Long-Term Effects of AAV1/SERCA2a Gene Transfer in Patients With Severe Heart FailureNovelty and Significance
Analysis of Recurrent Cardiovascular Events and Mortality
Rationale: The Calcium Up-Regulation by Percutaneous Administration of Gene Therapy In Cardiac Disease (CUPID 1) study was a phase 1/phase 2 first-in-human clinical gene therapy trial using an adeno-associated virus serotype 1 (AAV1) vector carrying the sarcoplasmic reticulum calcium ATPase gene (AAV1/SERCA2a) in patients with advanced heart failure. The study explored potential benefits of the therapy at 12 months, and results were previously reported.
Objective: To report long-term (3-year) clinical effects and transgene expression in the patients in CUPID 1.
Methods and Results: A total of 39 patients with advanced heart failure who were on stable, optimal heart failure therapy were randomized to receive intracoronary infusion of AAV1/SERCA2a in 1 of 3 doses (low-dose, 6×1011 DNase-resistant particles; mid-dose, 3×1012 DNase-resistant particles; and high-dose, 1×1013 DNase-resistant particles) versus placebo. The following recurrent cardiovascular and terminal events were tracked for 3 years in all groups: myocardial infarction, worsening heart failure, heart failure–related hospitalization, ventricular assist device placement, cardiac transplantation, and death. The number of cardiovascular events, including death, was highest in the placebo group, high but delayed in the low- and mid-dose groups, and lowest in the high-dose group. Evidence of long-term transgene presence was also observed in high-dose patients. The risk of prespecified recurrent cardiovascular events was reduced by 82% in the high-dose versus placebo group (P=0.048). No safety concerns were noted during the 3-year follow-up.
Conclusions: After a single intracoronary infusion of AAV1/SERCA2a in patients with advanced heart failure, positive signals of cardiovascular events persist for years.
Heart failure (HF) is one of the leading causes of death and hospitalization worldwide.1 Despite optimal therapy using a wide range of pharmacological, device, and surgical options, many patients deteriorate over time and experience an unfavorable clinical course. Considering the medical, economic, and quality of life consequences of these trends, novel treatment strategies for HF are urgently needed.
In This Issue, see p 2
Editorial, see p 12
In HF, the level and the activity of sarcoplasmic reticulum Ca2+ ATPase 2a (SERCA2a) is decreased, contributing directly to impaired cardiac contraction and relaxation. It has been shown that increasing the activity of SERCA2a via gene transfer increases contractility in isolated failing human cardiac myocytes2 and leads to an improvement in cardiac function and metabolism in animal models of HF.3–8
A phase 1 trial of 12 patients with advanced HF treated with 1 of 4 dose cohorts of recombinant adeno-associated viral vector serotype 1 (AAV1) containing human SERCA2a gene (AAV1/SERCA2a)9,10 established preliminary safety, which led to a phase 2 randomized, double-blind, placebo-controlled trial. In this study, the effectiveness of 3 different dose cohorts of AAV1/SERCA2a versus placebo was evaluated in 39 patients with advanced HF. At 12 months of follow-up, in the high-dose group versus placebo, there were improved signs and symptoms of HF, functional status, biomarker profile, and left ventricular function.11 A significant decrease in recurrent cardiovascular (CV) events (myocardial infarctions, hospitalizations related to HF, episodes of worsening HF) adjusted for correlated terminal events (CV death or the requirement for left ventricular assist devices [VAD] or cardiac transplant) was evident in the patients who received high-dose AAV1/SERCA2a compared with those who received placebo (hazard ratio=0.12; P=0.003).11
The full study was 3 years in length and included the amount of time in the 12-month active observation period plus the amount of time in the long-term follow-up period. After the 12-month active observation period or termination because of an event such as CV death, left VAD, or transplant, both placebo and AAV1/SERCA2a-treated patients were entered into long-term follow-up. All surviving patients have completed the 12-month active observation and the long-term follow-up periods of the study, with the last patient follow-up occurring in September 2012. This report presents the information from patients in the phase 2 trial through final 3-year follow-up postadministration.
