Embryonic Stem Cell Differentiation to Cardiomyocytes

Richard Lee
Brigham and Women’s Hospital, Boston, MA

Embryonic stem cells and tissues are derived from the inner cell mass of the blastocyst. The embryonic stem field is unique because under highly-specialized conditions the cells will continue to proliferate and be multi-potent. Although these highly-specialized conditions are very useful in the laboratory, they are also confining; the specialized manner in which the cells are grown restricts the ability to steer them down one pathway versus another pathway.

The typical manner in which embryonic stem (ES) cells have been differentiated involves their being cultured on embryonic fibroblasts or treated with leukemia inhibitory factor, which maintains the cells in the undifferentiated state. A key distinction for human stem cells is that they are not maintained in the undifferentiated state by leukemia inhibitory factor. The secreted product from human ES cells that allows them to be undifferentiated is not known.

In the classic differentiation protocol, a hanging drop is the most typical way for aggregating these cells. The cells are suspended and aggregated in a small cluster for a number of days and form the embyroid bodies, which can be plated on a monolayer-type dish, after which the cells will differentiate.

Another question is what is the embryoid body. Although it is thought that these cells talk with each other to promote differentiation, there is little understanding of why this process occurs. The typical efficiency of plating the embryoid body is quite low for obtaining aggregates of ES cells, a little less than 1%. ES cells have a tendency to differentiate to a neuronal phenotype, for reasons that are not really understood.

To drive the efficiency of the differentiation process, Lee and colleagues used the myosin heavy chain promoter and made a stem cell line that reported with green fluorescent protein. The concordance was very high, showing this to be an easy way to pick out cells that had differentiated to a cardiac myocyte from large numbers of samples. Using a number of different markers, including sarcomeric myosin and sarcomeric alpha-actinin, it was confirmed that these were cardiac myocytes.

In a trial run in which 800 compounds were screened in a drug library, Vitamin C was the only positive compound that reproduced. It is unknown why the ES cells in monolayers exposed to increasing concentrations of ascorbic acid resulted in about a 5-fold increase in the amount of cardiac myocytes in an ES culture. This is still not highly efficient, it can reach about 5-6% differentiation. It does not seem to be related to the anti-oxidant properties of ascorbic acid. Vitamin C has a complex mode of action that still is not elucidated after 30 years of research.

The ascorbic-induced GFP+ cells were positive for sarcomeric alpha-actin. In the presence of ascorbic acid, it drove the myocyte formation with a number of markers, including sarcomeric actin and myosin. Ascorbic acid also drove the formation of the gene expression for ANF, beta myosin heavy chain, alpha myosin heavy chain, Nkx2.5, and GATA4, all of which were confirmed by real-time PCR.

This was a curious finding, because it meant that a single compound can shift embryonic cell differentiation. It is challenging to determine why this might happen, when considering the cascade of events that are required for cardiogenesis in the developing embryo. It is becoming quite clear that relatively simple manipulations of ES cells can steer cells from one pathway down another pathway, including the cardiogenesis pathway. Some factors that can drive cardiogenesis include embryoid body formation, TGF-beta family members (IGF-1, FGFs, Oxytocin, EPO, retinoic acid, DMSO, dynorphin), co-culture with visceral-endodermal cells, electrical stimulation, hydrogen peroxide, and some new compounds. Some people are purporting the ability to convert 90% of ES cells into cardiac myocytes, but this data is very new and needs to be confirmed in many laboratories to ensure it is a general effect on many different types of ES cell lines.

In spite of the complexity of developmental programs that causes a cardiomyocyte to form in vivo, which includes signals of BNP pathways, FGF pathways, among others, and the necessity for signals to turn on and off, it is interesting that some relatively simple manipulations in culture can allow for differentiating stem cells.

Many large questions remain in cardiac regeneration. A key question is whether or not it is possible to recruit endogenous cells or whether it will be necessary to inject cells. For cell therapy, will human ES cells be needed? Strategies for rejection must be studied. Will there be a need for universal donor cells, or will some stem cells be permissive with regard to rejection? These questions must be answered before stem cells can be used therapeutically.

Perhaps the greatest challenge ultimately will be that injecting cells and getting them to live may not lead to long-term function of the heart. In all species there is a very fine vascular structure that is very intimately related to the myocytes. Ultimately, because of the problems of oxygen diffusion in the heart, and probably other factors, it is necessary to achieve a 3-dimensional structure of the myocardium that will need to be restored by regenerative strategies.

