Author + information
- Received September 12, 2017
- Revision received November 16, 2017
- Accepted November 20, 2017
- Published online January 22, 2018.
- Philippe Menasché, MDa,b,c,∗ (, )
- Valérie Vanneaux, PharmDd,e,
- Albert Hagège, MDb,c,f,
- Alain Bel, MDa,
- Bernard Cholley, MDb,g,
- Alexandre Parouchev, PhDd,e,
- Isabelle Cacciapuoti, MScd,e,
- Reem Al-Daccak, PhDh,
- Nadine Benhamouda, MSci,
- Hélène Blons, PhDj,
- Onnik Agbulut, PhDk,
- Lucie Tosca, PhDl,
- Jean-Hugues Trouvin, PharmDm,n,
- Jean-Roch Fabreguettes, PharmDo,
- Valérie Bellamy, BAScc,
- Dominique Charron, MDp,q,
- Eric Tartour, MDb,c,i,
- Gérard Tachdjian, MDl,
- Michel Desnos, MDb,c,f and
- Jérôme Larghero, PhDd,e,q
- aDepartment of Cardiovascular Surgery, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
- bUniversity Paris Descartes, Sorbonne Paris Cité, Paris, France
- cNational Institute of Health and Medical Research (INSERM) U970, Hôpital Européen Georges Pompidou, Paris, France
- dCell Therapy Unit, Assistance Publique-Hôpitaux de Paris, Hôpital Saint-Louis, Paris, France
- eINSERM, Clinical Investigation Center in Biotherapies (CBT-501) and U1160, Institut Universitaire d’Hématologie, Hôpital Saint-Louis, Paris, France
- fDepartment of Cardiology, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
- gDepartment of Anesthesiology and Intensive Care, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
- hINSERM U976, Institut Universitaire d’Hématologie, Hôpital Saint-Louis, University Paris Diderot, Sorbonne Paris Cité, Paris, France
- iDepartment of Biological Immunology, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
- jINSERM Mixed Research Units (UMR)-S1147, National Scientific Research Center (CNRS) Non CNRS Structure 5014, Sorbonne Paris Cité, Department of Biochemistry, Pharmacogenetic and Molecular Oncology Unit, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
- kSorbonne Universités, Université Pierre et Marie Curie, University Paris-6, Institut de Biologie Paris-Seine, UMR CNRS 8256, Biological Adaptation and Ageing, Paris, France
- lAssistance Publique-Hôpitaux de Paris, University Paris Sud, Histology-Embryology-Cytogenetics, Hôpitaux Universitaires Paris Sud, Clamart, France
- mSchool of Pharmacy, University Paris Descartes, Paris, France
- nCentral Pharmacy, Pharmaceutical Innovation Department, Assistance Publique-Hôpitaux de Paris, Paris, France
- oCentral Pharmacy, Clinical Trials Department, Assistance Publique-Hôpitaux de Paris, Paris, France
- pHuman Leukocyte Antigen and Médecine, Hôpital Saint-Louis, INSERM U976, Paris, France
- qUniversity Paris Diderot, Sorbonne Paris Cité, Paris, France
- ↵∗Address for correspondence:
Dr. Philippe Menasché, Department of Cardiovascular Surgery, Hôpital Européen Georges Pompidou, 20, rue Leblanc, 75015 Paris, France.
Background In addition to scalability, human embryonic stem cells (hESCs) have the unique advantage of allowing their directed differentiation toward lineage-specific cells.
Objectives This study tested the feasibility of leveraging the properties of hESCs to generate clinical-grade cardiovascular progenitor cells and assessed their safety in patients with severe ischemic left ventricular dysfunction.
