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

Disclosures

Brain. 2020;143(12):3574-3588. 

In This Article

Results

Patients

Over 200 patients from the Hadassah MS Centre and Unit of Neuroimmunology were prescreened for inclusion in the trial. Of these, 136 patients did not meet the inclusion criteria and 16 refused to participate. In total, 48 patients (21 female, 27 male), were included with a mean EDSS score at inclusion of 5.60 ± 0.88, mean age of 47.63 ± 9.72 years, and mean disease duration of 12.70 ± 7.5 years (Table 1 and Figure 1). The higher proportion of male versus female patients in our sample (which is not in line with the gender distribution of multiple sclerosis) may be explained by the 'biased' inclusion of only patients with progressive and active multiple sclerosis, who were non-responders to conventional multiple sclerosis treatments.

Forty-one patients had secondary progressive multiple sclerosis and seven had primary progressive multiple sclerosis. Twenty of the patients had either superimposed relapses or MRI activity (progressive multiple sclerosis with activity). Most patients (77%) had been previously treated with two or more of the accepted immunotherapeutic drugs for multiple sclerosis. All immunomodulatory treatments were stopped at least 3–6 months before the screening visit. Detailed demographic data (including previous treatments) of each patient are presented in Supplementary Tables 1 and 2. At baseline, the three treatment-groups showed no significant differences in EDSS score, gender, disease course and duration, proportion of active patients and previous treatments (Table 1 and Supplementary Table 1).

One patient (Patient 005) withdrew his consent 3 weeks after the first treatment. Compliance with the trial was excellent, and only nine of the scheduled 528 visits were missed.

The 'blindness' of the study was assessed by a questionnaire filled by the patients and the involved physicians at the end of the trial. According to this, 17 of 48 patients guessed correctly at least one of the treatment assignments. The correct answers from the treating and evaluating physicians were 17% and 22%, respectively.

Primary End Points

Safety. Three serious adverse events occurred during the study resulting in patient hospitalization. Two were related to relapses of multiple sclerosis, and one was due to an upper respiratory infection (not related to the treatment) that was resolved after a course of antibiotics. No other serious adverse events were observed during the 14 months of the trial. The full list of adverse events is presented in Table 2. All of them were transient/short-lasting and there was no carry-over adverse event between the two cycles of treatment (see also the detailed description of adverse events in Supplementary Table 3).

Clinical Efficacy. The intention to treat analysis of the statistical analysis plan predetermined primary efficacy end point, showed that the percentage of patients with treatment failure (increase in EDSS score of 1 point for patients with baseline EDSS values of ≤5.0 and of 0.5 degree for baseline EDSS > 5.0), confirmed by two consecutive evaluations at the end of each treatment cycle (i.e. at 6 months or 12 months), was significantly lower in the MSC-IT and MSC-IV groups (6.7% and 9.7%, respectively) compared with the sham-treated patients (41.9%) (Table 3 and Figure 2) (P = 0.0003 between MSC-IT and sham treatment, P = 0.0008 between MSC-IV and sham treatment, chi-squared test). Among the sham-treated patients, 76.7% experienced a deterioration in at least one Functional Systems score, and only 31% and 27.6% experienced similar deterioration in the MSC-IT and MSC-IV groups, respectively (P = 0.0002 and P = 0.0004, chi-squared test) (Table 3).

Figure 2.

