The practice of medicine, particularly cardiovascular medicine, in the 21^st century will be increasingly driven by the improved understanding of the molecular mechanisms of disease. Nabel illustrated this with a hypothetical case study in the year 2015 of a male patient presenting for a routine exam, concerned about his risk for early onset of colon cancer and coronary artery disease, due to the early death of his parents from these diseases. A history and physical show he is in good health, exercises regularly, has a blood pressure of 110/80 mmHg, heart rate of 65 and a normal cardiovascular (CV) exam. His cardiac risk factors are male gender and elevated total cholesterol and LDL.

DNA genotyping is done, as this is very common in the year 2015, revealing no elevated risk for colon cancer based on his normal p53 and p16 genes. A 15% risk for Alzheimer's disease due to beta-amyloid protein abnormalities is revealed and a 45% increased risk for coronary artery disease (CAD) as he is a heterozygote for LDL-receptor mutation or deficiency. Further testing is done. Exercise tolerance testing reveals an ST-segment elevation in the anterior leads at only 5 minutes into testing and a low heart rate of 122. Magnetic resonance (MR) angiography is performed, as cineagiography is no longer used, revealing a proximal LAD lesion. Using real-time MR, a DNA-coated stent, encoded for an antithrombotic or antiproliferative gene, is placed in his LAD. He is then treated with aspirin and a fifth-generation statin, and a yearly colonoscopy is recommended.

In this invited lecture, Nabel discussed the molecular mechanisms of some vascular diseases as an entree to these types of therapeutic approaches, focusing on human vascular proliferative disease and the role of cell cycle proteins in regulating vascular cell proliferation. A novel DNA binding protein was also reviewed.

Atherosclerosis being a systemic disease with complications occurring at local sites of the circulation led to the concept of using catheters commonly used for diagnostics and therapeutics to deliver recombinant genes into the blood vessel. The pathogenesis of the atherosclerotic lesion has been well characterized. In brief, when smooth muscle cells (SMC) are stimulated by mitogens after vascular injury, the SMC proliferate within the media of the artery, migrate up through the internal elastic lamina into the intima where they continue to proliferate and contribute to the lesion through the elaboration of the extracellular matrix as well as proliferation.

Excessive cell proliferation characterizes a number of human diseases loosely called human vascular proliferative diseases. These include coronary artery stenosis, in which stents have treated elastic recoil but in-stent restenosis remains high. Coronary and peripheral artery bypass grafts, failure of ateriovenous fistulas in hemodialysis patients, superficial femoral artery stenosis, and peripheral artery stenosis are also vascular diseases characterized by abnormalities in cell proliferation.

To elucidate the molecular mechanisms regulating vascular smooth muscle cell (VSMC) growth in proliferative disease, Nabel's group has recently focused on the cell cycle. Since VSMC are plastic they can re-enter the cell cycle, in contrast to cardiac myocytes. Normally quiescent with low proliferative indices in G serum, when VSMC are stimulated to divide by mitogens they enter the G1 phase of the cell cycle. Progression through the G1 phase is regulated by the assembly and phosphorylation of cyclin and cyclin-dependent kinases, predominantly cyclin D, CDK-4, and CDK-6, as well as cyclin E and CDK-2. These are critical for normal progression through the G-1 checkpoint in the S phase in which DNA synthesis occurs and there is a commitment to mitosis. The G1 checkpoint is critical for the phosphorylation of the retinoblastoma gene product and releasing E2F and other transcription factors.

Cyclin-dependent kinase inhibitors (CKIs) are endogenous inhibitors of the cyclin-CDK complexes and two families have been identified recently. The CIP/KIP family is comprised of p21, p27, and p57, and the Ink family is comprised of p15, p16, p18, and p19.

Characterization of the proteins p21 and p27

The protein p27, first identified and cloned by Toyoshima and Hunter using a yeast two-hybrid approach using cyclin D as bait, controls cell proliferation in response to many normal mitogenic stimuli. p27 safeguards against excessive cell proliferation in specific pathological settings, such as vascular injury, experimental glomerulonephritis, retinal dysplasia, and pituitary tumors.