Methods Six patients (median age 66.5 years [interquartile range (IQR): 60.5 to 74.7 years]; median left ventricular ejection fraction 26% [IQR: 22% to 32%]) received a median dose of 8.2 million (IQR: 5 to 10 million) hESC-derived cardiovascular progenitors embedded in a fibrin patch that was epicardially delivered during a coronary artery bypass procedure. The primary endpoint was safety at 1 year and focused on: 1) cardiac or off-target tumor, assessed by imaging (computed tomography and fluorine-18 fluorodeoxyglucose positron emission tomography scans); 2) arrhythmias, detected by serial interrogations of the cardioverter-defibrillators implanted in all patients; and 3) alloimmunization, assessed by the presence of donor-specific antibodies. Patients were followed up for a median of 18 months.
Results The protocol generated a highly purified (median 97.5% [IQR: 95.5% to 98.7%]) population of cardiovascular progenitors. One patient died early post-operatively from treatment-unrelated comorbidities. All others had uneventful recoveries. No tumor was detected during follow-up, and none of the patients presented with arrhythmias. Three patients developed clinically silent alloimmunization. All patients were symptomatically improved with an increased systolic motion of the cell-treated segments. One patient died of heart failure after 22 months.
Conclusions This trial demonstrates the technical feasibility of producing clinical-grade hESC-derived cardiovascular progenitors and supports their short- and medium-term safety, thereby setting the grounds for adequately powered efficacy studies. (Transplantation of Human Embryonic Stem Cell-derived Progenitors in Severe Heart Failure [ESCORT]; NCT02057900)
The current treatment of heart failure rests on 2 major pillars: drugs; and interventional, surgical, or catheter-based procedures. These therapies are either palliative or, at the other extreme, radical (cardiac replacement). In the past 2 decades, an intermediate strategy has emerged that aims at repairing the diseased heart by transplanting cells.
The phenotype of the “ideal” cells is still unsettled. However, there has been a recent shift toward cells committed to a cardiac lineage. In this setting, clinical trials have tested right atrial c-Kit cardiac stem cells (1) and right ventricle–derived cardiospheres (2) that have a derivation from adult tissues in common. We adopted a different strategy consisting of starting “upstream” from human embryonic stem cells (hESCs) and leveraging their intrinsic pluripotentiality to drive their fate toward a cardiovascular lineage. Several experimental studies have validated the concept that ESC-derived cardiomyocytes improve the function of infarcted hearts (3). Taking advantage of the knowledge that ESCs can be used at a defined stage of differentiation, we selectively generated hESC-derived cardiovascular progenitor cells, primarily defined by the coexpression of the stage-specific embryonic antigen-1 (SSEA-1) and cardiac transcription factor Isl-1 markers, whereas the recognition that biomaterials can efficiently promote cell engraftment (4) led us to incorporate these progenitors in a fibrin scaffold that was delivered onto the epicardium of the infarct area. The safety and efficacy outcomes of this approach, supported by almost a decade of preclinical studies in small and large animal models, have been sufficient (5) to allow regulatory approval for this safety trial (6).
Patients and procedures
Six patients were selected on the basis of 3 major inclusion criteria: 1) left ventricular (LV) systolic dysfunction reflected by an ejection fraction (EF) ≥15% and ≤35%, as assessed by echocardiography; 2) a history of myocardial infarction that had occurred at least 6 months before screening; and 3) an indication for surgical coronary revascularization. The other inclusion and exclusion criteria are reported in the Online Appendix. The protocol was approved by the Ethics Committee of Paris Descartes University (2010-A00794-35), and written informed consent was obtained from all patients.
The technique of cellularized patch preparation (7) is detailed in the Online Appendix. Briefly, expanded pluripotent hESCs were committed to the cardiovascular lineage by a 4-day exposure to 2 cytokines, immunomagnetically sorted for a positive expression of SSEA-1 (a marker for loss of pluripotency) and finally mixed with fibrinogen and thrombin to form a gel (Figure 1A), which could be easily manipulated.
