MS DMTs that hit the regenerative note

Santiago Ramon y Cajal

Who said that the brain doesn’t regenerate? Well, it was the neuroscience Dogma as far back as philosophy. But, not in certain quarters; Santiago Ramon y Cajal (1852-1934), widely believed to be the father of modern neuroscience in his treatise ‘Degeneration and Regeneration of the Nervous System’, alludes to the potential of the brain to regenerate in the ‘general theoretical interpretation of the phenomena of nervous regeneration’.

He noted two important things: 1) Law of continuous growth or the tendency of the axon to grow constantly in a longitudinal direction, and 2) The longitudinal growth of nerve fibres is especially localized in their free ends, where the cone of growth is situated, a constant organ of the axon in embryos as well as in nerves undergoing pathological regeneration. In short, the terminal axon/neuron is differentiated in a special way to facilitate the growth, create exploratory sprouts and orient the axon suitably in space.

What he didn’t know at the time was that situated within and near the growth cone are special proteins that elongate and orient the axon – unimaginatively termed growth cone proteins!

Neuronal growth cone proteins

As part of my side job, I’ve been looking at these in MS for the past 10 years and am now writing down here some of my observations.

The protein that drew my interest was NCAM (or neural cell adhesion molecule). It is found mainly in the nervous system, normally in the transmembrane or attached to the surface of the axon. Neuroscientists have been researching NCAM for decades. An MS researcher, Angelo Massaro, felt that NCAM had an important role to play in CNS remyelination. He even found that levels of NCAM were boosted in the CSF of MS patients receiving corticosteroids during a relapse!

Using CSF analysis, I found that NCAM was also less abundant in neurodegenerative conditions, such as Alzheimer’s disease, Parkinson’s disease, motorneurone disease. Levels by comparison were much higher in those without a neurological disorder. This led me to postulate that the soluble version of NCAM found in the CSF must be acting as trophic factor for axons/neurons creating environment favorable for growth/plasticity.

What surprised me was that levels of NCAM in the CSF dropped serially from CIS (first presentation), through RRMS, and into SPMS. Since, I published my observations, others have also found the same though not in all instances (different primary antibodies in their assays)! But, what has been annoying me to date like a festinating thorn in a hard to reach place, was whether you could boost CSF NCAM levels through treatment? Trying to find collaborators from pharma who were willing to test their drugs on this proved impossible. That was not until Prof Jan Lycke’s group from Sweden suggested that they have serial CSF’s before and after natalizumab, fingolimod and mitoxantrone treatments an year apart, and were willing to collaborate.

The samples were analyzed blinded (i.e. not knowing the timing of the samples or their treatments) and you can see the remarkable findings below in the figures.

Natalizumab showing a rise in CSF NCAM levels in majority
Mitoxantrone showing a rise in CSF NCAM levels in majority
Fingolimod failing to show a rise in CSF NCAM levels in majority

I will leave you to interpret these findings on your own…

Acta Neurol Scand. 2019 Jan 18. doi: 10.1111/ane.13069. [Epub ahead of print]

Cerebrospinal fluid NCAM levels are modulated by disease modifying therapies.

Axelsson M, Dubuisson N, Novakova-Nyren L, Malmeström C, Giovannoni G, Lycke J, Gnanapavan S

Little is known about what leads to recovery between relapses in multiple sclerosis (MS), particularly following treatment. In the past, it has been demonstrated thatsoluble neural cell adhesion molecule (sNCAM), a putative biomarker of neuroplasticity, increased following steroid treatment in the CSF of MS subjects undergoing acute relapses. Taking this a step further, we have evaluated the effect of disease-modifying treatment (DMTs) on CSF sNCAM levels in various subtypes of MS.

We measured CSF sNCAM levels at baseline and after 12-24 months of DMT in 69 patients, 49 relapsing-remitting MS (RRMS), 20 progressive MS(PMS), and 24 healthy controls (HC) using an in-house ELISA. Of this, 31 patients had received natalizumab, 17 mitoxantrone and 21 fingolimod. Changes in disability were measured using EDSS and disease severity by MSSS. In conjunction, CSF NfL levels were also measured.