Studies have tried to link the function of p27 with a variety of cancers. Although p27 is not mutated or deleted within human epithelial tumors, it is downregulated in breast, colon, and prostate cancer, likely via enhanced degradation by the ubiquination system. This occurs by shuttling of p27 from the nucleus to the cytoplasm where it undergoes ubiquination by proteosome complexes. This shuttling is mediated by the Jab1 gene product p38, as recently identified by Tomoda and colleagues.

p21 is an inducible CKI that alters their activity in phase G1, and is the human homologue of the FAR1 and SIC1 gene products defined in yeast. It is also known as WAF1, CIP1, or SDI1, and exists in catalytically active complexes with PCNA, cyclins, and CDKs. p21 is induced upon DNA damage by p53 and mediates G1 cell cycle arrest, at which point DNA repair is accomplished before DNA synthesis in the S phase. In SMC, p21 also has p53 independent effects. p21 null mice exhibit B cell lineage abnormalities and late onset tumors, in contrast to the p53 null mice that have early onset tumors.

Elucidating the role of p21 and p27 in vascular disease

The mitogenic pathways by which growth factor receptors and signal transduction pathways allow SMC proliferation has been a large focus of vascular biologists. Less understood is what turns off cell proliferation during the later phases of arterial repair.

Nabel and colleagues hypothesized that p21 and p27 may be important negative regulators of cell growth. To address this question, they first showed that p27 is constitutively expressed in SMC within the media as well as by occasional SMC and endothelial cells in the intima in an uninjured pig artery. After balloon catheter injury to the artery, p27 is rapidly downregulated and is at a very low level until day 7, during the time when cell proliferation in the lesion is increasing. By day 14, p27 expression is seen again and increases over time. At day 60 when the lesion is quiescent with low proliferation indices p27 continues to be constitutively expressed. In the lower region of the intima, SMC express p27, procollagen, and TGF-beta.

In contrast, p21 is not constitutively expressed in the normal artery in this model, and its expression is only seen in the later phases of arterial repair at day 21. At day 60 p21 expression is turned off.

Working hypothesis for p21 and p27

The initial working hypothesis, based on these studies, was that p27 is expressed in a normal quiescent artery potentially to keep cells out of cycle. However, upon vessel injury, p27 is rapidly downregulated and a number of mitogens are produced by SMC, platelets, and other cells that stimulate cell proliferation. Proliferation peaks at about one week in this animal model and then rapidly declines.

Studies of p21 and p27

Expression, cell culture, and in vivo studies of the CKIs p21 and p27, discussed hereafter, have shown they are expressed in both human and porcine tissue, but the overexpression of these lead to negative growth regulation. The mechanism is likely mediated through differential regulation of CDK-2.

The CKIs p21 and p27 inactivate cyclin-CDK complexes in the G1 phase, leading to cell cycle arrest, and thus function in growth regulation and wound repair. p27 is constitutively expressed in normal arteries, is downregulated after arterial injury, becomes upregulated during the later phases of arterial repair, and is inversely correlated with VSMC proliferation. p21 acts in concert with p27 in the later phase of wound repair.

The KIP/CIP and Ink CKIs differentially regulate CDK-2 and CDK-4 in VSMC. This leads to differences in the inhibition of VSMC proliferation in vitro and in vivo. Expression of p27 and p21 in VSMC inactivated CDK2 and CDK 4 activity. These different molecular mechanisms could account for observed differences in vivo.

Expression Studies of p21 and p27

In the animal model Ross and colleagues defined an autocrine pathway by which extracellular matrix secreted by SMC can feed back onto the SMC via the collagen receptor, alpha-2 beta-1 integrin, to cause an upregulation of p27 and p21, simultaneous G1 arrest, and downregulation of cyclin E and cyclin A. Thus, the CKIs are endogenous negative regulators of SMC that can turn off cell proliferation by promoting G1 arrest in the later phases of arterial repair.

