characterization of human-induced pluripotent stem cell–derived

Characterization of Human-Induced Pluripotent Stem Cell–Derived Cardiomyocytes: Bioenergetics and Utilization in Safety Screening

Payal Rana,* Blake Anson,† Sandra Engle,‡ and Yvonne Will*,1

*Compound Safety Prediction, Pfizer Global R&D, Groton, Connecticut; †Cellular Dynamics International, Inc., Madison, Wisconsin; and ‡Pluripotent Stem Cell Biology Laboratory, Pfizer Global R&D, Groton, Connecticut

1To whom correspondence should be addressed at Compound Safety Prediction, Pfizer Global R&D, Eastern Point Road, Groton, CT 06340. Fax: (860) 686-2353. E-mail: [email protected].

Received May 14, 2012; accepted July 21, 2012

Cardiotoxicity remains the number one reason for drug with-drawal from the market, and Food and Drug Administration issued black box warnings, thus demonstrating the need for more predic-tive preclinical safety screening, especially early in the drug discov-ery process when much chemical substrate is available. Whereas human-ether-a-go-go related gene screening has become routine to mitigate proarrhythmic risk, the development of in vitro assays predicting additional on- and off-target biochemical toxicities will benefit from cellular models exhibiting true cardiomyocyte charac-teristics such as native tissue–like mitochondrial activity. Human stem cell–derived tissue cells may provide such a model. This hypothesis was tested using a combination of flux analysis, gene and protein expression, and toxicity-profiling techniques to char-acterize mitochondrial function in induced pluripotent stem cell (iPSC) derived human cardiomyocytes in the presence of differing carbon sources over extended periods in cell culture. Functional analyses demonstrate that iPSC-derived cardiomyocytes are (1) capable of utilizing anaerobic or aerobic respiration depending upon the available carbon substrate and (2) bioenergetically clos-est to adult heart tissue cells when cultured in galactose or galac-tose supplemented with fatty acids. We utilized this model to test a variety of kinase inhibitors with known clinical cardiac liabilities for their potential toxicity toward these cells. We found that the kinase inhibitors showed a dose-dependent toxicity to iPSC cardio-myocytes grown in galactose and that oxygen consumption rates were significantly more affected than adenosine triphosphate pro-duction. Sorafenib was found to have the most effect, followed by sunitinib, dasatinib, imatinib, lapatinib, and nioltinib.

Key Words: iPSC cardiomyocytes; carbon source; long-term culture; mitochondrial function; kinase inhibitors; bioenergetics; cardiotoxicity screening.

Drug-induced adverse cardiovascular events are the number one cause of drug withdrawal, dose limitations, or black box warnings (Lawrence et al., 2008). The first drug withdrawn from the market for cardiac liability was sterotiline in 1991 due to QT prolongation. Several additional drugs have been with-drawn because of adverse effects on cardiac electrophysiology,

biochemistry, or physiology including fen-phen, terfenadine, cisapride, rofecoxib, pergolide, sibutramine, and propoxy-phore. These drugs were used to treat a variety of disorders, acted through diverse mechanisms, and possessed different structural chemical backbones. Furthermore, members of many different drug classes such as the anthracyclines, nonsteroidal anti-inflammatory drugs, and nucleoside reverse transcriptase inhibitors, while not withdrawn, carry black box warnings limiting their use (Dykens and Will, 2007). Oftentimes, toxic effects were not observed in preclinical or regulatory animal studies (Lasser et al., 2002; Lexchin, 2005). Therefore, the Food and Drug Administration has expanded requirements for cardiotoxicity testing during the preclinical drug development.

An ideal place to start cardiovascular assessment is early in the drug discovery process, when chemical matter is available and Structure Activity Relationship (SAR) can be achieved. Several hurdles currently exist to this approach, including a paucity of relevant tissue cells and imperfect translation of surrogate in vitro and animal model data to predicting cardiotoxicity in humans (Mandenius et al., 2011). In recent years, the utility of stem cells in a variety of different biomedical applications has been the subject of increasing scientific research as they can be maintained and expanded indefinitely in culture and differentiated into a variety of terminal and progenitor cell types (Chapin and Stedman, 2009; Davila et al., 2004). Human cardiomyocytes derived from induced pluripotent stem cells (iPSCs) exhibit many of the characteristics of in vivo cardiac myocytes, including appropriate ion channels, receptors, transporters, and syncytial and contractile activities, and expected electrophysiological and biochemical responses upon exposure to environmental stimuli (Cohen et al., 2011; Guo et al., 2011; Ma et al., 2011). Thus, iPSC-derived cardiomyocytes hold potential as a relevant cellular model for in vitro assessment of human cardiotoxicity (Anson et al., 2011).

