The dal-GenE trial: A case study of precision medicine in the cardiovascular setting
by Donald M Black MD, Therese Heinonen DVM, Marc Pfeffer MD and Jean-Claude Tardif MD
The dal-GenE study signals a new precision medicine paradigm for clinical trials that focus on assessments of cardiovascular (CV) endpoints in large clinical databases. In this example, we found that the retrospective analysis of the dal-OUTCOMES study indicated that patients in a genetically-identified subset (namely, recent acute coronary syndrome (ACS) event and the specific genotype AA at SNP rs1967309 of the ADCY9 gene) who are treated with the CETP inhibitor dalcetrapib were associated with reduced recurrent CV events. Key aspects of the dal-GenE trial included the requirement for a robust companion diagnostic, close attention to the predicate trial (focused on enrolling the same patient population), and frequent interaction with regulators. Subsequently, the study was efficiently recruited ahead of schedule and may function as a template for predictive studies to test pharmacogenetic hypotheses. We anticipate that this paradigm could lead to new therapy hypotheses and follow-up randomized controlled trials to determine the safety and efficacy of novel treatments.
While cardiovascular (CV) disease remains the principal cause of morbidity and mortality globally, significant progress has been made in reducing the risk of an initial CV event (such as myocardial infarction (MI) or stroke) and the recurrence of these events over the past 60 years.1 Much of that progress has leaned heavily on our increased understanding of risk factors, in particular, LDL-cholesterol levels. Reduction in LDL-cholesterol levels through therapeutic medical intervention, primarily the widespread use of statins, has played an important role in reducing the development of atherosclerotic plaque and subsequent CV events.2
The development of statins paralleled the understanding of molecular biology that,in turn, provided a window to the detailed pathogenesis and a path to potential treatment strategies. The elucidation of the causes of familial hypercholesterolemia was pioneered by Goldstein and Brown; they highlighted an inextricable link between genetic and non- genetic markers for which they were awarded the Nobel Prize in 1985. These researchers found that high serum cholesterol levels were due to the inability of some patients to remove cholesterol from the circulation, which led to cholesterol-rich lesions even in young patients. They established a genetic link between CV risk in patients with an inborn error of metabolism and dysfunctional or absent receptors.
Following the success of statins in reducing LDL-cholesterol levels, and thereby the risk of CV events in the treated “at risk” populations, many researchers switched their focus to HDL-cholesterol, another serum lipid which has an inverse relationship to cardiovascular risk. Unfortunately, existing HDL-cholesterol raising drugs (primarily nicotinic acid fibrates) did not consistently reduce CV events, despite raising total HDL-cholesterol levels by as much as 50%. Subsequent investigations indicated that the total amount of HDL-cholesterol may be much less important than the functionality of the HDL. Simply having an excess of cholesterol- laden HDL particles did not appear to result in an increased transport of cholesterol to the liver for expulsion from the body, as originally hypothesized. Research then shifted to assessing the functionality of HDL and whether some of the various types of HDL were more efficient in reducing atherosclerosis than others. Also, populations of patients with high levels of HDL-cholesterol and reduced CV disease were further evaluated to better understand the role of genetics in regulating lipid metabolism.
The inhibition of cholesteryl ester transfer protein (CETP) was first considered as a therapeutic target in 1990, following the description of a gene mutation in Japan that confirmed protection against premature atherosclerosis that was associated with elevated levels of HDL cholesterol and CETP deficiency.3 The subsequent development of CETP inhibitors focused on maximally increasing HDL cholesterol. The failure of a series of compounds developed by large pharmaceutical companies, however, cast a pall over the entire idea of inhibiting CETP4 While the effects of statins seemed to approximate 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMG-CR) genetic variants revealed in factorial Mendelian Randomization, the effects of CETP inhibitors (CETPi) were not consistent with CETP variants.5 The enormous investment in this class of drugs (estimated to be over $5 billion) seemed to be one of the most expensive failures in pharmaceutical development.
