Biochemical studies performed by this research group with heparin-binding EGF in refractory heart failure were reviewed by Seiji Takashima, MD, of Osaka University School of Medicine. G-protein coupled receptor (GPCR) agonists induce cardiac hypertrophy and cause shedding of HB-EGF via metalloproteinase activation, leading to transactivation of the epidermal growth factor. In mice with cardiac hypertrophy due to aortic banding or GPCR agonists, KB-R7785 inhibited HB-EGF shedding and attenuated cardiac hypertrophic changes. HB-EGF and its signal transduction are specific important molecules among the EGF family ligands for cardiac hypertrophy.

Work in KKAY mice led to the hypothesis that KB-R7785, a new inhibitor of metalloproteinase, attenuates cardiac hypertrophy by preventing shedding of HB-EGF shedding. Further work showed that KB-R7785 inhibited HB-EGF shedding and attenuated hypertrophic changes in rats with cardiac hypertrophy. Using the new antibody number 19, they showed that the effect is specific to HB-EGF, since any effect by angiotensin II, endothelin-1 and phenylephrine (PE) were blocked (Figure 1).

Cloning the shedding enzyme of HB-EGF showed that it bound to KB-R7785 with high affinity. The activated domain was used for hybrid screening with the heart cDNA library, from which ADAM (a disintegrin and metalloproteinase) 12 was identified. A mutant strain of metalloproteinase domain mutant 1 was prepared and compared to wild type ADAM12. KB-R7785 inhibited shedding in MOCK and the mutant strain, while increased shedding was seen in the wild type. Then, transactivation was studied in control, wild type and mutant rat cardiomyocytes. Transactivation with PE was unchanged in the control cardiomyocytes, but inhibition occurred in the dominant negative mutant cells. Blockade did not occur with point mutation. Then to determine the target of KB-R7785, they showed that Biotinylated KB-R7785 bound to wild type ADAM12 within the ADAM family, but did not bind to ADAM12 mutation.

Together, this work identified ADAM12 as the shedding enzyme for HB-EGF shedding in cardiomyocytes (Figure 2), that ADAM12 binds to KBR-7785, and that KB-R7785 specifically inhibits HB-EGF shedding. In the aortic banding mouse model, KB-R7785 is easy to administer without any hemodynamic changes. KB-R7785 attenuated cardiac hypertrophy caused by GPCR agonists after 7 days of PE or 14 days of angiotensin II.

HB-EGF binding seems to be involved with worsening hypertrophy and HB-EGF signaling might be involved with the alteration of cardiac function. Inhibiting ADAM12 or HB-EGF, perhaps with angiotensin II or PE, may be a beneficial therapeutic approach. However, more than one pathway could be involved in the process of developing heart failure, limiting the efficacy of these approaches, and further study is needed to assess these possibilities.

A novel view of inflammatory mechanisms potentially involved in the development of heart failure holds that various stimuli, including autoimmunity, infection and mechanical overload and ischemia, induce production of inflammatory cytokines, which negatively influence contractility and contribute to the remodeling process in the failing myocardium, resulting in heart failure. Hiroshi Okamoto, MD, Hokkaido University Graduate School of Medicine, Sapporo, Japan reviewed work by his and other research groups of immunosuppressive therapy.

Anti-TNF alpha therapy causes a significant dose-dependent decrease in ejection fraction, improved left ventricular end-diastolic and end-systolic volumes and mass, and functional status. Immunomodulation in heart failure remains controversial and large-scale trials are needed to confirm results in smaller trials evaluating anti-cytokine, immunoabsorption, immunosuppressive and intravenous immunoglobulin therapies.

The blockade of costimulatory molecules as an important strategy for treating autoimmune heart disease is indicated by work Seiko and colleagues showing that antigen-specific T-cells expressing cytotoxic lymphocytes (CTL) and natural killer (NK) cells infiltrated the heart concomitantly with B7-1, B7-2 and CD-40 in cardiac myocytes in patients with dilated cardiomyopathy (DCM) and myocarditis.

T-cells require two distinct signals for full activation. The first is provided by the engagement of the T-cell receptor (TCR) with MHC on antigen-presenting cells (APCs) and the second signal by engagement of one or more T-cell surface receptor with their ligands on APCs. CD-28, cytotoxic lymphocyte antigen 4 (CDA4) and B7 play major roles in T-cell activation. CTLA-4 effectively inhibits this ligation as immunoadhesion and induces antigen-specific T-cell responses.