Native cardiac tissue is energetically active (Pedersen, 2009) and undergoes a developmental transition from using glycoly-sis to fatty acid based oxidative phosphorylation (OXPHOS)

toxicological sciences 130(1), 117–131 (2012)doi:10.1093/toxsci/kfs233Advance Access publication July 27, 2012

© The Author 2012. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please email: [email protected]

Downloaded from https://academic.oup.com/toxsci/article-abstract/130/1/117/1667750by gueston 11 April 2018

mailto:[email protected]

as the major and preferred energy source (Allard et al., 1994; Kakinuma et al., 2002; Lopaschuk and Jaswal, 2010). In mature cardiac tissues, mitochondria are the primary sources of cellular bioenergetics, where they account for at least 20% of the myocyte volume and provide the constant energy required to meet the electrical and mechanical demands (Brand and Nicholls, 2011; Hill et al., 2009; Wallace, 2005). In contrast, in vitro cellular models are generally cultured in elevated non-physiological levels of glucose without fatty acids and thus gly-colytic. Most cells possess metabolic flexibility to shift their relative reliance on glycolysis versus OXPHOS in response to nutrient availability and thus rely on glycolysis in the presence of glucose despite having fully functional mitochondria (Gohil et al., 2010; Marroquin et al., 2007; Rossignol et al., 2004). Although human iPSC-derived cardiomyocytes possess many attributes of native myocytes, little is known about their gly-colytic/OXPHOS profile, their preferred substrate for energy utilization, or their bioenergetic adaptability. To address this unknown, mitochondrial function in iPSC cardiomyocytes was characterized when maintained in culture media containing dif-ferent levels and combinations of glucose, galactose, and fatty acids. Substrate effects were assayed by measuring adenosine triphosphate (ATP) production, the relative contributions of glycolysis, and OXPHOS to the bioenergetic profile, as well as mitochondrial and glycolytic gene expression profiles. Once the model was fully characterized, a variety of kinase inhibitors with known clinical cardiotoxicity liabilities were tested.

MATERIALS AND METHODS

Consumable Materials

Cell culture reagents were obtained from Sigma (St Louis, MO) or Invitrogen (Carlsbad, CA) unless noted otherwise. iPSC-derived cardiomyocytes (iCell Cardiomyocytes), plating medium, and carbon source–free basal media (DMEM) with serum were obtained from Cellular Dynamics International (CDI; Madison, WI). The CellTiter-Glo kit was purchased from Promega (Madison, WI). Hoechst 33342 and propidium iodide dyes were purchased from Invitrogen. RNeasy micro kit was obtained from Qiagen (Valencia, CA). The Genomic DNA Elimination Kit, custom PCR (polymerase chain reaction) arrays, cDNA synthesis kit, and SYBR Green/ROX master mix were purchased from Qiagen/SABiosciences (Frederick, MD). XF96 sensor cartridges, XF96-well plates, calibration buffer, and calibration plates were purchased from Seahorse Biosciences (Billerica, MA). Antibodies were sourced as follows: BCL2L1 from Abcam (Cambridge, MA); GAPDH, superoxide dismutase 1 (SOD1), COX 6A, and COX 4 from Cell Signaling Technology (Danvers, MA); Mitoprofile total OXPHOS human antibody cocktail (for mitochondrial complexes I–V) from Mitosciences (Eugene, OR); secondary goat anti-mouse and antirabbit antibodies from LI-COR Biotechnology (Lincoln, NB). Whole normal heart lysate was purchased from Novus biological (Littleton, CO). All test compounds were purchased from Sigma Aldrich, completely dissolved in 100% dimethyl sulfoxide (DMSO) to generate 30mM stock solutions. DMSO titrations were performed and concentrations < 1.5% were found to be nontoxic to the iPSC cardiomyocytes (data not shown). Furthermore, the final DMSO concentration in each experiment was 0.5%.

Carbon sourced media and cell culture. Basal media with serum was supplemented with 1mM sodium pyruvate, 5mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), 25 µg/ml gentamicin, and 2mM glutamine.

This media was further modified for different substrate conditions by adding proper carbon source such as for high-glucose (HG) media (4.5 mg/ml glucose), low-glucose (LG) media (1 mg/ml glucose), galactose (10mM), LG and fatty acid (1 mg/ml glucose and 100µM oleic acid + 50µM palmitic acid), and galactose and fatty acid (10mM galactose and 100µM oleic acid + 50µM palmitic acid).

iPSC-derived cardiomyocytes were thawed according to the manufacturer’s protocol and plated into 96-well Seahorse plates at 20,000 cardiomyocytes/well plating media. Two days after thawing, plating media was exchanged for HG, LG, LG + FA (low-glucose plus fatty acids), GAL (galactose), or GAL + FA (galactose plus fatty acids). Media was changed every other day, and cardio-myocytes were maintained at 37°C, 5% CO

2, and 95% humidity for at least 7,

14, or 21 days. In general, iPSC cardiomyocytes began to beat spontaneously within 24–48 h of thaw and by day 6 the entire population formed an electri-cally connected syncytial layer that beat in synchrony.