Genetic variation as the possible source of the variable effect of HDL-elevating drugs, especially CETPi, had been proposed by researchers prior to the initiation of the CETPi studies. Therefore, unlike the development programs for LDL-lowering drugs, all of the CETP developers had planned for a genotypic evaluation by collecting and banking DNA samples from consenting patients. Despite neutral findings of dal-OUTCOMES in 2012, F. Hoffmann La-Roche decided to conduct a large CV endpoint study to evaluate their CETPi, dalcetrapib, in patients with a recent acute coronary syndrome (ACS). They supported work at the Montreal Heart Institute to perform a post-hoc pharmacogenomic evaluation using a genome-wide approach on banked DNA samples. Pharmacogenomic, genome- wide association studies (GWAS) as an approach to drug development had replaced the previous single-gene discovery approach (pharmacogenetics).6 Of more than 5 million common genetic variants analyzed or imputed, a single region with genome-wide significance was found to be associated with CV events, but only in the patients receiving dalcetrapib. The GWAS identified SNPs in the adenylate cyclase type 9 (ADCY9) gene on chromosome 16 identified a single genotyped-SNP that passed the significance threshold — rs1967309. The association between rs1967309 ADCY9 and CV events was robust to further adjustments for risk factors and prior CV events.7
As can be seen in Figure 1, the results demonstrated that patients who were homozygous for the minor allele (AA) at variant rs1967309 in ADCY9 gene experienced a 39% reduction in the pre-specified composite endpoint of coronary heart disease death, resuscitated cardiac arrest, non-fatal MI, unstable angina with objective evidence of ischemia, atherothrombotic stroke, and unanticipated coronary revascularization with dalcetrapib compared to placebo. Patients homozygous for the genotype GG at rs1967309 and treated with dalcetrapib had a 27% increase in CV events versus placebo. Patients with GA showed no significant change in risk as compared to placebo. Further support was provided by post-hoc analysis in a second study comparing dalcetrapib to placebo. Some ADCY9-determined sub-populations had a reduction in intima-medial thickness (IMT), also demonstrating increasing risk associated with allele frequency. Subsequent analyses indicated that the differences in effect were not related to changes in drug metabolism between the genetic subtypes.8 These observations led to the challenge of determining whether the findings were by chance or a true medical advance.
These findings sparked renewed interest in dalcetrapib. After obtaining the world- wide rights to develop and commercialize dalcetrapib from Hoffmann La-Roche, DalCor Pharmaceuticals launched the dal-GenE study in 2016 to test the hypothesis that dalcetrapib treatment could reduce the risk of CV morbidity and mortality in patients with a recent ACS event, but enrolling solely patients with the AA genotype at variant rs1967309 in the ADCY9 gene.9
The dal-GenE study design
The prospective dal-GenE study is an ongoing randomized controlled clinical trial designed to confirm the findings identified through the retrospective dal-OUTCOMES GWAS evaluation. The overall design elements of dal-GenE have been composed to confirm the response to dalcetrapib is genotype related, namely, the AA genotype receiving the cardiovascular benefit originally expected of all treated patients. Since patients with other genotypes appeared to experience no benefit (AG), or potential harm (GG) following dalcetrapib therapy, the design of the dal-GenE study reflects in large part the design of dal-OUTCOMES, with the careful exclusion of patients with the AG or GG genotype at variant rs1967309 ADCY9.9 SNP rs1967309 is an Ancestry-Informative Mutation (AIM) and has been included in some commercial genealogy applications. The prevalence of rs1967309 ADCY9 AA homozygotes was subsequently supported by results of other CETPi studies and is consistently between 16-20% of the populations recruited.10
The clinical trial was intended to emulate typical medical practice and provide a minimum of barriers. Patients diagnosed with a recent ACS event (MI or unstable angina) provided informed consent to have blood drawn and to evaluated for the presence of the AA genotype at variant rs1967309 in the ADCY9 gene. Testing for the rs1967309 ADCY9 SNP was completed at one of three regional investigational testing sites with a clinical trial assay provided by Roche Molecular Systems, (RMS). Genetic analysis was conducted using the RMS cobas@4800 platform to perform automated sample preparation and detection of ADCY9 genotypes.
Other study entry criteria were designed to be as inclusive as possible, based on the broad safety experience with CETPi, and dalcetrapib in particular. Patients were randomized to receive investigational drug only when all entry criteria were confirmed, including the centralized assessment of the genotype. Patients were required to follow appropriate evidence-based guidelines for management of cardiovascular risk factors such as cholesterol and blood pressure. An equal number of patients were randomized to dalcetrapib or placebo, and patient follow-up included regular assessments at clinic visits. Study success depends on a significant delay in the “time-to-first-event” out of a composite of cardiovascular events that are considered most clinically meaningful—cardiovascular death, MI, stroke, and resuscitated cardiac death. The sample size of 6000 patients was considered appropriate, assuming an expected relative risk reduction of 22%, which is a bit more than half of the total benefit (39%) seen in the dal-OUTCOMES retrospective GWAS AA population. This provides the ability to detect as low as a 15% risk reduction, which, in this population, is broadly considered to be clinically meaningful.