In coronary arteries of patients who had undergone cardiac transplantation, Nabel's group found expression of p27 within SMC in the media and the intima. However, p21 was not seen in diffuse intimal lesions and was rarely seen in early atherosclerosis, but was commonly seen in late atherosclerosis, often in association with p27. In contrast the Ink CKI, p16, was not present in any of the coronary lesions, similar to their findings in the pig model where p16 was expressed only very early after vascular injury for unapparent reasons. Expression of p21 and p27 in SMC is also seen in new capillaries formed within the atherosclerotic plaque, and within macrophages within the intimal lesion in non-replicating cells.

Cell Culture Studies of p21 and p27

A series of in vitro cell culture studies were undertaken to understand the mechanism of CKI regulation. p16, p21, or p27 were expressed in porcine SMC using adenoviral vectors. Expression of p21 and p27 led to complete inhibition of cell growth, while p16 led to only partial inhibition.

A series of kinase assays then showed that the Ink and the CIP/KIP family members inhibited phosphorylation of cyclin D and CDK-4. Only p21 and p27, in contrast to p16, inactivated CDK-2 kinase activity in SMC. The differential regulation of cyclin E and CDK-2 activity in SMC is thought to account for the powerful effect of the CIP/KIP family members in regulating SMC growth, as opposed to the Ink family member p16.

In Vivo Studies of p21 and p27

Using an in vivo animal model of gene transfer, these observations were studied further. Balloon-injured pig arteries infected with either a p16 or p27 adenoviral vector showed that p27 inhibits SMC proliferation. At 3 weeks there was a reduction in lesion size by p27 infected arteries, as opposed to p16 or control arteries treated with either saline or control adenoviral vector.

The effects seen with p27 mirror the effects previously seen in response to p21, where overexpression of p21 also limits the development of the lesion at 3 weeks compared to the two control groups.

A mouse model of vascular injury and one of atherosclerosis formation are being used to test whether p21 and p27 are protective against vascular disease. These studies are ongoing, but preliminary results show that p27 null mice display abnormalities of growth regulation. The major phenotype is benign hyperplasia, predominantly in the endocrine organs (thyroid, ovaries, adrenal, and pituitary glands). VSMC quiescence is maintained by p27 and p21. The absence of these CKIs lead to accelerated VSMC growth and lesion formation during arterial repair.

Vascular injury studies

To address vascular injury, a p21 knockout, a p27 knockout, and a p21/p27 knockout is being studied using a C57Bl6 mouse as a control. Each was backcrossed against the C57Bl6 for 10 generations to remove any strain differences. The p27 knockout mouse is larger than the control mouse at an equivalent age because the absence of p27 leads to benign hyperplasia of many organs.

The development of a mouse model to test the response to vascular injury was undertaken. In the femoral artery wire, injury led to a very brisk lesion within several weeks. By 3 weeks in the C57Bl6 mouse, there was a very brisk intimal lesion and by 4 weeks the vessel was nearly occluded; this is not thrombus formation as wire injury denudes endothelium but does not produce thrombus.

The preliminary findings at two weeks post-vascular injury show a very small intimal lesion forms in the C57Bl6 mouse, while the largest intimal lesion is seen in the p27 null mouse. A representative photomicrograph from the p27 null mouse shows the SMC have a very dysplastic, unorganized feature within the lesion, not the concentric array of SMC interspersed with elastic tissue. This dysplastic appearance is very characteristic of that seen in some of the endocrine organs in the p27 null mice.

In contrast, the lesion size in a p21 null mouse is about one-half that seen in the p27 at two weeks. However, at 4 weeks the p21 catches up to the p27 null mouse. Interestingly, at 4 weeks in the p27 knockout the cellular proliferation is exuberant enough to cause a nearly complete occlusion of the blood vessel. In the combination p21/p27 knockout there is no difference between the single or the double knock-out mouse at 2 or 4 weeks.

Atherosclerosis formation studies

To test their hypothesis about atherosclerosis formation, they crossed the CKI deletion against an apoE null predisposing the animal towards atherosclerosis development, and created a series of double and triple null animals (p21/apoE null, p27/apoE null, p21/p27/apoE null). The C57Bl6 served as the control.

