xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EVA3MMCEL_2WKRKM INGEST_TIME 2017-07-11T22:03:39Z PACKAGE AA00037413_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
Cardiomyocyte differentiation and purification in Barth Syndrome induced pluri potent stem cells Iv C, Santostefano K, Terada N Abstract Barth Syndrome (BTHS) is a mitochondrial disease which develops secondary to functional deletion of the tafazzin (TA Z) gene. TAZ mutation leads to defective synthesis of cardiolipin, a phospholipid of the mitochondrial inner membrane important in oxidative phosphorylation. While diagnosis and treatment of symptoms has led to increased rates of survival and quality of life there is no cure for Barth Syndrome and cardiac dysfunction via heart failure or arrhythmia can be fatal. Human induced pluripotent stem cells (iPSCs) were used to evaluate disease mechanisms, through the e stablish ment of an in vitro human model fo r Barth Syndrome. Using three different methods, iPSCs derived from BTHS patients were differentiated into cardiomyocytes to offer insight into the molecular mechanisms underlying heart failure of BTHS patients. Cardiomyocytes were en rich ed using two app roaches, by targeting VCAM1 for magnetic nanoparticle sorting and by exploiting cardiac capacity for lactate metabolism to eliminate non cardiomyocytes. Using the more recent method of traction force microscopy, contractile output will be calculated in fut ure experiments to test the overall hypothesis that BTHS iPSC derived cardiomyocytes w ill exhibit weaker cardiac contractility compared to a n unaffected control, resulting from less efficient energy production caused by cardiolipin abnormalities. Introd uction Barth Syndrome (BTHS) is a rare X linked mitochondrial genetic disease. BTHS has an incidence of only 1 in 200,000 male infants, but it has a low survival rate due in large part to the high risk of heart failure. This condition generally has fema le carriers with a 50% chance of transmission to their sons, although there have been incidences of new causative mutations and one recently documented case of Barth Syndrome in a female patient.  In addition to the potential for heart failure, BTHS pa tients experience a range of symptoms which include dilated cardiomyopathy, neutropenia, hypotonia, muscle weakness, reduced skeletal muscle deve lopment, growth retardation, exercise intolerance and 3 methylglutaconic aciduria, with cardiomyopathy being t he primary cause of death 
In this condition, there is a recessive loss of function mutation of the TAZ gene (also called G4.5 located at the distal portion of Xq28 ), which encodes tafazzin, a protein most highly expressed in cardiac and skeletal mus cle cells. Tafazzin is a phospholipid acyltransferase, responsible for remodeling the mitochondrial phospholipid, cardiolipin. Cardiolipin is the signature lipid component of the mitochondrial inner membrane; it functions to stabilize the membrane comple xes of the electron transport chain with in the mitochondrial inner membrane during oxidative phosphorylation. TAZ mutation results in defective cardiolipin maturation leading to a decrease in cardiolipin levels, abnormal cardiolipin fatty acyl compositio n, and overall, inefficient energy metabolism and subsequently decreased energy levels. [1 ,2 ] This is considered to lead to mitochondrial dysfunction, as well as the various symptoms of BTHS although the mechanisms underlying heart failure are not yet un derstood. Historically, BTHS patients died from heart failure or bacterial infection by the age of 3. A dvances in diagnosis techniques as well as the monitoring and treating of clinical symptoms have led to an improvement in survival rate and life expect ancy.  However, a cure for Barth Syndrome has yet to be found, and mechanism s of the disease are still poorly understood. Thus, there is sufficient ongoing research focused on Barth Syndrome. Cardiac dysfunction via heart failure or cardiac arrhythmi a remains fatal, therefore research involving cardiac tissue is most critical and relevant to studying the disease. Multiple animal models and even a yeast model have been developed to study Barth Syndrome, but they are not adequate for evaluating specifi cally how TAZ mutations lead to dysfunctional cardiac tissues.  Unfortunately, access to live BTHS cardiac tissue is extremely limited. For this reason, human induce d pluripotent stem cells from patients with Barth Syndrome were used in this study to create an in vitro human model for investigating the disease. The heart has little, if any, regenerative capacity. During development, secondary to the establishment of binucleated cardiomyocytes, most of the mature myocardial cells withdraw from the ce ll cycle and become terminally differentiated.  The discovery of human pluripotent stem cells (hPSCs) has allowed researchers to focus on finding methods to direct pluripotent stem cell differe ntiation to the cardiac lineage, using both human embryonic stem cells (ESCs) and human induced pluripotent stem cells (iPSCs) to do so iPSCs are a type of cell in which pluripotency has been artificially generated by dedifferentiating mature somatic cells through the expression of certain exogenous transcriptio n factors most commonly the combination of Oct4, Sox2, Klf4, and c Myc  iPSCs allow the study of cells which are not easily accessible by biopsy, such as cardiomyocytes and neurons.