The study overview of the Calcium Up-regulation by Percutaneous Administration of Gene therapy InCardiac Disease (CUPID) phase 2 trial has been described in detail.9,11 Briefly, CUPID phase 2 was a randomized, double-blind, placebo-controlled, dose-ranging study that compared the use of intracoronary infusion of 3 different doses of AAV1/SERCA2a versus placebo. All patients were receiving and continued to receive optimal medical and device therapy for their HF. The primary end point in phase 2 was prospectively evaluated at 6 months, with additional prespecified analysis through 12 months.
Patients enrolled were required to have New York Heart Association Class III/IV HF symptoms for a minimum of 3 months before screening, left ventricular ejection fraction ≤35%, maximal oxygen uptake (VO2 max) ≤20 mL/kg per minute, implantable cardiac defibrillators, and were on stable (≥30 days) optimal outpatient therapy for HF. Patients were enrolled into the study and randomized to either placebo, low-dose AAV1/SERCA2a (6×1011 DNase-resistant particles [DRP]), mid-dose AAV1/SERCA2a (3×1012 DRP), or AAV1/SERCA2a high dose (1×1013 DRP). The study was approved by institutional review boards and institutional bio-safety committees at each site, and written informed consent was obtained from all patients before screening.
Patients were required to have undetectable neutralizing antibodies (NAb) against AAV1 (titer <1:2), which can block entry of the AAV1 into cardiac myocytes. All patients exposed to AAV, either via a therapeutic gene therapy agent such as AAV1/SERCA2a or by exposure to infection by wild-type virus, will seroconvert to become NAb positive. The development of AAV1 NAbs does not affect the initial transduction of heart muscle and generally occurs by week 6 after AAV1/SERCA2a administration. There is no indication from this or any of the other AAV human trials that a humoral response has any negative consequences other than as relates to the inability to re-administer because of the presence of NAbs.
Product Source and Administration
AAV1/SERCA2a consisting of AAV serotype 1 encoding for human SERCA2a driven by the Cytomegalovirus promoter and with AAV2 internal terminal repeats was manufactured by Celladon Corporation, La Jolla, CA, and was administered by antegrade epicardial coronary artery or bypass graft infusion for a 10-minute period in a cardiac catheterization laboratory.
Recurrent CV events and terminal events (CV death, transplant, or VAD), were adjudicated by a blinded clinical endpoint committee during the 12-month active follow-up. Any patient who received either a VAD or heart transplant was discontinued from the study and followed for long-term safety only.
Prespecified recurrent CV events included myocardial infarction, worsening HF, or HF-related hospitalization. Patients receiving positive inotropes were not terminated from the study but were considered to be associated with worsening HF, and therefore not counted as separate events.
The observation period was 12 months or up to early termination because of death, transplant, or VAD placement; thereafter, recurrent CV and terminal events were collected for an additional 24 months using nonadjudicated patient self-reported history and, in some cases, medical records collected by a designated healthcare provider at the investigative center.
During the long-term follow-up period, the healthcare provider contacted the patients every 6 months by telephone for a structured questionnaire on health status. Information collected was primarily by patient recall, although in some cases patients continued to receive primary care within the medical system of the investigative center, which provided access to medical records for a more complete reporting. The structured questionnaire was designed to elicit and record new findings on the occurrence of specific HF-related CV and terminal events and emergence of new clinical conditions, as described in Table 1.