All surgical procedures were performed with the patient under normothermic blood antegrade or retrograde cardioplegic arrest, except for 1 patient, whose left anterior descending coronary artery was bypassed off-pump on the beating heart. On completion of the distal coronary artery anastomoses, a piece of autologous pericardium, matching approximately that of the fibrin patch (20 cm2), was sutured around one-half the circumference of the infarct area, thereby creating a “pocket” into which the fibrin patch was slid (Figure 1B). The pericardial flap was then folded over and finally stitched to the remaining one-half of the infarct circumference (Figure 1C). The transplanted segments were marked on a 17-segment map matching the echocardiographic division of the left ventricle. Corticosteroids (methylprednisolone) were given intraoperatively at a total dose of 240 mg, split into 2 equal injections, and an additional 240-mg dose was given over the first 24 post-operative hours as a 60-mg injection every 4 h. All patients were also given cyclosporine (the dosing was adjusted on serial measurements of drug trough levels with a target concentration of 100 to 150 ng/ml) and mycophenolate mofetil (2 g/day). The duration of this treatment, initially planned for 2 months, was shortened to 1 month from the second patient onward. The basis of immunomonitoring was the repeated post-operative detection of human leukocyte antigen (HLA) antibodies against donor-specific antibody (DSA) and interferon-γ on Elispot assays (Diaclone, Besancon, France).
The primary endpoint was safety, assessed on: 1) intraoperative events; 2) arrhythmias, assessed by serial interrogations of the implantable cardioverter-defibrillators (ICDs) inserted in all patients pre-operatively, with a special focus on the detection of the most relevant events such as ventricular fibrillation or sustained ventricular tachycardia; 3) cardiac teratoma or remote tumor tracked by whole body computed tomography (CT) and fluorine-18 deoxyglucose positron emission tomography (PET) scans performed at baseline and repeated at 6 months (PET scan) and 12 months (CT scan) post-operatively; 4) MACE, defined as the composite of cardiovascular- and noncardiovascular-related death, myocardial infarction, congestive heart failure, resuscitated sudden death, and stroke occurring until 1 year of follow-up; and 5) complications of the immunosuppressive treatment. Detection of circulating donor cell-specific DNA was also performed at 4 and 10 days post-operatively (Online Appendix).
The secondary endpoints were feasibility and efficacy. Feasibility was assessed by the ability of generating hESC-derived cardiovascular progenitor cells meeting the following pre-specified criteria: sterility; viability >90%; purity reflected by a percentage of SSEA-1+ cells >95%; identity, reflected by polymerase chain reaction–based down-regulation (<0.1%) of the pluripotency marker Nanog and up-regulation (>5%) of the cardiac transcription factor Isl-1 relative to the undifferentiated hESC line; and morphology of the patch after the polymerization step.
Efficacy was assessed by transthoracic 2-dimensional echocardiography at baseline and then at 1, 3, 6, and 12 months post-operatively (Online Appendix). Global LV function was assessed on EF and volumes, and assessment was completed by scoring the wall motion of the cell-treated segments. Patients were also assessed for their functional status by clinical examinations, the 6-min walk test, and the Quality of Life Scale (using the Minnesota Living with Heart Failure Questionnaire) at 1 year post-operatively.
Given the small sample size, continuous variables were conservatively compared by nonparametric tests. A mixed model analysis of variance (ANOVA) on ranks was used to compare LV volumes, EF, and the wall motion score of the cell/patch-treated segments across the several time points at which these data were recorded during the follow-up period. Data are reported as median values with the interquartile range (IQR) or mean ± SD. A p value <0.05 was considered statistically significant. The study is registered with ClinicalTrials.gov (Transplantation of Human Embryonic Stem Cell-derived Progenitors in Severe Heart Failure [ESCORT]; NCT02057900).
Six patients (1 female and 5 male) were enrolled from May, 2013 to December, 2016. The median age was 66.5 years (IQR: 60.5 to 74.7 years). At the time of this report, the median follow-up was 18 months (IQR: 8.5 to 21.5 months).