They are produced from skin biopsies or blood samples taken from pa tients, which are then cultured with specific growth factors in vitro to prompt redifferentiation into specialized cells. These specialized cells derived from iPSCs are ideal for research purposes because they retain the genetic information of the patient s from which they come. The relatively recent successful induction of both ESCs and iPSCs into cardiomyocytes has led to increased efforts to improve efficiency and reproducibility of differentiation protocols, so that cardiomyocytes may be produced on a clinically relevant scale.  There are three major sources for producing de novo cardiomyocytes: hPSCs, adult heart derived cardiac progenitor cells, and directly reprogrammed fibroblasts. This study focuses on the use of hPSCs to generate cardiomyocy tes. The differentiation process from pluripotency through various mesoderm derivatives to the cardiovascular lineage is highly specific and time sensitive. The most successful cardiac induction protocols involve stage specific activation and inh ibition of the major signaling pathways of cardiac development activin, BMP, FGF, and WNT with precise temporal windows, simulating key steps in cardiomyogenesis in the early embryo, using several growth factors and small molecules to do so.  In the protocol s utilized in this study the addition of activin A, bone morphogenic protein 4 (BMP4) and basic fibroblast growth factor (bFGF ) at the start of induction is essential to the activation and specification of mesoderm differentiation, to form cardiac mesode rm. M edia are not changed for several days to allow the secreted factors t o accumulate. From here, vasc ular endothelial growth factor (VEGF) or Dickkopf related protein 1 (Dkk1, a WNT inhibitor) may be used as cardiac specification factors to produce car diac progenitor cells with the addition of L ascorbic acid improving the efficiency of progenitor production At this point ad ditional growth factors are no longer needed, and the differentiation process continues on its own, producing immature cardiomyo cytes and eventually mature cardiomyocytes with mature electrophysiology and calcium handling, capable of spontaneous contr actions in culture conditions. [ 3,4 ] A diagram summarizing cardiac differentiation in this study is shown below ( Figure 1 ). Figure 1. Overview of Cardiac Induction. Because the differentiation process is so time sensitive, cardiac induction protocols generally produce a mixture of cardiomyocytes, endothelial cells, smooth muscle cells, other
mesoderm derivatives, and in some cases un differentiated stem cells. In order to use de novo cardiomyocytes in both clinical and laboratory settings, isolation of a pure population of cardiomyocytes is necessary. Contamination of the cardiomyocyte population with undifferentiated hPSCs could res ult in the formation of a teratoma after transplantation back in to patients Traditional purification methods rel y on genetic modifications involving insertion of a selection cassette within the host genome, which may also result in tumorigenesis.  Th is study seeks to try two newer non invasive methods of cardiomyocyte en rich ment, or purification. The first method uses an antibod y for VCAM1, a cardiac cell surface marker, to isolate enriched populations of cardiomyocytes through magnetic nanoparticl e sorting. [ 5 ] The second method takes advantage of the high capacity of fetal cardiomyocytes to use lactate as a major energy source in oxidative phosphorylation, creating a toxic environment for unwanted cell populations.  Pure populations of de no vo cardiomyocytes have numerous applications. In the case of damaged heart tissue (i.e. post myocardial infarction), donor cardiomyocytes may be used in the future to exchange scar tissue with grafts of new functional myocardium in cell replacement therap y. Another important application for de novo cardiomyocytes is drug discovery; studies have shown that hPSC derived cardiomyocytes (hPSC CMs) will respond to cardioactive drugs with the expected reaction, signifying that de novo cardiomyocytes may be used for drug discovery, development, and toxicology screening. Finally, disease modeling and personalized medicine are other avenues for investigation; by deriving cardiomyocytes from iPSCs of patients with various cardiac diseases, electrophysiological and molecular analyses may be done to study the molecular mechanisms behind each disease. [ 3 ] Therefore, studying Barth Syndrome iPSC derived cardiomyocytes (iPSC CM s ) allows direct observation of how genetic mutations manifest in cardiac muscle cells in vit ro Upon differentiating BTHS iPSCs into cardiomyocytes and en rich ing the cultures, it will be possible to study the mechanisms underlying heart failure of BTHS patients. If mitochondrial and cellular function can be characterized, then it may be possibl e to investigate how cardiolipin abnorm alities lead to heart failure.  Hypothesis and Aims Through the use of Barth Syndrome iPSC derived cardiomyocytes the aim of the overall study is to examine mitochondrial dysfunction, potential abnormalities in cellular energy and redox status, and subsequent cardiomyocyte dys function under non stressed and stressed
conditions to provide an understanding of the mechanisms that cause cardiac dysfunction in clinical BTHS. Specifically for this paper, the main ob jective will be to evaluate cardiomyocyte dysfunction of BTHS iPSC CMs. Experiments will be conducted to test the hypothesi s t hat BTHS iPSC CMs will exhibit weaker cardiac contractility in comparison to control hPSC CMs due to less efficient energy produ ction caused by cardiolipin abnormalities. T o test this, t he specific aims are as follows: (1) T o use established protocols for cardiac induction and modify them as necessary to obtain cardiomyocytes derived from BTHS iPSCs, control iPSCs, and control ES Cs ( 2) To apply non invasive experimental methods for cardiomyocyte purification in order to obtain en rich ed cardiomyocyte populations. ( 3) To compare cardiac contractility in functional testing of BTHS iPSC CM s with control iPSC CMs and ESC derived car diomyocytes (ESC CMs) using traction force microscopy  Methods hPSC Cell Culture The control human ESC line used was the Rockefeller Univers ity Embryonic Stem Cell Line 2, RUES2 ( NIH approved hESC 09 0013) Two different control human iPSC lines w ere used for this study One iPSC line was derived from the human fibroblast strain, IMR90 (ATCC, line CCL 186) from human fetal lung myofibroblasts The other iPSC line was generated from a newborn foreskin fibroblast primary cell line BJ (ATCC, line C RL 2522 ). The BTHS iPSC line, BTH003, was produced before the start of the study from human dermal fibroblasts of a patient with Barth Syndrome, through retroviral transduction of transcription factors Oct4, Sox2, c Myc, and Klf4 By genotyping using H otStarTaq DNA polymerase it was previously established that BTHS patient 003 possessed a missense mutation G T at position 170 in Exon 2 of the TAZ gene causing a change in the amino acid sequence from arginine (CGA) to leucine (CTA).  