Therapeutic angiogenesis for coronary artery disease: clinical trials of proteins, plasmids, adenovirus and stem cells

Incidence

Coronary artery disease (CAD) continues to be the leading cause of morbidity and mortality in North America and Europe. This is despite significant advances in interventional cardiology and pharmacological treatments. Currently more than 12 million people in the USA have CAD, and more than 7 million have had a myocardial infarction. Chronic stable angina is the initial manifestation of CAD in approximately half of all presenting patients, and it is estimated that 16.5 million Americans currently have stable angina [1]. Surgical interventions are increasing, there were 1.3 million in-patient cardiac catheterizations, more than 1 million percutaneous transluminal coronary angioplasty (PTCA), and over 0.5 million coronary artery bypass graft (CABG) procedures performed in the year 2000 in the USA [2]. The economic cost of CAD in the USA is estimated to be more than US $50 billion per year.

Pharmacology

Four main pharmacological approaches are currently in use for treating ischemic heart disease [3], including:

Newer procedures including antiapoptosis reagents, antioxidants, preconditioning agents, and growth factors such as insulin growth factor (IGF)-1, have demonstrated strong cardiopro­tective actions in numerous preclinical studies, and may be important next-generation anti-ischemia therapies [4–7].

Interventional cardiology

Invasive procedures to treat severe CAD and angina including angioplasty with stent implantation and CABG have been increasingly used for the past two decades [8]. These proce­dures are usually employed only when occluded arteries do not respond to thrombolytic therapy or when the ischemia is unstable and unrespon­sive to pharmacology. In CABG, sections of the patient’s saphenous vein or internal thoracic artery are grafted around the occluded artery to bypass the occlusion and re-establish blood flow to the ischemic tissue. For single or double coro­nary arteries, CABG has an operative success rate of more than 95%, and the patency of thoracic artery grafts exceeds 90% at 10 years. Clearly, other coronary arteries of patients receiving bypass grafts remain at risk, but morbidity and mortality are significantly improved following these procedures. Percutaneous transluminal coronary angioplasty (PTCA) has a similar 95% operative success rate, significantly improves morbidity and mortality, but suffers from a high rate of restenosis. This latter problem may be sig­nificantly alleviated through catheter-mediated implantation of stents at the occlusion site [9]. Stent implantation reduces elastic recoil, plaque dissection, and the rate of restenosis. Conven­tional stents also suffer up to 20% restenosis and often require repeat procedures, but this is dra­matically reduced by using drug-eluting stents [10]. Polymeric stents eluting sirolimus (Rapa­mune^®) or paclitaxel (Taxol^®) have been proven to reduce or eliminate intimal hyperplasia and in-stent resenosis in multiple clinical trials, and are rapidly replacing conventional stents.

In combination with anticoagulant therapy and conventional drugs, invasive procedures are effective in reperfusing ischemic tissue and sal­vaging myocardium. They have contributed sig­nificantly to the vast improvement in the life expectancy of patients with heart disease that has occurred over the past two decades. On the nega­tive side, these procedures are expensive and carry a risk of reperfusion damage [11–13]. In addition, a significant number of patients with ischemic heart disease are not candidates for revasculariza­tion and/or receive incomplete revascularization. CAD is a progressive disorder that is usually dif­fuse involving multiple regions of the vasculature and multiple stenoses. CAD is also usually associ­ated with comorbid conditions such as diabetes, obesity, hypertension and old age, the effects of which are not restricted to specific regions of the vasculature. More than 10% of patients with symptomatic CAD are not suitable candidates for either PCTA or CABG [14,15]. These patients are the primary candidates for the relatively new pro­cedure called therapeutic angiogenesis. This pro­cedure involves the delivery of pro-angiogenic growth factors to the myocardium of patients with CAD to stimulate collateral vessel produc­tion and hopefully resolve the ischemia. Parallel procedures are being developed to treat peripheral artery disease [16].