The age of the infarctions ranged from 6 months to 30 years, and the median values for pre-operative LVEF, LV end-diastolic volume, and LV end-systolic volume were 26% (IQR: 22% to 32%), 155 ml (IQR: 148 to 161 ml), and 117 ml (IQR: 99 to 126 ml), respectively. Two and 4 patients received a single and a double bypass graft, respectively. At least 1 mammary arterial conduit was used in all cases. The epicardial delivery of the cell-laden patch and its coverage with the pericardial flap were uneventfully performed in a few minutes.
Highly purified (97.5% [IQR: 95.5% to 98.7%]) SSEA-1+ Isl-1 cardiovascular progenitors (Figure 2) were delivered at a median dose of 8.2 million (IQR: 5 to 10 million) (Online Table 1). Immunologically, at steady state, SSEA-1+ progenitors expressed HLA I antigens but were largely negative for HLA II and costimulatory or regulatory molecules (Online Figure 1). Pre-operative mixed lymphocyte reaction assays demonstrated that SSEA-1+ progenitors are weakly immunogenic. Four of the patients failed to elicit a significant allogeneic T-cell response against HLA-mismatched SSEA-1+ progenitors, whereas the remaining 2 patients only showed a weak T-cell–proliferative response (Online Figure 2).
All patients had an uneventful post-operative course, except for the first patient who died shortly after the operation. This 77-year-old man with a large anterior myocardial infarction was in New York Heart Association functional class III and had an LVEF of 20%, severe triple-vessel coronary artery disease, and several comorbidities including chronic obstructive pulmonary disease, atrial fibrillation, type 2 diabetes, peripheral arterial disease, and obesity. Seven days after an uneventful surgical procedure, he needed reintubation for cardiorespiratory distress and died 2 days later. At autopsy, the 700-g heart displayed a sequela of the old infarction and severe LV hypertrophy. The patch was identified without any noticeable inflammatory response. The Data Safety Monitoring Board concluded that this death was not treatment related and allowed the trial to continue.
For the 5 operative survivors, outcomes measures are reported at 1 year, except for the last patient, whose most recent follow-up visit was at 6 months. None of these patients has experienced ventricular arrhythmias, such as fibrillation or sustained tachycardia, as assessed by the repeated interrogations of the ICDs. Nonsustained ventricular tachycardia of more than 5 beats was recorded in only 1 patient at 1 month post-operatively. Similarly, whole body PET and CT scans did not show abnormalities suggestive of cardiac teratoma or off-target adverse proliferation. Only 1 patient, who was 81 years old at surgery and had a pre-operative LVEF of 22%, had major adverse cardiac events in the form of recurrent episodes of heart failure, which ultimately led to his death, precipitated by a mechanical fall, 22 months post-operatively.
Three patients developed donor-specific low-level DSA. There were no complications related to the immunosuppressive regimen. The mean daily dose of cyclosporine was 163 mg ± 51 mg/day and resulted in serum drug levels averaging 102 ± 46 ng/ml. This relatively low level of immunosuppression was confirmed by the antiviral response assessed by an ex vivo interferon-γ Elispot assay with immunodominant viral peptides (Online Appendix). There was no donor cell–derived DNA detected in patients’ plasmas at a ratio reaching significance (allelic ratio >0.005).
The 4 patients assessed at the 1-year follow-up time point reported a symptomatic improvement reflected by a decrease in New York Heart Association functional class from a baseline median value of III (IQR: 3.0 to 3.0) to I/II (IQR: 1.0 to 2.2), a decrease in the Minnesota Living with Heart Failure Questionnaire from 69 (IQR: 67 to 72) to 42 (IQR: 34.7 to 50) and a median increase in the 6-min walk test of 23.5 m (IQR: 19.0 to 20.0 m). During follow-up, there was a modest decrease in LV volumes and an increase in LVEF, which rose from 26% (IQR: 22% to 32%) at baseline to 38.5% (IQR: 33.5% to 41%) at 1 year, but their evolution over time was not significant by ANOVA (Online Table 2). Conversely, the wall motion of the cell/patch-treated segments significantly improved during follow-up (p = 0.004 by the mixed model ANOVA on ranks) (Central Illustration), with a score that decreased from 4.2 ± 0.8 at baseline to 2.5 ± 0.4 at 1 year. Of note, 3 of the 4 patients who contributed these 1-year data had no revascularization of the treated segments; the bypass conduits were placed in remote myocardial areas. In the most recent patient, the wall motion score of the nonrevascularized cell-treated segments also decreased from 4.3 to 3.0 at the 6-month study point.