The ACS population is commonly used in cardiovascular studies, and was the population studied in the original dal-OUTCOMES study. “Despite use of highly advanced strategies, including prompt coronary revascularization, intense statins, and antiplatelet agents, patients hospitalized for ACS have a high risk for recurrent events, and death, MI or stroke occurs in 7-10% within a year.”11 Compared to the general population, targeting a more-sensitive patient group assumes greater benefit and therefore smaller sample size. The study population is composed of patients with ACS diagnosed in the prior 3 months. The original dal-OUTCOMES study was designed to maximize the anti- atherosclerotic effects that appear to be more readily modifiable during the early post-ACS phase of the disease, as demonstrated in statin studies, including shorter post-ACS interval, such as MIRACL.12
As important as a possible CV risk decrease in the AA population treated with dalcetrapib is the possible increased risk in the GG population inadvertently treated. Therefore, the robustness of the genetic test device to exclude patients is critically important. The performance of the clinical trial version of the companion diagnostic had been carefully characterized prior to utilization in the trial. Genetic testing was performed at three central laboratories globally, in accordance with the local regulatory authorities, and supervised jointly by DalCor and RMS. Testing results were provided on average within 2-4 days of blood draw in all participating countries.
A total of 6149 patients from 810 investigational sites located across 31 countries located in North America, South America, Europe, Middle East, Africa, Australia and New Zealand were randomized into this ongoing clinical trial. The expected timeline of the dal-GenE study, based on the sub-group analysis of the dal-OUTCOMES study, indicated that the dal- GenE study was likely to last for 2 years after recruitment was completed in order to reach the target of 582 endpoint events (however, there was a great deal of uncertainty as to how long recruitment would last). Genetic testing on a broad scale as part of a clinical trial for cardiovascular patients had not been done previously, especially in a population post-MI, with many other urgent medical issues.
The management activities applied in preparation for this challenge included significant efforts in study site education and training, along with a bi-phasic consenting option designed to allow rapid and streamlined genotyping consent. These management aspects were particularly important given the high anticipated study screen selection rate inherent in a precision medicine trial targeting a population with a genetic match only 1 in every 5 or 6 assessments. In a cardiovascular research setting, such a screen selection rate was unprecedented.
Overall screening almost 45,000 patients with a recent ACS event in order to randomize over 6,000 patients with the AA genotype at variant rs1967309 in the ADCY9 gene for dal-GenE was accomplished more rapidly than screening 30,000 patients with a recent ACS event but without genotyping in the dal-OUTCOMES study, even with fewer investigational sites. The reasons for this success may be multifactorial, including the management aspects noted previously, as well as the enthusiasm on the part of clinical trial investigators, staff, and patients in precision medicine as assessed through limited polling of participating trial sites.
Discussion: The dal-GenE study
The dal-GenE trial is designed as a prospective study to confirm the findings of the retrospective analysis of the genotyped population in the larger dal-OUTCOMES study. Many of the design aspects reflect decisions made in the original study conducted by Hoffmann La- Roche, e.g., a template of the target patient population (ACS); the target geographic distribution; and laying the groundwork by familiarizing some clinical sites and regulators about the safety and tolerability of the investigational drug.
The dal-GenE study needed to demonstrate performance execution that was better than dal-OUTCOMES to allow for an efficient testing of the hypothesis by a small, start-up company. While some central ethics committees have reported higher enrollment and compliance in genetic-associated trials, the dal-GenE study increased enrollment at least 30% over dal- OUTCOMES, attesting to the interest in the community. Many success factors hinged on a seamless execution between the clinical sites and the companies developing the drug and the device (companion diagnostic), and their contracted organizations. Technical features of screening included: developing a robust test for the SNP; sites trained to submit the sample; logistics, including shipment across borders; plural linkages from sample acquisition through to randomization to drug supply to electronic data capture in 31 countries involving over 800 clinical sites.