hPSCs (RUES2, IMR 90, BJ, and BTH003) were mainta ined in feeder free culture on M atrigel coated dishes (1:100 dilution ) in mTeSR1 (Stem Cell Technologies ) medium. C ulture medium was changed daily. Cells were kept at 37 o C in 5% CO 2 hPSCs were passaged every 6 7 days using either Ethylenediaminetetraacetic acid (also called EDTA, Life Technologies) or the StemPro EZPassage Disposable Stem Cell Passaging tool (Life Technologies) For EDTA dissociation EDTA was added directly to the cell culture dishes following remov al of the culture medium and gentle washing with PBS (Mg 2+ and
Ca 2+ free, Corning Cellgro) ; after incubating with EDTA for 3 4 minutes at ro om temperature, colonies began to separate ; EDTA was then neutralized by adding mTeSR1 medium, and the mixture was p ipetted drop wise into new culture dishes. For manual dissociation, after gently washing with PBS, the culture medium was replaced with fresh mTeSR1, and colonies were cut to roughly equal size using the EZ Passage roller; cells were gently rinsed with cu lture medium to lift cut colonies from the plate; the mixture was then pipetted drop wise into new culture dishes. Differentiation of hPSC derived Cardiomyocytes (adapted from Carpenter et al)  Four or five days before the start of differentiation, BJ iPSC and BTH003 iPSC lines were each manually dissociated into colonies of roughly equal size using the EZPassage tool in Essential 8 medium (Life Technologies) onto M atrigel coated plates. On day 0, the start of the induction, Essential 8 was replaced wi th a base medium containing StemPro 34 SFM (Life Technologies) with 1% Penicillin/Streptomycin (Life Technologies), 2 mM L glutamine (Life Technologies) 400 M Monothioglycerol ( Sigma ), and 50 g/ml Ascorbic Acid (Sigma). At this time, 0 hours, the base medium was supplemented with 50 ng/ml Activin A for 4 hours; then 5 ng/ml Activin A, 10 ng/ml BMP4, and 5 ng/ml bFGF (Life Technologies) for 44 hours. On day 2, the medium was refreshed with new base medium, supplemented with the same cytokine concentrations from hours 4 to 48, with no change in medi um for another 48 hours. On day 4, cells were cultured in the base medium with no supplementary cytokines with a change in medium every other day from this point on. Cell masses were expected to show spontaneous contractions by day 10. Differentiation of hPSC derived Cardiomyocytes ( adapted from Ye et al) [ 9 ] Four days before starting the induction, the IM R90 iPSC line was dissociated into single cells by exposure to Accutase ( Stem Cell Technologies ) for 7 minutes at 37 o C to generate a single cell suspension Cells were seeded at a density of 250,000 cells per cm 2 onto growth factor reduced M atrigel in mTe SR1 medium Fresh mTeSR1 was added daily over the next 4 days. C ell cultures were supplemented with 1 M ROCK inhibitor (also called Y 27632 dihydrochloride, Tocris) to encourage single cell survival Once the monolayer was confluent, cells were considered ready fo r cardiac differentiation. On d ay 0, the start of the induction, i PSCs were cultured for 24 hours with RPMI 1640 medium ( Corning Cellgro ) supplemented with B27 without insulin (Life Technologies ), 50 ng/ml Activin A ( Life Technologies ), and 25 ng/ml BMP4 ( Life Technologies ). On d ay 1, the cardiac induction medium was c hanged to RPMI 1640 medium supplemented with B27 without insulin and 10 ng /ml VEGF ( Life Technologies ) and
was left unchanged for 72 hours. On d ay 4, the cells were maintained in RPMI 1640 medium supplemented with B27 complete ( Life Technologies ), to be c hanged every other day. Cells were expected to start spontan eously beating on d ay 10 or 11. Differentiation of hPSC derived Cardiomyocytes ( adapted from Uosaki et al) [ 5 ] Two to three days before the start of differentiation, RUES2 ESC, BJ iPSC, and BTH0 03 iPSC lines were each dissociated into a monolay er by exposing colonies to EDTA for 5 6 minutes at room temperature Cells were seeded at a density of 100,000 cells per cm 2 growth area onto M atrigel coated plates in mTeSR1 medium supplemented with 1 M ROCK inhibitor. On the day before induction, hPSCs were cov ered with a M atrigel overlay ( 1:60 dilution in DMEM:F12, Life Technologies ). At the start of the induction, mTeSR1 medium was replace d with RPMI+B27 base medium containing RPMI 1640 medium supplemented with B27 without insulin and 2mM L glutamine. At this time, d ay 0, the base medium was supplemented with 100 ng/ ml Activin A for 24 hours. On d ay 1, the RPMI+B27 was refreshed and then supplemented with 10 ng/ml BMP4 and 10 ng/ml bFGF with no change in medi um for 96 hours. On d ay 5, for 48 hours, the medium was replaced with RPMI+B27 base medium supplemente d with 100 ng/ml Dkk1 ( Life Technologies ). On d ay 7, cells were cultured in RPMI +B27 base medium without supplementary cytokines; from this point on, the medium was refreshed every other day. Spontaneous contractions were expected on d ay 8 or 9. hPSC derived Cardiomyocyte Maintenance and Passaging During maintenance of the hPSC CMs at day 21 of the Uosaki protocol [ 5 ], the RPMI+B27 maintenance medium was changed to include the B27 complete supplement. hPSC CMs were passaged on day 39 of the Uosaki protocol [ 5 ] using two different enzymes onto plates coated with four different mate rials. Half of the cells were exposed to Accumax (Sigma) for 5 minutes at 37 o C and then diluted with DMEM:F12 and 10% FBS (Atlan ta Biologicals), centrifuging at 1,000 rpm for 3 minutes. The other half were exposed to 0.25% Trypsin EDTA (Corning Cellgro) for 5 minutes at 37 o C and then inactivated with DMEM:F12 and 10% FBS, centrifuging at 1,000 rpm for 3 minutes. Plates were coated with 0.6 g fibronectin (Life Technologies) per ml H 2 O, 10 g laminin (Sigma) per ml DMEM:F12, and 0.1% gelatin (Millipore), incubated at 37 o C for 2 hours, and also with M atrigel (1:100 dilution), incubated at room temperature for 2 hours. Cells were resuspended i n "EB20" medium , containing DMEM:F12, 20% FBS, 1% 100x NEAA ( Life Technologies ), and 0.1 mM mercaptoethanol ( Life Technologies ). Cells were seeded at 50,000 cells per cm 2 growth area, in small clumps of cells. After 24 hours, the medium was change d back to the maintenance RPMI 1640 medium
