Participants and follow-up
From 05/10/2017 to 02/01/2020 a total of 70 patients with ALS were screened for eligibility, of whom 63 were randomly assigned to a trial group: 21 to rapamycin 2 mg/m2/day, 21 to rapamycin 1 mg/m2/day and 21 to placebo (Fig. 1). Seven patients dropped out from the study during the treatment period, whereas 15 did not conclude the follow up (after treatment). Two patients did not take at least 80% of the study drug as planned per protocol. As these two patients dropped out between week 18 and 30, Per Protocol (PP) and Intention To Treat (ITT) analyses differ at week 8 and 18 only, whereas the results of the two analyses were identical at week 30 and 54.
Source data are provided as a Source Data file.
Baseline demographic and disease characteristics are summarized in Table 1. Baseline biological features of the trial participants are summarized in Supplemental materials, Table S1.
Rapamycin effect on Treg cells
Although 56 patients reached treatment end, only 50 blood samples from 50 participants were available for cells populations analysis at week 18 due to the COVID-19 pandemic. Among patients treated with rapamycin 1 mg/m2/d, 28% of patients presented an increase of Treg cells by at least 30%, versus 12% of patients in the placebo group (Relative Risk [RR] 2.36, 97.5% confidence interval [CI] 0.42 to 13.12; p = 0.236) (ITT analysis, Table 2); the positive response was observed in 20% of patients treated with rapamycin 2 mg/m2/d (RR1.70, 97.5%CI 0.26 to 11.21, p = 0.522). PP analysis yielded similar results (Table S2).
In an adjusted analysis, after correction for sex, time from onset, ALSFRS-R slope at baseline, and edaravone treatment, based on the comparison of differences of clinical significance between groups at baseline, odds ratio (OR) of a positive response was non-significantly increased in Rapamycin treatment arms (Table 2). At week 18, 6/17 (35%) placebo-treated patients experienced any increase in Treg cells compared with 10/18 (56%) patients treated with rapamycin 1 mg/m2/d (RR 1.57, 97.5%CI 0.66–3.77; p = 0.2291) and 10/15 (67%) patients treated with rapamycin 2 mg/m2/d (RR 1.89, 97.5%CI 0.81–4.38; p = 0.0765).
At treatment end patients treated with rapamycin 1 mg/m2/d experienced a mean increase of Treg cells by 0.41, patients treated with 2 mg/m2/d showed stable values (−0.05), whereas patients in placebo showed a decrease in Treg cells by −0.46 from baseline (mean difference [MD] 0.87, 97.5% CI −0.36 to 2.09; p = 0.109 in rapamycin 1 mg/m2/d group and MD 0.41, 97.5%CI, −0.88 to 1.69; p = 0.465 in rapamycin 2 mg/m2/d group) (Table 3). While in the placebo arm Treg monthly variation during and after treatment showed a nearly constant decline (mean −0.10, 95%CI: −0.25 to 0.05, p = 0.182 during treatment; mean −0.04, 95% CI: −0.09 to 0.02, p = 0.216 after treatment), during treatment there was a trend towards an increase in patients who were treated with the lower dose of the drug compared to those who received the placebo (MD 0.17, 97.5%CI: −0.06 to 0.41, p = 0.100 in rapamycin 1 mg/m2/d arm; MD 0.08, 97.5%CI: −0.16 to 0.33, p = 0.450 in rapamycin 2 mg/m2/d arm), that was not statistically significant (Table S3). In a post-hoc analysis on a limited number of samples, we found that after in vitro stimulation with anti-CD3/CD28 functionality of Treg cells resulted similar before and after therapy and not different to those from age and sex-matched healthy controls (Fig. S1).
Treatment impact on blood cell subpopulations
We next examined the change from baseline to each time point (week 8, 18, 30, 54) of the activation and homing capabilities of different T, B, NK cell subpopulations, comparing treatment and placebo arms (Fig. 2, Table S4); no correction was applied for multiple tests (55 outcomes were examined), therefore the following results on secondary outcomes should be interpreted cautiously.