Clinical and radiological effects of MSC-IT and MSC-IV transplantation versus sham treatment in progressive multiple sclerosis. (A) Beneficial effect of MSC treatment on the progression of multiple sclerosis as evidenced by the changes in EDSS. Number of patients in each treatment subgroup (MSC-IT, MSC-IC, and sham treatment) who deteriorated in EDSS score or were stable or improved in EDSS, during each 3- or 6-month period (pooled data from the two cycles of treatment). P < 0.0001 (3 months) and P = 0.0003 (6 months) for MSC-IT versus sham treatment (chi-squared test); P = 0.0085 (3 months), P = 0.0008 (6 months) for MSC-IV versus sham treatment (chi-squared test). (B) Treatment with MSC induces beneficial effects on the progression of multiple sclerosis, as evidenced by the changes in EDSS score. Longitudinal follow-up of the mean changes in EDSS scores in each of the major treatment group (MSC-IT, MSC-IV, and placebo/sham treatment), during the run-in pretreatment period, and at 3 and 6 months following each cycle of treatment. (P < 0.0001 for MSC-IT versus sham at 3 months and P < 0.0001 at 6 months; the corresponding P-values for MSC-IV versus sham, were 0.001 at 3 months and 0.0002 at 6 months, Wilcoxon signed-rank Test.) (C) Differential clinical effects in each of the six treatment subgroups during the two phases of the trial (longitudinal changes in the EDSS score). Longitudinal follow-up of the mean EDSS scores in each of the six treatment subgroups (1A: MSC-IT/MSC-IT, 1B: MSC-IT/Placebo, 2A: MSC-IV/MSC-IV, 2B: MSC-IV/Placebo, 3A: Placebo/MSC-IT and 3B: Placebo/MSC-IV), during the run-in pretreatment period, and the two cycles of the study. Comparison between the 12 months of treatment versus the run-in period: P < 0.001 (for the repeated MSC-IT treatment), (Wilcoxon signed-rank test) P = not significant (for the single MSC-IT treatment), (Wilcoxon signed-rank test) P < 0.001 (for the repeated MSC-IV treatment), (Wilcoxon signed-rank test) P = not significant (for the single MSC-IV treatment), (Wilcoxon signed-rank test) P = 0.023 (between repeated versus single MSC-IT treatment), (Mann-Whitney test) P = 0.50 (between repeated versus single MSC-IV treatment), (Mann-Whitney test). (D) Incidence of clinical relapses and gadolinium-enhancing lesions in MRI. Number of relapses and gadolinium-enhancing lesions (presented as black dots for sham treatment, black squares for MSC-IV, and triangles for MSC-IT) in each patient of the three treatment subgroups (MSC-IT, MSC-IV, and sham treatment) during both cycles (two 6-month periods). (Pooled data from the two cycles of treatment.) Cyc-1 = first cycle; Cyc-2 = second cycle. P < 0.0005 for relapses in the MSC-IT group versus sham treatment. P = 0.052 for relapses in the MSC-IV group versus sham treatment. There was a good correlation between the incidence of relapses and the presence of gadolinium-enhancing lesions in each of the treatment groups. Mean number of gadolinium-enhancing lesions on MRI in the two cycles of treatment pooled together: 0.17 ± 0.47 in the MSC-IT group, 0.97 ± 1.93 in the MSC-IV group, and 0.55 ± 1.03 in the sham-treated group (P = 0.062 for MSC-IT versus sham treatment, P = 0.90 for MSC-IV versus sham treatment, P = 0.077 for MSC-IT versus MSC-IV, Wilcoxon test). (E) MSC treatment induces beneficial effects on the motor network on functional MRI. Changes in mean z-values of the motor network on functional MRI in the group of patients treated with MSC-IT or MSC-IV versus the sham-treated group, during both cycles of the study. An increase was observed in the mean z-score in the MSC-IT group (annualized change of 0.108 ± 1.06 and 0.156 ± 0.68 at 3 and 6 months, respectively; pooled data), and a deterioration in the sham-treated group (−0.504 ± 1.06 and −0.288 ± 0.61 at 3 and 6 months, respectively). The changes in the MSC-IV group were +0.036 ± 0.88 at 3 months, and −0.06 ± 0.816 at 6 months (P = 0.0675 and P = 0.042 for MSC-IT versus sham treatment at 3 and 6 months, respectively; P = 0.031 and P = 0.077 for MSC-IV versus sham treatment at 3 and 6 months, respectively).

