Beneficial Effects of Autologous Mesenchymal Stem Cell Transplantation in Active Progressive Multiple Sclerosis

Panayiota Petrou; Ibrahim Kassis; Netta Levin; Friedemann Paul; Yael Backner; Tal Benoliel; Frederike Cosima Oertel; Michael Scheel; Michelle Hallimi; Nour Yaghmour; Tamir Ben Hur; Ariel Ginzberg; Yarden Levy; Oded Abramsky; Dimitrios Karussis


Brain. 2020;143(12):3574-3588. 

In This Article


With an aim to evaluate the safety and clinical efficacy of autologous MSC transplantation in progressive multiple sclerosis, our study revealed positive results in all predefined primary end points. No serious, treatment-related adverse effects were observed and significantly fewer patients in the MSC-IT and MSC-IV groups experienced treatment failure (based on the EDSS changes), as compared with the sham-treated group (Table 4). Significant changes favouring MSC-IT treatment over sham treatment were also observed in the ambulation index, sum of functional scores, 25-foot timed walking test, 9-hole peg test, PASAT, and OWAT/KAVE cognitive tests, and the rate of change in T2 lesion load on MRI, as well as newer biomarkers, such as OCT (retinal nerve fibre layer) and functional MRI (motor network). Repeated intrathecal injection of MSCs at Month 6 significantly boosted the effects observed during the first cycle of treatment. Beneficial (but less significant) effects were also observed in the MSC-IV group. Overall, the robust effects of MSC transplantation on various parameters that reflect neurological dysfunction and especially on multiple sclerosis activity, may indicate the involvement of (central and peripheral) immunomodulatory and possibly also neuroprotective mechanisms.

These benefits seem to be of particular clinical significance, as they were observed in patients with progressive multiple sclerosis unresponsive to conventional immunotherapies, and for which limited treatment options exist. To have a crude estimation of the magnitude of our findings versus previously published randomized trials in progressive multiple sclerosis (although it is not scientifically sound to compare different trials with different designs), the difference in the EDSS change between the placebo arm and the MSC-IT group in our study was 0.5 (in favour of the MSC-IT treatment) as compared with an EDSS difference between placebo and active treatment of 0.36 in a mitoxantrone trial and 0.16 in a biotin trial (Hartung et al., 2002; Tourbah et al., 2016; Montalban et al., 2017).

Previous trials using various types of MSC in multiple sclerosis have been small (mostly open-label) and mainly showed the clinical safety of MSC-administration, providing in parallel some preliminary indications of potential clinical benefits (Karussis et al., 2010; Yamout et al., 2010; Connick et al., 2012; Cohen et al., 2018; Harris et al., 2018; Riordan et al., 2018). A triple blinded study, in which 30 patients were infused (11 placebo, 10 low-dose and 9 high-dose) intravenously with adipose-MSC showed an inconclusive trend of efficacy (Fernandez et al., 2018). Another small double blind, placebo-controlled trial (Llufriu et al., 2014) in nine multiple sclerosis patients treated intravenously with MSC, showed a positive trend in the primary end point of MRI gadolinium-enhancing lesions, but no significant differences in the secondary end points of EDSS and multiple sclerosis functional scores. In a phase Ib study (Lublin et al., 2014a), human placental MSCs (PDA-001) were injected intravenously in 16 patients in a double blind dose-ranging trial. Most patients were stable or improved after the treatment.

Despite the development of highly efficient and more targeted immunotherapies, two major unmet needs in the treatment of multiple sclerosis still exist. First, the need for treatment to suppress compartmentalized and meningeal inflammation in the CNS, which seems to drive tissue injury and progression of disability (Magliozzi et al., 2007; Lucchinetti et al., 2011; Ruggieri et al., 2015; Eshaghi et al., 2018; Reich et al., 2018). These compartmentalized inflammatory and degenerative activities seem to be less responsive to the majority of immunomodulatory drugs, accounting for the relatively poor efficacy of the majority of registered multiple sclerosis therapies in progressive multiple sclerosis, especially in patients with no prominent inflammatory activity, with the exception of very few (Hartung et al., 2002; Shirani et al., 2016; Tourbah et al., 2016; Montalban et al., 2017).

Second, the need for a treatment that may substantially promote regeneration–remyelination. Generally, the CNS loses its capacity for efficient regeneration and remyelination over time. This is especially pronounced in chronic neuroinflammatory and neurodegenerative diseases such as multiple sclerosis, possibly due to an insufficiency of growth factors or defective mobilization of intrinsic CNS stem cells/oligodendrocyte progenitors (Ben-Hur et al., 2005; Karussis and Kassis, 2007; Einstein and Ben-Hur, 2008; Karussis et al., 2013).