An encouraging factor is that cardiac myocytes seem to have some intrinsic information that tells them to form this relation between the myocyte and the vasculature. When mixing cardiac myocytes with endothelial cells, they will spontaneously form vascular structures, as endothelial cells do in 3-dimenstional cultures. Then the myocytes will wait for the vascular structures to form and then grow on the outside. It looks very much like a process similar to vasculogenesis. Lee and colleagues believe thatmyocytes have some instructions that allow them to help form this structure. However, with regenerative strategies it will be important to ensure that the myocytes retain those same instructions. An injected cell will need to find its correct location relative to the vasculature or it will have to create its own vasculature. Ultimately, getting cells to survive is the beginning, and more research will be needed on the end structural solution.

Regeneration Therapy for Heart Failure Due to Nonischemic Cardiomyopathy

Genzou Takemura
Gifu University School of Medicine, Gifu, Japan

Beneficial effects of treatment with autologous bone marrow cell (BMC) transplantation in doxirubicin cardiomyopathy in rabbits and of treatment with granulocyte-colony-stimulating factor (G-CSF) in hereditary cardiomyopathy in hamsters (Um-X7.1) was shown in work by Takemura and colleagues. 

The beneficial effects on nonischemic chronic heart failure (CHF) were accompanied by prevention of myocyte degeneration and death, activation of MMPs, suppression of TNF-aalpha, and tissue regeneration.
The results provide evidence that bone marrow implantation and G-CSF may be a novel therapeutic strategy for heart failure due to nonischemic cardiomyopathy.

Doxirubicin cardiomyopathy in rabbits

The cardiomyopathic model was created by an 8-week intravenous injection of doxirubicin into rabbits. Bone marrow cells (BMCs) were aspirated from the iliac bone 2 weeks later and mononuclear cells were directly injected into about 10 points of the left ventricular (LV) wall. Hemodynamic and histological assessments and molecular biological assay were performed 4 weeks later.

Parameters of LV function, LV developed pressure, positive and negative dP/dT, ejection fraction, and fractional shortening, were significantly improved by the BMC implantation (Figure 1). LV compensatory hypertrophy was significantly attenuated by the BMC implantation, as shown by a smaller heart to body weight ratio and a thinner LV wall thickness (Figure 2).

The percentage of fibrotic area was significantly reduced in the BMC implantation group. Extracellular matrix metalloproteinases (MMP-2, MMP-9) were significantly increased in the hearts treated with BMC. The cardiotoxic cytokine, TNF-alpha, was significantly reduced on Western blot in the BMC implantation group.

Hereditary cardiomyopathy in hamsters

G-CSF can proliferate and differentiate granulocytes, and also release bone marrow stem cells into peripheral blood flow. This group recently reported that treatment with G-CSF after acute myocardial infarction improved cardiac function. However, it is unknown whether G-CSF has a beneficial effect in nonischemic heart failure, such as dilated cardiomyopathy (DCM), a major cause of morbidity and mortality in CHF. Thus, this group examined the long-term effects of G-CSF on survival, cardiac function, and cardiac histology in the hamster model of nonischemic heart failure.

The UM-X7.1 hamster is a representative animal model of autosomal recessive cardiomyopathy and vascular dystrophy. Because of the lack of the delta-sarcoglycan gene, it develops progressive myocyte death beginning at about 4 weeks of age and worsens over time. At 20 weeks, cardiac hypertrophy is seen, followed by progressive ventricular remodeling, myocardial fibrosis, and CHF. About 50% of these animals die by 30 weeks of age. Importantly, 1 family and 2 sporadic cases of human DCM were identified who presented with mutations in the delta-sarcoglycan gene, comparative to this animal model.

Intraperitoneal injection of G-CSF (10 mg/kg/day) was started at 15 weeks of age in male UM-X7.1 hamsters (n=16) and continued until 30 weeks of age. In the control group (n=15) the animals were treated with the same amount of distilled water in a similar manner.

In other experiments with mice, this group already confirmed that cardiac myocytes under pathological conditions express G-CSF receptor.       

Survival at 30 weeks was significantly improved with G-CSF treatment compared to controls (100% versus 53%, respectively; p<0.0001). The surviving animals underwent echocardiography at 30 weeks of age. The control hamsters showed severe LV enlargement and signs of decreased cardiac function, a low LVEF and percent fractional shortening, while each of these was significantly improved in the G-CSF treated animals (Figure 3).

