Intramyocardial cell-based therapy with Lomecel-B during bidirectional cavopulmonary anastomosis for hypoplastic left heart syndrome: the ELPIS phase I trial

Abstract

Aims: Hypoplastic left heart syndrome (HLHS) survival relies on surgical reconstruction of the right ventricle (RV) to provide systemic circulation. This substantially increases the RV load, wall stress, maladaptive remodelling, and dysfunction, which in turn increases the risk of death or transplantation.

Methods and results: We conducted a phase 1 open-label multicentre trial to assess the safety and feasibility of Lomecel-B as an adjunct to second-stage HLHS surgical palliation. Lomecel-B, an investigational cell therapy consisting of allogeneic medicinal signalling cells (MSCs), was delivered via intramyocardial injections. The primary endpoint was safety, and measures of RV function for potential efficacy were obtained. Ten patients were treated. None experienced major adverse cardiac events. All were alive and transplant-free at 1-year post-treatment, and experienced growth comparable to healthy historical data. Cardiac magnetic resonance imaging (CMR) suggested improved tricuspid regurgitant fraction (TR RF) via qualitative rater assessment, and via significant quantitative improvements from baseline at 6 and 12 months post-treatment (P < 0.05). Global longitudinal strain (GLS) and RV ejection fraction (EF) showed no declines. To understand potential mechanisms of action, circulating exosomes from intramyocardially transplanted MSCs were examined. Computational modelling identified 54 MSC-specific exosome ribonucleic acids (RNAs) corresponding to changes in TR RF, including miR-215-3p, miR-374b-3p, and RNAs related to cell metabolism and MAPK signalling.

Conclusion: Intramyocardially delivered Lomecel-B appears safe in HLHS patients and may favourably affect RV performance. Circulating exosomes of transplanted MSC-specific provide novel insight into bioactivity. Conduct of a controlled phase trial is warranted and is underway.Trial registration number NCT03525418.

Keywords: Hypoplastic left heart syndrome; Lomecel-B; allogeneic medicinal signalling cells; mesenchymal stem cells; mesenchymal stromal cells.

© The Author(s) 2023. Published by Oxford University Press on behalf of the European Society of Cardiology.

Figures

Structured Graphical Abstract

HLHS, hypoplastic left heart syndrome; MSCs, medicinal signalling cells; RV, Right Ventricular; RA, Right Atrium; SVC, Superior Vena Cava; RPA, Right Pulmonary Artery; LPA, Left Pulmonary Artery

Figure 1

Study timeline and study flow, phase I of the ELPIS study. (A), Study timeline. (B), Study flow. Patients who met the eligibility criteria underwent BDCPA and intramyocardial injections, then were followed up with CMR at 6 and 12 months.

Figure 2

Cardiac function at baseline, 6 months follow-up, and 12-month follow-up in Lomecel-B treated patients (n = 10). (A), Right ventricular ejection fraction did not change over the follow-up period compared to baseline. (B), Right ventricle stroke volume index did not change over the follow-up period compared to baseline. (C), Right ventricle diastolic diameter did not change over the follow-up period compared to baseline. (D), Right ventricle systolic diameter did not change over the follow-up period compared to baseline. (E), Global longitudinal strain did not change over the follow-up period compared to baseline. (F), Tricuspid regurgitation appeared to decrease over the follow-up period compared to baseline.

Figure 3

Tricuspid regurgitation (TR) at baseline, 6-month, and 12-month. (A), TR was assessed by CMR-reader interpretation as mild, moderate, or severe Lomecel-B treated patients (n = 10). (B), TR regurgitant fraction (TR RF) was quantitatively measured from CMR for combined run-in and phase 1 patients. Reported as mean +/− standard deviation. * Indicates P < 0.05.

Figure 4

Baseline and follow-up data monitoring patients throughout ELPIS phase I trial Lomecel-B treated patients (n = 10). (A-B), Serum BNP increased following day two post-treatment but decreased following week twenty-four post-treatment. No changes were statistically significant. Error bars indicate standard deviation. (C), Somatic growth of patients at baseline and at follow-up. Weight for age z-score increased from baseline to 6 months. Weight for age z-score did not differ between baseline and 6 or 12 months. (D), Length for age z-score increased from baseline to 6 months. The length for age z-score significantly decreased from 6-month to 12-months (P = 0.0136) (E), HLA class I reactions significantly decreased from baseline to 26 weeks post-treatment. (F), HLA class II reactions showed a trend towards decreasing after treatment, but this did not reach statistical significance. Only patients with HLA data at baseline and 26 weeks are represented (n = 8). * Indicates P < 0.05. BNP, brain natriuretic peptide.

Figure 5

Unsupervised analysis of MSC-derived exosome RNA. (A), Plasma exosomes from days 2 and 7 were assessed for miRNA and total RNA content with microarray. RNA expression was related to CMR 6 months-to-1 year fold change values. (B), Volcano plot of RNA expression fold change from days 2 to 7. Few miRNAs (circles) and RNAs (squares) upregulated in days 2 and 7 are highlighted in orange and blue, respectively. Principal component analysis of days 2 and 7 miRNAs (C), and RNAs (D). Merged dynamic clustering from weighted correlation network analysis (WGCNA) detected 25 modules of highly correlated day 7 miRNAs (E) and RNAs (F). Module correlation to CMR outcomes (6 months-to-1 year fold change) are depicted in heatmaps (red = high, positive correlation). Modules significantly correlated to an CMR outcome (P < 0.1) are marked. (G), Enriched gene ontology (GO) parent terms of WGCNA RNA modules negatively correlated to TR RF (M2 and M4) and gene targets of miRNA modules positively correlated to TR RF (M12, M20, M24). (H), A network of experimentally validated gene targets of miRNAs positively correlated to TR RF (M12, M20, M24). Enriched terms are coloured by cluster (Metascape).

Figure 6

Partial least-squares regression (PLSR) modelling of day 7 exosome RNA and identification of RNA signals related to TR RF. A two-component model was trained using the top 200 RNA variables of importance for the model projection (VIPs): R2(RNA) = 0.927, R2(CMR Outcomes) = 0.865. (A), Total number of RNAs in the model was reduced to 200 from the initial set of 16 351. A higher proportion of the VIP RNAs are miRNAs. (B), Scores plot of components 1 and 2 from the PLSR model trained with day 7 top 200 VIP RNAs. Components 1 and 2 explain 80% and 13% of the variance, respectively. (C), Loadings plots of components 1 and 2 from the PLSR model show the VIP RNAs covarying with CMR outcomes. (D), PLSR model predictions of CMR outcomes correlated with overserved measurements. RNAs related to TR RF were determined from WGCNA and PLSR analyses. Pearson correlation P-values for RVEF, RV SV, RV Mass, GLS, and TR RF were P < 0.0001, P = 0.0067, P = 0.0005, P = 0.0881, and P = 0.0021 respectively. (E), Venn diagram of WGCNA modules significantly correlated to TR RF (P < 0.001, miRNA M12, M20, M25 and RNA M1, M5) and PLSR 200 VIP RNAs. 54 TR RF-related RNAs identified from the intersection. (F), Top 10 positive and negative weighted coefficients from TR RF-related RNAs were determined with PLSR. (G), Pathway analysis of 54 TR RF-related RNAs (Metascape).