Experience and Insights
The Precision Medicine Initiative, is an “emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” The concept has been around for a long time (including the frequent example of ‘typed’ blood) and has been successfully utilized in the field of oncologic drug development.13 The acceptance of genetic testing has significantly increased with the commercial availability of genetic testing, primarily for ancestral research.14
The discovery of the link between the ADCY9 gene and response to dalcetrapib therapy represents a unique opportunity in the field of cardiovascular disease. Improving the benefit -risk equation (as well as benefit-cost) is the primary driver for predictive medicine.
Companion Diagnostics and link to therapies
Several FDA guidance documents indicate that one avenue to approval requires simultaneous development of a therapeutic product and a companion diagnostic. The labeling of each product (in this case a small molecule oral drug and an in vitro PCR test) must include information on the selection and monitoring of patients. A draft guidance in 2016 provided further information on the pathway to development.15 Under the draft guidance, therapeutic developers have the responsibility to ensure that a diagnostic device is available and approved by FDA. Many other regulatory authorities have adopted a similar view. In a few cases, a development paradigm may embrace a “complementary diagnostic” as a non- mandatory test to identify patient populations that respond particularly well to a therapeutic.16 In the case of dalcetrapib/rs1967309 ADCY9, however, the test is as critical for selecting AA genotype patients that are likely to benefit from treatment as it is for excluding patients that could suffer adverse events (GG patients).
The FDA and medical product companies have focused on pairing the correct diagnostic with the appropriate drug (even if the respective labels for drugs and devices may diverge in some areas). In theory, the rs1967309 ADCY9 test could be categorized as a “class” companion diagnostic if sufficient evidence could indicate that other drugs in the class had a similar benefit in the same patient population, however, the evidence is not consistent in this regard. The collective experience with oncology products has been that the lack of documented drug-device linkage can lead to confusion in the medical community, hence, market surveillance is a consideration to assure that an FDA-approved test is utilized in all cases. This is especially important when the genetic marker (rs1967309) is Ancestry-informative and therefore available outside of the medical setting, wherein guidance may be lacking a risk/ benefit analysis.
Could ADCY9 be an important test for treatment with other CETP inhibitors? There is contradictory evidence as to the range of effect. While the potency of dalcetrapib was questioned (it has the least impact on HDL- and LDL-cholesterol of the CETPi that have completed phase 3), a recent analysis indicates that it is similarly effective in reducing new onset diabetes mellitus in the study population.12 Nevertheless, large CV endpoint studies for evacetrapib and anacetrapib were recently completed, and while both programs were subsequently terminated, evacetrapib was associated with a reduction in total mortality that was close to statistical significance and anacetrapib reduced major cardiovascular events by 9% over a 4-year period.
The lack of efficacy seen with CETP inhibitors in large studies (despite reductions in LDL-C) has led to multiple theories. Known genetic heterogeneity in HDL response, and pre- supposing a difference in HDL functionality, has led to the genetic assessments by the sponsor of the dal-OUTCOMES study. Fortuitously, all of these large CV studies included genetic sampling as part of the protocol. The findings of dal-OUTCOMES were directionally supported by findings in the evacetrapib analysis which showed a non-significant trend for a genotype- dependent reduction in CV endpoints, similar to a Mendelian Randomization assessment.17 While the anacetrapib analysis showed no effect by genotype, we note that the patient populations, as well as the pharmacology of the drugs, are also very different.10
Pharmacological interventions often need to be tailored to the population at risk, as well as defining the population most likely to benefit. The original approach to the CETP inhibitors may have been too simplistic. Upon observing a significant rise in HDL-cholesterol levels in initial dal-OUTCOMES studies, a significant effect on atherosclerosis and consequent reduced CV events was expected. While improved outcomes were not found in the general population tested, our retrospective analysis found that a subset associated with a specific SNP (rs1967309 on ADCY9) was a surprise responder group. Prior to starting any CETP inhibitor studies, there was already a perspective of genetic heterogeneity affecting metabolism, including dysfunctional HDL, especially in association with an MI. Therefore, we ascribe greater credence to the finding due to CETP association with lipid metabolism and atherosclerosis, rather than a positive finding in a disease with a different pathogenesis.
While we undertook an ambitious goal for a small start-up biotechnology company, we chose not to attempt immediately any further pre-clinical or clinical programs, as they might not provide better assurance of success. Rather, we executed a major medical morbidity and mortality study to determine the true results. Based on these results, we postulate that a positive result might be more likely to repurpose drugs that failed on efficacy in other indications but may be viable leads based on the association of genetic variation and expected patient response.18
Donald M. Black, MD, MBA, FACC Dr. Black is Chief Medical Officer for DalCor Pharma UK Ltd. Previously, he held general manager or research positions at Becton-Dickinson, GE, Merck, and Warner-Lambert Co. where he was responsible for the clinical development of Lipitor.