supplemented with B27 complete. Spontaneous contractions were observed in the small clumps of cells, as early as 24 hours after passaging. After this initial trial, in subsequent cardiomyocyte passaging, cells w ere dissociated using Accumax for 5 minutes at 37 o C onto fibronectin coated plates. Purification of hPSC derived Cardiomyocytes using Lactate [6 ] Glucose free DMEM (no glucose, pyruvate, or glutamine, Corning Cellgro ) medium was supplemented with 4 mM lac tate (Sigma) made by dil uting lactate stock solution (60 g per 100 ml) in sterile 1 M Na HEPES ( Corning Cellgro ) At differentiation day 21, hPSC derived cardiomyocytes from the Uosaki protocol [ 5 ] were treated with the lactate selection medium, with one well from each cell line remaining untreated for comparison purposes. The selection medium was refreshed every day to remove dead, unwanted cells. After exposing cell s to the selection medium for 96 hours, the medium was changed back to cardiac maintena nce medium containing RPMI 1640 and B27 complete. Purification of hPSC derived Cardiomyocytes with Magnetic Nanoparticl e Separation [ 5 ] For magnetic nanoparticle separation the EasySep B iotin S election kit (Stem Cell Technologies) was used, and the ma nual protocol using "The Big Easy" Silver EasySep magnet was followed. At differentiation day 21 of hPSC CMs from the Uosaki protocol [ 5 ] a single cell suspension was prepared by exp osing cells to Accumax for 10 minutes at 37 o C and resuspending dissocia ted cells in a round bottom tube at a concentration of 2.5 x 10 7 cells per 250 l of the recommended medium: PBS (Mg 2+ and Ca 2+ free ), with 2% FBS and 1 mM EDTA 100 l Anti Human CD32 FcR Blocker, the blocking antibody per ml of suspension was added to the round bottom tube followed by 2 g/ml biotin anti hum an CD106 antibody (Bio Legend) the primary antibody; the mixture was incubat ed at room temperature for 15 minutes The biotin selection cocktail was added at 100 l/ml incubating for 15 minutes a t room temperature. 50 l/ml magnetic nanoparticles were then mixed in vigorously incubating for 10 minutes at room temperature. The cell suspension was brought up to a total volume of 5 ml by adding additional recommended medium and was then placed in the magnet for 5 minutes. Afterwards, the magnet was inverted for 2 3 seconds, with the flow through fraction poured off into a separate container. At this point, the positively selected, magnetically labeled hPSC CMs should have remained in the tube, an d unlabeled unwanted cells should have been in the flow through. The tube was removed from the magnet and filled with 5 ml recommended medium after which the tube was placed back into the magnet for 5 minutes. This process was repeated for a nother 5
mi nute separation in the magnet, with the flow through again poured off and the tube removed from the magnet. Final l y, c ells were resuspended in RPMI 1640 supplemented with B27 without insulin and 5% FBS and replated onto laminin and 0.1% gelatin coated pla tes. P rior to this positive CD106 selection process, a negative selection was performed to remove dead cells that nonselectively attached to the nanoparticles Both t he flow through cells from the positive selection and the nonselectively bound cells w ere replated onto laminin and 0.1% gelatin coated plates Quantitative Real Time PC R (qPCR) Total RNA was extracted from various cells and tissues using the RNAqueous kit (Ambion) for phenol free total RNA isolation. The Turbo DNA free kit (Ambion) was u sed for DNase treatment to remove contaminating DNA from the RNA preparation. OD260 /OD280 ratio was measured for each sample using a Gen5 plate reader with a Take3 microspot plate (Bio Tek). First strand cDNA synthesis was performed under RNase free cond itio n s using the High Capacity cDNA Reverse Tr anscription kit using random primers (Applied Biosystems). R eaction tubes were incubated in the thermal cycler at 25 o C for 10 minutes, 37 o C for 120 minutes, 85 o C for 5 seconds, and 4 o C overnight. For quantita tive analysis, cDNA was diluted to 0.5 ng/ml across all samples and was then used as the template in quantitative real time polymerase chain reaction (qPCR) assays using Power SYBR Green dye (Applied Biosystems). The primers (Nanog, Nkx2.5, vWF, and a SMA) were designed using Primer3 software to comply with suggested guidelines of the SYBR Green PCR Master Mix Protocol (Applied Biosystems). All samples were run in duplicate or triplicate, with standard curves generated for each primer set on each PCR run. Reactions were performed using Step One Plus (Applied Biosystems) at 95 o C for 10 minutes then 40 cycles alternating between 95 o C for 15 seconds and 60 o C for 1 minute Expression level was calculated through the comparative C T method, using either actin or GAPDH for normalization. Immuno fluorescence C ells were exposed to Accumax for 5 minutes at 37 o C and seeded at 25,000 cells per cm 2 onto sterile coverslips within a 12 well plate Cells were then left to grow for 96 hours to guarantee protein r e expression after passaging. All media were used at room temperature unless otherwise specified. On the coverslips, cells were rinsed with 1 x PBS, then fixed with 4% PFA (Sigma ) for 10 minutes. Cells were washed four times with 1x PBS for 5 minutes eac h time on a rotator and then permeabilized with 0.1% Triton X (Sigma) in 1x PBS for 5 minutes. Cells were again washed with 1x PBS four times for 5 minutes on a rotator. Cells were then
blocked for 1 hour in 10% rabbit anti mouse serum ( Sigma ) in 1x PBS T (0.05% T riton X ) The primary antibody, mouse monoclonal sarcomeric actinin (1:100 dilution, Sigma), d iluted in 10% serum in 1x PBS T, was added and kept overnight at 4 o C in a humid environment, with the cell side in contact with the antibody. Covers lips were washed four times with 1x PBS for 5 minutes on a rotator. The secondary antibody, AlexaFluor 488 (green dye, 1:500 dilution, Life Technologies), was added in 10% serum in 1x PBS T for 1 hour in a dark, humid environment, with the cell side in co ntact with the antibody. Cells were washed three times with 1x PBS for 5 minutes on a rotator then once with H 2 O for 5 minutes. Coverslips were air dried overnight in the dark. The next day, coverslips were mounted to slides with VectaShield HardSet mou nting medium with DAPI for 15 20 minutes, and stained cells were detected by fluorescence microscopy ( Zeiss ). Results Cardiomyocyte differentiation was attempted using three different directed differentiation protocols ( Figure 2 ).
