Stem cell therapy for the prevention and treatment of cardiac dysfunction holds significant promise for patients with ischemic heart disease. Excitingly early clinical studies have demonstrated safety and some clinical feasibility, while at the same time studies in the laboratory have investigated mechanisms of action and strategies to optimize the effects of regenerative cardiac therapies. One of the key pathways that has been demonstrated critical in stem cell-based cardiac repair is (stromal cell-derived factor-1) SDF-1:CXCR4. SDF-1:CXCR4 has been shown to affect stem cell homing, cardiac myocyte survival and ventricular remodeling in animal studies of acute myocardial infarction and chronic heart failure. Recently released clinical data suggest that SDF-1 alone is sufficient to induce cardiac repair. Most importantly, studies like those on the SDF-1:CXCR4 axis have suggested mechanisms critical for cardiac regenerative therapies that if clinical investigators continue to ignore will result in poorly designed studies that will continue to yield negative results.
Over a decade ago we hypothesized that stem cell-based repair of ischemic tissue is a natural response to tissue injury but that it is clinically inefficient because of the short-lived nature of the molecular signals regulating the process, not a lack of stem cells. At that time it was of great interest to us that stem cells in the blood stream homed to the myocardium in animal models of acute myocardial infarction (AMI), but not in models of chronic ischemic cardiomyopathy. These observations led us to investigate potential regulators of stem cell homing. We eventually identified stromal cell-derived factor-1 (SDF-1, aka CXCL12) as the key regulator of stem cell migration to sites of tissue injury.1 More recently studies have extended the relevance of the SDF-1:CXCR4 axis by demonstrating its critical importance in cardiogenic specification during development.2
As that initial observation we have demonstrated that transient engineered-cell-based3 or plasmid-based4 overexpression of SDF-1 in ischemic cardiomyopathy improved cardiac function. Furthermore, we have demonstrated that delivery of mesenchymal stem cells engineered to overexpress SDF-1 at the time of AMI leads to improvement in cardiac function.5 Research by our laboratory and others have demonstrated that mechanism of action of SDF-1 overexpression in AMI and chronic heart failure (CHF) are clearly multifactorial including both systemic and direct trophic effects.4, 5, 6 The initiation of endogenous stem cell-based repair appears blunted because of the natural short-term expression of SDF-1 at the time of AMI.1 Furthermore, there appears to be commonalities associated with the mechanisms of action of SDF-1 in AMI and CHF as well as some distinct differences. We were the first to show that SDF-1 leads to the recruitment of cardiac stem cells to the infarct and infarct border zone;6 an observation now verified by multiple groups.7 Correlating with the increased expression of SDF-1 in AMI and CHF is an increase in recruitment of bone marrow-derived progenitor cells and an increase in vascular density in the infarct border zone.3, 4, 8 In AMI there is a decrease in infarct expansion, decrease in cardiac myocyte apoptosis and an increase in surviving cardiac myocytes in the infarct zone.5, 8 Given the low level of cardiac myocyte apoptosis in ischemic cardiomyopathy, there is no decrease in cardiac myocyte apoptosis with the delivery of SDF-1; however, there is evidence for remodeling of the scar matrix.4
Most recently we have demonstrated that the effects of mesenchymal stem cells on the recruitment of cardiac stem cells, inhibition of cardiac myocyte death and improvement in cardiac function correlate directly to the ability of SDF-1 produced by the mesenchymal stem cell to bind its receptor CXCR4.8 These studies are consistent with recent reports that heterogeneity in SDF-1 expression defines the vasculogenic potential of adult cardiac progenitor cells.9 These findings further demonstrate the critical importance of SDF-1 in the initiation and maintenance of the endogenous stem cell repair process that we hypothesized existed almost10 10 years ago.1, 11
On the basis of the findings of our preclinical studies and the validation of our findings by multiple groups, we have recently completed a successful multicenter Phase I dose-escalation clinical study sponsored by Juventas Therapeutics, Inc., in which we examined the effects of SDF-1 overexpression in patients with class III CHF via the delivery of naked DNA plasmid encoding SDF-1 (JVS-100) (ClinicalTrials.gov-NCT01082094) using the BioCardia Helical Infusion Catheter. The findings of this study were presented at the 2011 American Society of Gene and Cell Therapy (Seattle, WA, USA).12
Plasmid gene delivery to the myocardium