The figure displays only a selection of the most interesting outcomes (55 cell subpopulation were examined and 11 inflammasome/cytokines, without accounting for multiple outcomes). In detail from left to right: changes from baseline to week 18 (a, n = 32 patients) and 30 (g n = 26 patients) in activated (CD38+, HLA-DR+) CD4 + T cells; changes from baseline to week 18 (b, n = 29 patients) and 30 (h, n = 21 patients) in activated (CD38 + , HLA-DR+) CD8 + T cells; changes from baseline to week 18 (c n = 22 patients) and 30 (i, n = 19 patients) in memory switched B cells (B.C. Mem. Sw.); changes from baseline to week 18 (d, n = 19 patients) and 30 (j, n = 19 patients) in classical monocytes (Class. Mono.); changes from baseline to week 18 (e, n = 27 patients) and 30 (k, n = 20 patients) in IL18 mRNA level (IL18 mRNA); changes from baseline to week 18 (f, n = 49 patients) and 30 (l, n = 39 patients) of plasmatic IL18. Comparison were performed using linear regression models that include indicator variables for treatment arms as the independent variables. For the comparison of Rapamycin and placebo arms, a P value of 0.05 or less was considered to indicate statistical significance and uncertainty in results was expressed with the 95% confidence interval (CI). For the comparisons between Rapamycin 1 mg/m2/d or 2 mg/m2/d arms and the placebo arm, a P value of 0.025 or less was considered to indicate statistical significance and uncertainty in results was expressed with the 97.5% CI, to account for multiple arms comparison with the Bonferroni method. CIs were calculated based on the exact t distribution. All statistical tests were two-tailed. * means p < 0.05 for R and p < 0.025 for R1 and R2; ** means p < 0.01 for R and p < 0.005 for R1 and R2 arms. Error bars represent ± standard deviation. Source data are provided as a Source Data file.
At week 18 patients treated with rapamycin 1 mg/m2/day showed a not significant trend toward a reduction of activated (CD38+, HLA-DR+) CD8 + T lymphocytes (MD −0.92, 97.5% CI −1.88 to 0.05; p = 0.032), and intermediate monocytes (CD14+, CD16dim monocytes, MD −13.48, 97.5% CI −28.17 to 1.22; p = 0.038), with respect to placebo group from baseline to treatment end; in the same frame time this group showed an increase of the percentage of memory switched B cells (defined as IgM-, IgD-, CD21+,CD24+, CD27+, CD38- B cells, MD 0.82, 97.5% CI 0.28 to 1.37; p = 0.002) and of classical monocytes (CD14+, CD16- monocytes, MD 17.76, 97.5% CI 0.91 to 34.60; p = 0.019) (Fig. 2a–d)(Table S4).Similar results were detected at week 30 (Fig. 2g–j). In a post-hoc analysis on a limited number of samples, we found that, with respect to controls, at baseline ALS patients showed a trend towards higher percentages of Th1 CD4 + T cells that remained similar after 18 weeks (Fig. S2).
Rapamycin effect on inflammasome
Patients treated with rapamycin showed lower mRNA relative expression of pro-inflammatory cytokine IL-18, which is a readout of inflammasome activation (MD −0.45, 97.5%CI −1.09 to 0.18; p = 0.101 for rapamycin 1 mg/m2/d group and MD −0.60, 97.5%CI −1.18 to −0.01; p = 0.022 for rapamycin 2 mg/m2/d group), and consistently a marked reduction in plasmatic IL-18 protein (MD −107.80, 97.5%CI −187.12 to −28.48, p = 0.002 for rapamycin 1 mg/m2/d group and MD −103.00, 97.5%CI −183.51 to −22.48; p = 0.004 for rapamycin 2 mg/m2/d group) from baseline to treatment end with respect to placebo-arm patients (Fig. 2e, f, Table S5).These effects were lost at subsequent measurements during follow up (week 30 and 54) (Fig. 2k, l). Due to the high number of tests (11 outcomes were examined) these results should be interpreted cautiously.
Longitudinal assessment of neurofilament
In the placebo arm an overall decrease in neurofilament levels was observed since the first follow-up by serial measurements in the serum (MD from baseline in serum pNfH at week 8: −174.95 ± 602.19; week 18: −289.35 ± 788.33; week 30: −482.29 ± 927.19; MD from baseline of serum NfL at week 8: −11.61 ± 96.07; week 18: −25.76 ± 93.46; week 30: −33.21 ± 101.57) (Tables S6, S7). The decrease of serum pNfH and NfL levels from baseline to week 18 that was found in the placebo group was not observed in patients treated with rapamycin. This difference among treatment arms (pNfH MD 399.03, 95%CI 81.77 to 716.28; p = 0.015 and NfL MD 34.92, 95%CI 0.47 to 69.36; p = 0.047) was lost after treatment end (Tables S6, S7). Similar results were obtained from measurement of CSF pNfH and NfL at week 18 (pNfH MD 866.57, 95%CI −105.15 to 2345.72; p = 0.075 and NfL MD 4756.10, 95%CI 808.45 to 11531.5; p = 0.051, respectively) (Table S8).