The major finding of this study is that hESCs can be differentiated in clinical-grade cardiovascular progenitors, which seem to be safe once transplanted in patients with severe ischemic LV dysfunction.
The potential clinical usefulness of hESCs has 2 major assumptions as its basis. First, cardiac-committed cells seem more efficacious than those from an extracardiac lineage for repairing a chronically damaged heart (8,9), thereby supporting the intuitive idea that cardiac repair is best served by cells phenotypically close to those they are intended to replace or rescue. Second, adult sources of cardiac cells are still fraught with uncertainties because c-Kit cells are rather credited to give rise to vascular cells (10), whereas it was recently announced (Capricor Therapeutics, May 12, 2017) that the ongoing phase III trial testing cardiospheres was unlikely to meet its primary endpoint. In this context, the investigation of hESC cardiac cells seems fully justified, in particular because they offer the unique possibility to control the sequential steps of their differentiation down the cardiovascular lineage and thus to select precisely the commitment stage at which they can be transplanted. Our choice of using cells at an early stage was dictated by the observations that progenitor cells have a plasticity endowing them with a trilineage differentiation potential (cardiomyocytes, endothelial and smooth muscle cells) (5) and seem to display a greater cardiac-repairing capacity than their more mature counterparts (11).
The present study confirms that the scalability and pluripotentiality of hESCs can be leveraged to a clinically usable product. The I6 cell line used for this trial was first expanded under good manufacturing practice conditions. Only 2 passages were required for generating a 600-million cell Master/Working Cell Bank. In-process controls showed that the cells had retained their identity, as demonstrated by the persistent expression of pluripotency markers, did not show any evidence for contamination by viruses or adventitious agents, and remained genetically stable.
The second phase of lineage-specific differentiation requires exposure of the cells to defined cues and has as its basis the principle of recapitulating in vitro the major signaling events in play during embryonic development of the targeted tissue. These instructive signals now seem to be under control, as demonstrated by the clinical trials in which pluripotent ESCs have been driven toward oligodendrocytes for spinal cord injury repair (12), retinal pigment epithelial cells for ocular diseases (13), or insulin-producing cells for diabetes (14). In our trial, cells were exposed to a 2 growth factor–based protocol that committed them to a mesendodermal/cardiac lineage. The major fingerprint of these cells was the expression of the transcription factor Isl-1, thereby reflecting an early progenitor stage committed to the cardiac and vascular lineages (15). To optimize the early retention and survival of the transplanted cells, the cells were delivered in an epicardial patch, an approach shown to be superior to intramyocardial injections (16), and fibrin was selected as the patch material on the basis of its documented efficacy as a cell carrier (17). For the same purpose, the patch was covered by an autologous pericardial flap used as a natural bioreactor intended to provide trophic factors and necessary supportive noncardiac cells (18) to the underlying cellular graft.
Within the time frame of the study, the primary endpoint (i.e., safety) is considered to have been met because there were no complications that could be specifically ascribed to the cellular graft. Three of these complications were particularly scrutinized.