Similarly, cardiac catheterization revealed significantly improved LV function and congestion in the G-CSF treated animals versus controls (Figure 4). The heart was smaller in the G-CSF treated animals at necropsy. The heart to body weight ratio was significantly smaller in the G-CSF treated animals. Histologically, there were numerous foci in the fibrotic area and abundant interstitial fibrosis in the untreated hearts. Fibrosis in the ventricles was significantly reduced in the G-CSF treated hearts. The percent area of fibrosis in the ventricles was 8.6% in the treated animals versus 20% in the untreated animals at 30 weeks. The percent area of fibrosis was 7.3% in animals at the age of 15 weeks before treatment, which was not significantly different from that of the G-CSF treated animals at 30 weeks of age.

Electron microscopy analysis of the heart revealed degenerative changes of the myocytes with various grays in the untreated group. Many myocytes in the untreated hearts showed severe vacuole degeneration. The vacuoles contained degraded subcellular organelles, such as mitochondria. However, these were greatly attenuated in the G-CSF treated animals. In another example, the cardiac myocyte was affected by severe autophagic degeneration with scanty myofibrils in the periphery. It is assumed that such a cell can no longer function normally, but it is unknown whether such degeneration is linked to cell death.

Thus, sarcolemma integrity was analyzed by peritoneal injection of Evans blue dye 24 hours before sacrificing the animals. The dye was exuded by cardiomyocytes that preserve normal sarcolemma permeability, but is taken up by the cardiomyopathic cells with leaky cell membrane. Such cells can be recognized to be dead. The control hearts revealed extensive dye uptake by cardiomyocytes, while the uptake was significantly ameliorated in the G-CSF treated hearts.

Immunohistochemistry revealed a significantly lower incidence of cardiac myocytes positive for cathepsin D and for ubiquitin in the G-CSF treated hearts than in the control hearts. Cathepsin D and ubiquitin are markers of autophagy. Under confocal microscope, cathepsin D positive cells were found that were taking Evans blue dye simultaneously. This suggests continuity between autophagic degeneration and autophagic cell death. Overall, these findings indicate a protective effect of G-CSF against autophagic degeneration and death of the cardiomyocytes.

The possibility of myocardial regeneration was checked in the present model. Bone marrow was aspirated from the femur bones of 15-week-old hamsters and labeled with fluorescent dye and autologously returned into the bone marrow spaces. The animals were treated with G-CSF or distilled water for 2 weeks. Cells that were double positive for cells that were muscle-specific troponin I and DiI, which are considered BMC-derived cardiomyocytes, were found, although the incidence was small (<0.1%) in the G-CSF treated hearts, but none in the untreated hearts. However, they could not find DiI-labeled endothelial cells or vascular smooth muscle cells in the hearts of either group.

The present study revealed a significant reduction of fibrosis in the G-CSF treated hearts. They noted a significant increase MMP-2 and MMP-9 in the G-CSF treated hearts, compared with the control hearts. TNF-alpha is not only one of the representative toxic cytokines that can directly depress cardiac function, but also contributes to angiotensin II-mediated fibrosis of organs through upregulation of angiotensin II type-1 receptor. In the present model, they found that long-term therapy with G-CSF resulted in a significant reduction in the cardiac TNF-alpha level in the 30-week-old hamsters. Overall, such an increase in MMP activity and TNF-alpha downregulation in the heart caused by G-CSF might have contributed to thereduction of collagen content in the G-CSF treated hearts.  

The hypothesized mechanisms for responsible for the beneficial effects of the G-CSF treatment on CHF due to DCM are shown in Figure 5. The first mechanism is protection of cardiomyocytes from autophagic degeneration and death. The second mechanism, MMP upregulation, would cause degradation of excessive extracellular matrix and reduction in ventricular fibrosis, and then may reduce wall stiffness and contribute to improvement of cardiac function. The third mechanism, TNF-alpha downregulation, would relieve cardiotoxic action of TNF-alpha on the heart and may contribute to functional improvement. The fourth mechanism is cardiomyocyte regeneration would result in preservation of the contractile force of the myocardium.

Application of Embryonic Stem Cells for Vascular Regeneration Medicine

Hiroshi Itoh
Kyoto University School of Medicine, Kyoto, Japan

Previously this group demonstrated that VEGF-R2 (Flk1)-positive cells derived from mouse embryonic stem cells (ES cells) can differentiate into both endothelial cells and mural cells and organize into blood vessels in vitro and in vivo. They called these cells vascular progenitor cells (VPC). They have also demonstrated that implanted VPC derived from mouse ES cells could be incorporated in tumor vessels and augment tumor blood flow in a tumor angiogenesis model, suggesting the therapeutic potential of VPC for vascular regeneration.

