Randomized Comparison of Allogeneic Vs. Autologous Mesenchymal Stem Cells for Non-lschemic Dilated Cardiomyopathy: POSEIDON-DCM Trial

Abstract

Background

While human mesenchymal stem cells (hMSCs) have been tested in ischemic cardiomyopathy, few studies exist in chronic non-ischemic dilated cardiomyopathy (NIDCM).

Objectives

The POSEIDON-DCM trial is a randomized comparison of safety and efficacy of autologous (auto) vs. allogeneic (allo) bone marrow-derived hMSCs in NIDCM.

Methods

Thirty-seven patients were randomized to either allo- or auto-hMSCs in a 1:1 ratio. Patients were recruited between December 2011 and July 2015 at the University of Miami Hospital. Patients (age: 55.8 ± 11.2; 32% female) received hMSCs (100 million) by transendocardial stem cell injection (TESI) in ten left ventricular sites by NOGA Catheter. Treated patients were evaluated at baseline, 30 days, 3-, 6-, and 12-months for safety: serious adverse events (SAE), and efficacy endpoints: Ejection Fraction (EF), Minnesota Living with Heart Failure Questionnaire (MLHFQ), Six Minute Walk Test (6MWT), MACE, and immune-biomarkers. This trial is registered with ClinicalTrials.gov, #NCT01392625.

Results

There were no 30-day treatment-emergent (TE)-SAEs. 12-month SAE incidence was 28.2% (95% CI: 12.8, 55.1) in allo, and 63.5% (95% CI: 40.8, 85.7; p=0.1004) in auto. One allo-group patient developed an elevated donor specific cPRA. EF increased in allo by 8.0 units (95% Cl: 2.8, 13.2; p=0.004), and in auto: 5.4 units (95% Cl: −1.4, 12.1; p=0.116, allo vs. auto p=0.4887). 6MWT increased for allo: 37.0 meters (95% Cl: 2.0 to 72.0; p=0.04), but not auto: 7.3 meters (95% Cl: −47.8, 33.3; p=0.71, auto vs. allo p=0.0168). MLHFQ score decreased in allo (p=0.0022), and auto (p=0.463; p=0.172). The MACE rate was lower in allo vs. auto (p=0.0186). Tumor necrosis factor alpha (TNF-α) decreased (p=0.0001 for each), to a greater extent in allo vs. auto at six-months (p=0.05).

Conclusion

These findings demonstrate safety and support greater, clinically meaningful efficacy of allo-hMSC vs. auto-hMSC in NIDCM patients. Pivotal trials of allo-hMSCs are warranted based on these results.

Introduction

Non-ischemic dilated cardiomyopathy (NIDCM) is a progressive disorder with no current cure, often culminating in heart transplantation (1,2). Cell-based therapy for heart disease is a promising new treatment strategy undergoing evaluation (3–8), with a major challenge and opportunity in developing allogeneic therapy (9). Bone marrow-derived MSCs may be a viable source of allo cells, as they lack major histocompatibility class II and co-stimulatory molecules rendering them immune-evasive (8,10). Allo-MSCs also have immunomodulatory effects (8), which could have therapeutic importance in NIDCM, a disorder with a major component of immune dysregulation as an underlying etiology (11,12). While allo-hMSCs offer a major opportunity as an off-the-shelf therapeutic, they may lack the efficacy of auto-hMSC therapy as some preclinical data indicate a higher risk of immunological clearance (13).

Compared to auto, allo cell therapy has great potential for developing readily available, disease-free cell products in a cost-effective manner, an important issue for disorders with high incidence. In the field of heart failure (HF), hMSCs exert antifibrotic and pro-regenerative effects leading to improved ventricular function and architecture in patients with antecedent myocardial infarction (6,8,9). As MSCs have powerful and sustained anti-inflammatory effects (8,14) and stimulate restoration of endothelial health (15), they could be of substantial therapeutic importance in conditions such as NIDCM. Accordingly, we performed a randomized trial (16) testing the hypothesis that allo-hMSCs therapy represents a safe and efficacious alternative to auto-hMSCs in patients with NIDCM. We also tested the prediction that immunomodulation and endothelial restoration contribute to the therapeutic effects of hMSCs in patients with NIDCM.

Methods

Study Design and Enrollment

The POSEIDON-DCM study titled “A Phase I/II, Randomized Pilot Study of the Comparative Safety and Efficacy of TESI of auto-hMSCs vs. allo-hMSCs in patients with NIDCM” - was conducted under FDA IND #14419. Study details are posted on clinicaltrials.gov/POSEIDON-DCM (#NCT01392625), and study design was published (16).