Figure 2. Schematic representation d etailing differences between cardiomyocyte inductions. Protocols were revised as needed from (A) Carpenter et al [8 ], (B) Ye et al , and (C) Uosaki et al [ 5 ]. This diagram presents the many similarities and minute differences between i nduction protocols including the estimated start of contractions. The first cardiac induction attempt was based on a protocol used by Carpenter et al , modified only in that cells were cultured in colonies of similar size rather than in a monolayer. Before the induction, control BJ iPSCs (5 samples) and BTH003 iPSCs (5 samples) exhibited typical undifferentiated morphology characterized by a large nucleus and little cytoplasm By day 2, following initial cytokine treatment, the centers of colonies ha d died, leaving a ring of cells. Differences in morphology typical of differentiation were identified, including elongated holly shaped cells characteristic of early mesoderm derivatives and a 3 dimensional geographic growth pattern. On day 8, a distinct ive dark cobblestone pattern, characteristic of endothelial cells w as observed as well as 3 dimensional clumps of round cells with possible cardiac morphology. By day 15, large dense regions were discernible around the colony center along with extensive o utgrowth of cells. Cardiomyocytes are known to spontaneously contract in cell culture, with the presence of beating implying successful differentiation, but no contractions were observed in any of the samples, suggesting that iPSC CMs had not been produce d The second cardiac differentiation attempt was based on a protocol outlined by Ye et al [9 ]. Only IMR90 iPSC s (6 samples) w ere used in th is induction. Prior to initiating differentiation, control iPSCs maintained typical stem cell morphology Initia l attempts at establishing a monolayer were unsuccessful, possibly due to the lack of cell attachment after passaging, low cell viability caused by passaging reagents, and the use of older M atrigel and mTeSR1 medium. After adjustments were made and ROCK i nhibitor was added at the start to help establish single cell attachment and survival, a confluent monolayer was observed ( Figure 3A ) and the rest of the cardiac induction protocol was followed as reported Following treatment with Activin A, there was s ubstantial cell death, with rou nding up occurring as cells died; large patches detached with additional cytokine treatment ( Figure 3 B ). By day 7, holly shaped cells were
observed and groups of cells became more 3 dimensional ( Figure 3C ). Maintenance of c ells with RPMI 1640 and B27 complete was continued, and by day 15, in addition to the characteristic cobblestone endothelial cells 3 dimensional clumps of cells with possible cardiac morphology w ere observed ( Figure 3D ) However, t here w ere no visible co ntraction s in any of the 6 samples Figure 3. Different phases of the Ye et al cardiac induction. 10x objective microscopic images from (A) day 3, (B) day 3, (C) day 7, and (D) day 15. RNA isolated from a differentiated RUES2 sample from day 21 of a previous ly successful Carpenter induction was compared using qPCR analysis to differentiated IMR90 samples from day 17 of the Ye cardiac induction ( Figure 4 ). Nkx2.5 was used as a marker for cardiac lineage cells, and Nanog was used as an indicator for pl uripotency, expressed if any pluripotent cells remained within the culture. In both cases, undifferentiated RUES2 was used as the control (RQ = 1). All samples showed Nkx2.5 expression, indicating development of some cardiac mesoderm derivatives, althoug h at relatively lower levels in the Carpenter sample
than in the Ye samples Additionally, relative to the undifferentiated control ESC sample, there was little to no Nanog expression in the Ye samples, signifying that essentially all cells had differenti ated ; however Nanog expression in the Carpenter sample had decreased but did not disappear altogether, impl ying that undifferentiated ESCs remained in the culture. Both the lower Nkx2.5 expression and higher Nanog expression in the Carpenter sample compar ed to the Ye samples may have been caused by the greater difficulty in exposing all cells to the necessary cytokines when cultured in colonies versus a monolayer. Overall however while qPCR analysis indicates that the Ye induction was slightly more succe ssful regarding gene expression than the previous Carpenter induction, neither cardiac induction had been successful this time in producing cardiomyocytes since cells did not exhibit spontaneous contractions in culture. Figure 4 qPCR r e sults : Ye inducti on compared with previous Carpenter induction. Undifferentiated RUE S 2 was used as a control ( RQ value = 1 ) and all results were normalized to actin (A) Expression levels of Nkx2.5, a cardiac marker, in hPSC derived cells. (B) Expression levels of Nanog, a pluripotency marker, in hPSC derived cells.