The absence of apoE accelerates intimal formation. Immunohistochemistry shows the C57Bl6 lesion is predominantly an SMC-rich lesion, with macrophages in the adventitia and a paucity of T cells. In contrast, in the apoE null mouse, there is a combination of an SMC-rich and inflammatory-rich lesion, with a larger number of macrophages within the intima compared to the C57Bl6, and occasional T cells in the adventitia and intimal lesion. The studies in the apoE crosses are underway.

To understand the regulation of p27 itself within SMC, Nabel's group used a yeast two-hybrid approach, with a human B cell library, to potentially clone novel DNA binding proteins. The carboxy terminal domain of p27 was used as bait to avoid pulling out a number of the cyclins and CDKs.

Initially, about 50 clones were assayed based on the strength of their beta-gal staining. Nabel discussed the clone initially called C21 and subsequently named hKIS. C21 bound p27 in the carboxy terminal domain and had very strong beta-gal staining.

Since C21 is 99% similar to a protein previously identified in the rat, serine/threonine kinase and called kinase interacting stathium (KIS), they concluded that C21 was a human homologue to the rat KIS and named it hKIS. hKIS is a 46.5 kDA protein consisting of an NH2 terminal serine/threonine kinase consensus region and a carboxy terminal region that binds p27.

Through a series of GST proteins they found that hKIS specifically bound p27, in contrast to other CIP/KIP CKIs, p21 and p57, as well as the Ink CKI p16. A series of biochemical studies have shown that hKIS functions to phosphorylate p27 in the N terminal region. A series of mutational studies showed the phosphorylation site is at position 10 on serine. When serine is substituted for by alanine at position 10 phosphorylation of p27 is abrogated by hKIS.

hKIS is located predominantly in the nucleus as identified by GFP studies, while cellular extracts show that p27 is located in the cytoplasm and the nucleus. The binding of hKIS to p27 is predominantly nuclear.

In SMC, CKI can be expressed alone or in combination with hKIS to look at the effects on cell cycle. When hKIS is overexpressed, normal cell cycle progression is maintained. When p27 is overexpressed, G1 arrest is produced. However, when p27 and hKIS are co-expressed while controlling for transfection levels, hKIS abrogates cell cycle arrest by p27 allowing progression into G2M. In contrast, p21 alone arrests the cell cycle and overexpression of p21 and hKIS continues cell cycle arrest.

hKIS working hypothesis

Nabel's group hypothesizes that hKIS is a binding protein that uniquely regulates p27 function. hKIS binds p27 through the carboxy terminal domain and then phosphorylates and inactivates p27 upstream of a cyclin-CDK binding site.

Further, they hypothesize that hKIS may be an upstream regulator of p38, important for inactivating p27, and preparing it for translocation out of the nucleus into the cytoplasm and degradation. As a regulatory protein for p27, hKIS can bind and thus inactivate p27 and potentially prepare it for binding by the Jab 1 gene product, p38. p38 has been identified to be important for translocation of p27 from the nucleus out to the cytoplasm where it is then degraded by ubiquitin-conjugation system within proteosomes. The studies with hKIS are ongoing.

Several approaches may be taken to investigate the methodologies for vascular gene transfer to develop approaches for treating human vascular proliferative diseases. Nabel's group has predominantly focused on regulators of the cell cycle. Other investigators have pursued other approaches, including the overexpression of a mutant form of the Rb gene product by Leiden and transcription decoys to E2F by Dzau.

Nabel's group used a thymidine kinase (TK) approach early on and she reviewed these studies as proof of principle that SMC proliferation could be disrupted during vascular injury and translated into therapeutic application.

The rationale for using TK is that it can achieve a bystander effect in cells, and hence create a larger biological effect in a greater number of cells than the actual number of cells transfected. This was important when the studies were begun five years ago because the efficiency of gene transfer, even with adenoviral vectors, was not overwhelming in the vasculature.

Normally TK is not expressed in mammalian cells. When it is overexpressed using a vector approach it encodes for an enzyme that phosphorylates the prodrug ganciclovir (GCV) or acyclovir (ACV), nucleoside analogues, into a phosphorylated form that is incorporated into replicating DNA, causing chain termination and cell death. By a poorly described bystander effect, byproducts are then diffusable to adjacent cells where again they are incorporated into replicating DNA, and can kill adjacent dividing cells in which the gene is not expressed.