Patients provided written informed consent (IRB approval #20100968). All patients were recruited between December 2011 and July 2015 at the University of Miami Hospital (Figure 1). Thirty-seven patients were randomized to either auto- or allo-hMSCs in a 1:1 ratio. Following cardiac catheterization and cell injections, patients remained hospitalized for a minimum of two days, and then followed at two weeks post-catheterization, and two-, three-, six- and twelve-months for safety and efficacy assessments. An electronic data entry system was used for randomization and data collection. Although this was an open-label study, all data analysis was masked to those assessing all study endpoints and statistical analysis was performed by a third party for unmasking. Detailed methods are explained in the online method. The NHLBI Gene and Cell Therapy Data and Safety Monitoring Board provided safety oversight of the trial.

Figure 1. Study Flow chart.

Study screening, enrolment and randomization in detail showing 1:1 ratio of allo and auto with 30 day TE-SAEs and patients lost to follow up. TE-SAEs = treatment emergent serious adverse events.

Patient Population

Patient eligibility was determined after confirmation of diagnosis of NIDCM with an EF less than 40% and either a left ventricular end diastolic diameter greater than 5.9 cm in male subjects and greater than 5.6 cm in female subjects or a left ventricular end diastolic volume index > 125 ml/m², as previously described (16).

Study Procedures and Timeline

Baseline assessments included chemistry and hematology laboratories, echocardiography, and chest, abdominal and pelvic computed tomography scans. Cardiac imaging was performed (9).

hMSCs for Cell Therapy

All allo-hMSCs and auto-hMSCs were manufactured at the University of Miami ISCI (8,16). Allo hMSCs were derived from Caucasian male donors mean age 25.4 ± 3.3 years and were between 80 to 90% viable at time of TESI.. The auto-hMSC were from 11 males with a mean age of 58.0 ± 9.9 and six females with a mean age of 55.0 ± 12.4 years.

Transendocardial Stem Cell Injection

Injection sites were selected to prioritize safety of the TESI procedure and to distribute sites throughout the accessible myocardial territories. Considerations for site selection included avoidance of the ventricular apex, and optimization of catheter stability prior to needle extension.

Study Endpoints

The primary safety endpoint was the incidence of any treatment-emergent SAEs (TE-SAEs occurring after treatment within 30 days (16). Secondary safety endpoints included AEs, ectopic tissue formation, and Forced Expiratory Volume In One Second (FEV1). Secondary efficacy endpoints include incidence of Major Adverse Cardiac Event, LV structure and function, patient quality of life measured by New York Heart Association (NYHA) and MLHFQ, 6MWT, and maximal oxygen consumption (VO₂), endothelial function, and immunologic status.

Endothelial Function

Endothelial function was assessed at baseline and three-months post allo-hMSC or auto-hMSC. Endothelial progenitor colony forming units (EPC-CFUs) from peripheral blood samples and Flow-Mediated Vasodilation (FMD) brachial artery diameter measurements and FMD% were performed. A subset of the patients’ results for endothelial function was previously described (15).

Immune Monitoring

Calculated panel reactive antibodies (cPRA) were measured at baseline and at six-months using Luminex 200. Serum TNF-α was measured using Human TNF-α ELISA high sensitivity kit (eBiosciences). Lymphocytes were stained for T-cells markers of activation, Late/exhausted T-cells, B-cell subsets (Switched memory and Late/Exhausted B-cells) and TNF-α by B-cells. All samples were acquired using the LSR-Fortessa-HTS analyzer (BD Pharmigen) and analyzed with the FlowJo V10 software.

Statistical Analysis

The sample size of 18 per treatment arm was chosen to be appropriate for a phase I/II study so that if the true TE-SAE event rate were 25%, the probability of observing at least one event per treatment arm would be 99%. All patients who received study injection were included in analysis. Continuous variables were summarized using the following descriptive statistics: n (non-missing sample size), mean, standard deviation (or median and interquartile range (IQR) as appropriate), maximum and minimum. The frequency and percentages (based on the non-missing sample size) of observed levels were reported for all categorical measures. Outcomes which were collected at multiple follow-up visits were analyzed using a mixed model for repeated measures to compare treatment groups with treatment group considered an effect as well as a group-by-time interaction. Within group effects were described using model-estimated contrasts. Outcomes that were highly skewed were analyzed using ranked analysis of covariance adjusting for baseline, at each follow-up assessment and within group effects were described using a Wilcoxon Signed Rank Sum test. Categorical variables were compared between groups using Fisher’s exact tests. Analysis of time-to-event data was done using a two-sided Log-Rank test, censoring those who did not experience an event at their last known follow-up day. All statistical tests were performed at α=0.05 using two-sided tests. All data analyses and statistical computations were conducted with SAS, version 9.3 (Cary, NC).

