progression, further undermining the mechanistic rationale for this claim (see Tolebrutinib PPMS trial is negative, 15-Dec-2025). I would counter the latter argument that the PPM population in the Perseus trial differs significantly from the nrSPMS population in the Hercules trial. Ultimately, the FDA could not identify a clinically identifiable population for whom the benefits of tolebrutinib outweighed the “serious and unusually high risk of severe DILI”.
My initial thoughts
This is not the first or last time this has happened with MS DMTs and the FDA. They did the same with natalizumab (Tysabri), alemtuzumab (Lemtrada), and cladribine (Mavenclad) based on PML, secondary autoimmunity & other adverse events (infusion reactions, infections, …) and cancer risk with each DMT, respectively. All these drugs eventually got to market. I am confident there is a way forward.
Defining a more responsive subgroup
The FDA is clearly worried about giving tolebrutinib a wide label for smouldering MS. This is because when you look for it using sensitive measures, for example, cognitive and other neurological stress tests, the majority of people with MS have smouldering disease. This means that the use of tolebrutinib would likely spread like wildfire to most subgroups of MS, and hence, the risk of life-threatening liver toxicity would increase. The FDA determined that, in its present form, a favourable benefit-risk profile could not be established for any patient subpopulation. The challenge for Sanofi is to identify a subgroup that would benefit most from the drug.
Interestingly, in a post-hoc analysis, subjects with early-relapsing MS who had paramagnetic rim lesions (PRLs) responded better to tolebrutinib than those without PRLs. We have yet to see the same data for the PPMS cohort. But if PRLs serve as a prognostic marker in this population, it would make a compelling case for using this biomarker.
Please see the MS-Selfie newsletter on ACTRIMS 2025 Highlights (3-Mar-2026), which covers this analysis.
As you are aware, PRLs are one biomarker of smouldering MS lesions, and their presence is associated with poorer outcomes. PRLs are considered a “red flag” in MS because they represent chronic, smouldering inflammation that actively destroys brain tissue long after the initial attack has passed. While a typical MS lesion stabilises and “heals”, PRLs behave like a slow-burning fire that continues to expand outward, causing progressive damage.
The defining feature of a PRL is a dark ring seen on specific MRI scans (susceptibility-weighted imaging). This ring corresponds to a dense accumulation of iron-laden microglia at the edge of the lesion. Instead of clearing debris and promoting repair, these immune cells remain chronically activated. Iron accumulation causes oxidative stress, which is toxic to neurons and oligodendrocytes (myelin-producing cells). This “toxic rim” slowly eats away at the healthy tissue surrounding the lesion. A proportion of PRLs are also slowly expanding lesions (SELs).
Because these lesions do not shut down as standard lesions do, they are a primary driver of neurodegeneration (brain shrinkage). The ongoing inflammation at the rim severs nerve fibres or axons. Once an axon is lost, it generally cannot be replaced. Patients with multiple PRLs typically show accelerated brain atrophy (shrinkage), particularly in the deep grey matter, which is strongly correlated with physical and cognitive disability.
PRLs are a major biological cause of smouldering MS and PIRA (Progression Independent of Relapse Activity). Traditionally, disability was thought to accumulate via acute relapses (attacks). PRLs cause disability to worsen between attacks. A person with MS may not have a new relapse for years, yet their walking or cognition slowly deteriorates because these smouldering lesions are quietly destroying tissue in the background.
PRLs are associated with poor outcomes because they are difficult to treat. The inflammation in a PRL is trapped behind the blood-brain barrier (BBB). Most standard DMTs work by stopping immune cells from entering the brain from the blood. Since the immune cells in a PRL are already inside the brain and self-perpetuating, these drugs are less effective at stopping smouldering inflammation. Tolebrutinib is the first DMT to show an effect in smouldering MS, presumably by targeting PRLs and other intrathecal processes driving smouldering MS, which is why this FDA response is so devastating for people with MS.
The FDA are also behind the curve when it comes to smouldering MS and are questing whether or not it is a real entity. I suggest they ask people with MS what smouldering MS is. If they read some of the comments on MS-Selfie, they would recognise that this is clearly the unmet need in treating MS and most people with MS, particularly those with advanced MS who identify as having smouldering MS. Clearly, we need a public engagement campaign to highlight what a problem smouldering MS is for people with MS. The FDA does listen to the patient community. I am told it was patient testimonies that won the day for getting natalizumab licensed. Any ideas in this regard are welcome?
Could the FDA license tolebrutinib for the subgroup of people with MS with smouldering MS and PRLs?