Therapeutic angiogenesis

Risk factors for both coronary and peripheral artery diseases include hereditary, age, diabetes, hypertension, lifestyle, for example smoking, and serum lipid composition [17–19]. Evidence from both animal and patient studies indicate that several of these risk categories, in particular age, diabetes and hyperlipidemia are associated with depressed levels of angiogenic growth fators including vascular endothelial growth factor (VEGF) [19]. Growth factor deficiency and depressed angiogenic potential may therefore contribute to the establishment and progression of arterial disease. Preclinical studies initiated in the early 1990s provided conceptual proof for therapeutic angiogenesis, and supported the implementation of clinical trials of VEGF and fibroblast growth factor (FGF) [16,20–22]. VEGF and FGF are both multigene families, there are at least six distinct VEGF members and more than 22 FGF proteins. Most of the trials have focused on VEGF-A and -C, and FGF-1 and -2. Although these trials have generally been positive, the results are mixed in terms of the degree of therapeutic gain, and the dramatic responses seen in animal studies have thus far not been repro­duced clinically. There have been more than 1000 patient studies testing the different isoforms of VEGF or FGF protein or genes, and approximately 100 patient studies testing bone-marrow-derived stem cells. Here the clinical trial results shall be reviewed and discussed with hind­sight what may realistically have been expected from the delivery methods used, and how these may be improved.

Clinical trials

Fibroblast growth factor

Protein therapy

In the first therapeutic angiogenesis clinical study FGF-1 protein was injected directly into the left ventricles of 20 patients with three-vessel CAD that were undergoing venous bypass grafts [23]. For each patient, purified recombinant FGF-1, 10 µg/kg body weight, was injected at multiple sites close to the left anterior descending coronary artery (LAD) and distal to its anastomosis with the internal mammary artery (IMA). Transfemo­ral intra-arterial digital subtraction angiography 12 weeks after surgery showed pronounced accu­mulation of contrast medium at the site of injec­tion, and extending peripherally around the LAD for 3–4 cm distal to the IMA/LAD anastomosis. At the site of injection a capillary network could be visualized sprouting from the coronary artery into the myocardium. Stenoses distal to the anas­tomosis also appeared to be bridged by neovascularization. None of these effects were observed in patients receiving heat-denatured FGF. Follow-up 3 years later confirmed the safety and apparent efficacy of FGF-1 therapy. The capillary network seen at 12 weeks persisted on angiography, and echocardiography suggesting an improved left ventricular ejection fraction (LVEF) [24].

Pilot Phase I/II trials testing various dose and delivery methods have been reported for both FGF and VEGF. In a Phase I randomized double-blind placebo-controlled study, Laham and col­leagues implanted slow-release (3–4 weeks) heparin alginate microspheres containing 10 or 100 µg FGF-2 in the epicardial fat overlying via­ble ischemic myocardium of 24 patients [25]. Nuclear and magnetic resonance imaging (MRI) perfusion scans were performed at onset and again at 90 days. All patients in the 100 µg FGF-2 treat­ment group reported improved symptoms at 90 days, whereas three of the seven control group patients experienced persistent symptoms, and two required repeat revascularization. Nuclear perfusion imaging showed a significant reduction in the size of the ischemic target region in the 100 µg FGF-2 group, but not in the 10 µg FGF-2 group. As in the first study, the benefits of FGF-2 therapy were reported to be maintained at 3-year follow-up [26]. Positive results from several pilot open-label, dose-escalation studies of intracoro­nary single-bolus FGF-2 delivery were reported that supported a larger Phase II trial. In the pre­liminary studies, a total of 91 patients received intracoronary infusions of FGF-2 ranging from 0.33 to 100 g/kg [27–29]. Hypotension becomes dose limiting at 48 µg/kg, and transient mild thrombocytopenia and proteinuria occurred in some subjects. One study reported improvements of LVEF, target wall thickening and myocardial perfusion as well as quality of life (angina fre­quency, treadmill exercise tolerance) at 3- and 6-month follow-up [27].

In the FGF Initiating RevaScularization Trial (FIRST), 337 patients were enrolled in a multicenter, double-blind, Phase II trial that examined three different concentrations (0.3, 3, and 30 g/kg) of single intracoronary infusions of FGF-2 versus placebo controls [30]. Efficacy was evaluated by exercise tolerance test, myocardial nuclear perfusion imaging, and quality of life questionnaires. Irrespective of dose, the mean change in exercise tolerance test time was not sig­nificantly different after 90 or 180 days between treatment and control groups. Angina frequency was significantly reduced at 90 days, but not 180 days and there was no significant difference in stress nuclear imaging. Follow-up data at 32 months showed a trend for slightly sustained improvement in the treatment group [26].