The last cardiac induction protocol used a previous report by Uosaki et al [ 5 ] as the foundat ion with modifications as p reviously mentioned in methods the main adjustment being that cells were cultured in mTeSR1 before induction, rath er than conditioned MEF CM medium The biggest difference in this protocol as compared to the others was the use of a M atrigel overlay to prevent cell death by keeping cells from detaching en masse In the first cardiac induction using this method, c ontrol RUES2 ESCs (1 sample), control BJ iPSCs (4 samples), and BTH003 iPSCs (2 samples) were used. Prior to the st art of differentiation, hPSCs showed morphology typical of undifferentiated stem cells The cardiac differentiation protocol was followed as directed, and a confluent monolayer was observed by day 1. As expected, there was less cell detachment after ini tial cyt okine treatment than occurred using the other two methods but cell death was not nonexistent. On day 5, longer holly s haped cells with dark cytoplasm and less visible nuclei were observed in some samples while other wells contained lighter round shaped cells characteristic of cardiac derivatives By day 11, the cobblestone growth pattern of endothelial cells and a build up of 3 dimensional clumps of rounder cells were observed although there was no sign of beating in any of the samples. Finall y on day 20, spontaneous contractions were observed in all 4 BJ sa mples in flat areas with rounded cells ; u nfortunately, the RUES2 and BTH003 samples had been contaminated with fungi. The BJ iPSC CM samples were kept for maintenance and purification, and the other samples were discarded. On day 21, a separation using magnetic nanoparticles w as attempted on the 4 BJ iPSC CM samples. Vascular cell adhesion molecule 1 (VCAM1, also called CD106) was the cardiac cell s urface marker used to identify i PSC CM s [ 5 ] through the use of biotin anti human VCAM1 antibody Ideally the magnetic beads w ould recognize biotin, which w ould be conjugated in a tetrameric antibody complex with an antibody for VCAM1 ( Figure 5 ) ; subsequently using a magnet, cells expressing VCAM1 could be selected for. First, a negative selection was performed to remove cells (i.e. dead cells with "stickier" membranes ) that would attach nonselectively to the magnetic beads; approximately 1 x 10 7 nonselectively bound cells were removed in thi s way and kept in a separate container After this, the VCAM1 sorting Figure 5. Tetrameric Antibody Complex (Stem Cell Technologies EasySep manual)
protocol was followed as directed. At the end of the purification process, VCAM1 sorted cells were replated on both laminin and 01% gelatin plates, to determine whether there was a pre ference for attachment on either coating Additionally, both the flow through cells from positive selection and the nonselectively bound cells with no antibody from negative selection were replated, to be analyzed in an impending gene expression study to ensure ne ither group contained i PSC CMs. All samples were observed for signs of contractions after replating ; if everything went as planned only wells with VCAM1 sorted cells would contain beating cardiomyocytes. However, in the following week, there we re no observable contractions in any of the groups and it was determined that the BJ iPSC CMs had been lost during the purification process. This may have been caused by low replating cell density low cell viability, or lack of re attachment after dissoc iation. Figure 6. qPCR results from VCAM1 nanoparticle separation. Expression levels of vWF (endothelial cell marker), aSMA (smooth muscle marker), and Nkx2.5 (cardiac cell marker) were compared in negatively selected cells with no antibody (used as the control, RQ value = 1), cells from the flow through of the positive selection, and VCAM1 sorted cells. All results were normalized to GAPDH At day 28, RNA was isolated for qPCR analysis ( Figure 6 ). Nkx2.5 was again the cardiac marker used. Primers for vWF, an endothelial cell specific marker, and smooth muscle actin (aSMA), specific for vascular smooth muscle cells, were also used to determine the success of the en rich ment. Standard amplification curves for the vWF primer were not well spaced or complete, indicating that vWF was not a good prime r set for this analysis; nevertheless from the data it can be tentatively determined that endothelial expression was similar across all groups. Surprisingly, VCAM1 sorted cells contained the lowest cardiac Nkx2.5
expression with smooth muscle aSMA expres sed mor e than three times that amount, and cells from the flow through expressed the highest amount of Nkx2.5 Likely, these results are due to the fact that RNA was collected over a week after replating rather than immediately after the separation. Thu s, assuming that many of the cardiomyocytes and other i PSC derived cells did not attach after replating, this RNA analysis was misrepresentative of the original samples. As such overall, the only thing that can be d efinitively concluded from qPCR analysi s is that cardiomyocyte enrichment did not definitely occur. This does not mean enrichment did not occur at all; it only means it cannot be proven one way or the other. In the second cardiac induction using th e Uosaki method [ 5 ], control RUES2 ESCs (4 s amples) control BJ iPSCs (5 samples) a nd BTH003 iPSCs (3 samples) were used. The cardiac differentiation protocol was followed in the same way as the first induction. On day 1, BJ and BTH003 iPSCs showed a confluent monolayer while RUES2 ESC samples c ontained many holes throughout. On day 7, BTH003 and RUES2 samples revealed a similar morphology to the first induction, while BJ samples exhibited a much more 3 dimensional, geographic growth pattern. By day 11, one RUES2 group already contained a spont aneously beating clump of cells. By day 17, there were 3 dimensional clumps of beating cells, sitting on top of the monolayer, in all 4 RUES2 samples ( Figure 7 A ) ; the holes in the monolayer appeared to have filled in with elongated cells characteristic of smooth muscle cells. O ne of the BTH003 samples contained a single contracting region wi th a large flat surface area, completely different in morphology from the RUES2 ESC CMs but more similar to BJ iPSC CMs of the previous induction This difference in morphology is observable when comparing Figures 7 B and 7 C This time, BJ samples did not contain any contracting areas. From this induction, a cardiomyocyte en rich ment using lactate selection medium was attempted Through oxidative phosphorylation, ca rdiomyocytes are able to generate energy efficiently from numerous substrates, including glucose, fatty acids, and lactate. Fetal cardiomyocytes have even greater capacity for lactate metabolism, taking advantage of the lactate rich environment of the pla centa during development. In a previous study, it was found that culturing pluripotent stem cell derivatives with lactate rich medium containing no glucose, pyruvate, or glutamine led to survival of only cardiomyocytes.  Thus, on day 21 the hPSC deri ved cardiomyocytes were purified using this lactate selection medium T reatment was applied as directed to samples (RUES2, BJ, BTH003) regardless of whether contractions were present or not leaving one sample of each type untreated for comparison. After 96 hours with the selection medium had passed, it was obvious that some cells had died, especially the elongated cells characteristic of smooth muscle cells ( Figures 7B and 7C ). However, many of
the remaining cells were not beating, implying that some no n cardi omyocytes had survived the lactate purification After en rich ment, the previously observed flat region of beating BTH003 iPSC CMs significantly increased in size and several other contracting areas were identified in both of the treated BTH003 sam ples indicating that lactate selection medium may potentially be used as an identification tool for cardiomyocytes. Unfortunately treated BJ samples still did not demonstrate signs of contracting and unexpectedly did not show much cell death after lacta te treatment In general however the lactate en rich ment was successful in some purification of hPSC CMs, although adjustments would need to be made to obtain completely pure cardiomyocyte cultures Figure 7 Images from videos of contracting hPSC CMs. 10x objective microscopic v ideos of contractions were taken of RUES2 ESC CMs at days 19 and 25, (A) 2 days before and (B) after the 96 hour lactate exposure. Videos of BTH003 iPSC CMs were ta ken only (C) after en rich ment on day 25. Areas circled in red e xhibited spont aneous contractions in culture.