The major risk of ESCs is teratoma originating from residual pluripotent cells that have not responded to the lineage-specific instructive cues and have therefore retained a state of potential uncontrolled proliferation. This highlights the importance of the purification step and of the quality controls to ensure that the differentiated progeny intended for clinical use has been efficiently purged from teratoma-forming cells. The technical challenge associated with the removal of these contaminating cells is strongly lineage dependent. Thus, in the clinical trial of macular degeneration and Stargardt’s macular dystrophy, differentiation of hESCs into retinal pigment epithelial cells resulted in a >99% pure population and as such did not require a specific selection step (19). Similarly, in the study testing ESC-derived oligodendrocytes in patients with spinal cord injury, the percentage of endodermal, mesodermal, and pluripotent cells in the differentiated population was <1% (12). Conversely, the highest specification rate achieved with our cardioinstructive protocol was 64%, thereby requiring a selection step.
To this end, several fail-safe strategies have been developed (reviewed by Hentze et al. ), of which only surface marker-based cell sorting is currently suitable for translational applications. We therefore used immunomagnetic columns to isolate magnetically labeled cells expressing SSEA-1 taken as a surrogate marker for the loss of pluripotency (21). In this safety trial, the maximal dose of cells allowed for delivery was 10 million. Setting the purity threshold at 95% implied that the highest amount of SSEA-1–negative cells that a patient could receive was 500,000. When injected in immunodeficient mice, this dose did not cause intracardiac or off-target tumors (7). These in vivo data were indeed consistent with our in vitro observations that: 1) SSEA-1– negative cells express a mixture of genes including some related to pluripotency, but the latter are also present in other cells such as mesenchymal cells (22) or cord-derived cells (23), which are used clinically and have not raised safety concerns so far; and 2) the SSEA-1+ cells were free from genetic abnormalities, thereby providing an additional safety net against oncogenesis (24). As an ultimate safety check, individual lot release required the expression of the pluripotency marker Nanog to be <0.1% relative to that of the parental undifferentiated cells, an objective that was reached in the 6 patients. Taken together, these data likely explain the lack of abnormalities on the post-operative PET and CT scans.
A second potential complication of cardiac cell therapy is the occurrence of arrhythmias, which have been reported following the intramyocardial delivery of large doses of ESC-derived cardiomyocytes in a nonhuman primate model of myocardial infarction (25). In our study, repeated interrogations of the ICDs inserted pre-operatively in all patients failed to detect clinically significant arrhythmias in any patient throughout the 1-year follow-up. One possible reason could be the use of an epicardial patch reportedly less arrhythmogenic than multiple intramyocardial punctures (26), possibly because it results in limited, if any, physical integration of the grafted cells into the myocardial tissue; this should, in turn, mitigate the risk of their random distribution at the contact interfaces with the host cardiomyocytes, with the resulting inhomogeneity of action potential propagation across these interfaces and the ultimate occurrence of reentries (27). However, in future studies including patients already fitted with an ICD, more extensive documentation of the arrhythmic status during the weeks preceding the surgery would be helpful to interpret better the significance of potential post-operative events.
A third safety concern with ESCs is alloimmunization. On the basis of HLA phenotyping and mixed lymphocyte reaction assays, the SSEA-1+ progenitors were found weakly immunogenic. Our results are in accord with the inherent immune features of various stem/progenitor cells, which shift their signaling capacity linked to T-cell activation toward delivery of signals that likely promote their development, maintenance, and functioning rather than their rejection (28). However, SSEA-1+ progenitors expressed natural killer (NK) cell–activating receptor NKG2D ligands. However, an inflammatory environment, such as that of injured myocardium, can be paradoxically protective against NK cell killing, in part through a higher expression of HLA class I molecules on the transplanted cells (29), and such an up-regulation was actually observed when our SSEA-1+ progenitors were stimulated in vitro by interferon-γ (Online Figure 1). Put together, these data rationalized our choice of a low-dose immunosuppressive regimen. The modest degree of induced immunosuppression was confirmed by the lack of post-operative changes in T-cell responses to viral peptides demonstrated by Elispot assays. This regimen avoided drug-associated side effects, which, by contrast, represented the major complications in the ESC macular degeneration trial (13). The disadvantage of this protocol is that 3 patients developed DSA (against HLA I, HLA II, and, in one case, both) but always at mean fluorescence intensities far below the reported cytotoxicity threshold (30). These alloimmunization events were clinically silent. To optimize the risk-to-benefit ratio further, the immunosuppressive treatment was limited to 2 months and subsequently 1 month, with the premise that the transplanted cells would be likely short-lived and act predominantly through paracrine signaling. The priority was then to optimize their early retention to give them enough time to release signaling molecules. This hypothesis is supported by the observation that although allogeneic cells are expectedly eliminated more rapidly than their syngeneic counterparts, their transient presence shortly after implantation is sufficient to yield equivalent long-term functional benefits (31).