Gene therapy

Adenoviral delivery of the FGF-5 or -4 genes by intracoronary infusion in a porcine model of myocardial ischemia were shown to alleviate ischemia and improve function [31,32]. This pro­vided support for the Angiogenic Gene Therapy trial (AGENT), a dose-escalating, double-blinded, Phase I/II trial of adenovirus (Ad) 5 encoding FGF-4 [33]. Atotal of 79 patients were randomized to receive either placebo, or one of five doses of Ad5-FGF-4 (3.3 × 10⁹ – 3.3 × 10¹⁰ total viral particles) also delivered by intracoronary infusion. At 12-week follow-up, exercise tolerance was not significantly improved in the test groups compared with placebo, and there were no differences in stress-induced wall motion scores as measured by echocardiography. Although subgroup analyses suggested some improvement of the most severely effected patients, these effects were also very moderate.

Vascular endothelial growth factor

Protein therapy

Several small pilot trials attested to the safety and tolerance of low dose recombinant VEGF administration by coronary infusion to patients with severe CAD [34–36]. These studies showed a trend for improved coronary flow and supported larger Phase II trials. The VEGF in Ischemia for Vascular Angiogenesis (VIVA) trial [37] was a multi­center, randomized, double-blind, placebo-con­trolled study of intracoronary and intravenous recombinant (r) VEGF-A infusion. A total of 178 patients were randomized to receive low dose (17 ng/kg), high dose (50 ng/kg), or placebo. VEGF was delivered by an intracoronary infusion followed by intravenous infusions at 3-day inter­vals. No significant improvements were reported in treadmill time or angina class in the treatment groups at 60 days or 1 year compared with pla­cebo. Interestingly, both treatment and placebo groups were significantly improved relative to con­trols indicating the powerful influence of placebo on CAD symptoms.

Gene therapy

Positive results, although in some cases marginal, were obtained in multiple small trials of VEGF (VEGF-C and VEGF-A variants) delivered as plasmid or adenovirus [38–40]. Typically these studies demonstrated slightly improved left ventricle ejection fractions and angiograms in treatment groups, and reduced anginal episodes at short-term follow-up. Losordo and colleagues [39] injected plasmid human (ph) VEGF₁₆₅ directly into the ischemic myocardium of five male patients with angina and documented CAD. A total of 125 µg of plasmid DNA was injected in 4 × 2mL aliquots into the anterolateral left ventricular free wall via a mini left anterior thoracotomy. No apparent short- or long-term side effects were associated with the procedures. The average postoperative hospital stay was 3.8 days. During 60 days of follow-up, all patients reported significant reduction in angina; LVEF was unchanged or improved, and coronary angiography showed improved Rentrop scores in all treated patients. The low dose of DNA in these studies is noteworthy, only 125 µg of VEGF plasmid was delivered to the ischemic myocardium; preclinical studies have used up to 4 mg of plasmid encoded VEGF to treat limb ischemia [41,42]. In a larger nonrandomized uncontrolled, dose-escalating trial 25 patients received phVEGF₁₆₅ and were followed for 180 days. Safety and tolerance were confirmed, and improvements in collateral filling and single pho­ton emission computed tomography (SPECT)-sestamibi perfusion was demonstrated at 60 days. In an attempt to target and contain transgene expression in the ischemic tissue, Vale and col­leagues used an electromechanical mapping (NOGA) catheter to identify viable tissue in ischemic hearts, and inject phVEGF₁₆₅ directly into the ischemic region. In the first, single-blinded pilot study phVEGF₁₆₅, delivered by intramyocardial injections in three patients, was compared with three patients that received pla­cebo injections [38]. Significantly improved per­fusion scores were reported in the treatment group at 90 days, and a reduction in angina fre­quency and nitroglycerin use at 1-year follow-up. These results were supported in a larger Phase II double-blind placebo-controlled study in which 19 patients were assigned to phVEGF-C or pla­cebo both introduced using the NOGA tech­nique [43]. Treatment groups showed improvements in exercise tolerance, anginal class, and a reduction in the area of ischemic myocar­dium at 12-week follow-up.