The mean EDSS score deteriorated in the sham-treated group and was improved in the MSC-IT and MSC-IV groups during both treatment cycles (P = 0.0002 and P = 0.007, respectively, versus sham treatment; Mann-Whitney test) (Figure 2 and Table 3). Two patients showed improvement in EDSS during the first cycle of treatment with MSC-IT and 11 during the second cycle (ranging from 0.5 to 1.0 degrees). The respective numbers of patients with improvement in the MSC-IV group, were three and six in the two cycles; one patient showed improvement in the sham treatment group (Table 4). There was no clear association between the treatment-induced benefit with the status of activity of the disease (presence of relapse or MRI activity) during the year prior to inclusion and the run-in period (Table 4). At inclusion, seven patients had active disease in the placebo group, six in the MSC-IT group and seven in the MSC-IV group. During the second cycle of the treatment, only one patient in the MSC-IT group showed activity compared with 12 patients in the sham-treated group (Table 4 and Supplementary Table 4). During only the first treatment cycle (comparison of parallel groups before the crossover, n = 16 in each group), as seen in Supplementary Table 4, the annualized EDSS change during the year prior to the inclusion and the run-in period (total 14 months) was similar in the subgroup of the patients with activity and those without, in all three groups, and was significantly reduced after MSC-IV and MSC-IT treatment, similar to the active and the inactive patients (e.g. in the MSC-IT group the annualized ΔEDSS was reduced from +0.88 to −0.11 in the active patients and from +1.01 to +0.14 in the inactive subgroup). However, since the majority of the included patients had some activity of the disease, either in the year prior to the inclusion or in the run-in period (in total, before starting MSC transplantation 27 patients were active and 21 non-active and three more became active during the first placebo cycle, further increasing the number of patients with active disease to 30 of 48), it seems that MSC transplantation benefits predominantly active progressive multiple sclerosis.

The changes in ambulation scores and in the sum of all Functional System scores, followed the same trend, strongly favouring MSC-IT and MSC-IV treatments over sham treatment (Table 3 and Table 5) (the detailed EDSS, Functional System and ambulation scores of each individual patient at each visit, are provided in Supplementary Tables 4–9). These clinical effects do not seem to have been influenced by steroidal treatment (which was administered due to relapse, deterioration or MRI activity). Actually more patients in the sham-treated group received steroids (11:4 in the first cycle and 7 in the second, versus 6 in the MSC-IV group and 4 in the MSC-IT group, during the two cycles of the study) (detailed data on steroid treatment of each patient in Supplementary Table 4).

For the changes in EDSS and Functional System scores, the MSC-IT was superior to the MSC-IV treatment (Table 3 and Table 5). Notably, half of the patients treated twice with MSC-IT, had a confirmed disability improvement (CDI) at the end of the whole trial (12 months), and none of them exhibited confirmed disability progression.

An additional analysis using the 'any treatment scenario' during the whole 12-month duration of the study (i.e. combining all treatment groups together; n = 48), the EDSS score, showed a significant deterioration during the pretreatment run-in period (from 5.58 ± 0.88 at screening to 5.88 ± 0.80 at baseline visit, P < 0.001, Wilcoxon signed-rank test), and a stabilization during the whole 12 months of the two treatment-cycles (5.88 ± 0.80 at baseline versus 5.87 ± 0.9, on final visit at 12 months). The difference between the change in EDSS in the run-in period versus the change from baseline to the 12-month value, for all 48 patients, was statistically significant (P < 0.001, Wilcoxon signed-rank test). The patients treated twice intrathecally with MSC had the best outcome at 12 months and the difference between single versus double intrathecal treatment, was statistically significant in all the major efficacy parameters (EDSS, sum of functional systems, ambulation score) (Table 5).