Other biological outcome measures
Monthly changes of selected biological outcome measures during and after treatment, across arms confirmed an increase of classical monocytes/CD14+ (MD 5.42, 97.5%CI 2.19 to 8.65, p = 0.0003) and memory switched B cells/CD45+ (MD 0.20, 97.5%CI 0.06 to 0.33, p = 0.0018) and a decrease of intermediate monocytes/CD14+ (MD −3.28, 97.5%CI −5.29 to −1.28; p = 0.0004) and IL-18 (MD −24.94, 97.5%CI −41.58 to −8.30; p = 0.009) in the rapamycin 1 mg/m2/day arm. The change from baseline to each time point of the phosphorylation of the S6RP was not different between rapamycin arms and placebo arms; this test was performed on a limited number of samples (Table S9). Changes from baseline to each time point in creatinine and albumin, CK, vitamin D, were not different in rapamycin and placebo arms.
Secondary clinical outcomes
Absolute changes from baseline to each time point in ALSFRS-R total score in patients treated with rapamycin or placebo is showed in Table S10, whereas ALSFRS-R variation before, during and after treatment is showed in Table 4. Patients treated with rapamycin 1 mg/m2/d showed a mean monthly difference of 0.50 points in the ALSFRS-R total score with respect to placebo during treatment (97.5%CI −0.32 to 1.32; p = 0.172), and of −0.12 after treatment (97.5%CI −0.75 to 0.51; p = 0.671). The difference between the rapamycin 1 mg/m2/d group and the placebo group in the change in monthly variations from before to during treatment was 0.46 (97.5%CI −0.21 to 1.13; p = 0.174). Figures 3 and S3 show the mean and individual rate of decline of the ALSFRS-R total scores.
Source data are provided as a Source Data file.
Correlation analyses was performed on changes in ALSFRS-r and neurofilament levels to investigate whether a relation existed between clinical and biological outcomes. In the placebo arm, an inverse correlation was found between the change (week 18–baseline) in serum and CSF neurofilament light levels and the change (baseline – week 18) in ALSFRS-R (Pearson’s r coefficient: −0.41, 95% CI 0.09 to −0.74, p = 0.106 between serum NFL changes and ALSFRS-R, and r coefficient: −0.58, 95% CI −0.01 to −0.86, p = 0.049 between CSF NFL changes and ALSFRS-R), indicating a decrease in neurofilament while increasing disease progression. Correlation analyses among the change (week 18–baseline) in serum and CSF neurofilament light levels and the change (baseline – week 18) in ALSFRS-R within rapamycin arm showed no correlation: Pearson’s r coefficient was 0.18 (95% CI −0.50 to 0.17, p = 0.308) between ALSFRS-R and serum NFL change (w18–w0), and 0.32 (95% CI −0.65 to 0.10, p = 0.131) between ALSFRS-R and CSF NFL changes (w18–w0). Similar results were obtained with pNfH (Fig. 4). Comparison among correlation coefficients showed a nearly significant difference between rapamycin and placebo arm in serum NfL (p = 0.0608); the difference was less pronounced for pNfH (p = 0.1429).
In detail from left to right, upper panels: changes from week 18 to baseline in serum pNfH in rapamycin (a) and placebo arm (b), in relation to progression rate calculated as the monthly decline in the ALSFRS-R from baseline to week 18. From left to right, lower panels: changes from week 18 to baseline in serum NfL in rapamycin (c) and placebo arm (d), in relation to progression rate calculated as the monthly decline in the ALSFRS-R from baseline to week 18. Individual differences in neurofilament concentration between week 18 and baseline are plotted as colored symbols (Rapamycin arms in red; placebo arm in blue). The shaded areas represent the 95% confidence intervals around the model estimates. The lines and confidence intervals are drawn from the actual distributions of linear model fits. All statistical tests were two-tailed. Source data are provided as a Source Data file.