Therapeutic angiogenesis using adenoviral delivery of VEGF was tested in Phase I and II trials. In the former, 21 patients received Ad-VEGF₁₂₁ by direct intramyocardial injection, 15 patients received the therapy in conjunction with CABG and six received it as sole therapy [40]. Three patients in the former group each received 4 × 10⁸, 4 × 10^8.5, 4 × 10⁹, 4 × 10^9.5, and 4 × 10¹⁰ pfu Ad–VEGF₁₂₁ and the latter patients each received 4 × 10⁹. Injections were directed to the regions of reversible ischemia, determined by ^99mTc-sestambi perfusion scans. Trends toward improvement in angina classification and treadmill exercise testing were seen at 6 months in all groups, and there was a suggestion of improved exercise tolerance in the gene therapy-only group. The study also reported improve­ments in wall motion at stress in the region of vector administration and increased collateral vessels in the majority of patients. There were no significant adverse effects or increase in systemic VEGF associated with the treatments; however, antiadenoviral antibodies were detected in some subjects. The Kuopio Angiogenesis Trial (KAT) was a randomized, double-blinded trial of intrac­oronary infusion of VEGF₁₆₅ cDNA at the time of PTCA and stenting [44]. A total of 109 patients were included in the study and 90% received stents. Thirty-seven patients received Ad–VEGF (2 × 10¹⁰ pfu), 28 patients received phVEGF (2 mg plasmid DNA in liposomes), and 38 con­trol patients received Ringer’s lactate. The follow-up period was 6 months. No adverse responses were reported and myocardial perfusion was significantly improved (only) in the Ad–VEGF treatments.

Conclusions from protein & gene therapy

An optimistic evaluation of the combined results from VEGF and FGF trials could support the concept of therapeutic angiogenesis, and con­clude that the procedures used to implement the therapy are flawed. There are two broad possibili­ties that cannot yet be distinguished. One is that the target tissues respond poorly to the exogenous growth factors because multiple components of vessel regeneration are missing. A second possibil­ity is that cardiac muscle in the diseased tissue is responsive to pro-angiogenic factors, but the delivery methods do not provide adequate tem­poral or directional cues for the establishment of new collateral networks with mature functional vessels. In either case, some of these flaws may be overcome by introducing pluripotential stem cells into the diseased tissues with the capacity to dif­ferentiate into host tissues and simultaneously deliver pro-angiogenic cytokines.

Cell therapy

Preclinical trials

Endothelial progenitor stem cells (EPCs) have been isolated from the peripheral blood and bone marrow of animals and humans and have been shown to incorporate into active sites of angiogenesis in models of ischemia [45]. EPCs promote angiogenesis in ischemic hearts of mice, rats, dogs, and pigs [46–48]. EPCs or bone marrow mononuclear cells (BM-MNCs) are usually delivered to the ischemic tissue as intramuscular grafts by direct injection. Numerous studies have demonstrated the potential of these cells to pro­vide therapeutic benefit for myocardial ischemia. Intramyocardial injection of BM-MNCs follow­ing myocardial infarction (MI) resulted in decreased infarct size and increased regional blood flow in a porcine model [49]. Similar results were reported in a rat model of MI where human BM-MNC–CD34⁺ endothelial progenitors were injected into the tail vein [50]. The potential of bone marrow stromal cells (BMSCs) to develop into cardiomyocytes in vivo was also examined in rats [51]. Isolated BMSCs were expanded and labelled with the LacZ reporter gene ex vivo and infused into the aorta 2 weeks after coronary artery ligation. Histological analysis 4 weeks later indicated that the donor cells differentiated into cardiomyocyte-like cells, fibroblast, and endothelial cells in eight out of 12 recipient rats. Further studies confirmed the potential of BM-derived stem cells to produce functional cardiac cells as well as blood vessels [52]. Lineage-negative (lin^-), c-kit⁺ bone marrow cells, a population of hematopoietic cells enriched for hematopoietic stem and progenitor cells, were isolated from male transgenic mice expressing enhanced green fluorescent protein (EGFP). MI was induced by coronary artery ligation and 3–5 h later donor lin^-; c-kit⁺ hematopoietic cells were injected into the area adjacent to the infarct. Examination nine days later indicated that myocardial regen­eration was obtained in 12 out of 30 recipient mice. In a second strategy, endogenous bone marrow cells were mobilized by infusion of gran­ulocyte-colony stimulating factor (G-CSF) and stem cell factor. This treatment increased the in vivo population of stem cells able to contrib­ute to myocardial repair [53]. Cytokine treatment significantly improved survival following MI and resulted in the formation of vascular structures and myocytes occupying an average of 76 ± 11% of the infarcted myocardium. Hemodynamics and left ventricle functions were significantly improved and scar formation was reduced. The potential for bone marrow or circulating progen­itor cells to transdifferentiate into cardiac myo­cytes and repair infarcted tissue remains controversial, as two other groups reported no evidence that such cells could support myocar­dial regeneration, or indeed survive long-term in the host [54,55]. Two additional recent studies have demonstrated the existence of cardiac stem cells in adult mouse hearts [56,57]. These cells may have the potential to regenerate and repair infarcted myocardium and may be the basis for future therapy.