RUES2 and BTH003 hPSC CMs were passaged on day 39, and BJ iPSC samples were discarded. On day 4 1 RUES2 ESC CMs and BTH003 iPSC CMs were dissociated onto coverslips for staining ( Figure 8 ). For immunofluores cence staining actinin, a sarcomeric actin binding protein with a specific striation pattern for skeletal and cardiac muscle cells, was targeted using AlexaFluor 488 (green); DAPI, a nuclear counterstain, was used to identify nuclei to indicate cell pre sence The differe nce in morphology between RUES2 ESC CMs and BTH003 iPSC CMs is again apparent in the staining ( Figure s 8A and 8B ). Figure 8 Immunofluorescent staining for Actinin (green). Presence of blue nuclei (DAPI counterstain) alongside gree n fluorescence indicates the presence of hPSC CMs. Lactate purified RUES2 ESC CMs are shown at (A) 20x and (C) 63x, exhibiting 3 dimensional clumps of cardiomyocytes. Lactate purified BTH003 iPSC CMs are shown at (B) 20x and (D) 63x, exhibiting a one dim ensional plane of interconnecte d, more widely spread cardiomyocytes. While slightly blurry, images (C) and (D) show how actinin manifests in the striations of cardiac muscle cells.
Figure 9. qPCR results from Uosaki cardiac induction and lactate purification. All results were normalized to GAPDH. (A) Expression levels of Nkx2.5, a cardiac marker, in the hPSC derived cel ls. Human heart cDNA was used as the control (RQ value = 1). (B) Expression levels of Nanog, a pluripotency marker, in the hPSC derived cells. Undifferentiated RUES2 was used as the control (RQ value = 1). Additionally on day 41, RNA was extracted for qPCR analysis ( Figure 9 ). For cardiac marker Nkx2.5, cDN A from adult human heart tissue obtained from a colleague was used as the control (RQ = 1). For pluripotency marker Nanog, undifferentiated RUES2 cDNA was used as the control (RQ = 1). The four exp erimental samples showed varying degrees of Nkx2.5 expression, enough so that development of cardiomyocytes can be successfully concluded in each sample, when also taking contraction presence into consideration In all cases, samples demonstrated lower Nk x2.5 expression than the adult human heart sample, although it is difficult to tell whether this was due to lower overall expression or lower percentage of cardiomyocytes; most likely both are true. Lactate purified RUES2 ESC CMs expressed Nkx2.5 at about 25% of the level of human heart tissue, whereas unpurified RUES2 ESC CMs expressed slightly more than half of that. Lactate purified BTH003 iPSC CMs expressed Nkx2.5
at about 1% of the level of the human heart tissue, whereas unpurified BTH003 iPSC CMs a lso expressed slightly more than half of that although it is difficult to tell from the figure Flow cytometry would be needed to know conclusively but the overall increase of Nkx2.5 expression from unpurified samples to lactate purified samples implies that the cardiomyocyte en rich ment was fairly successful. Moreover compared to control ESC expression, there was little to no Nanog expression in the four experimental samples, suggesting that essentially all cells had differentiated and were no longer p luripotent. Overall, qPCR results as well as video recordings of cardiomyocyte contractions ( Figure 7 ) suggest that this was a successful cardiac differentiation and a somewhat successful cardiomyocyte purification Discussion and Conclusions There ar e v arious protocol options available for differentiating human pluripotent stem cells into cardiomyocytes. Directed differentiation from pluripotency through multipotent mesodermal cells to terminally differentiated cardiomyocytes focuses on stage specifi c activation and inhibition of cardiomyogenic signaling pathways. Numerous attempts to direct differentiation into the cardiac lineage have provided the current understanding of the stages of cardiac development, pictured below ( Figure 10 ).  Many of t he pathways shown should be familiar, as they may have been targeted by the cardiac differentiation protocols used in this paper Figure 10 Overview of current knowledge of factors involved in hPSC cardiac differentiation. From Burridge et al. 