Results

Patient population

Thirty-seven patients were randomized to either auto- or allo-hMSCs in a 1:1 ratio. All patients were recruited between December 2011 and July 2015. Thirty-four patients received study injection; 16 and 18 patients in auto- and allo-hMSC, respectively. Three patients did not receive the study injection. One patient withdrew consent before treatment. Another patient was recruited but did not receive treatment due to automatic implantable cardioverter-defibrillator placement (n=1), and one patient died before treatment (n=1). The mean age of injected participants was 55.8±11.2, 29% were female, and 35% were Hispanic (Table 1). The mean years of NIDCM diagnosis before the TESI was 6.1±6.2 years for allo and 6.9±7.3 years for auto patients (p=0.5 between groups). Fifty percent of patients had NYHA class II symptoms, mean baseline global EF was 26.5±9.64%, mean 6MWT was 422±86.8 M, and median baseline MLHFQ scores were 36 (IQR 18.0, 64.0).

Table 1.

Baseline Characteristics.

Baseline Characteristics	Cell Type
	Allo (n=18)	Auto (n=16)
Age at injection (years)	54.4 (11.5)	57.4 (11.0)
Gender
Male	14 (77.8%)	10 (62.5%)
Female	4 (22.2%)	6 (37.5%)
Years of NIDCM diagnosis before TESI	6.05 (6.2)	6.93 (7.3)
History of adriamycin chemotherapy	1 (5.5%)	2 (12.5%)
AICD or BIV/CRT	15 (83.3%)	14 (87.5%)
Ethnicity: Hispanic or Latino	4 (22.2%)	8 (50.0%)
Race: White	16 (88.9%)	14 (87.5%)
End Diastolic Diameter (cm)	7.2 (1.3)	7.1 (1.7)
History of Coronary Interventions	2 (11.1%)	1 (6.3%)
Previously Referred for AICD Placement	15 (83.3%)	14 (87.5%)
History of Atrial or Ventricular Arrhythmia	5 (27.8%)	1 (6.3%)
History of Hypertension	7 (38.9%)	3 (18.8%)
New York Heart Association Class
Class I - No Limitation	4 (22.2%)	6 (37.5%)
Class II - Slight Limitation of Physical Activity	9 (50.0%)	8 (50.0%)
Class III - Marked Limitation of Physical Activity	5 (27.8%)	2 (12.5%)
History of Congestive Heart Failure	11 (61.1%)	7 (43.8%)
History of Valvular Heart Disease	3 (16.7%)	2 (12.5%)
History of Smoking	11 (61.1%)	6 (37.5%)
History of Diabetes	0 (0.0%)	1 (6.3%)
Peak VO2 (mL/kg/min)	17.9 (±5.2)	16.0 (±5.1)
Six Minute Walk Test (meters)	427.2 (±67.4)	416.4 (±105.7)
Forced Expiratory Volume in one second (%)	81.1 (±22.2)	80.8 (±23.8)
MLHF: Median (IQR)	38.0 (23.0–64.0)	30.5 (16.5–60.5)
LV Size and Function
Ejection Fraction (%)	27.6 (±9.0)	25.2 (±10.5)
Left Ventricular End Diastolic Volume (ml): Median(IQR)	326.6 (259.1–348.6)	280.8 (239.2–360.4)
Left Ventricular Systolic Volume (ml): Median (IQR)	239.8 (179.4–331.2)	191.1 (167.3–317.8)
End Diastolic Diameter: Median (IQR)	73.4 (66.1–80.0)	70.0 (61.1–80.2)
End Diastolic Long Axis Diameter: Median (IQR)	105.6 (102.0–117.0)	98.8 (92.9–111.0)
Sphericity Index	0.5 (±0.08)	0.6 (±0.14)

Values presented as n, n (%) or mean (standard deviation); AICD=Automatic Implanted Cardioverter-Defibrillator; BiV=Biventricular Pacing.

Safety

The intervention was safe in all TESI recipients, with no TE-SAEs within 30-days. Furthermore, the incidence of AEs by 30-days did not significantly differ by cell type (p=0.3820; Table 2). Moreover, SAE rates were infrequent through day 30 and similar in both groups (p=0.6238; Table 2). Accordingly, the study met the primary safety event endpoint documenting the safety of TESI in patients with NIDCM.

Table 2. Safety Summary by 30-Days Post-TESI.

Two deaths occurred post-injection in auto: one experienced a fatal subdural hematoma due to trauma on day 152 post-injection, and a second patient died of NIDCM 291 days post-injection. Both events were considered unrelated to auto treatment. NIDCM = Non-Ischemic Dilated Cardiomyopathy.