Clinial trials

Several pilot Phase I clinical trials reported positive results of autologous BM-MNC transplant into patients with ischemic heart dis­ease. In the first study, ten patients undergoing PTCA were transplanted with autologous BM-MNC via a balloon catheter placed into the inf­arct-related artery during balloon inflation. At 3-months follow-up the infarct region was signifi­cantly decreased compared with standard ther­apy, and wall motion in the ischemic region and hemodynamics were improved [58]. In a second study, BM was aspirated from the iliac crest of six patients undergoing CABG, and MNCs were isolated [59]. The day after isolation, immediately after CABG, 1 × 10⁶ AC133⁺ cells for each patient were injected by ten injections of 0.2 mL each along the circumference of the infarct border. All patients survived the procedures and follow-up was implemented 9–16 months later. Transthoracic echocardiography revealed a gain in LVEF and improved diastolic ventricular dimensions in four patients, and SPECT scans indicated markedly improved perfusion of the previously nonperfused or hypoperfused infarct zone in five patients. The authors acknowledge that they could not definitively attribute the pos­itive effects to cell transplantation because all patients received simultaneous CABG. A third study included eight patients with stable angina that was refractory to maximum medical ther­apy. The ischemic regions were identified by electromechanical mapping, and a mixed popu­lation of autologous iliac crest bone marrow cells was injected into the ischemic region through the NOGA catheter [60]. No adverse side effects of the procedures were reported. At 3-months follow-up, patients reported a reduction of angi­nal episodes and reduced use of nitroglycerine. MRI scans revealed no change of LVEF, but sig­nificant improvement in radial wall motion and thickening during pharmacological stress, and slight improvement in perfusion parameters.

Assmus and colleagues compared intracoronary infusion of BM-MNC with periph­eral blood-derived progenitor cells in 20 patients with recent MI. They reported significant improvements of LVEF, wall motion in the infarct zone and end-systolic left ventricular (LV) vol­umes in both cell-therapy groups compared with a reference group [61]. Echocardiography and quantitative F-18 fluorodeoxyglucose-positron emission tomogrphy revealed significantly improved contractile function, coronary blood flow reserve, and increased myocardial viability. In a similar study, Perin and colleagues enrolled 21 patients with end-stage ischemic heart disease in a prospective, nonrandomized, open-label study of BM-MNC delivered by injection into ischemic myocardium by NOGA catheter [62]. Ischemic viable myocardium was identified by electromechanical mapping and patients under­went 2 and 4 month follow-up. The study reported significant improvement of LVEF (20 to 29%; p = 0.003) and mechanical function of injected segments (p = 0.0005), as well as signifi­cant reduction of end-systolic volume (p = 0.03). In the MAGIC cell randomized clinical trial, 27 patients with MI who underwent PTCA and stenting were randomized into three groups [63]. The first group received peripheral blood derived stem cells and daily subcutaneous injections of G-CSF to mobilize the stem cells. The second group received only G-CSF and the third group was a control group, follow-up was 6 months. The cell treatment group displayed significantly improved treadmill exercise capacity (450s to 578s; p = 0.004), reduction of myocardial perfusion deficit (11.6 to 5.3%; p = 0.02), and increased LVEF (48.7 to 55.1%; p = 0.005) at 6 months compared with baseline. G-CSF administration was discontinued because it appeared to increase the incidence of in-stent restenosis.