There were no statistically significant differences between patients treated with rapamycin and placebo as far as PEG (19.0% in the placebo group, 14.3% in the rapamycin group, p = 0.664) or NIV positioning (28.6% in the placebo group, 19.0% in the rapamycin group, p = 0.468) are concerned. During the study 7.1% of patients treated with rapamycin and 4.8% of patients in the placebo group had died (p = 0.672). There was only one IV in the placebo group. The most common cause of death was respiratory failure, accounting for three of the four deaths, a finding consistent with the natural history of ALS.
A post-hoc analysis on tracheostomy-free survival with last observation set on 31st December 2021, showed that 52.4% of patients treated with rapamycin and 61.9% of patients in the placebo group had died or underwent tracheostomy, not a statistically significant difference (p = 0.356) (Table S11, Fig. S4). There were no significant differences in the mean absolute change from baseline to each time point in the FVC (at week 18, MD 3.20, 97.5%CI −9.798 to 16.19, p = 0.571 for rapamycin 1 mg/m2/d group, and MD −4.57, 97.5%CI −17.56 to 8.42, p = 0.419 for rapamycin 2 mg/m2/d group) (Table S12, Fig. S5). There were no significant differences in the mean absolute change from baseline to each time point in the ALSAQ40 scores (Tables S13, S14). ALSAQ40 main questions scores are represented in Fig. 5. Correlations among clinical outcome measures are presented in Table S15.
A–E These show the treatment-dependent mean scores of ALSAQ40 physical mobility, Activity Daily Living (ADL) and independence, eating and drinking, communication, and emotional functioning main questions, respectively, from baseline to study end. Panel F shows ALSAQ40 mean total score from baseline to study end (red = R1, rapamycin 1 mg/m2/d, violet = R2, rapamycin 2 mg/m2/d, blue = P, placebo; Intention to treat population). Source data are provided as a Source Data file.
Safety and drug adherence
A total of 23 over 42 individuals (55%) in the rapamycin group and 11 over 21 individuals (52%) in the placebo group had one or more AEs during the trial (Table S16). The total number of reported AEs was 23 for placebo arm, and 46 for the rapamycin arms. Severe AEs (SAEs) were 7 in the placebo group (30.4% of total AEs in that group), and 9 in the rapamycin groups (19.6% of total AEs in those groups) (Table S17). Among the totality of AEs, four caused treatment discontinuation (one in the placebo group and three in the rapamycin groups) (Table S17).
Individuals with SAEs were 19% both in the placebo and in the rapamycin groups (Table S18).
Events occurring at a greater frequency in the rapamycin group were primarily skin and subcutaneous tissue disorders (erythema, pruritus, rash, conjunctivitis, dermatitis, eczema), then gastrointestinal disorders, injuries, respiratory disorders, headache and psychiatric disorders (Table 5).The majority of SAEs were represented by complications related to disease progression, such as dysphagia and hospital admissions to undergo PEG positioning, or respiratory failure/pneumonia. One subject committed suicide.
Of note, there was a case of acute hepatitis probably related to the study drug, that occurred in a subject allocated to the rapamycin 2 mg/m2/day treatment arm, who had a very high peak of rapamycin at first blood dosage and for whom dose reduction applied. No permanent consequences had been reported from this event.
A total of 11% of the participants dropped out during the study treatment, 5% in the placebo group and 14.3% in the rapamycin group. During follow up time, 24% of patients in each group abandoned the study. Events leading to discontinuation of the treatment are presented in Table S19. Data on adherence to the trial regimen are summarized in Table S20.
Drug dosage assessment
Plasma levels of rapamycin at different time points for each treatment arm are displayed in Fig. S6. While patients who were allocated to rapamycin 1 mg/m2/day treatment had a quite stable plasma dosage, patients allocated to the other treatment arm, during the first dosages presented an initial peak, with plasma levels above the upper limit of therapeutic range and also of the safety threshold (15 ng/ml), that next required drug re-dosing either to 1 mg/m2/day arm or to the minimal dose of 1 mg/day. Finally, examining CSF samples of patients at week 18 by LC-MS/MS a clear peak corresponding to Sirolimus was not detected neither in the CSF of treated patients, nor in the CSF of placebo (Fig. S7).