A recombinant adenovirus AdexCTLA-4IgG was constructed by Okamoto and colleagues due to the limited half-life of CTLA-4 and difficulty maintaining sufficient cell concentration (Figure 3). Characteristics of AdexCTLA-4IgG include a dose-dependent cell concentration maintained at a sufficient dose 180 days after only one intravenous (IV) administration. AdexCTLA-4IgG-mediated gene expression occurs mainly in liver cells and maintains a sufficient serum concentration in vivo.

Host immune responses have limited the use of recombinant adenoviruses for gene therapy, in part due to cytotoxic lymphocyte activation. In the liver, there is little if any inflammatory response after AdexCTLA-4IgG administration. Complete blockade of anti-viral antibody production is achieved with AdexCTLA-4IgG administration. The serum concentration of CTLA-4IgG was significantly elevated after the second administration in mice treated with AdexCTLA-4IgG. However, CTLA-4IgG is not detected in mice pre-treated with AdexACE because of antibody production. Thus, AdexCTLA-4IgG alone can completely prevent anti-viral antibody production in vivo, leading to a persistent and efficient serum concentration of CTLA-4IgG.

The effects of AdexCTLA-4IgG before and after the onset of experimental autoimmune myocarditis (EAM) was studied by Okamoto and colleagues. Although EAM is transferable by activated T-cells, the effect of T-cell activation blockade on the prevention of EAM was unknown. In the AdexLacZ-treated rats, 25% of EAM rats died of severe myocarditis and heart failure, while none died in the AdexCTLA-4IgG-treated groups. Expression of the co-stimulatory molecules B7-1, B7-2, CD40, CD28, and GAPDH was enhanced in the AdexLacZ treated EAM rats. Enhanced expression of B7-1, B7-2 and CD40 was observed on cardiac myocytes on immunohistochemical staining. Thus, myocytes as APCs may play a critical role in the engagement with T-cells.

All rats treated with AdexLacZ showed discoloration of the surface and cardiac enlargement and developed typical autoimmune regions composed of inflammatory cells. The AdexCTLA-4IgG-treated rats showed microscopic abnormalities, with little infiltration of inflammatory cells in the myocardium. On day 14, minimal myocarditis was observed. The heart weight to body weight ratio in the AdexCTLA-4IgG-treated rats was significantly lower than in AdexLacZ-treated rats, and nearly equivalent to that in normal rats. These findings indicate the benefits of adenovirus-mediated co-stimulatory signal blockade after the onset of myocarditis.

Cytokine expression in the EAM rats is shown in Figure 4. AdexCTLA-4IgG almost completely inhibited cytokine expression, in contrast to AdexLacZ. The onset and progression of EAM was prevented by AdexCTLA-4IgG concomitantly with the suppression of cytokine activation.

Ischemic/reperfusion injury after cardiac transplantation was evaluated using an in vivo system in which AdexCTLA-4IgG or AdexLacZ was injected via a recipient vein 5 days prior to transplantation. Four hours of ischemia after pre-treatment with AdexLacZ resulted in a novel 40% infarcted area and a novel 30% affected area evaluated by extent of inflammatory cell infiltration. Pre-treatment with AdexCTLA-4IgG reduced the size of the infarcted and affected area, as shown in Figure 5.

In the absence of AdexCTLA-4IgG treatment, there was severe inflammatory cell infiltration and slight but significant myocyte apoptosis occurred. Pre-treatment with AdexCTLA-4IgG suppressed inflammation and the number of TUNEL-positive cells. The mechanism by which AdexCTLA-4IgG protects against reperfusion injury could involve inhibition of the immune response and increased resistance of cardiomyocytes to apoptosis.

In humans unable to mount immune responses against various bacteria and bio-insults after AdexCTLA-4IgG administration, long-lasting CTLA-4IgG expression may be problematic. A new adenovirus vector containing the AdexCre system was constructed, which terminated expression of AdexCTLA-4IgG. Hepatic CTLA-4IgG production was stopped. This new vector enabled termination of in vivo gene expression at the desired time. FK506, CsA and anti-cytokines impaired T-cell receptor signal transduction leading to pan-T-cell suppression, including na_ve T-cells and antigen-specific T-cells. In contrast, CTLA-4IgG specifically blocked the costimulatory signaling between antigen-treated cells and T-cells only when antigen is present. Moreover, AdexCTLA-4IgG caused persistent and efficient expression in vivo for a long period that was terminated at the desired time, but it has not been clinically evaluated. Potential immunomodulator therapies including AdexCTLA-4IgG are shown in Figure 6 and possible treatment modalities of AdexCTLA-4IgG are summarized in Figure 7.