Conclusions & future perspective

Overall, the clinical trials of VEGF and FGF delivered as protein, or by plasmid or adenoviral vectors have fallen short of expectations. The suc­cess of neovascularization of diseased tissues is determined by numerous factors, some of which are specific for the underlying pathology. Vessel development in the embryo involves three distinct steps referred to as vasculogenesis, angiogenesis, and arteriogenesis [64]. These steps involve multi­ple sequentially activated factors and receptors as well as negative regulatory factors that are mutu­ally dependent and coordinated. Vasculogenesis involves the assembly of endothelial cells into a primary vascular plexus followed by the incorpo­ration of smooth muscle cells, monocytes and per­icytes to form mature cells and mature contractile vessels. Angiogenesis involves vascular sprouting from pre-existing vessels, vascular fusion and intussusception, to form a functional capillary network. Vasculogenesis and angiogenesis in both embryo and adult tissue involves recruitment of endothelial and smooth muscle cells from pools of circulating progenitor cells that originate in the bone marrow. The homing signals that activate and promote the targeting of these cells involve concentration gradients of cytokines and ang­iogenic growth factors such as VEGF, hepatocyte growth factor, stromal-derived factor (SDG)-1, and IGF-1. In arteriogenesis, the vascular network is further remodeled to form large collateral arter­ies. This process involves additional sprouting, longitudinal migration, proliferation, and recruit­ment of more endothelial and smooth muscle cells. The lamina elastic interna is degraded dur­ing the process of arteriogenesis; monocytes and macrophages are recruited to the vessel walls, and the vessel diameter can increase by two- to 20-fold. Each step in the generation of the mature vessel network is modulated by different sets of factors that are interrelated and precisely regulated both spatially and temporally (Box 1).

The success of therapeutic angiogenesis in the treatment of CAD relies on the ability of the inter­vention to reactivate both angiogenesis and arterio­genesis, and to generate mature and stable conducting vessels. Most of the animal studies, and all of the patient studies to date have used sin­gle angiogenic factors; VEGF or FGF delivered as recombinant protein, or in plasmid or adenoviral vectors. In each case the duration of exposure of the ischemic tissue to growth factor is probably too short to allow arteriogenesis and may stimulate only a short burst of angiogenesis that cannot pro­duce mature vessels. Studies from the author’s lab­oratory argue in favor of this. Using a rabbit ischemic hindlimb model of peripheral artery dis­ease it was found that Ad-VEGF stimulated a rapid proliferation of endothelium during the first week of treatment that coincided precisely with a transient rise of serum-VEGF. The new capillaries were not stable and rapidly disappeared, probably by apoptotic death 1–3  weeks after treatment [Gounis et al. 2004, in press]. Other studies confirm that the exposure time to VEGF critically determines whether or not stable conducting vessels are pro­duced [65], and a recent report showed that chronic VEGF expression in rat hindlimb delivered by adeno-associated virus (AAV) promoted arterio­genesis as well as angiogenesis [66]. This latter find­ing was confirmed in the rabbit model [Webster, Unpublished data]. Preclinical studies have also shown that AAV vectors can be delivered with high effi­ciency to the heart [67].

Permanent delivery vehicles such as AAV with tightly regulated expression, and targeting of the angiogenic genes to the ischemic regions of tissue, may alleviate the deficiencies of current therapeu­tic angiogenesis protocols (Figure 1). Such opti­mized delivery of appropriately regulated factors may stimulate arteriogenesis and provide chem­oattractive signals for circulating peripheral blood derived mononuclear cells (PB-MNCs). Clinical trials of BM-MNCs currently look more promis­ing than trials with proteins or genes although this may change if the genes are delivered opti­mally. The relative efficacy of gene procedures versus stem cell procedures cannot be determined because none of these procedures have been run in parallel. It seems likely that the optimal therapy will include both stem cells and genes delivered in a manner that will allow the genes to target and possibly regulate the activity of the cells. The future of this technology may indeed rest in stem cells appropriately engineered with disease-responsive and regulated genes [16,22]. We have shown previously that a plasmid vector engi­neered with a hypoxia-regulated promoter is selectively expressed in the ischemic heart [68]. Additional fine-tuning of this regulation may allow the use of permanent delivery vehicles such as AAV in the treatment of ischemic heart and limb disease. The fine-tuning is required to fully silence expression of the transgene (VEGF, FGF) under normal perfusion conditions and activate it sufficiently to provide therapy during ischemia. This group has recently developed such a vector [69,70].