Figure 1 shows the differential kinetics of mouse ES cells from VPC to vascular cells. VEGF-R2(-)and VE-cad (-) undifferentiated ES cells differentiate into VEGF-R2(+) VE-cad(-) VPC. VPC can differentiate into VEGF-R2(+), VE-cad (+) endothelial cells. In contrast, the same VPC can differentiate into VEGF-R2(-) VE-cad(-) and alpha smooth muscle actin positive (SMA) mural cells.

Study design

To investigate the potential of ES cell-derived VPC for clinical application, they examined whether VPC occur in the primate cynomolgus monkey and human ES cells.

ES cell markers in mice ES cells are SSEA-1(+) SSEA-3(-) SSEA-4(-) TRA1(-), while in primates they are SSEA-1(-) SSEA-3(+) SSEA-4(+) TRA1(+). In terms of morphology, mice ES cells grow in rounded clumps with indistinct cell borders and primate ES cells grow in flat colonies with distinct cell borders. Regarding l eukemia inhibitory factor (LIF) deficiency, in mice ES cells pluripotent cells can remain undifferentiated, but not in primate ES cells. Even on a feeder cell layer, all primate pluripotent cells grow very poorly when dissociated to single cells, whereas mouse ES cell lines can be cloned with relatively high efficiency in the presence of LIF. There are also differences in attachment to extracellular matrix.

In contrast to mice ES cells, monkey undifferentiated ES cells express VEGF-R2. Flow cytometry analysis revealed that about 70% of ES cells were positive for VEFF-R2 (Figure 2). On immunostaining, ES cells colonies cultured on MEF layer were stained positively for VEGF-R2. On flow cytometry, VEGF-R2 expression was detected in about 70% of undifferentiated monkey ES cells at day 0, but VEGF-R2 expression diminished by day 6 of differentiation on OP9 feeder layer. VEGF-R2 expression re-appeared in about 8% of the total cells after 8 days of differentiation. At day 10, VE-cad(+) endothelial cells appeared.

To compare monkey VEGF-R2(+) cells at day 0 and day 8, alkaline phosphatase activity, which was reported to be detected in undifferentiated ES cells, was examined. Alkaline phosphatase activity was clearly detected in undifferentiated monkey ES cells at day 0, but not cells at day 8 cultured on OP9 feeder layer.

The re-culture of VEGF-R (+) monkey ES cells at day 0 did not differentiate into PECAM1 positive endothelial cells after 5 days of re-culture on OP9 feeder layer. At day 8, VEG-R2(+) cells differentiated into PECAM1, VEcadherin, and eNOS positive cells after 5 days of re-culture on OP9 feeder layer. When re-culturing on a collagen IV coated dish with 10% serum for 5 days, the VEGF-R2(+) cells differentiated at day 8 into alpha-SMA, calponin, and smooth muscle myosin heavy chain positive cells. The addition of VEGF-R2(+) cells to culture on collagen IV coated dish resulted in the appearance of PECAM1 positive cells that were surrounded by SMA positive cells. A 3-dimensional culture of VEGF-R2(+) cells at day 8 in collagen IA gel showed aggregates from which cells migrated outward and formed tube-like structures within 3 days.

Human ES cells were positively stained for SSEA4 and were positive for alkaline phosphatase activity. VEGF-R2(+) cells in human ES cells at day 0 were positive for TRA1-60 activity, but during differentiation on OP9 feeder layer VEGF-R2(+) TRA1-60(-) cells appeared. On re-culture, the endothelial cell marker CD34, VEcadherin, PECAM1, and eNOS positive cells appeared after 8 days re-culture of TRA1-60(-) VEGF-R2(+) VEcadherin(-) cells on a collagen IV coated dish with 10% serum and VEGF 50 ng/ml. PECAM1 negative cells were positively stained for alpha-SMA. When cultured without VEGF, nearly all of the cells differentiated into alpha-SMA positive and calponin positive mural cells. A 3-dimensional culture of VEGF-R2(+) TRA1-60(-) cells in collagen A1 gel at day 8 showed that undifferentiated VEGF-R2(+) cells could not aggregate on the floating culture, unlike the mice and monkey cells (Figure 3). They mixed gene cells in collagen A1 again and re-cultured, after which some of the cells formed tube-like structures within 5 days.

Conclusion

Their findings indicate that VEGF-R2(+) cells differentiated on the OP9 feeder layer can act as vascular progenitor cells in primates. Further, the differentiation kinetics of VPC in the primate and mice ES cells are different. They identified VPC in primates. The VPC-derived cells can be highly expanded in vitro, and would be a promising material for vascular regeneration therapy.       