Research by Yano and colleagues demonstrates that excess beta-stimulation induces PKA-mediated hyperphosphorylation, inducing the dissociation of FK-binding proteins (FKBP) from the ryanodine receptor (RyR), conformational change of RyR, and then Ca²⁺ leak. This series of events within the RyR might cause Ca2+ overload, and in turn cardiac dysfunction and heart failure. Importantly, they demonstrated that FKBP-mediated stabilization by JTV519 or a beta-blocker prevents development of heart failure. Their results may represent new strategies to prevent and treat heart failure.

Stabilizing the Ca²⁺ release channel of the cardiac sarcoplasmic reticulum, often called the ryanodine receptor, addresses a chief pathogenic mechanism for various types of dysfunctions in heart failure. FKBP are tightly coupled with the RyR and have a channel stabilizing effect in skeletal and cardiac muscles. A partial loss of RyR-bound FKBP12.6 causing a prominent abnormal Ca²⁺ leak leading to conformational changes was demonstrated by this group in a dog model of pacing-induced heart failure. Presumably, this abnormal Ca²⁺ leak causes Ca²⁺ overload, and hence results in diastolic and systolic dysfunctions.

PKA-mediated hyper-phosphorylation of RyR causes dissociation of FKBP12.6 from RyR, in turn causing an increased sensitivity to Ca²⁺-induced activation and defects in channel functions, as demonstrated by Marx and colleagues. This suggests that failing hearts lack normal FKBP12.6-mediated channel regulation. Hence this group assessed FKBP12.6-mediated stabilization of RyR as a novel therapeutic strategy for heart failure.

In an experiment model of heart failure produced by 4 weeks of rapid RV pacing, FK506 binds to FKBP and dissociates it from RyR, inducing a prominent Ca²⁺ leak in normal SR. In failing SR, spontaneous Ca²⁺ leak occurs. A new benzothiazepine derivative, the cardioprotective agent JTV519, which inhibits intracellular Ca²⁺ overload due to ischemia-reperfusion, completely inhibited the Ca²⁺ leak.

In normal SR, FK506 induced an increase in MCA (mechylcoumarin acetate) fluorescence, which was inhibited by JTV519. In failing SR, there was virtually no further increase in MCA fluorescence. Interestingly, JTV519 decreased the level of MCA fluorescence, suggesting that JTV519 restores the normal conformational state of RyR in failing SR. The MCA fluorescent change was much faster than the Ca²⁺ leak, indicating the MCA fluorescent change precedes Ca²⁺ leak. Chronic administration of JTV519 decreased end-diastolic pressure (EDP), tended to increase +dP/dt of LVP, shortened Tau, and strikingly reduced the LV chamber size, compared to failing dogs with 4 weeks of right ventricular pacing. Treatment with JTV519 reduced left ventricular (LV) chamber size and attenuated the development of LV remodeling.

The ability of beta-blockade to prevent heart failure by restoring FKBP12.6-mediated stabilization of RyR was then studied. Propranolol, 0.05mg/kg/day for chronic administration, decreased heart rate, but had no effect on LV pressure and contractility. This dose of propranolol seemed to exert a negative chronotropic action, but not a negative inotropic action. A higher dose caused death due to heart failure in the canine model. Chronic infusion of propranolol decreased LVEDP, shortened Tau, reduced LV size and increased fractional shortening. Under normal conditions, Ca²⁺ leak was induced by the addition of FK506. In failing SR, spontaneous Ca²⁺ leak was observed, and FK506 had no further effect on the Ca²⁺ leak. In the SR taken from propranolol-treated dogs, there was virtually no spontaneous Ca²⁺ leak, and FK506-induced Ca²⁺ leak appeared like in normal SR. MCA fluorescence change was induced by FK506 in normal SR, while no change was observed in failing SR, as conformational change might have already occurred in failing SR. In propranolol-treated SR, FK506-induced MCA fluorescent change was partially restored. Back phosphorylation was lower in failing RyR than in normal RyR, indicating that RyR was hyperphosphorylated in failing SR, whereas it was restored towards normal in propranolol-treated SR. Ca²⁺ uptake and expression of Ca²⁺ ATPase were decreased similarly, regardless of the treatment with propranolol, in both groups.

