Allogeneic mesenchymal stem cell therapy with laromestrocel in mild Alzheimer's disease: a randomized controlled phase 2a trial

Abstract

Alzheimer's disease (AD) is characterized by progressive cognitive decline, severe brain atrophy and neuroinflammation. We conducted a randomized, double-blind, placebo-controlled, parallel-group phase 2a clinical trial that tested the safety and efficacy of laromestrocel, a bone-marrow-derived, allogeneic mesenchymal stem-cell therapy, in slowing AD clinical progression, atrophy and neuroinflammation. Participants across ten centers in the United States were randomly assigned 1:1:1:1 to four infusion groups: group 1 (placebo; four monthly infusions, n = 12); group 2 (25 million cells, one infusion followed by three monthly infusions of placebo, n = 13); group 3 (25 million cells; four monthly doses, n = 13); and group 4 (100 million cells; four monthly doses, n = 11). The study met its primary end point of safety; the rate of treatment-emergent serious adverse events within 4 weeks of any infusion was similar in all four groups: group 1, 0% (95% CI 0-26.5%); group 2, 7.7% (95% CI 0.2-36%); group 3, 7.7% (95% CI 0.2-36%) and group 4, 9.1% (95% CI 0.2-41.3%). Additionally, there were no reported infusion-related reactions, hypersensitivities or amyloid-related imaging abnormalities. Laromestrocel improved clinical assessments at 39 weeks compared to placebo, as measured by a composite AD score (secondary end point was met: group 2 versus placebo change: 0.38; 95% CI -0.06-0.82), Montreal cognitive assessment and the Alzheimer's Disease Cooperative Study Activities of Daily Living. At 39 weeks, Laromestrocel slowed the decline of whole brain volume compared to placebo (n = 10) by 48.4% for all treatment groups combined (groups 2-4: P = 0.005; n = 32) and left hippocampal volume by 61.9% (groups 2-4, P = 0.021; n = 32), and reduced neuroinflammation as measured by diffusion tensor imaging. The change in bilateral hippocampal atrophy correlated with the change in mini-mental state exam scores (R = 0.41, P = 0.0075) in all study patients (N = 42). Collectively these results support safety of single and multiple doses of laromestrocel treatment for mild AD and provide indications of efficacy in combating decline of brain volume and potentially cognitive function. Larger-scale clinical trials of laromestrocel in AD are warranted. ClinicalTrials.gov registration: NCT05233774 .

Fig. 1. Consort diagram depicting study design.

Full analysis set (FAS): N = 120; per-protocol population (PP): N = 41; ITT population: N = 49; mITT population: N = 48.

Fig. 2. Clinical assessments.

Groups 1–4 are color coded: group 1 (placebo; black), group 2 (25 M × 1; blue), group 3 (25 M × 4; salmon) and group 4 (100 M × 4; magenta). Two-sided P values are associated with testing for statistically significant differences in least-squares (LS) means between laromestrocel and placebo obtained from a mixed model for repeated measures (MMRM) (F-statistic). No adjustment for multiple comparisons was made. Statistical significance was pre-set at P < 0.10 for the CADS scores. For CADS, the placebo group showed a trending decline in CADS scores at 39 weeks (a), while treatment groups showed numerical improvement relative to placebo, reaching statistical significance at 39 weeks (group 2, P = 0.091; n = 11) compared to placebo (n = 10). Statistical significance was met for MoCA at 39 weeks versus placebo (n = 10) (group 2, P = 0.009; n = 11) (b) as well as for combined groups 2–4, P = 0.015; n = 32). MMSE-2 (group 4, P = 0.067, n = 10) (c) and CDR-SB (group 4, P = 0.270; n = 10) (d) scores showed numerical improvement for all treatment groups at week 39 compared to placebo (n = 10), while ADAS-cog13 measures (combined groups 2–4, P = 0.780; n = 33) were not statistically different from placebo (n = 10) (e). f–i, For QoL measures, statistical significance was met with group 4 for ADCS-ADL at 39 weeks (group 4, P = 0.04; n = 10) compared to placebo (n = 10) (f), while group 2 (P = 0.228, n = 12) and group 3 (P = 0.103, n = 11) showed numerical improvement relative to placebo (n = 10). Numerical improvement was observed in group 4 at 39 weeks for QoL-AD (caregiver) (P = 0.142, n = 10) compared to placebo (n = 10) (g), but no improvement was found in QoL-AD (study participant) scores (combined groups 2–4, P = 0.797, n = 33) compared to placebo (n = 10) (h) while a numeric trend towards improvement was found in group 4 (P = 0.275, n = 10) at 39 weeks for ADRQL scores compared to placebo (n = 10) (i). Error bars represent ±s.e.m.

Fig. 3. Volumetric MRI assessments of AD patients treated with laromestrocel versus placebo.