Based on their well-described properties (Ben-Hur et al., 2005; Freedman et al., 2010; Freedman and Uccelli, 2012; Karussis et al., 2013; Scolding et al., 2017), stem cells may represent a 'logical' treatment approach to achieve those unmet needs and induce neuroprotection and enhance endogenous remyelination. Moreover, stem cells are strong immunomodulators that may potentially downregulate the localized and compartmentalized inflammation upon their migration to the CNS (Magliozzi et al., 2007; Lucchinetti et al., 2011; Reich et al., 2018). Several studies have shown that embryonic, neuronal, and other adult stem cells can induce beneficial clinicopathological effects in animal models of neurological diseases, including multiple sclerosis (Pluchino et al., 2003, 2009; Ben-Hur et al., 2004; Zappia et al., 2005; Aharonowiz et al., 2008; Karussis and Kassis, 2008; Kassis et al., 2008; Harris et al., 2012). MSCs are commonly used for such therapies, as they have several practical advantages for clinical use over other types of stem cells: (i) they can be easily cultured and expanded in large quantities; (ii) they can be obtained from the patient, thus eliminating the need for a donor, the risk of rejection, or the need for chemotherapy; and (iii) they seem to be safe and carry low risks of malignant transformation. During the past decade, MSC treatments have been applied to various neurological diseases in small or pilot open-label trials (Karussis et al., 2010; Yamout et al., 2010; Connick et al., 2012; Lee et al., 2012; Llufriu et al., 2014; Lublin et al., 2014a; Petrou et al., 2016; Steinberg et al., 2016; Cohen et al., 2018; ,Fernandez et al., 2018), with promising indications.

The putative mechanism of action of MSCs in neurological diseases is controversial. Some investigators claim that the most prominent effects are mediated through peripheral immunomodulation (Karussis and Kassis, 2007, 2008; Uccelli et al., 2007, 2008, 2011; Karussis et al., 2008; Kassis et al., 2011). We have long advocated that neuroprotective and neurotrophic mechanisms also play a crucial role, as supported by our findings in animal models and pilot trials of multiple sclerosis and amyotrophic lateral sclerosis (Karussis et al., 2010; Petrou et al., 2016). We speculate that intrathecal injection, which brings a higher proportion of the injected cells into close proximity with damaged areas of the CNS, may induce more robust effects than intravenous injection. In the current study, intrathecal transplantation of MSCs was indeed shown to be superior to intravenous administration in several efficacy parameters (Table 3). If peripheral immunomodulation was the dominant mechanism, we would expect the opposite to be true. Moreover, the clinical beneficial effects were observed more prominently in patients with active disease (i.e. with relapses or MRI activity) but also in those without activity (Table 4). The improvements observed in OCT and motor networks on functional MRI, and the trend towards increased brain volume may indicate the involvement of neurotrophic or even neuroregenerative mechanisms, but to demonstrate whether autologous MSCs may facilitate remyelination or regeneration, a larger trial with appropriate biomarkers in progressive patients with disability worsening without disease actvity, is warranted. On the other hand, the less pronounced effects of the treatment on gadolinium-enhancing lesions on MRI, advocate against a major contribution of peripheral immunomodulatory mechanisms.

The strengths of our trial include: (i) the inclusion of patients with active progressive multiple sclerosis, for which existing immunotherapies are usually ineffective; (ii) the double-blind design, making this the first randomized controlled trial comparing intrathecal versus intravenous methods of MSC administration and single versus repeated treatment; and (iii) the robust, though short-term, clinical benefits observed in several disease activity parameters, including newer biomarkers, such as functional MRI-network connectivity, OCT, and cognitive testing. The limitations of our study include the small number of patients in each group, the short duration, and the crossover design (in the second cycle), which may have introduced a 'carry-over' effect from the first cycle of treatment.

An additional possible limitation of our findings could be related to the severe progression of half of the patients during the run-in period before the treatment. This may be related to the patients' population (highly active patients non-responders to a mean of 2.58 previous treatments) and the discontinuation of the immunotherapies before the study. A rebound effect could partially explain this deterioration and the higher than in other cohorts progression rate of the placebo group, especially after treatment with fingolimod, which was the last treatment in 21 of 48 patients (with a rather equal distribution in the three treatment groups). Some of the beneficial effects therefore could be theoreticaly related to a 'regression to the mean' phenomenon. However, such regression, although may have affected the clinical changes at some degree (especially in the first months of the study), cannot—to our view—explain the finding of more robust benefits during the second cycle of treatment and of the improvements in several clinical and paraclinical parameters at various time-points, during both phases of the trial.