Safety Summary by Time Post-TESI		Cell Type
		Allo (n=18)		Auto (n=16)
	n	% (95% CI)	n	% (95% CI)
Day 30 Post-TESI
Incidence of AE	7	38.9% (20.8–64.7)	4	25.0% (10.2–53.7)
Incidence of SAE	2	11.1% (2.9–37.6)	1	6.3% (0.9–36.8)
Incidence of TE-SAE	0	0.0% (0.0–18.5)	0	0.0% (0.0–20.6)
Incidence of MACE	0	0.0% (0.0–18.5)	0	0.0% (0.0–20.6)
Day 180 Post-TESI
Incidence of AE	12	66.7% (45.5–86.3)	11	68.8% (46.4–88.6)
Incidence of SAE	4	22.2% (9.0–48.9)	4	25.0% (10.2–53.7)
Incidence of MACE	1	5.6% (0.8–33.4)	3	18.8% (6.5–47.5)
Incidence of Death	0	0.0% (0.0–18.5)	1	6.3% (0.9–36.)
Day 365 Post-TESI
Incidence of AE	12	66.7% (45.5–86.3)	12	75.0% (52.8–92.2)
Incidence of SAE	5	28.2% (12.8–55.1)	10	63.5% (40.8–85.7)
Incidence of MACE	3	20.3% (6.8–52.1)†	9	57.1% (34.9–81.2)
Incidence of Death	0	0.0% (0.0–18.5)	2	12.5% (3.3–41.4)

Note: %s are Kaplan-Meier event rates. Exact binomial CIs were used for zero counts.

^†indicates between group p ≤ 0.05

TESI was technically successful in 97.05% (n=33) of patients. One patient experienced ventricular tachycardia after the 9^th injection, and did not receive the last injection. No patients experienced significant post-procedural pericardial effusion.

Long Term Adverse Events, Rehospitalization, MACE and Ectopic Tissue Formation

The 12-months post-TESI SAE incidence was 28.2% (95% CI: 12.8, 55.1) in allo and 63.5% (95% CI: 40.8, 85.7, p=0.1004) in auto. Post-TESI, two auto and one allo patients received heart transplantation; and LV assist devices were implanted in one auto and one allo patients. Two deaths occurred post-injection in auto: a fatal subdural hematoma due to trauma on day 152 post-injection, and a death due to NIDCM 291 days post-injection. Both events were considered unrelated to study treatment.

The 12-month all-cause re-hospitalization rate was lower in the allo group vs. the auto recipients: 28.2% (95% CI: 12.8, 55.1) in allo and 70.0% (95% CI: 47.0, 89.8) in auto (p=0.0447). Similarly MACE over 12-months was lower in allo compared with auto. The 12-month MACE rate in the allo group was 20.3% (95% CI: 6.8, 52.1) and 57.1% (95% CI: 34.9, 81.2) in the auto group, p=0.0186. No ectopic tissue formation was identified in either group over one year by CT.

Left Ventricular Function

At baseline, average EF was 26.5±9.6%, median LV End Diastolic Diameter (LVEDD) was 70.4mm (IQR: 64.1, 80.0). EF increased significantly in allo by 8.0 units (95% CI: 2.8–13.2, p=0.004), but not in auto: 5.4 units (95% CI: −1.4, 12.1, p=0.116) at 12-months (p=0.49 between group; Figure 2A, Figure 3). This magnitude of EF increase resulted in EF rising to above 40% in 46.7% of the allo (Central Illustration) and 22.2% of the auto patients. Stroke Volume (SV, Figure 2B), End Diastolic Volume (EDV, Figure 2C), and End Systolic Volume (ESV, Figure 2D) did not significantly decrease from baseline. End diastolic long axis diameter decreased 0.5mm (95% CI: −6.4, −0.6 p=0.04) from baseline to 12-months in allo and 1.7mm (95% CI: −7.3, 3.9) in auto (p=0.73; Figure 2E). Sphericity index, End Diastolic Diameter or End Systolic Diameter did not change from baseline to 12-months in either group (data not shown).

Figure 2. Patient changes in Left Ventricular Structure and Function from baseline to 12-months Post-hMSC Injection after Allogeneic (denoted by: ●) and Autologous (denoted by ▲) hMSC therapy.

(A) EF increased from baseline in allo but not in auto: 8 EF units (95%CI: 2.8–13.2 p=0.04). (B) Stroke volume was not significantly increased in either groups at 12-months. (C, D) Neither group showed significant improvement in EDV or ESV. (E) Structural remodeling was evident in the reduction in long axis diameter in allo from baseline to 12-months (p=0.03). * indicates p ≤ 0.05 within group, and † indicates between group p ≤ 0.05. EF = ejection fraction, EDV = end diastolic volume, ESV = end systolic volume.

Figure 3. Global Ejection Fraction.

Graphic representation of global EF from baseline to 12-months. In green allo shows 7 out of 15 patients increase above 40% or better at 12-months while in purple auto only 2 out of 9 patients had an increase of 40 or better at 12-months. EF = ejection fraction.

Central Illustration. Allogeneic Mesenchymal Stem Cell Therapy for Non-Ischemic Dilated Cardiomyopathy.

(A) Cardiac computed tomography shows decreased global EF at baseline. (B) Global EF has significantly improved at 12-months post TESI. EF = ejection fraction, TESI = Transendocardial Stem Cell Injection.