Three-Year Follow-Up of the Safety and Feasibility of Intramyocardial Bone Marrow Mononuclear Cell Implantation in Patients with Ischemic Heart Disease

Kimikazu Hamano
Yamaguchi University School of Medicine, Ube, Japan

Bone Marrow Cell Implantation (BMCI) has been performed by Hamano and colleagues since 1999. Focused cell sources include bone marrow mononuclear cells (BM-MNCs), peripheral blood (EPC, CD34 cells), umbilical blood (EPC, CD34 cells), and ES-derived cells (EPC, EC, SMC). BM-MNCs were used in their work because they consist of many endothelial progenitors and cytokine-producing cells, without the problems of immunological rejection and ethical conflict for clinical application.

The mechanisms of therapeutic angiogenesis by BMCI include the induction of angiogenesis by cytokine-producing cells. Endothelial progenitor cells induce vasculogenesis, resulting in new collateral vessels result. However, the process of vasculogenesis is not clarified, and the contribution of angiogenesis and vasculogenesis are not clearly defined.

Study design

The inclusion criteria were patients with severe ischemic heart disease (IHD) scheduled to undergo coronary artery bypass grafting (CABG), at least 1 perfusion that showed ischemia and coronary artery stenosis that was unsuitable for traditional PCI or CABG, and no wall thinning of an old myocardial infarction.

BM-MNCs were prepared and then injected into ungraftable vessels. To date, BMCI has been used in 8 patients, ranging in age from 50 to 73 years, who had 2-4 grafts performed. Figure 1 shows the patient characteristics. Initially, the target area was the small area of the left circumflex, but now the total circumflex area is treated. The number of cells initially was 5x10⁷ and is now 5x10⁸.

Study results

The targeted perfusion area was improved in 5 of the 8 patients. Wall motion of the target area was improved in 2 of 8 patients. In the initial 4 cases, only a small area of the circumflex was treated, therefore it is difficult to judge improvement in left ventricular (LV) function.

In a representative case, the coronary angiogram showed no graftable vessels in the circumflex area, and tight stenosis in the left anterior descending (LAD) (Figure 2). The pre-operative echocardiography showed reduced wall motion in the circumflex. The LAD was bypassed and cells injected into the circumflex area. After treatment, wall motion in the circumflex area was improved. Post-operative scintigraphy showed no ischemia in the circumflex area. The coronary angiogram after treatment showed right filling of the circumflex area. Scintigraphy performed 2 years after treatment showed no ischemic change in the circumflex area (Figure 3).

In another case, before surgery several stenoses could be seen in the LAD and no graftable vessels in the circumflex area. Scintigraphy showed ischemia in the anterior and lateral walls, indicating the posterior wall was already dead. After treatment, no ischemia was seen in the anterior or lateral walls. The LAD was bypassed and cells injected into the circumflex. Scintigraphy at 1 year after treatment showed no ischemia in any part of the left ventricle. Pre-operative left ventriculography showed reduced wall motion in the circumflex area, however, after treatment the wall motion was improved.

To quantify the perfusion of the circumflex, they measured the perfusion index at the circumflex area and compared this to the control group, comprising patients who did not have bypass performed in the circumflex area. In the BMCI group, the perfusion index in the circumflex increased after treatment, while it was not changed or reduced in the control group.

To evaluate the safety of this treatment, mass formation or calcification was assessed by CT, chest radiography, and echocardiography. Through the first 3 years, no abnormality was found in any patient. Also, no significant or lethal arrhythmias were found in any patients. There was PVC in 1 patient. At 3 years follow-up, no significant arrhythmias had occurred.

In summary, remarkable improvement of regional perfusion and clinical symptoms were observed in about 70% of patients after BMCI treatment, which continued throughout the 3-year follow-up period. No systemic or local side effects related to BMCI treatment was recorded by the 3-year follow-up examinations.

Efficacy study in mice

Can BMCI regenerate injured myocardium effectively? They cultured BMC with a low concentration of TGF-beta, so the cell shape was changed somewhat and it expressed myosin. BMC were harvested from GFP-transgenic mice, which were pre-treated with TGF-beta (Figure 4). Acute myocardial infarction models were then injected with pre-treated BMC, untreated BMC, or PBS alone. LV function, echocardiography, and histology were assessed. In the pre-treated group, cells survived and expressed myosin. However, in the untreated group, some cells survived but did not express myosin. In the pre-treated group, the percent fractional shortening was significantly improved compared to the untreated group (Figure 5). Therefore, they conclude that BMCI should be a feasible and safe method to induce therapeutic angiogenesis for the treatment of ischemic heart disease, and is also a possible method for regenerating injured myocardium.