Jean-Claude Tardif, MD Dr. Tardif is the Director of the Research Centre at the Montreal Heart Institute and Professor of Medicine at the University of Montreal. He holds the Canada Research Chair in translational and personalized medicine and the University of Montreal endowed research chair in atherosclerosis.
Marc Pfeffer, MD, PhD Dr. Pfeffer is Dzau Professor of Medicine at Harvard Medical School; Senior Physician in Cardiology at Brigham and Women’s Hospital; Boston, Massachusetts.
Therese Heinonen, DVM Dr. Heinonen is Vice President of Clinical Operations at DalCor Pharma UK Ltd. She was a member of the Parke-Davis and Pfizer Global Research and Development cardiovascular senior management teams, and she founded the International Partnership for Critical Markers of Disease.
- Johansson, , Rosengren, A., Young, K., Jennings, E. Mortality and morbidity trends after the first year in survivors of acute myocardial infarction: a systematic review, BMC Cardiovascular Disorders (2017) 17: 53
- Mega, JL, Stitziel, NO, Smith, JG, Chasman, DI, et Genetic Risk, Coronary Heart Disease Events, and the Clinical Benefit of Statin Therapy, 2015, Lancet 385(9984) 2264-2271. Doi:10.1016/S0140- 6736(14) 61730-X.
- Inazu, A, Brown, ML, Hesler, , Agellon LB, et al., Increased High- Density Lipoprotein Levels caused by common Cholesteryl-Ester Transfer protein gene mutation, 1990 N Engl J Med 323:1234-8.
- LaMattina, , The Blockbusters that wasn’t. Forbes, Oct.17, 2017.
- Armitage J, Holmes MV, Preiss D Cholestryl Ester Transfer Protein Inhibition for preventing Cardiovascular Events J Am Coll Cardiol 2019 73; 477-87
- Chambliss, AB, Chan, DW, Precision medicine: from pharmacogenomics to pharmacoproteomics Clin Proteom (2016) 13:25, DOI 1186/s 12014-016-9127-8
- Tardif, J-C, Rheaume E, Perreault L-P L, et Pharmacogenomic Determinants of the Cardiovascular Effects of Dalcetrapib. Circgenetics. 2015; 114.0000663.
- Black, DM; Bentley D, Chapel S, Lee, J et. Clinical Pharmacokinetics and Pharmacodynamics of dalcetrapib Clin Pharmacokinet 2018 57: 1319-1367
- Tardif, JC, Dube, M-P, Pfeffer, , et al. Study design of dal-GenE, a pharmacogenetics trial targeting reduction of cardiovascular events with dalcetrapib Am Heart J 2020; 222: 157-165; see also dalGen-E NCT02525939: https://clinicaltrials.gov/ct2/show/NCT02525939
- Hopewell, JC, Ibrahim M, Hill M, Shaw PM, et al. and on behalf of the HPS3/TIMI55- REVEAL Collaborative Group, Impact of ADCY9 genotype on response to anacetrapib. Circulation 2019, ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.119.041546
- Schwartz GG Olsson, AG, Balantyne, CM, et Rationale and design of the dal-OUTCOMES trial: Efficacy and Safety of dalcetrapib in patients with acute coronary syndrome Am Heart J 2009; 158: 896-901; See also Dal-OUTCOMES NCT00658515 https://clinicaltrials.gov/ct2/show/NCT00658515
- Schwartz GG, Olsson AG, Ezekowitz MD, et al, Effects of atorvastatin on early recurrent ischemic events in acute coronary syndromes: the MIRACL study: a randomized controlled trial JAMA 2001; 285: 1711-8.
- Hamburg, M Paving the Way for Personalized Medicine: FDA’s Role in a New Era of Medical Product Development
- Jackson, M Genealogy of a Gene Princeton University Press, 2016 148-150
- FDA Principles for Co-Development of an In Vitro Companion Diagnostic Device with a Therapeutic Product 2016
- Scheerens , Malong A., Bassett K., et al Current Status of Companion and Complementary Diagnositcs, 2017, Clin Transl Sci 10:84-92.
- Ference, Personal communication, 2016
- Winslow, Wall Street Journal, Startup Aims to revive failed drug, 19 April 2016 – 1461038641