Functional Status, Quality of Life, and Pulmonary Function

The 6MWT distance significantly increased in allo: 37.0 meters (95% CI: 2.0, 72.0, p=0.04) at 12-months compared to baseline, but did not significantly change in auto: 7.3 meters (95% CI: −47.8, 33.3, p=0.71) (Figure 4A). The between group difference was 67.7 meters (95% CI: 16.4, 118.9; p=0.0116) and 46.53 meters (95% CI: −5.5, 98.5; p=0.0770) from baseline to six or 12-months, respectively (overall comparison between allo and auto p=0.0168). There were no significant differences in the maximum VO₂ at 6 or 12-months compared to baseline in either group (data not shown). Forced expiratory volume in one second improved in allo by 3.7% (95% CI: −0.08, 0.30, p=0.2423) compared to a decrease of 3.8% (95% CI: −0.36, 0.06, p=0.16) among auto at 12-months. However, the between-group difference at 12-months was 0.29L (95% CI: 0.01, 0.56, p=0.0430; Figure 4B).

Figure 4. Cardiac Functional Outcomes and changes in endothelial function after Allogeneic (denoted by: ●) or Autologous (denoted by ▲) hMSC therapy.

Graphic representation of estimated mean of the functional Outcomes (A) Allo but not auto therapy showed an improvement in the 6MWT (p=0.01 at six-months and p=0.04 at 12-months). (B) FEV1 improved by 0.11L (p=0.04) in allo compared to auto. (C) NYHA classification improved in 66.70% and 27.3% in allo and auto, respectively. (D) MLHFQ score improved in allo but not auto 12-months relative to baseline (p=0.002). (E) Allo but not auto show improvement in the FMD (p=0.0005) at 3-months relative to baseline. (F) The production of EPC-CFU was significantly greater (p=0.01) in allo compared to auto. * indicates p ≤ 0.05 within group, and † indicates between group p ≤ 0.05. 6MWT = six-minute walk test, NYHA = New York Heart Association, MLHFQ = Minnesota Living with Heart Failure Questionnaire, FMD = Flow Mediated Vasodilation, EPC-CFU = Endothelial Progenitor Cell Colony Forming Unit.

Functional capacity and quality of life showed greater improvement in allo compared to auto. At 12-months, allo showed 66.7% improvement in NYHA whereas only 27.3% improved in auto (between treatment group change in NYHA class: p=0.0527). Two patients in auto worsened by 12-months (Figure 4C). The median MLHFQ score at baseline was allo: 38 (IQR: 23, 54) and auto: 30.5 (IQR: 16.5, 60.5). The MLHFQ improved in both groups (Figure 4D) over 12-months (allo: p=0.0022; auto: p=0.1719).

Endothelial Function

EPC-CFU significantly increased in allo (p=0.0107) compared to auto (Figure 4E). Likewise, allo increased the %FMD at three-months as compared to auto (p=0.09). The %FMD was allo: 4.5% (IQR: 2.9, 7.4) and 6.4% (IQR: 5.1, 12.3) (p=0.0005), and auto: 6.4% (IQR: 3.7, 10.0), and 5.8% (IQR: 4.4, 10.0) (p=0.8457) at baseline and three-months respectively (Figure 4F).

Immune Monitoring

cPRA results showed that 67% of allo and 92% of auto recipients had no reaction to low cPRA (0–20% cPRA). Twenty-seven % of allo and 8% of auto had a moderate cPRA (21–79% cPRA), and one subject (7%) receiving allo MSCs had a high cPRA response (+80% cPRA; Table 3).

Table 3. Recipient cPRA Change 6-months post-TESI.

Change of cPRA from baseline to six-months. One patient senzitized to greater than 80 percent cPRA after 6-months. cPRA = Calculated Panel Reactive Antibodies.

cPRA Risk	Cell Type
	Allo (n=15)	Auto (n=13)
No reaction to Low risk (0 – 20% cPRA)	10 (66.7%)	12 (92.3%)†
Moderate risk (21 – 79% cPRA)	4 (26.7%)	1 (7.7%)†
High risk (+80% cPRA)	1 (6.7%)	0 (0%)

^†indicates between group p ≤ 0.05

Elevated baseline levels of TNF-α decreased from baseline to 6-months in both groups (allo: −10.6 ± 1.6 pg/ml, p<0.0001, auto: −6.8 ± 1.4 pg/ml, p<0.0001, Table 4; between group, p=0.05). Temra T-cells (exhausted T-cell phenotype) were reduced in both groups with a greater decrease in allo (allo: −15.9±5.4%, p<0.0001, auto: −9.3±3.3%, p<0.0001, Table 4; between group, p=0.0111; Table 4). Suppressed %Switch Memory B-cells (predictive biomarker for antibody response) at baseline were significantly increased at six-months in both groups with a greater improvement in allo (allo: +10.2±4.9%, p<0.0001 vs. auto: +4.3±3.9%, p=0.0014, between group, p<0.0001; Table 4). Finally, intracellular TNF-α expression in B-cells was also decreased at 6-months relative to baseline in both groups (allo: −11.2±3.3%, p<0.0001, auto: −8.5±3.0%, p<0.0001; between group, p=0.174; Table 4). In contrast, Late/exhausted B-cells decreased significantly in both groups (allo: −5.4±1.03%, p<0.0001, auto: −5.9±2.5%, p=0.003; between group difference, p=0.57; Table 4). Early T-cell activation (allo: −5.57±1.03%, p<0.0001, auto: −2.92±1.5%, p=0.02; between group, p=0.08; Table 4) decreased to similar degrees in both group. Whereas late/chronic T-cell activation did not significantly decrease in either group (allo: −2.3±1.3%, p=0.4, and auto: −3.4±2.7%, p=0.7).