The TK approach has now been tested in a number of animal models of restenosis and vascular proliferative disease by a number of laboratories in the US with similar findings across groups. It is a fairly potent way to disrupt VSMC proliferation. Using heterologous promoters is non-permissive in that they will also disrupt proliferation of other cell types. Nabel showed in rabbit hyperlipidemic arteries, for example, that macrophage proliferation is inhibited.

Modifying the TK vector for human therapy

Nabel's group studied whether they could modify and enhance the TK vector before attempting potential human therapy. They incorporated an internal ribosomal entry site to obtain expression of two genes from the vector. Guanylate kinase (GK) was used to enhance phosphorylation of TK. The TKciteGK vector was then tested with a variety of promoters in an adenoviral construct to test the strength of the heterologous promoters, RSV, EF1-alpha, and CMV in SMC. The two pro-drugs ACV and GCV were also tested.

In testing SMC in culture, the largest percentage of cells were killed with the combination of the adenoviral vector with CMV and GCV. This was then used in an animal model of vascular proliferation using a channel balloon with a stent mounted on the balloon catheter. When the catheter is inserted within the artery, the stent is deployed, and the vector can be infused through the balloon catheter through very small pores within the channel of the balloon. The vector is then mechanically introduced into the stretched artery. Thus, two components that lead to the pathophysiology of restenosis is treated: elastic recoil through stent deployment and the proliferative component through the vector TKciteGK. A number of other anti-proliferative vectors could be used.

Nabel's group has tested this approach in peripheral pig arteries using an EF1-alpha and a CMV promoter and two different pro-drugs, ACV and GCV. Saline controls were used. The smallest intimal lesions in the artery after 3 weeks were obtained with the combination of the CMV promoter with the TKciteGK vector in combination with GCV.

This seemingly straightforward and basic pharmacology is the type of testing of promoter-vector-drug delivery devices that will be absolutely essentially as these technologies are brought forward to the clinical arena.

Collaborating with another laboratory, Nabel has recently looked at a SMC-specific promoter, SM22-alpha, which was used to drive expression of two different reporter genes as well as the TKciteGK vector. These vectors have been tested in a series of cell culture and animal models.

With the SM22-alpha promoter in cell culture there is faithful expression in pig aortic SMC, pig jugular venous SMC, and mouse aortic SMC. But, not in non-SMC lines in NIH3T3 cells or pig endothelial cells. In contrast, the heterologous promoter RSV drives expression of the reporter gene in all cell types.

The expression being driven by an SM22-alpha promoter, in arterial SMC at least, is at least 3 log-fold lower than by an RSV or a CMV promoter. This suggests that the levels of expression achievable in vivo would be far less with a cell specific promoter. The cell specific promoter might offer increased safety advantages for in vivo arterial applications.

In a pig balloon injured-femoral artery overexpressing the vectors , AdSM22-alpha-hpAP, AdRSV-hpAP, at 5 days the reporter gene is driven by the RSV promoter and there is expression by SMC in the upper regions of the media, the subintima, and by endothelial cells. In contrast the SM22-alpha promoter drives exclusive expression of SMC in the media, the lower regions of the intima and by endothelial cells. Thus, expression with a SMC-specific promoter is faithful in vivo, although the levels of expression are much lower. These are the types of approaches required to bring these molecular concepts forward into safe treatments for human vascular disease.

In patients with cardiovascular disorders, it will be increasingly important to understand and apply a molecular understanding of the disease for both diagnostics and therapeutics.

This requires a very good understanding of the genetic structure and function of the different proteins that regulate cell growth and cell differentiation. It is possible to dissect out molecular mechanisms of vascular disease using expression studies and knockout studies in mice, and these can used as a basis for developing genetic therapies.

A word of caution was delivered by Nabel. As researchers proceed with these approaches, it is important to focus not only on doing good science in the sense of having a very good understanding of the therapeutic genes, but also focusing on understanding the safety and potential toxicity of the vectors and the delivery devices. It is essential to take the time necessary to go forward with these studies in a very careful and logical manner. This might require a few more years to achieve the goal, but in the final analysis this will benefit the research and, with time, patients.