Marks and colleagues also demonstrated that PKA-mediated hyperphosphorylation was restored by treatment with the beta1-selective blocker, metoprolol, in the same canine model. They further showed that the expression of phospatase in RyR, that is PP2A and PP1, was decreased in heart failure. This downregulation of PP1 and PP2A may contribute to the hyperphosphorylation of RyR. But, they were restored to normal in metoprolol-treated heart. Consistent with the finding by Yano and colleagues, the expression of FKBP12.6 was decreased in heart failure and was reversed with metoprolol.

Several proposed mechanisms for the improved cardiac function in chronic heart failure with beta blockers, include inhibition of beta-adrenergic receptor kinase, upregulation of SERCA and inhibition of metalloproteinase. The inhibition of hyperphosphorylation of RyR and the subsequent suppression of Ca²⁺ leak might also lead to an improvement of cardiac function through an inhibition of intracellular Ca²⁺ overload.

The concept of transplantation of exogenous cells to compensate for myocyte loss as an alternative treatment for end-stage heart failure has been explored since 1992. Allogenic, xenogenic and autologous cell sources have been explored, with autologous sources showing the most promise. Work by Tomita and colleagues sand others in cell-based therapy was reviewed by Shinji Tomita, National Cardiovascular Center Research Institute, Suita, Japan.

In a landmark study by Soonpaa and colleagues, fetal cardiomyocyte transplantation enabled synchronous myocardial contraction due to formation of Gap junction. Concerns about immunorejection of fetal cardiomyocytes led to study of autologous cell transplantation. Skeletal myoblast transplantation has been studied experimentally and clinically, with Menasche and colleagues in France showing their viability. However, since cell transplantation is done as part of a revascularization procedure of the coronary artery, determining which is more effective is difficult. The detailed analysis of the ten completed cases is awaited.

Bone marrow cells (BMC) for transplantation are strong candidates in heart failure, particularly mesenchymal stem cells. Myogenic cells were differentiated in a study by Wakitani and colleagues. Beating cardiomyogenic (CMG) cells were developed from BMC by Makino and colleagues of Keio University in Tokyo, Japan. Tomita and colleagues used BMC for autologous transplantation and reported detectable Troponin I and improved heart function. Bone marrow stroma cells were transplanted into a porcine model by Tomita and colleagues after which cardiomyocyte tissue was detectable. One month after post-myocardial infarction and stroma cell transplantation, perfusion and wall thickening were improved.

Human mesenchymal stem cell transplantation into fetal sheep early in gestation resulted in site-specific differentiation in work by Liechty and colleagues. In terms of which part of the BMC are responsible for this improved heart function, Orlic and colleagues demonstrated that transplantation of Lin c-Kit⁺ cell after myocardial infarction resulted in good differentiation and improved heart function.

Angiogenesis with endothelial progenitor cells (EPC) and bone marrow mononuclear cell has been studied. Clinical trials of BMC transplantation examining the use of adult stem cells, bone marrow mononuclear cells, EPC, mesenchymal stem cells and cardiomyocytes are underway in Japan. Advantages of BMC are shown in Figure 8.

An in vivo study using a co-culture system in which cardiomyocytes were the host and green fluorescent protein expressing mouse bone marrow cell (GFP-BMC) the donor demonstrated synchronous contraction in the cardiomyocytes after two days. Tomita and colleagues also showed myosin heavy chain expression from day 1 and connexin43 and ANP from day 2 and cardiac-specific Troponin I after day 4.

Clinically, cell-based therapy has been used in patients waiting heart transplantation in Japan. Six patients on a left ventricular assist device recovered and were weaned form the device. This may change the environment for heart transplantation. However, a number of issues must be resolved for clinical application (Figure 9). Side effects such as arrhythmia must be addressed. Clarification of the mechanism and the best type of human cell is required. Improved evaluation methods in humans are needed, new cell sources must be identified, and facilities that comply with good manufacturing principles are required.