Table 4. Effects of hMSCs on Cellular and Humoral Immunity at 6-months post-TESI.

Serum TNF-α was assessed by ELISA. All lymphocyte populations were assessed by flow cytometry and expressed as a percent from gated (B-cell parameters gated on CD19⁺ and T-cell parameters gated on CD3⁺)

Immune Biomarkers	Allo		Auto
	Baseline	6-months	Baseline	6-months
Serum TNF-α (pg/ml)	13.5 ± 1.5	2.3 ± 0.2 **,†	11.8 ± 0.9	4.8 ± 0.8**
%Early T-cell Activation (CD3+, CD69+)	13.4 ± 4.4	7.9 ± 3.3**	13.3 ± 4.8	10.6 ± 5.4*
%Late/Chronic T-cell activation (CD3+, CD25+)	8.9, IQR (5.9, 10.4)	3.4, IQR (3.0, 5.0)*	8.1, IQR (5.2, 14.1)	4.4, IQR (3.6, 11.1)
% Temra (CD3+, CD45RA+, & CCR7−)	33.7 ± 7.6	17.9± 5.2**,†	30.1 ± 10.8	21.3 ± 7.5**
%Late/Exhausted B-cells (CD19+, CD27− & IgD−)	19.3, IQR (17.9, 28.6)	14.7, IQR (14.2, 17.3)**,†	20.0, IQR (17.7, 32.0)	15.5, IQR (14.0, 19.0)*
%Switched Memory B-cells (CD19+, CD27 high & IgD−)	10.0 ± 3.6	20.2 ± 3.1**,††	9.6 ± 3.1	14.1 ± 4.5*
% B -ells expressing Intracellular TNF-α	32.1 ± 7.3	20.4 ± 6.3**	28.2 ± 6.8	19.7 ± 4.8**

^*indicates p ≤ 0.05 within group;

^**indicates p ≤ 0.001 within group;

^†indicates between group p ≤ 0.05,

^††indicates between group p ≤ 0.001.

TNF-α = Tumor Necrosis Factor alpha, ELISA = Enzyme Linked Immunosorbent Assay.

Discussion

The POSEIDON-DCM study was a randomized comparison of allo-hMSCs vs. auto-hMSCs in patients with NIDCM (16). The results indicate the safety and feasibility of TESI for both allo and auto, with no ectopic tissue formation detected with either cell type. Importantly, allo produced a constellation of clinically meaningful effects of greater magnitude than auto, including significant improvement in EF, 6MWT and MLHFQ scores. In addition, endothelial function was improved only in allo, TNF-α suppression was greater with allo, and several findings showed evidence of clinical efficacy including improved NYHA class, lower MACE and hospitalization rates at one year in allo. Together these findings show a substantial magnitude of clinical responsiveness in patients with NIDCM, a group with major unmet medical needs.

The findings here provide evidence of a clinically relevant effect substantially larger than previous trials addressing cell therapy in patients with ischemic cardiomyopathy (4–7,17). Whereas cell therapy in ischemic cardiomyopathy produces reduction in infarct scar size (6,9), increases in EF have been difficult to show in this population (6,17). In the present study, EF increased by 8 units in allo. Also nearly half of allo patients increased EF to levels above 40% (Figure 3), an accepted cut-off for the diagnosis of NIDCM (1,18). HF with recovered EF is a recognized syndrome that carries an improved prognosis relative to HF with persistently low EF (19). As such, if the present results are replicated in a larger trial, they represent a major clinical advance for NIDCM, a condition affecting individuals of all ages and accounting for approximately half of the heart transplants (2,20).

In this study, two pathophysiologic features were identified which could underlie the effects of allo-hMSCs. First, hMSCs exerted a significant restoration of endothelial dysfunction, implicated as an underlying contributor to the failing circulation in ischemic cardiomyopathy and NIDCM (15,21). Second, hMSCs reduced the elevated TNF-α levels in the study population. Elevated levels of TNF-α, a crucial pro-inflammatory cytokine tied to progression of heart disease, are implicated in modulating both cardiac contractility and peripheral resistance (22,23).