To study the persistence of the transferred gene, quantitative polymerase chain reaction (qPCR) analysis (AltheaDx, Inc, San Diego, CA) was performed on biopsy samples of myocardial tissue from any patient who died, required placement of a VAD, or received a heart transplant. A TaqMan qPCR assay was developed and qualified to detect and quantify a specific sequence unique to AAV1/SERCA2a. The assay amplifies a 107 nucleotide sequence of the SERCA2 gene of the AAV vector. The number of copies of AAV1/SERCA2a detected in ≤1 microgram of genomic DNA extracted from each tissue was quantified using serial dilutions of a plasmid containing the target sequences as standards. The lower limit of detection of the assay was 20 single-stranded copies of AAV1/SERCA2a per μg DNA; the lower limit of quantification was 200 single-stranded copies of AAV1/SERCA2a per μg DNA. The 5′- and 3′- PCR primers, along with the fluorescently labeled probe are located in a region that spans the SERCA2a gene insert. The primer and probe sequences were:
SERCA2a forward primer: 5′- AAC CCT CCC ACA AGT CTA AAA TC -3′
SERCA2a reverse primer: 5′- CTC GGC TTT CTT CAG AGC AG -3′
The SERCA2a probe: 5′-6FAM AGC ATC GTT CAC GCC A MGBNFQ -3′
Recurrent CV and Terminal Events
The study patient disposition in phase 2 is summarized in Figure 1. The characteristics of recurrent CV events and terminal events during the full 3 years of the study are illustrated in Figure 2. This figure represents the clinical course of a patient starting from the date of infusion of AAV1/SERCA2a or placebo. Each line represents a single patient. If a patient underwent a VAD implant or heart transplant (terminal events) during the first 12 months, the patient was discontinued from the 12-month active observation period and entered into long-term follow-up at the time of surgery. Patients without terminal events completed the 12-month active observation period and continued in long-term follow-up for a total of 3 years. CV events occurring during the first 12 months or up to the time of discontinuation for a terminal event were adjudicated by a clinical endpoint committee. Terminal and recurrent CV events captured in long-term follow-up were limited to death, heart transplant, VAD implantation, and worsening HF and were not adjudicated. Recurrent CV and terminal events are depicted by symbols; a star at the beginning of a line represents a patient with an anti-AAV1 NAb titer that was <1:2 (negative) at screening and ≥1:2 (positive) at baseline.
Prespecified recurrent CV and terminal events in AAV1/SERCA2a-treated versus placebo patients were either delayed or had a substantially reduced frequency or both. As shown in Figure 2, patients in the placebo group had multiple CV events early during the follow-up after the infusion and continued experiencing these events throughout the ensuing 3 years. The low- and mid-dose groups had few events in the first 6 months of follow-up. Beyond 6 months, these 2 groups had an increase in CV events that was sustained throughout the rest of the 3 years. The high-dose group had relatively few CV and terminal events throughout the 3 years of follow-up.
As in the 12-month analysis,11 the risk of recurrent CV events during 3 years of follow-up was evaluated for the high-dose group versus placebo using the joint frailty model.12 Events were not adjudicated during the long-term follow-up period, and only the month of the event occurrence was available. For analysis purposes, the events were assigned to the end of the month in which they occurred. This assumption is conservative because more events occurred in the placebo group than in the high-dose group. The joint frailty model accounts for multiple (recurrent) CV events for each patient and evaluates the risk of recurrent CV events adjusted for correlated terminal events. The joint frailty model also takes into account the following:
Increased risk of terminal events (such as death) for patients with multiple recurrent HF hospitalizations (informative censoring).
Increased risk of repeated HF hospitalizations for patients who were previously hospitalized.
Differential follow-up times across the patients and treatment groups because of the timing of terminal events.
Random differences among patients’ histories (the number of recurrent events per patient may vary from 0–5).
On the basis of the joint frailty model, the hazard ratio at 3 years for the AAV1/SERCA2a high-dose group versus placebo for recurrent CV events adjusted for correlated terminal events and individual patients’ susceptibility to recurrent and terminal events (frailties) was 0.18, P=0.048, representing an 82% reduction in risk with AAV1/SERCA2a (Figure 3). It is interesting to note that the traditional time-to-first of CV event or terminal event analysis based on the proportional hazards model demonstrated a substantially different risk reduction for high-dose AAV1/SERCA2a versus placebo (hazard ratio=0.32; P=0.082). The study data satisfied the joint frailty model assumptions, whereas the proportional hazards model assumptions were violated. Refer to the Online Data Supplement for the summary of recurrent and terminal event analyses based on various statistical models and data assumptions.