Two-sided P values are associated with testing for statistically significant differences in LS means between laromestrocel and placebo obtained from a MMRM (F-statistic). No adjustment for multiple comparisons was made. a, vMRI images depicting changes in brain volume in group 1 (placebo) versus group 4 (high repeated dose) at week 39 (increase, yellow; decrease, blue). Insets in a are higher magnification regions of the boxed regions for groups 1 and group 4. b, Whole brain volume steadily decreased by ~1.2% in the placebo group during the 39-week trial period, whereas treatment groups showed statistically significant slowing of whole brain atrophy (group 3, P = 0.006, n = 11; group 4, P = 0.009, n = 10) compared to placebo (n = 10). c, Gray matter volume (left) declined in placebo by about 2.2% at week 39, whereas treatment group 4 (P = 0.069, n = 10) showed numerical improvement compared to placebo (n = 10). d, Lateral ventricles enlarged in placebo over time, but all treatment groups trended toward improvement at all time points, compared to placebo (n = 10), reaching statistical significance for group 3 (right ventricle) at week 39, (P = 0.042; n = 11). e, Bilateral hippocampal volume similarly decreased by approximately 1.6% in placebo (n = 10) while treatment groups showed less decline (group 2, P = 0.029, n = 11; group 3, P = 0.028, n = 11). f, Left temporal cortex showed progressive decline in volume in the placebo group (n = 10), but partial rescue with group 2 (P = 0.042; n = 11) and group 4 (P = 0.028; n = 10). g, Left medial temporal cortex also showed progressive decline in volume in the placebo group (n = 10) while treated groups showed slower decline (group 3, P = 0.001; n = 11; group 4, P = 0.032; n = 10). Units are indicated as raw non-normalized units ratio to intracranial volume (×10⁻³). h, Patients with AD receiving placebo showed increasing cingulate cortex diffusion tensor imaging (DTI) values for mean diffusivity (MD) indicating increasing inflammation and disease progression, whereas all groups receiving laromestrocel showed numerically improved MD at 39 weeks, reaching statistical significance for group 2; P = 0.048; n = 11. Error bars represent ±s.e.m.

Fig. 4. Reduction of bilateral hippocampal, whole brain atrophy correlates with improved MMSE scores.

a,b, Pearson correlation analysis showing correlation between bilateral hippocampal (a) and whole brain (b) atrophy reduction and improved MMSE-2 scores at week 39 (R = 0.41; P = 0.0075; n = 42 and R = 0.35; P = 0.0227; n = 42, respectively). Correlation of reduced bilateral ventricular enlargement with improved MMSE-2 scores was also evident (R = −0.35; P = 0.0213; n = 42) (c). Blood serum collected at the indicated time points was analyzed by Meso Scale Discovery (MSD) for TIE2 (d). Two-sided P values are associated with testing for statistically significant differences in LS means between laromestrocel and placebo obtained from a MMRM (F-statistic). No adjustment for multiple comparisons was made. Soluble TIE2 levels are represented as raw CFB and were reduced in treatment groups compared to placebo, particularly during the treatment period (to week 16) and reached statistical significance in group 3 at week 4 (P = 0.010; n = 13), week 8 (P = 0.020; n = 12) and week 16 (P = 0.045; n = 12). TIE2 levels were significantly increased in the placebo group (n = 11) at week 26 compared to baseline (P = 0.047; n = 11). Error bars represent ±s.e.m.

Extended Data Fig. 1. Volumetric MRI assessments of thalamus and frontal cortex of AD patients treated with laromestrocel vs placebo.

Two-sided p values are associated with testing for statistically significant differences in LS means between laromestrocel and placebo obtained from a MMRM (F-statistic). No adjustment for multiple comparisons was made. Statistically significant slowing of right thalamus atrophy compared with placebo (N = 10) was observed in Group 4 at week 39 (p = 0.026; N = 10), with numeric slowing in Groups 2 and 3, while no statistically significant change was observed in left thalamic volume (a, b). Group 4 showed statistically significant slowing of frontal cortical atrophy on both the left (p = 0.028) and right sides (p = 0.002; N = 10) compared with placebo (N = 10) (c, d). Error bars represent ± SEM.

Extended Data Fig. 2. Example flow cytometry workflow for CD19 (negative marker) and CD105 (positive marker) for identity assessment of laromestrocel.

Ungated cells (a, b) were subcategorized through detection of live cells using the live/dead marker 7-AAD (c, d). Single stained cells were distinguished by gating 0.01% of non-stained cells (e, f). Laromestrocel lots must be positive for cell surface markers, CD105, CD90, and CD73 ( ≥ 95%). Laromestrocel must be negative ( ≤ 2%) for CD45, CD11b, CD19, and ≤5% for CD34. Abbreviations: SSC-H, side scatter height; FSC-H, forward scatter height; SSC-A, side scatter area.