The use of entanercept to inhibit TNF-α in HF did not improve mortality and hospitalization rates (24), however hMSC therapy has the advantage of reducing several pro-inflammatory cytokines and favoring an anti-inflammatory profile. In our trial, TNF-α was significantly reduced 6-months post-TESI by both allo and auto, but allo were more effective.

Several reasons might account for providing allogeneic MSCs greater efficacy relative to autologous. These include age of the donors and possible adverse impact of the disease milieu (e.g. the pro-inflammatory phenotype) (25). Alternatively, preferential response to allo vs. auto hMSCs may reflect enhanced endogenous repair, an important mechanism underlying hMSC cardiorepair (26). Further studies to delineate potency differences between autologous and allogeneic MSCs in NIDCM are underway.

In addition, our present findings include a detailed evaluation of humoral lymphocytes following hMSC therapy. hMSCs favorably altered several immunologic markers typically elevated in chronic inflammation such as Temra T-cells (exhausted phenotype) and Late/exhausted B-cells (27,28). As with TNF-α, hMSCs reversed the exhausted immune phenotype (Temra T-cells and Late/exhausted memory B-cells) (29). In contrast to previous studies, we show the important finding that hMSC therapy (in particular, allo) increased Switched Memory B-cells (predictive biomarker for a protective antibody response) (28). Restoration of immune competence may also have clinical relevance in these patients who are of higher risk for co-morbid infectious disease.

The cPRA yielded mostly no to low response with very few moderate response for both allo and auto. Only one patient in the allo group mounted an elevated cPRA response that included donor specific antibodies. This incidence is similar to two previous studies employing allogeneic MSCs or mesenchymal precursor cells (9,30). This cPRA response did not cause clinical immunologic rejection in the patient. Ongoing immunologic monitoring is warranted during the conduct of larger pivotal trials.

There are limited previous cell-based therapy trials in NIDCM showing improved cardiac function and/or quality of life. The TOPCARE-DCM trial, showed that intracoronary delivery of auto bone marrow cells improved LVEF three-months after cell administration (31). Similarly, another study demonstrated an improvement in EF and sustained improvement in quality of life three years post-treatment with autologous bone marrow cells (32,33). Although several other groups have shown that cell based therapy appears to reduce the incidence of heart transplantation and/or mortality in NIDCM, there is variability in the selected cell for therapy and their distinct cell surface markers such as CD34+ (34). Perin et al. in a study including both ischemic cardiomyopathy and non-ischemic cardiomyopathy delivered bone marrow-derived mesenchymal precursor cells characterized by surface antigen expression of STRO-1, STRO-3, CC-9, and HLA class I and II antigens. This study, however, did not address differences in responses between ischemic cardiomyopathy and NIDCM.

To date, no previous study has compared bone marrow-derived allo-hMSC and auto-hMSC in patients with NIDCM. In the POSEIDON (The PercutaneOus StEm Cell Injection Delivery effects On Neomyogenesis) pilot study, we showed that allo and auto are safe and do not increase SAEs or immunologic reactions with allo therapy. Furthermore, we showed reverse remodeling of the LV chamber dimension (shown by reduction of the long axis), decreased scar size, improved EDV and sphericity index and increased EF with low-dose allo-hMSC (9). The limitations of this study included the lack of a placebo group, which was by trial design. Another limitation, was the loss of patients due to withdrawal of consent, or loss to follow-up. This trial is limited by small sample size, which was prospectively determined based upon a pre-set threshold for 30-day TE-SAE rate. The sample size limits interpretation of efficacy results; nevertheless the findings are of significant value in designing and determining sample size for future pivotal trials. Future phase II/III studies will incorporate a placebo-controlled, double-blind design.

Conclusions

This study tested the safety and efficacy of allo vs. auto in NIDCM patients. This study revealed a highly acceptable safety profile at 30 days in both groups, which are similar to previously published trials using TESI (6,9,32). Importantly, several lines of evidence support that allo is superior to auto in regard to efficacy including EF, 6MWT, MLHFQ and endothelial function. In addition, the pro-inflammatory/exhausted immune phenotype in patients receiving allo-hMSCs for NIDCM had a dramatic remodeling of the immune cells towards a less inflammatory/exhausted phenotype. These data provide important and clinically relevant insights into the therapeutic basis and effects of allo and auto for NIDCM patients. Clinical benefits of the magnitude shown here support the development of allo-hMSCs for treating NIDCM.

Online Methods

Study Design and Enrollment

POSEIDON-DECM was a Phase I/II, Randomized Pilot Study of the Comparative Safety and Efficacy of TESI of auto-hMSCs vs. allo-hMSCs in patients with NIDCM”, and was conducted under FDA IND #14419. The primary objective of this study was to test the safety of allo- and auto-hMSCs, and the second objectives were to establish efficacy for each cell type and to compare pre8 specified endpoints between allo and auto hMSCs (i.e. to test superiority or inferiority of allo- compared to auto-hMSCs).