Our findings suggest that myocardial signaling is critically important for stem cells to promote myocardial healing. Based on our understanding of SDF-1 signaling we believed that the intravascular delivery of stem cells at a time greater than several days following an ischemic myocardial event would be ineffective, because of a lack of an SDF-1 signal to home those cells to the ischemic region.1 Consistent with this hypothesis, results from the LATE TIME trial, in which patients were dosed with bone marrow-derived mononuclear cells or placebo 2–3 weeks after AMI, demonstrated no clinical benefits or patients receiving cells compared with the placebo-treated group. Given the design of the LATE TIME trial13 and our hypothesis that it would fail to show any benefit, we independently sought to investigate whether re-establishment of myocardial signaling through the overexpression of SDF-1 at a time after AMI similar to that studied in the LATE TIME trial would result in benefit. We further hypothesized that the benefit would be observed in the absence of stem cell infusion as there is constitutive stem cell release from the bone marrow that appears to be not only increased after AMI,14 but the stem cells appear to be more responsive to SDF-1 signaling.15, 16
Previous studies from our laboratory demonstrated that delivery of a nonviral DNA plasmid engineered to express SDF-1 resulted in cardiac benefit.4 This plasmid is named JVS-100. Before moving into larger animal models and human studies we set out to determine the key parameters that regulated plasmid dosing and potential application of different catheter delivery systems and approaches. For these studies we used the same JVS-100 plasmid construct with luciferase cDNA in place of SDF-1. This plasmid has been described previously and contained a CMV promoter and an RU5 enhancer element.4
Catheter-based delivery is the most commonly used strategy to deliver gene products to the myocardium. Despite several published reports spanning more 15 years of cardiovascular gene therapy clinical experience, we identified few references describing whether delivery was resulting in efficient gene uptake or if the administered doses resulted in significant gene expression. To better understand this relationship, we engineered JVS-100 to express the luciferase reporter plasmid (JVS-LUC). JVS-LUC was delivered to porcine hearts using different catheter systems. To characterize gene expression, 3 days post-delivery, we harvested the myocardium and assessed gene expression location and magnitude using a Xenogen chemiluminescence imaging system4 (Hopkinton, MA, USA) (Figure 1) or ELISA (Figure 2). Gene uptake and expression was observed using each of the delivery systems, namely: intra-coronary microsyringe injection (Figure 1a) retrograde infusion (Figure 1b) and endoventricular needle injection (Figure 2). For subsequent dosing studies and our clinical trial we chose to deliver JVS-100 using the BioCardia Helical Infusion Catheter. The BioCardia Helical Infusion Catheter allows for efficient delivery to the endocardium and has a helical needle that is rotated into the heart tissue to provide active fixation during therapeutic delivery, similar to the active fixation electrodes used in cardiac pacing.
As seen in Figure 2, using the BioCardia Helical Infusion Catheter, we varied concentration and volume of plasmid injection and found optimal delivery characteristics. These data demonstrate that one can increase gene expression over two orders of magnitude by increasing the volume of the injection to 0.5–1 ml per site. We also observed that plasmid concentrations above 0.5 mg ml−1 led to significantly greater gene expression. Based on these findings we fixed our volume of delivery at 1 ml and varied our plasmid concentration accordingly. It is interesting to note that these plasmid concentrations and volumes are significantly greater than those used in previous plasmid-based cardiovascular gene transfer studies.
Effect of SDF-1 overexpression at a time late after AMI
Based on these data we finalized our study to include the intramyocardial injection of 7.5 mg. (0.5 mg ml−1, 15 injections), 30 mg (2 mg ml−1, 15 injections) or 100 mg (5 mg ml−1, 20 injections) of plasmid encoding SDF-1. All injection volumes were 1 ml. For these studies we implemented a porcine AMI model in which a catheter is inflated in the left anterior descending coronary artery just after the first diagonal branch for 90 min. We injected SDF-1 plasmid 4 weeks after AMI in animals that had an ejection fraction <40% and an end-systolic volume >56 cc.
The data in Figure 3 demonstrate that we observed a trend towards a dose-dependent increase in troponin elevation in animals at 6 h after injection that was gone by 24 h.
We quantified SDF-1 plasmid copy number as a function of space and time after injection (Figure 4). We observed significant increases in SDF-1 plasmid in all part of the myocardium at 3 days post-injection. For example, while we observed generally less JVS-100 in the right ventricle, but still greater amounts at day 3 compared with days 30 and 90 after injection. As shown by the mean trend lines in Figure 4, there was significant loss of copy number of JVS-100 by 90 days after injection.