A total of 13 patients died during the course of the phase 2 study (Table 2). None of the deaths were considered to be related to AAV1/SERCA2a or to the administration procedure. There was a trend for better survival for patients in the high-dose AAV1/SERCA2a group, with the survival probability over time higher for patients in the high-dose group versus placebo (P=0.11, log-rank test), as shown in the overall survival curve in Figure 4. Of the 13 deaths in phase 2, most deaths were related to the underlying disease, with the cause of death listed a pump failure. Three of the 13 deaths in phase 2 were not related to HF. Of the non–HF-related deaths, 1 death (placebo control) had a primary cause of recurrent colon cancer. No information was available surrounding the other 2 deaths, both in the mid-dose AAV1/SERCA2a group, which were discovered in the Social Security database after repeated attempts to contact the patients were unsuccessful.
A total of 39 patients were enrolled in phase 2 of the study with an average of 29 months of follow-up postinfusion, representing a total of 1122 patient-months.
Patients with advanced HF have nonsustained and sustained ventricular arrhythmias and atrial fibrillation at baseline. In all the treatment groups, there were no new arrhythmias that appeared in the first year of follow-up or an increase in pre-existent arrhythmias.
Persistence of Vector Sequences
Because SERCA2a is an integral membrane protein, there is no known surrogate that might indicate vector persistence. Testing for the presence of the SERCA2a transgene in the cardiomyocyte target tissue requires invasive cardiac biopsy. Because this procedure is associated with substantial risk in patients with advanced HF, biopsy collection of tissue was not included as part of the CUPID trial. Therefore, opportunities to collect tissue for testing were limited to cases of death followed by swift notification and rapid mobilization; cardiac transplantation, where the native treated heart becomes available; implantation of VAD, where core tissue becomes available; and certain other cardiac procedures in which biopsy is practical under the circumstances. The qPCR assay for AAV1/SERCA2a DNA has demonstrated the persistence of the SERCA2a transgene out to month 31 in the target tissue of 1 patient and to month 22 in another. A third patient demonstrated the presence of vector DNA at month 23. All 3 patients with qPCR-positive vector DNA results were in the high-dose (1×1013 DRP) group. Results for both phase 19,10 and phase 211 tissue samples tested by qPCR are provided in Table 3.
The current study results demonstrate favorable safety and efficacy signals in this 3-year follow-up. No increases in adverse events, disease-related events, or laboratory abnormalities were observed in any of the AAV1/SERCA2a-treated patients compared with those receiving placebo. There was no indication of an increase in any new occurrences or exacerbation of pre-existing clinical conditions or previous disorders during long-term follow-up including malignancy(ies), neurological disorders, rheumatologic or other autoimmune disorders, hematologic disorders, or other unexpected illnesses associated with AAV1/SERCA2a administration. Specifically, there was no evidence of changes in the occurrence of ventricular arrhythmia. One concern with refilling the sarcoplasmic reticulum with Ca2+ in the setting of a leaky ryanodine receptor has been the possibility of inducing ventricular arrhythmias. Multiple studies,13–15 in fact, have now shown that gene transfer of SERCA2a decreases the incidence of ventricular arrhythmias in the setting of HF models and improves arrhythmogenic substrate and triggers. The results from the CUPID trial would seem to extend these preclinical observations to the clinical setting.