Patient Population

Patient eligibility was first determined by NIDCM diagnosis with less than 40% EF. Stability for at least 3 months of LV dysfunction was ascertained to exclude patients recovering from cardiomyopathy. All patients were treated with appropriate maximal medical therapy for heart failure. For biventricular pacing, the device was placed three months or greater before patient treatment with hMSCs. Myocarditis was excluded by endomyocardial biopsy.

Study Procedure and Timeline

All patients were evaluated at baseline, 6-months and 12-moths post TESI by computed tomography or magnetic resonance imaging and analyzed as previously described. Computed tomography was predominantly conducted due to implanted devices or scanable permanent pacemakers and implantable cardioverter defibrillators.

Transendocardial Stem Cell Injection (TESI)

Patients’ hearts were injected with stem cells on 10 left ventricular sites avoiding the apex. In very dilated ventricles this necessitated leaving some walls, which could not be reached by the catheter as untreated. In cases where a discrete myocardial scar was identified by electro-anatomic mapping, sites were also selected to encircle the scar border. Using the NOGA catheter hMSCs were delivered via TESI to ten LV sites with a voltage of 6–12 mV and wall thickness greater than 5 mm (1). TESI was discontinued according to the Clinical Research Protocol if the patient experienced: sustained hypotension, clinical signs or symptoms indicating coronary syndrome, cerebrovascular accident, two episodes of sustained ventricular tachycardia or ventricular fibrillation requiring cardioversion, or discovery of a new LV thrombus, aortic aneurysm, or aortic dissection.

Study Endpoints

The primary safety endpoint was the incidence of any TE-SAEs, defined as death, non-fatal MI, stroke, hospitalization for worsening HF, cardiac perforation, pericardial tamponade, sustained ventricular arrhythmias, occurring within 30 days of treatment (2). Secondary safety endpoints included AEs, ectopic tissue formation, and Forced Expiratory Volume In One Second (FEV1). Secondary efficacy endpoints included the incidence of MACE (defined as death, hospitalization for worsening HF, non-fatal recurrent MI, or angina requiring hospitalization), LV structure and function, patient quality of life measured by New York Heart Association (NYHA) and MLHFQ, 6MWT, and maximal oxygen consumption (VO₂), endothelial function, and immunologic status.

Immune-Biomarker Staining

Peripheral blood mononuclear cells (PBMCs) were cryopreserved in 90% fetal bovine serum and 10% DMSO. PBMCs were slowly thawed and rested for one-hour at 37°C and 5% CO₂ in complete RPMI-1640 media (supplemented with 20% fetal bovine serum, 1x non-essential amino acids, 1x L-glutamine, 1x sodium pyruvate, 1x penicillin-streptomycin, and 0.1% beta-mercaptoethanol). The PBMC resting time is essential to re-establish all cell surface markers. T-cell activation markers were surface stained with CD3-FITC, CD25-APC, and CD69-PE-Cy7. The B-cell subsets were surface stained with: CD19-PE, CD27-FITC and IgD-AlexaFluor700 and gated on CD19⁺. Switch Memory B-cells were CD27^high and IgD⁻ and exhausted B-cells were CD27⁻ and IgD⁻ (3). TNF-α expression by B-cells was evaluated by surface staining with CD19 and intracellular staining with TNF-α. The T cell subsets were stained with: CD3-FITC, CD8- PE, CD4-APC, CCR7-AlexaFluor700 and CD45RA-PacBlue. T effector memory re-expressing naïve marker CD45RA (Temra) were CD3⁺, CD45RA⁺ and CCR7⁽⁴⁾ (all antibodies: BD Pharmigen). Isotype controls were used for proper gating.

Perspectives.

Competency in Medical Knowledge

Allogeneic human MSCs offer a safe and readily available off-the-shelf therapeutic agent for patients with non-ischemic dilated cardiomyopathy. Furthermore, the ability of allo-hMSCs to modulate the immune system in NIDCM may contribute importantly to the improvements in cardiac function, quality of life, and endothelial function in this disorder.

Translational Outlook

Allo-hMSCs may have advantages over autologous cell therapy in NIDCM. The impact observed in the POSEIDON-DCM trial forms the basis for additional pivotal trials of this therapeutic strategy.

Abbreviations and Acronyms

6MWT

6 minute walk test

Allo-hMSC

allogeneic human mesenchymal stem cell

Auto-hMSC

autologous human mesenchymal stem cell

cPRA

calculated panel reactive antibodies

EF

ejection fraction

EPC-CFU

endothelial progenitor cell colony forming unit

FMD

Flow Mediated Vasodilation

HF

heart failure

LVEDD

left ventricular end diastolic diameter

MACE

Major Adverse Cardiac Event

MLHFQ

Minnesota Living with Heart Failure Questionnaire

NIDCM

non-ischemic dilated cardiomyopathy

NYHA

New York Heart Association

SAE

serious adverse events

TE-SAE

treatment-emergent serious adverse events