The data in Figure 5 depict echocardiographic parameters (ejection fraction and left ventricular end-systolic dimension) as a function of dose and time after SDF-1 delivery. There was no significant change in these parameters at 30 days after SDF-1 delivery (60 days after AMI). Consistent with the overexpression of SDF-1 catalyzing a repair process through the recruitment of stem cells from multiple sources, we observed significant increases in ejection fraction and decreases in left ventricular end systolic volumes at 60 days versus control after SDF-1 delivery in those animals that received 7.5 or 30 mg. No toxicity or benefits were associated with the 100 mg dose.
The improvements in echocardiographic parameters in the 7.5 and 30 mg dose groups were associated with significant increases in vascular density in the infarct border zone 30 days after treatment. The timing of the increase in vascular density relative to the improvement in cardiac function at 60 days suggests that tissue remodeling occurs in advance of improvements in cardiac function. Importantly, the lack of benefit in the 100 mg dose group was associated with no significant increase in vascular density (Figure 6).
Despite administering relatively high doses of plasmid to the myocardium, we did not observe significant evidence of plasmid at noncardiac sites (Table 1), nor was there observed toxicity at any dose as determined by an independent blinded pathologist.
As the early findings of Orlic et al.17, 18 the field of cardiovascular regenerative medicine has aggressively pursued the potential of stem cell therapy for the prevention and treatment of left ventricular dysfunction in patients with myocardial infarction. The field is evolving from the clinician investigator driven ‘have cell, will inject’ strategy to one in which there is a growing body of mechanistic biology that needs to be utilized for the purpose of scientifically rigorous clinical design. Whether it be pre-cardiac specification of mesenchymal stem cells to optimize stem cell function, induction of SDF-1 through mechanical means before stem cell infusion or gene therapy to induce stem cell recruitment, each of these approaches, and many others, are hypothesis driven based on an emerging understanding of the molecular mechanisms involved. If we fail to recognize and respond to the biology at hand we will not exploit the full potential of regenerative therapies.
There are clear hypotheses of the mechanisms of action that should be and need to be tested that will significantly advance the field and more importantly bring potentially efficacious therapies to patient populations at need.19, 20 Importantly and to the fields benefit, stem cell therapies and the mechanisms involved have translated very well from animal models to clinical studies,21 as well as other organ systems.22, 23
Driven by the hypothesis that stem cell-based repair of the heart is a natural process that is clinically limited secondary to the lack of molecular signaling, our mechanistic-based approach to regenerative medicine has led us to describe multiple potential tissue repair regulators, including SDF-1,1, 4 disabled-2,24, 25 MCP-3(ref. 26) and others. This approach has further aided us in understanding the role of these factors in cell-based strategies8 and defining the role of stem cell-based repair not only on preservation of cardiac myocytes5, 8 and left ventricular remodeling,3, 4 but also in preventing re-entrant arrhythmias.27, 28
The findings reported herein have multiple implications on the field of cardiovascular gene therapy. Perhaps most importantly our findings suggest that prior plasmid gene transfer studies may have been significantly under dosed. The dosing we used was 5–10 × that was used in previous studies.29, 30 Furthermore, the injectate concentration and volume clearly modulate myocardial gene expression and need to be carefully accounted for in future trials. We also invoked an RU5 enhancer element to increase peak gene expression which significantly increased the biological effect observed in early rodent studies.4 Taken together, these findings demonstrate that gene therapy delivery parameter optimization can significantly improve the level of gene expression and ultimately the biological effect. Finally, our observation of a correlation between plasmid concentration and mild troponin elevation suggests a potential clinical parameter through which we can monitor myocardial transfection.
The need to prevent or treat cardiovascular dysfunction is significant. The re-establishment of antegrade flow, the use of cholesterol-lowering agents, the addition of b-blockers and angiotensin converting enzymes-inhibitors have all significant improved the outcome in patients with ischemic heart disease.31 Regenerative therapies offer the potential to further optimize left ventricular remodeling, recruit additional myocardial work and restore blood flow to the ventricle all with the goal to maximize cardiac performance.4, 19, 32, 33 As shown here, through careful delineation of relevant molecular mechanisms we can define key molecular targets for which gene transfer can be utilized to induce endogenous repair systems, define parameters that optimize clinical trial design and develop therapeutics that will be potentially less complicated to implement, safer for the patients and less costly to the system.34
This work was funded by Juventas Therapeutics and the Skirball Foundation.