In the CUPID phase 2 study at 12 months of active observation, prespecified adjudicated CV-related events in high-dose AAV1/SERCA2a–treated versus placebo patients were delayed, had a substantially reduced frequency, or both.11 The incidence rates of terminal events during the course of the study in placebo patients were in line with the expected rates in this patient population with advanced HF. In this report, the findings in the previous publication of the 12-month active observation data have been extended to include the full 3 years of follow-up prespecified in the protocol. In the additional 2 years of long-term follow-up, the durability of reduced CV and terminal events in the high-dose cohort has been maintained. In addition, the survival probability over time trended higher for patients in all AAV1/SERCA2a groups compared with the placebo group but especially in the high-dose group. Finally, persistence of the vector DNA as assessed by qPCR was demonstrated in high-dose patients in whom a biopsy was feasible but not from patients in the placebo or lower-dose cohorts. Although this was a small trial, the data from the long-term follow-up portion of the study further support the initial findings during the 12-month active observation period, in which the high-dose AAV1/SERCA2a group versus placebo group demonstrated improved HF status and reduced CV and terminal events.11
In addition, this is the first clinical application of an AAV vector for cardiac gene transfer. By virtue of its exquisite capacity to transduce postmitotic cells, AAV is currently the vector of choice for CV applications. Second, it is interesting to learn that a single intracoronary infusion of AAV1/SERCA2a has been sufficient to provide a potential therapeutic long-term benefit in a condition where the transgene must be expressed inside the cells and where a high efficiency of transduction is required. In this respect, we have taken advantage of the AAV serotype 1, which has known tropism for muscle tissue, a property shared with AAV serotypes 6 and 9. Serotype 9 has been shown to have cardiac tropism in rodents where tail vein injection results in homogeneous cardiac myocyte transduction.16 However, in larger animals, AAV 1, 6, and 9 have similar efficiencies of cardiac transfer.17–19 Third, a requisite for efficacy of cardiac gene therapy in treating HF is prolonged expression of the transgene, a requisite that AAV vectors seem to meet.
In the original report of the 12-month active observation, an efficacy signal with the group of patients receiving the highest dose of AAV1/SERCA2a was observed.11 This suggests that improvements in vector and delivery method might further improve the efficacy of treatment. It is not clear, however, whether higher doses might trigger a cellular immune response against the virally transduced cells, an observation found in several other studies with AAV vectors in humans.20,21 This strategy will need to be tested in future trials.
The long-term results presented in this report are promising and support the possibility that AAV1/SERCA2a therapy may be an important new addition to the armamentarium for treating chronic systolic HF. However, the number of patients is small, and the favorable effects seen in CUPID need to be replicated in larger groups of patients. Further clinical studies are now underway including an international study in 200 patients, testing whether AAV1/SERCA2a (1×1013 DRP) versus placebo, randomized 1:1, is an effective therapy to reduce CV events in advanced HF.
We would like to thank Janice Pogoda, PhD, for biostatistical and programming support.
Sources of Funding
Funding for this study was provided by Celladon Corporation, La Jolla, CA.
K.M. Zsebo, J. Rudy, and K. Wagner are employed at Celladon Corporation and have ownership interest (includes stock options); B. Greenberg receives consultancy fees; and R.J. Hajjar has ownership interest (includes stock and stock options and rights in patents) in Celladon Corporation.
In August 2013, the average time from submission to first decision for all original research articles submitted to Circulation Research was 12.8 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.113.302421/-/DC1.
- Nonstandard Abbreviations and Acronyms
- adeno-associated virus serotype 1
- recombinant adeno-associated viral vector containing human SERCA2a gene
- Calcium Up-Regulation by Percutaneous Administration of Gene Therapy in Cardiac Disease
- DNase-resistant particles
- heart failure
- neutralizing antibody
- quantitative polymerase chain reaction
- sarcoplasmic reticulum Ca2+ ATPase a
- ventricular assist device
- Received August 14, 2013.
- Revision received September 20, 2013.
- Accepted September 24, 2013.
- © 2013 American Heart Association, Inc.