In summary, our results provide clear signals of short-term clinical efficacy and possible indications of neuroprotection, induced by the administration of autologous MSCs in patients with progressive multiple sclerosis. Our findings suggest a superiority of intrathecal over intravenous administration, and indicate a possible boost of the beneficial effects by a repeated injection of the cells. These data may contribute to the design of future trials with cell therapies and the use of objective biomarkers for the evaluation of neurodegeneration and neuronal regeneration. A larger, phase III study is warranted to confirm these observations and further evaluate the therapeutic potential of cellular therapy in neuroinflammatory and neurodegenerative diseases such as multiple sclerosis.

Research in Context and Evidence Before This Study

We searched PubMed up to 1 April 2019 for English language articles using the search terms: 'mesenchymal stem cells', 'multiple sclerosis' and 'clinical trial'.

We found six published clinical trials using treatment with MSC in multiple sclerosis during the past 10 years. Almost all of them were open pilot studies and only one small trial was randomized and controlled by a placebo group. These phase I and IIa studies mainly investigated the safety of intravenous or intrathecal MSC. One was performed in our centre (Karussis et al., 2010) and was an open safety phase I/II trial including 15 multiple sclerosis and 19 amyotrophic lateral sclerosis patients; MSCs were injected both intravenously and intrathecally. No serious side effects were reported and there were indications of clinical efficacy in terms of stabilization or improvement of EDSS and amyotrophic lateral sclerosis functional rating scale (ALSFRS) scores.

In the second, an open label trial by Yamout et al. (2010), MSCs were injected into 10 patients with multiple sclerosis intrathecally and the results showed hints of clinical, but no radiological, efficacy. The third was an open label study, by Connick et al. (2012), MSCs were injected intravenously into 10 progressive multiple sclerosis patients, with primary objective being safety, and the secondary objective being the effects on visual pathways. No serious side effects were observed and there was some evidence of structural, functional and physiological improvement in few of the visual end points. In a larger open trial by Cohen et al. (2018), 25 patients and eight controls were injected intravenously with MSC. The authors reported only data for the feasibility and safety of this treatment but no clinical data. In a small double blind, placebo-controlled trial by Llufriu et al. (2014), nine patients with multiple sclerosis were injected with MSC intravenously. The study used as primary end point the number of gadolinium-enhancing lesions and showed a positive trend in this primary end point, but no statistically significant differences in the secondary end points of EDSS and multiple sclerosis functional scores.

Further to these trials in which autologous unmodified bone marrow MSCs were used, there are four additional trials using different types of modified or enhanced MSCs. In the first, a phase Ib study by Lublin et al. (2014a), human placental MSCs (PDA-001) were injected intravenously into 16 patients in a double-blind dose-ranging trial. The primary end point was to rule out any negative clinical effect of the procedure. Most patients were stable or improved after the treatment. Fernandez et al. (2018) used adipose-derived MSCs and injected different doses of the cells into 19 multiple sclerosis patients and 11 controls. The authors showed the safety of the procedure and indicated possible beneficial clinical and radiological effects. Riordan et al. (2018), in an open study, injected umbilical cord tissue-derived MSCs into 20 multiple sclerosis patients intravenously using a protocol of multiple infusions. The procedure was shown as safe and induced some clinical improvements. Finally, Harris et al. (2018), in an open trial (n = 20), injected MSC-derived neural progenitors intrathecally, in three doses spaced 3 months apart. No serious adverse events were observed and there were indications of clinical benefits.

Added Value of This Study

Our current study was a phase II double-blind trial, randomized and controlled by placebo and included a significantly higher number of patients (n = 48) than in previous trials. It compared for the first time the efficacy of two methods of administration of the MSCs (intrathecal and intravenous) using a parallel-group and crossover design. In addition, we compared the effects of repeated (two) MSC infusions with those of a single injection.

This is the first trial—to our knowledge—that used several objective and novel biomarkers to evaluate clinical efficacy and indications of neuro-regeneration (timed walking, hand dexterity tests, cognitive tests, quantitative neuroradiological measurements, functional MRI, OCT, VEP and dynamic visual tests).

Implications of all the Available Evidence

Various types of MSCs were used in small (mostly open label) clinical trials in multiple sclerosis during the last decade. These studies showed the clinical safety of MSC-administration and provided some preliminary indications of potential clinical benefits. The protocols of preparation and administration of MSC and the patients' populations were highly variable. In our trial, which was randomized and placebo-controlled, we recruited a significant larger and well-defined group of patients with precisely defined active progressive multiple sclerosis.