At baseline, the mean sNCAM level was 268.7 ng/ml (SD 109 ng/ml) in MS patients compared with 340.6 ng/ml (SD 139 ng/ml) in HC, and PMS had significantly lower sNCAM (239.2 ng/ml, SD 123.0, p=0.019) compared to RRMS (269.4 SD 127.4, p=0.043). After natalizumab and mitoxantrone treatment, we observed an increase of mean sNCAM. However, in the fingolimod treated group, mean sNCAM decreased. There was no correlation was found with EDSS or MSSS, or NfL levels as a whole.

CSF sNCAM were found to be lower in MS than in HC and the lowest sNCAM levels were found in PMS. Following natalizumab and mitoxantrone treatment, we observed an elevation in sNCAM levels, an effect that was not observed following fingolimod treatment. These changes, however, did not appear to correlate with disability in the short-term or NfL levels.

About the author

Neuro Doc Gnanapavan


  • So very interesting. Helps to reassure myself I am on the right treatment despite the risks.

    Would be interesting to see same for Ocrevus. And understand why NFL don’t show the same.

    • Yes, would be interesting to see how the rest of the DMTs fare. The reason why the NFL don’t correlate is maybe that the mode and chronology of injury isn’t the same. NFL sits throughout the neurone whereas these proteins sit on the tip of the growth cone. I believe if the sample size was large enough there may be group associations, such as early vs late MS.

  • Any reason why they did not also assess rituximab treated patients, especially considering the large number of PwMS receiving ritux in Sweden?

  • Correlation or causality (cofounding factor for disability prpgression) in your opinion?
    We surely need larger samples, right?
    And ehat about alemtuzumab, HSCT?

    • It certainly broadens the picture in not only thinking about neurodegenertion alone, but also damage to the existing growth mechanisms.

      I have an ongoing study to recruit patients for biomaterial analysis in alemtuzumab. This will be over two as it’s treatment protocol. NCAM in blood is not as useful as CSF, so you’re highly reliant on CSF samples.

      • will you also try to prove that NfLs in plasma correlate with NfL in CNS? That would be a big big step forward for patients, as spinal tap is surely a deterrent. If we could use NfLs in plasma instead, we could measure this more frequently and be on top of smoldering inflammation much much quicker…

        • Yes, that’s the plan. We need to get some funding in to do this project, and coordinate with MRI appearances as well.

      • Thank you.
        Now Barts has collected quite a few LP samples over the years for various studies. Have these been frozen and/or could you revisit them to test for NCAM or would you have to do another round of collections again?

          • What do you mean by “conceive”? You have not been storing the spinal fluid you have been extracting at a great pain to your patients and expense to the tax payer?

          • The CSF were performed for NFL analysis for clinical use. They are there largely for double checking this analysis. Some of the patients have also consented to have their CSF stored for future research. It has nothing to do with the tax payer!

  • Does this explain why fingolimod did not work in PPMS? I would be interested to know what happens with siponimod.

  • “These changes, however, did not appear to correlate with disability in the short-term or NfL levels.”

    So no effect on progression?

    • These were associations that we looked for at the point of analysis. It maybe different if we look for associations an year later say. The same may be the case with NFL.

  • Natalizumab inhibits lymphocyte trafficking into the CNS. Fingolimod sequesters lymphocytes in lymph tissue. Both would prevent activated lymphocytes from entering the CNS. Why would NZB be more effective? Was there a measure of lymphocytes in patients’ CNS? Since HC showed higher levels then treated patients I would surmise that other factors such as innate immune activity are acting along with activated lymphocytes.

    • Fingolimod is not a global inibitor it affects some cells more than others notably naive and central memory cells. However the pathogenic cells would be in the effector memory population that are less affected by fingo. Natalizumab will also target monocytes

    • We didn’t measure CSF immune profile in this work. But I’m assuming this is now widely available for all they DMTs.

  • Mitroxantrone mean age 46
    1 RRMS, 16 PMS


    31 RRMS mean age 36


    17 RRMS, 4 PMS mean age 38

    So the Mitroxantrone group had ,more pms ,where older , still nCAM rise

    The Natalizumab patients where all rrms and where the youngest of the 3 group


    You know better than me, that age is a neurological disease and the remair mechanisms tend to get worse as we age

    Would like to see apples to apples 🙂

    Obriagdo luis

    • Yes, age does factor in regenerative mechanisms. Although, we haven’t looked at NCAM in healthy controls before. But, as you say based on the findings I believe the changes amount to an analysis effect exclusive to the drug. There is a reason why Mitoxantrone was used in those transitioning in their MS from RRMS to SPMS, and still continues to be used.



Recent Posts

Recent Comments