- Bonow RO
- Mann DL
- del Monte F,
- Williams E,
- Lebeche D,
- Schmidt U,
- Rosenzweig A,
- Gwathmey JK,
- Lewandowski ED,
- Hajjar RJ
- Del Monte F,
- Dalal R,
- Tabchy A,
- Couget J,
- Bloch KD,
- Peterson R,
- Hajjar RJ
- Jaski BE,
- Jessup ML,
- Mancini DM,
- Cappola TP,
- Pauly DF,
- Greenberg B,
- Borow K,
- Dittrich H,
- Zsebo KM,
- Hajjar RJ
- Jessup M,
- Greenberg B,
- Mancini D,
- Cappola T,
- Pauly DF,
- Jaski B,
- Yaroshinsky A,
- Zsebo KM,
- Dittrich H,
- Hajjar RJ
- Cutler MJ,
- Wan X,
- Plummer BN,
- Liu H,
- Deschenes I,
- Laurita KR,
- Hajjar RJ,
- Rosenbaum DS
- Lyon AR,
- Bannister ML,
- Collins T,
- Pearce E,
- Sepehripour AH,
- Dubb SS,
- Garcia E,
- O’Gara P,
- Liang L,
- Kohlbrenner E,
- Hajjar RJ,
- Peters NS,
- Poole-Wilson PA,
- Macleod KT,
- Harding SE
- Prunier F,
- Kawase Y,
- Gianni D,
- Scapin C,
- Danik SB,
- Ellinor PT,
- Hajjar RJ,
- Del Monte F
- Pacak CA,
- Mah CS,
- Thattaliyath BD,
- Conlon TJ,
- Lewis MA,
- Cloutier DE,
- Zolotukhin I,
- Tarantal AF,
- Byrne BJ
- Beeri R,
- Chaput M,
- Guerrero JL,
- Kawase Y,
- Yosefy C,
- Abedat S,
- Karakikes I,
- Morel C,
- Tisosky A,
- Sullivan S,
- Handschumacher MD,
- Gilon D,
- Vlahakes GJ,
- Hajjar RJ,
- Levine RA
- Fish KM,
- Ladage D,
- Kawase Y,
- Karakikes I,
- Jeong D,
- Ly H,
- Ishikawa K,
- Hadri L,
- Tilemann L,
- Muller-Ehmsen J,
- Samulski RJ,
- Kranias EG,
- Hajjar RJ
Novelty and Significance
What Is Known?
Gene transfer of sarcoplasmic reticulum Ca2+ ATPase2a (SERCA2a) to the heart using an adeno-associated type 1 vector (AAV1) has been shown to be clinically effective in patients with severe heart failure during a period of 12 months.
It is not known whether AAV gene transfer results in long-term persistence in the heart or whether the clinical effects last.
What New Information Does This Article Contribute?
We found that recombinant adeno-associated viral vector containing human SERCA2a (AAV1/SERCA2a) gene transfer results in persistence of vector sequence or SERCA2a gene up to 31 months in cardiac tissues from patients injected with a high-dose of AAV1/SERCA2a viral construct.
Clinical event rates are lower 3 years after gene transfer in the high-dose group compared with the placebo group.
There were no untoward effects after AAV1/SERCA2a delivery during the 3-year follow up.
Heart failure is characterized by decreased activity of the SERCA2a, contributing to abnormal calcium cycling and impaired cardiac contraction–relaxation. AAV for gene therapy have been found to be safe and effective for myocardial gene delivery in the setting of heart failure. A First-in-Man randomized, double-blind, placebo-controlled trial of 3 different dose cohorts of AAV1/SERCA2a versus placebo was evaluated in 39 patients with advanced heart failure. At 12 months of follow-up, in the high-dose group versus placebo, there were improved functional status and left ventricular function with a significant decrease in recurrent cardiovascular events. However, it is not known whether AAV1/SERCA2a has lasting functional effects. In this study, we show that AAV1/SERCA2a results in persistence of vector sequences up to 31 months in cardiac tissues from patients injected with high-dose AAV1/SERCA2a. We found beneficial effects on clinical event rates 3 years after gene transfer in the high-dose group with no untoward effects in this patient population. The long-term results presented in this report of the first gene therapy trial in heart failure are promising and support the possibility that AAV1/SERCA2a therapy may be an important new addition for treating chronic systolic heart failure.