In t his study, o riginally hPSCs were differentiated using the Carpenter et al  protocol as t his lab had been previously successful in obtaining contracting card iomyocytes using this method.  Other than past success, a major draw for using this protocol was the reduced Activin A concentration for 92 hours following a 4 hour high dose treatment I deally the toxic effects of Activin A should have been mitigated by this shortened high concentration exposure While cell survival was improved, there was st ill substantial cell death after 24 hours. Moreover there was low cardiomyocyte yield in the previously successful induction and no contractions at all in subsequent experiments of this paper Following suggestions made by a colleague, the Ye et al  protocol was next attempted, but there were similar issues with cell death. Treatment with Activin A and other growth factors was harsh on cells, and cell death led to large patches of detachment. The Uosaki et al  protocol was then chosen because it addressed this issue through the use of a M atrigel overlay to prevent mass cell detachment from cytokine treatment. This method produced a relatively high cardiomyocyte yield. Following successful cardiac induction i t was necessary to determine an effi cient method for cardiomyocyte en rich ment Initially magnetic nanoparticle separation was attempted because this was the en rich ment method used by Uosaki et al  Unfortunately it was unsuccessful potentially either due to compromised viability after enzymatically dissociating cells or due to a suboptimal sorting procedure. After again obtaining a culture of contracting cardiomyocytes, an en rich ment method outlined by Tohyama et al  relying on a lactate rich selection medium containing no glucose, pyruvate, or glutamine was next tried because it did not require dissociation to purify hPSC derived cardiomyocytes While the en rich ment did not result in a completely pure cardiomyocyte culture, it was successful in eliminating som e non cardiomyocyte cell types most noticeably smooth muscle cells. Purification of cardiomyocytes was not measured quantitatively, but it can be qualitatively observed from the results of real time PCR ( Figure 9 ). Regular maintenance as well as future assays for functional testing would require dissociation of cardiomyocytes, so t o determine the optimal conditions for cardiomyocyte passaging, RUES2 and BTH003 hPSC CMs were passaged using two enzymes (Accumax and 0.25% Trypsin EDTA) onto 4 different coatings (fibronect in, la minin, 0.1% gelatin, and M atrigel). Spontaneous impulses were observed in all sample s as early as 24 hours after passaging, and cells attached well across all coatings. I t was determined qualitatively that hPSC CM contractions seemed improved after en zym atic passage by Accumax. S ubsequent passaging was thus accomplished using Accumax for dissociation of cardiomyocytes onto fibronectin coated dishes
Overall, while much progress was made through the establishment of a fairly pure BTHS cardiomyocyte cult ure, this study remains incomplete because the hypothesis has yet to be addressed and remains an important avenue of further study. Future Directions To test the hypothesis of this paper that BTHS iPSC CMs will demonstrate weaker cardiac contractility c ompared to control hPSC CMs, secondary to less efficient energy metabolism, cells w ere plated onto hydrogels ( Figure 11 ) and will be analyzed in future through traction force microscopy analysis with help from Hazeltine et al. The normal cardiac response to an inc rease in blood flow is to linearly increase contractile output; certain diseased phenotypes fail to maintain this functional response. By providing a source for contraction resistance, hydrogel substrate stiffness was previously used successfully by Hazeltine et al  to simulate changes in contractile demand to study contractile output Fluorescent beads were added to provide a means for tracking hydrogel di splacement and thus movement by cardiomyocytes. Traction force micros copy uses algorithms in computational analysis to resolve bead displacement into cell tractions. Figure 11 provides an example of how cardiomyocyte contractions may be measured in this way, through the use of vectors and video analysis to calculate cardia c contractility. Figure 11. Images from videos of contracting hPSC CMs. V ideos of contractions in RUES2 ESC CMs produced from this study were taken both (A) with out and (B) with vectors to indicate strength and direction of cardiomyocyte contractions. After adjustments are made to fluorescent bead spacing contractile output will be measured by computer analysis in future studies
Many assays for evaluating mitochondrial function require pure cardiomyocyte cultures. While some cardiomyocyte en rich me nt was observed in this study following lactate purification, the protocol still needs to be modified in order to obtain an entirely pure cardiomyocyte population. It may be beneficial to apply the lactate selection medium for 8 days, as Tohyama et al  did, rather than for 4 days as applied in this study. Furthermore, en rich ing a cardiomyocyte population after enzymatic dissociation would be worth attempting, as cells are more spread out after passaging and would be more sensitive to treatment. As par t of the goal for the overall study, once an essentially pure population of cardiomyocytes has been obtained, additional assays for functional testing can be done to evaluate mitochondrial dysfunction in Barth Syndrome iPSC derived cardiomyocytes. Mitocho ndrial functioning can be investigated by compa ring BTHS iPSC derived cardiomyocytes with healthy control hPSC derived cardiomyocytes, more specifically in protein expression, ROS production, O 2 consumption, and ATP production. Once mitochondrial and cell ular function are characterized, it may be possible to investigate how cardiolipin abnormalities lead to heart failure, provid ing a greater understanding of the disease mechanisms of Barth Syndrome.
Literature Cited 1 Barth Syndrome Foundation. Overview of Barth Syndrome. Available at http://www.barthsyndrome.org/about barth syndrome Accessed July 3, 2014. 2 Santostefano K Terada N Correction of mitochondrial dysfunction in a human iPS cell model of Barth Syndrome. University of Florida, Depa rtment of Pathology, Immunology, and Laboratory Medicine. Unpublished grant proposal, 2011. 3 Burridge PW, Keller G, Gold JD, et al. Production of de novo cardiomyocytes: Human pluripotent stem cell differentiation and direct reprogramming. CELL STEM CELL 2012;10:16 28. 4 Williams LA, Davis Dusenbery BN, Eggan KC. Snapshot: Directed differentiation of pluripotent stem cells. CELL 2012;149:1174. 5 Uosaki H, Fukushima H, Takeuchi A, et al. Efficient and scalable purification of cardiomyocytes from human embr yonic and induced pluripotent stem cells by VCAM1 surface expression. PLoS ONE 2011;6(8):1 9. 6 Tohyama S, Hattori F, Sano M, et al. Distinct metabolic flow enables large scale purification of mouse and human pluripotent stem cell derived cardiomyocytes. C ELL STEM CELL 2013;12:127 137. 7 Lee MA, Santostefano KE, Soustek M, et al. Disease specific induced pluripotent stem cells and cardiomyocytes for Barth Syndrome. University of Florida, Department of Pathology, Immunology, and Laboratory Medicine. Unpublis hed undergraduate thesis, University of Florida, College of Liberal Arts and Sciences, 2013. 8 Carpenter L, Carr C, Cheng TY, et al. Efficient differentiation of human induced pluripotent stem cells generates cardiac cells which provide protection followin g myocardial infarction in the rat. STEM CELLS AND DEVELOPMENT 2012;21(6):977 986. 9 Ye L, Zhang S, Greder L, et al. Effective cardiac myocyte differentiation of human induced pluripotent stem cells requires VEGF. PLoS ONE 2013;8(1):1 10. 10 Hazeltine LB, Simmons CS, Salick MR, et al. Effects of substrate mechanics on contractility of cardiomyocytes generated from human pluripotent stem cells. INTERNATIONAL JOURNAL OF CELL BIOLOGY 2012;2012:1 13.