The small sample size, lack of blinded assessment, and confounding effect of the concomitant coronary artery bypass grafting obviously preclude any meaningful conclusion regarding efficacy. Even the slight improvement in wall motion of the cell-treated nonrevascularized myocardial segments should be interpreted cautiously and without any enthusiastic claim, in part because a visual echocardiographic assessment is more prone to yield positive outcomes than magnetic resonance imaging (32). Suffice to say that if cells may have contributed to some extent to the improved contractility of the grafted segments, it is likely through the secretion of biomolecules harnessing host-associated reparative mechanisms (33). These biomolecules seem largely clustered in extracellular vesicles, a hypothesis supported by our experimental findings that the cardioprotective effects of the SSEA-1+ Isl-1+ cardiovascular progenitors used in this trial can be recapitulated by the sole administration of the extracellular vesicles that they release (34). Of note, when we assessed the content of extracellular vesicles released by our clinical fibrin-embedded SSEA-1+ progenitors (n = 2 patches), we found a predominant expression of miR 302, which is reported to stimulate cardiomyocyte proliferation (35). These data open the way to a new paradigm of cell-based acellular therapy whereby pluripotent stem cell–derived differentiated derivatives could be assigned the new role of exclusive in vitro producers of a secretome considered as the “active substance.” Delivery of this sole cargo to patients could streamline the process with regard to manufacturing, reproducibility, storage, quality controls, off-the-shelf availability, and costs. An additional advantage of extracellular vesicle-based therapy would be the possibility of catheter-based delivery, thereby avoiding the invasiveness of a surgical procedure.
First, the conservatively low number of transplanted cells, ethically justified for this trial and with our preclinical nonhuman primate studies as a basis, could have skewed our reassuring safety data, which thus need to be confirmed by dose-escalating studies. Second, although experimental teratomas usually occur within the first weeks after delivery of undifferentiated cells, the delayed occurrence of a tumor cannot still be completely ruled out, and for this reason, our patients will be followed up for 5 years. One also cannot exclude that some highly differentiated tumors may escape fluorine-18 deoxyglucose uptake-based detection. Third, although efficacy was only a secondary endpoint, it is conceivable that better functional outcomes would have been achieved with more mature cardiomyocytes (36), although the greater reliance of these cells on oxidative phosphorylation, as opposed to glycolysis for earlier progenitors (37), may lessen their survival once they are implanted in hostile environments; comparative studies to identify further the optimal differentiation stage at which the cells should be grafted are thus clearly required. Fourth, the basis of the detection of circulating donor-derived DNA was a method validated in our laboratory to detect circulating tumor DNA, and we cannot exclude that if some cells had migrated off the patch, their low number could have fallen beyond the sensitivity of this method. Finally, the patch-based technique used in this trial is obviously limited to intraoperative applications. However, the concept of a cell-embedded scaffold can be extended to catheter-based delivery of injectable materials that polymerize in situ and, therefore, provide cells with a 3-dimensional structural support that can similarly enhance their early retention, provided material degradation products do not cause microembolization.
This trial demonstrates the technical feasibility of producing clinical-grade hESC-derived cardiovascular progenitors and supports their short- and medium-term safety after transplantation in patients with severe post-infarction LV dysfunction. More generally, an additional asset of our program has been the institution of a 3-pronged (amplification, differentiation, purification) good manufacturing practice–compliant production platform for hESCs and their differentiated derivatives that, in the future, should be upgraded through the use of feeder-free cultures in automated bioreactors. This approach could thus serve as a template for applications extending beyond heart failure to areas of regenerative medicine relying on pluripotent stem cell-derived lineage-specific differentiated cells.