Fifteen of 32 patients in the sham-treated groups experienced at least one relapse during both 6-month periods of the study, compared with only seven in the MSC-IV group and 2 of 32 in the MSC-IT group (46.9%, 21.9%, and 6.3%, respectively, P = 0.0002 for MSC-IT, and P = 0.035 for MSC-IV versus the sham treatment) (Table 3 and Figure 2). The mean annual relapse rates in both cycles of treatment were 0.06 ± 0.25, 0.28 ± 0.57, and 0.56 ± 0.67 in the MSC-IT, MSC-IV, and sham-treated groups, respectively (P = 0.0005, for MSC-IT versus sham treatment; P = 0.052, for MSC-IV versus sham treatment, Wilcoxon test).

Secondary End Points

MRI. The mean number of gadolinium-enhancing lesions per patient during the two cycles of treatment were: 0.55 ± 1.03 in the sham-treated group, 0.17 ± 0.47 in the MSC-IT group, and 0.97 ± 1.93 in the MSC-IV group (P = 0.062 for MSC-IT versus sham treatment; P = 0.90 for MSC-IV versus sham treatment; P = 0.077 for MSC-IT versus MSC-IV, Wilcoxon test) (Figure 2 and Table 3). The mean monthly rate of the T2-FLAIR lesion volume change compared to the rate during the run-in period was −0.024 ± 0.053 in the pooled group of patients treated with MSC-IT; −0.016 ± 0.036 in the MSC-IV group; and 0.003 ± 0.029 in the sham-treated group (P = 0.029 for MSC-IT versus sham treatment; P = 0.123 for MSC-IV versus sham treatment) (Table 3). The changes in total brain volume, although showing a trend in favour of the MSC-IT treatment (Table 3), were not statistically significant. This may be related to possible effects on brain volume of the intrathecal injection itself, and also to diurnal and water intake fluctuations. Notably, there was an increase of brain volume in all three groups (+0.21% for MSC-IT, +0.15% for MSC-IV and +0.41% for placebo) during the second cycle of the trial, potentially reflecting carryover effects of the first treatment and therefore, difficult to interpret.

Functional MRI. Testing of the motor networks revealed a significant annual increase in the mean z-score of the MSC-IT group (0.108 ± 1.06 and 0.156 ± 0.68 at 3 and 6 months, respectively; pooled data), and a decrease/deterioration in the sham-treated group (−0.504 ± 1.06 and −0.288 ± 0.61 at 3 and 6 months, respectively). The z-scores changes in the MSC-IV group were 0.036 ± 0.88 at 3 months and −0.06 ± 0.816 at 6 months (P = 0.0675 and P = 0.042 for MSC-IT versus sham treatment at 3 and 6 months, respectively; P = 0.031 and P = 0.077 for MSC-IV versus sham treatment at 3 and 6 months, respectively) (Figure 2).

No Evidence of Disease Activity. Seventeen of 29 (58.6%) MSC-IT-treated patients [for whom there were data for all parameters of no evidence of disease activity (NEDA)] had no evidence of disease activity (NEDA-3) during the two pooled 6-month treatment periods, as compared with 40.6% (13/32) in the MSC-IV group and 9.7% (3/31) in the sham-treated group, (P < 0.0001 for MSC-IT versus sham treatment and 0.0048 for MSC-IV versus sham treatment). The percentages of patients with NEDA-4 (Kappos et al., 2016)status (i.e. including annual brain volume loss of <0.4% according to MRI) were 44.8% in the MSC-IT, 28.1% in the MSC-IV, and 9.7% in the sham-treated groups (P = 0.005 for MSC-IT versus sham treatment and P = 0.12 for MSC-IV versus sham treatment) (Table 3).

Other Parameters. Statistically significant benefits were observed in the MSC-IT group in the 25-foot timed walking test, 9-hole peg test, OCT of the retinal nerve fibre layer, PASAT, and KAVE/OWAT cognitive tests. Trends of beneficial effects were also observed in VEP, the SDMT, and the proportion of CD4+CD25+highcells (increased) (Table 3).

processing....