I suspect not. In clinical practice today, measuring, identifying and quantifying paramagnetic rim lesions (PRLs) is moderately difficult. It is not yet a “push-button” result like measuring blood sugar, but it is becoming increasingly feasible in most MS centres with the right equipment and protocols. You cannot see paramagnetic rims on a standard “routine” MRI sequence (like T1 or T2/FLAIR) used for general MS monitoring. To see the iron rim, the radiologist must run a specific sequence called SWI (Susceptibility-Weighted Imaging) or T2*-weighted imaging. Most modern MRI scanners (specifically 3 Tesla machines found in major hospitals) can run this sequence, but they often do not turn it on automatically because it adds an extra 4–5 minutes to the scan time. Time is money in the neuroradiology unit. It is significantly easier to spot these rims on a powerful 3T MRI scanner. On older, standard 1.5T scanners, the rims can be faint or indistinguishable from noise.
Once the image is taken, a human (a radiologist) must find the rims. This is currently the biggest bottleneck. The “rim” is sometimes a subtle shadow rather than a distinct black ring. Distinguishing a true iron rim from a vein or a normal lesion edge requires a trained eye. In a patient with 50+ MS lesions, a neuroradiologist must manually check each lesion on the SWI scan to determine whether it has a rim. In a busy clinical workflow, they may simply report “multiple lesions” without characterising the rims unless specifically asked to do so. Let’s hope AI changes this. Until recently, there was no universal rule for “how thick” or “how dark” the rim must be to count. New guidelines (such as those from the NAIMS cooperative) are addressing this, but not all local neuroradiologists have yet adopted them.
While saying “Patient X has 3 rim lesions” is doable, measuring exactly how much those lesions have expanded (e.g., “The lesion grew 1.2mm this year”) is extremely difficult in routine practice. Most hospital software is designed to show new bright spots, not the subtle expansion of old dark spots. Research centres use AI algorithms (like “RimNet”) to automatically detect and measure these, but these tools are rarely integrated into the standard computer systems neuroradiologists use daily.
Would you be interested in knowing your PRL status? I suspect not unless something can be done about it, for example, switching from your current DMT to tolebrutinib.
Because of these issues, I would be surprised if the FDA licenses tolebrutinib for patients with baseline PRLs. I hope I am wrong.
Can we derisk the liver injury?
Based on the FDA response, the short answer is likely no—at least not effectively enough to satisfy the FDA. While we attempted to “derisk” tolebrutinib during clinical trials using strict exclusion criteria and frequent monitoring, the FDA has concluded that these mitigation strategies are insufficient to prevent severe liver injury. The primary hurdle is that tolebrutinib-induced liver injury appears to be idiosyncratic, meaning it is unpredictable. It is not strictly dose-dependent (meaning you can’t just lower the dose to fix it) and does not have clear biomarkers that predict who will get it before they start. It typically occurs within the first 90 days of treatment. In the clinical trials, the rate of severe liver injury (Hy’s Law cases) was significantly higher than in other MS therapies. There was one fatality following a liver transplant, which is a massive red flag for FDA approval.
What “derisking” strategies failed?
Sanofi implemented aggressive safety protocols during the Phase 3 trials after the initial safety signals appeared (leading to the 2022 clinical hold). These included excluding anyone with pre-existing liver risk factors or complications. Increasing liver function testing (LFT) frequency to weekly or bi-weekly during the first 3 months.
According to the FDA, even with these strict protocols, new cases of severe liver injury continued to occur. The FDA noted that while monitoring can catch injuries earlier, injuries can progress so rapidly that monitoring alone does not guarantee safety.
Very worrying for the field is that the FDA states that DILI is a known “class effect” of BTK inhibitors (the family of drugs tolebrutinib belongs to). However, the FDA noted that tolebrutinib’s risk profile appeared “unusually high” compared to others in the class and existing MS treatments. The question we all have is, will this affect the chances of fenebrutinib, remibrutinib and other follow-on BTKi? I suspect yes.
Other MS drugs (like alemtuzumab or natalizumab) carry high risks (autoimmunity, PML), and these risks can be effectively managed with REMS (Risk Evaluation and Mitigation Strategies) programs. I am not sure why the FDA thinks DILI from tolebrutinib is any different. I envisage a future in which people with MS who start tolebrutinib will have home liver function tests that they do daily for the first three months, so a rise in LFTs can be detected very early and addressed. This technology already exists in the form of lateral flow tests. I can’t see why the FDA can’t see beyond the trial protocol.
Paternalistic Medicine
The question I have asked in the past and do so again: “Who is taking the risk when someone with MS decides to start a treatment with a known severe adverse event profile?” The regulator, the healthcare system, the prescribing physician or the patient themselves? At the end of the day, if we are transparent, surely it is the patient who makes the decision, with support from their team and the healthcare system. Provided we have a responsive system to manage DILI, I see no difference between tolebrutinib’s DILI risk and the PML risk with natalizumab, or the secondary autoimmune risk with alemtuzumab.