COMPETENCY IN MEDICAL KNOWLEDGE: Implantation of cardiac-committed cells has been more successful than transplantation of noncardiac cells to augment the function of failing myocardium. Adult sources of cardiac-committed cells are more limited, however, than pluripotent stem cells that can be driven toward cardiac lineage. It is thus feasible to generate large numbers of cardiovascular progenitor cells and to incorporate them in an epicardially-delivered scaffold covered by the patient’s pericardium used as a natural bioreactor. This technique is associated with low risks of short- and medium-term adverse events.
TRANSLATIONAL OUTLOOK: A similar framework may be applicable to other tissue-specific phenotypes, broadening the spectrum of conditions that might benefit from differentiated derivatives of pluripotent stem cells.
The authors express their thanks to the team of the Clinical Research Department of Hôpital Européen Georges Pompidou for expert data monitoring and study management. They also thank Muriel Tafflet and Marie-Cécile Périer (Paris Cardiovascular Research Center, INSERM U 970) for their expert statistical assistance, the members of the Data Safety Monitoring Board (Professors R. Dion, R. Frank, Y. Juillière, and Ph. Noirhomme), and the 2 French regulatory agencies that contributed to make this trial possible (Agence de la Biomédecine and Agence Nationale de Sécurité du Médicament et des Produits de Santé).
This study was primarily funded by public grants from the French Ministry of Health (Programme Hospitalier de Recherche Clinique PCR11001, P100303) and Assistance Publique-Hôpitaux de Paris, which approved the protocol but did not participate in data collection, analysis and interpretation, and writing of the report. Additional funding was provided by charity grants from the Association Française contre les Myopathies and the Fondation Coeur et Artères, but neither of these bodies was involved in data collection or interpretation. The trial was sponsored by Assistance Publique-Hôpitaux de Paris (Direction de la Recherche Clinique et du Développement). All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- computed tomography
- donor-specific antibody
- ejection fraction
- embryonic stem cell
- human embryonic stem cell
- human leukocyte antigen
- implantable cardioverter-defibrillator
- left ventricular
- left ventricular ejection fraction
- positron emission tomography
- stage-specific embryonic antigen-1
- Received September 12, 2017.
- Revision received November 16, 2017.
- Accepted November 20, 2017.
- 2018 American College of Cardiology Foundation
- Chugh A.R.,
- Beache G.M.,
- Loughran J.H.,
- et al.
- Sharma S.,
- Mishra R.,
- Bigham G.E.,
- et al.
- Tulloch N.L.,
- Muskheli V.,
- Razumova M.V.,
- et al.
- Mummery C.L.,
- Zhang J.,
- Ng E.S.,
- Elliott D.A.,
- Elefanty A.G.,
- Kamp T.J.
- Trivanović D.,
- Jauković A.,
- Popović B.,
- et al.
- Beeravolu N.,
- Khan I.,
- McKee C.,
- et al.
- Hovatta O.,
- Jaconi M.,
- Töhönen V.,
- et al.
- Terajima Y.,
- Shimizu T.,
- Tsuruyama S.,
- et al.
- Chen H.S.,
- Kim C.,
- Mercola M.
- Al-Daccak R.,
- Charron D.
- Hocine H.R.,
- Costa H.E.,
- Dam N.,
- et al.
- Malliaras K.,
- Li T.S.,
- Luthringer D.,
- Terrovitis J.,
- et al.
- Kervadec A.,
- Bellamy V.,
- Harane N El,
- et al.
- Tao L.,
- Bei Y.,
- Zhou Y.,
- Xiao J.,
- Li X.
- Funakoshi S.,
- Miki K.,
- Takaki T.,
- et al.
- Foldes C.,
- Terzic A.