I wish the FDA would come and sit in a clinic with me to see how common smouldering MS is and how large the unmet need is for treating smouldering MS. I am aware this news is devastating for people with MS living in the US. Interestingly, tolebrutinib is already licensed in the United Arab Emirates, and hopefully, the EMA and MHRA will take a different approach to the FDA. I also have concerns for fenebrutinib and the other BTKi.
As a person with MS, are you not tired of regulators, payers, healthcare providers and HCPs deciding on what is too risky or not for you? I think the MS community should push back on this decision and find a path forward to get tolebrutinib to market for people with smouldering MS. Do you agree?
Abstract of interest
Background: Tolebrutinib is a brain-penetrant Bruton’s tyrosine kinase inhibitor (BTKi) that, in phase 3 pivotal trials, led to a 31% and 29% reduction in disability accumulation relative to placebo and teriflunomide in nrSPMS and RMS, respectively. MRI paramagnetic rim lesions (PRLs) are chronic active lesions associated with inflammation, demyelination, and axonal transection in the white matter and are correlated with disability accumulation and resistant to currently approved therapies. Because BTKi can modulate neuroinflammation driven by disease-associated microglia and B cells, we asked, in a post-hoc analysis, whether PRLs were associated with response to tolebrutinib.
Objectives: To evaluate PRLs observed at baseline as a prognostic and predictive biomarker for disability accumulation and treatment response in HERCULES and GEMINI.
Methods: HERCULES (NCT04411641), GEMINI 1 (NCT04410978), and GEMINI 2 (NCT04410991) were phase 3, double-blind trials of 60 mg tolebrutinib once daily. In HERCULES, participants were randomized 2:1 to receive tolebrutinib or placebo. In GEMINI, participants were randomized 1:1 to receive tolebrutinib or teriflunomide (14 mg once daily), each with matching placebo. 437 (39%) of the 1131 HERCULES participants and 631 (34%) of the 1873 GEMINI participants were from sites with imaging capabilities allowing evaluation of PRLs. Following prior studies, we analyzed the effect of tolebrutinib on time to onset of 6-month confirmed disability worsening (6-mo CDW) in participants with 0, 1-3, or ≥4 PRLs at baseline. PRLs were manually identified on SWI images generated from 3D-gradient echo phase images (6 echoes ranging from 4.9 to 41 ms, 0.8-mm isotropic resolution).
Results: Across both trials, 653 participants (61%) had PRLs, consistent with data from observational studies with high-sensitivity MRI sequences. In HERCULES, the proportion of participants with 0, 1-3, or ≥4 PRLs at baseline was 40%, 36%, and 24%, respectively, with similar proportions in GEMINI. In both HERCULES and GEMINI, the risk of 6-mo CDW increased as a function of baseline PRLs in the placebo and teriflunomide comparator groups, respectively. In HERCULES, tolebrutinib appeared to mitigate the risk of 6-mo CDW with greater effect in participants with more baseline PRLs, reducing the risk by 54% in participants with ≥4 PRLs. In GEMINI, a similar risk mitigation was observed with a 46% and 49% risk reduction in participants with 1-3 and ≥4 PRLs, respectively. In tolebrutinib-treated participants with PRLs in both HERCULES and GEMINI, the risk of 6-mo CDW was numerically similar to the corresponding risk in participants without PRLs.
Conclusions: This post-hoc analysis suggests that the impact of tolebrutinib treatment may be greater in those with higher number of PRLs, consistent with the CNS bioactive mechanism of action of tolebrutinib.
Specific conflict of interest
I sit on the tolebrutinib phase 3 trials steering committee. Therefore, you may want to take anything I say on this topic with a pinch of salt.
Accidental readers
If you have been forwarded this email and are not an MS-Selfie subscriber, please consider subscribing and helping MS-Selfie expand its resources for the broader MS community. MS-Selfie relies on subscriptions to fund its curated MS-Selfie microsite, MS-Selfie books, MS-Selfie Infocards, and other activities that extend beyond the MS-Selfie Substack newsletters.
Subscriptions and donations
MS-Selfie newsletters and access to the MS-Selfie microsite are free. In comparison, off-topic Q&A sessions are restricted to paying subscribers. Subscriptions are being used to run and maintain the MS Selfie microsite and other related activities, as I don’t have time to do this myself. You must be a paying subscriber to ask questions unrelated to the newsletters or podcasts. If you can’t afford to become a paying subscriber, please email a request for a complimentary subscription (ms-selfie@giovannoni.net).
Questions
If you have questions unrelated to the newsletters or podcasts, please email them to ms-selfie@giovannoni.net. Prof. G will try to answer them as quickly as possible.
Important Links
🖋 Medium
General Disclaimer
Please note that the opinions expressed here are those of Professor Giovannoni and do not necessarily reflect the positions of Queen Mary University of London or 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 healthcare professional, who will be able to help you.