Evaluation of iPSC-Derived Cardiomyocytes by Fluorescence Microscopy

For visual verification, plates were also subjected to fluorescence micros-copy (Nikon Eclipse TE 300). The excitation wavelength was 450–480 nm, and the emission was measured using a long pass 520 filter, which allowed us to see red florescence intensity.

Evaluation of Cell Numbers and Viability

Total cell counts and cell viability were utilized as non-specific indicators of cellular toxicity at 7, 14, and 21 days. Cardiomyocytes were stained with Hoechst 33342 at 30 µg/ml final concentration and propidium iodide at 15 µg/ml final concentration for 30 min at 37°C in clear black collagen-coated 96-well plates prior to acquiring cell counts on a ThermoFisher Scientific Cellomics array scan. Cellular viability was assessed by adding 100 µl CellTiter-Glo (Promega) reagent to each well, agitating for 10 min in the dark at room tem-perature, prior to acquiring luminescence levels on an EnVision plate reader (PerkinElmer, Waltham, MA).

Measurement of Oxygen Consumption Rate and Extracellular Acidification Rate of iPSC-Derived Cardiomyocytes

OXPHOS and glycolysis in iPSC-derived cardiomyocytes were assayed by measuring oxygen consumption rates (OCRs) and extracellular acidification rates (ECARs) in real-time using an XF-96 Extracellular Flux Analyzer (Seahorse Biosciences) per the manufacturer’s protocols. Low-buffered media for measuring ECAR was composed of DMEM without bicarbonates supplemented with 2mM glutamine, 30mM NaCl, 1mM sodium pyruvate, and carbon source to match the HG, LG, LG + FA, GAL, or GAL + FA (pH 7.4) carbon-sourced culture media used to maintain iPSC-derived cardiomyocytes. Briefly, following maintenance in carbon-sourced media for 7, 14, or 21 days, iPSC-derived cardiomyocytes were equilibrated in 150 µl HG, LG, LG + FA, GAL, or GAL + FA low-buffer media for 90 min at 37°C incubator without CO

2

prior to obtaining readings. Measurements of oxygen and pH were periodically made over 5 min and OCR and ECAR were obtained from the slopes of concentration change in these parameters versus time. To test the effect of a single compound, four baseline rate measurements of the OCR and ECAR were made using a 2-min mix and 5-min measure cycle. The compounds were then injected pneumatically into each well, mixed, and six OCR and ECAR measurements were made using the 2-min mix and 5-min measure cycle. To test the effect of two compounds, four baseline rate measurements of OCR and ECAR were taken, then the first compound was injected followed by three OCR and ECAR measurements, then the second compound was injected, which was followed by another four OCR and ECAR measurements.

RNA Isolation, RT2 profiler PCR Array, and Gene Expression Analysis

Total RNA was isolated using the RNeasy microkit according to the manu-facturer’s instructions with an additional genomic DNA elimination step via a Genomic DNA Isolation Kit. In brief, 500 ng RNA was incubated with genomic

118 RANA ET AL.


DNA elimination mixture (containing 5 × genomic DNA elimination buffer and nuclease-free water) for 5 min at 42°C. First strand cDNA was synthesized by incubating the product of the genomic DNA elimination reaction with 10 µl of the reverse transcriptase (RT) mixture (5 × [RT] buffer, primers and external control, and nuclease-free water) for 15 min at 42°C. Reactions were immedi-ately stopped by heating at 95°C for 5 min, and cDNA samples were further diluted with 91 µl of nuclease-free water. cDNA samples were then loaded on to the SABiosciences 384-well custom qPCR arrays (96 genes, 4 samples/plate) and run on a 7900HT Real-Time PCR System (Applied Biosystems, Foster City, CA) according to the following thermal cycle conditions: 95°C for 10 min, and 40 cycles at 95°C for 15 s and 60°C for 1 min. Each array con-tained wells that allowed to control for genomic DNA contamination, and first cDNA synthesis efficacy. Genes that had Ct values > 30 were not considered in the data analysis process. Gene expression levels were normalized to ACTB, GAPDH, HPRT1, PPIA, and RPS18 housekeeping genes in the custom array (CAPH10095) and B2M, HPRT1, GAPDH, ACTB, and RPL13A housekeep-ing genes in the human mitochondrial energy metabolism array (PAH-008).

Western Blotting

Protein was extracted from iPSC-derived cardiomyocytes using the mam-malian cell lysis kit with protease inhibitor (Invitrogen) and quantified using the Bio-Rad Bradford assay (Life Science, Hercules, CA). Ten micrograms of protein were loaded per well on 8–12% Bis-Tris gel (Invitrogen) and run until the dye front migrated out of the gel followed by transfer onto nitrocellulose membrane with iBlot gel transfer device (Invitrogen). The membranes were treated in blocking buffer for 2 h (Rockland Immunochemicals, Gilbertsville, PA), washed three times for 15 min, and then incubated overnight at 4°C with primary antibodies. Membranes were then washed 3 times for 15 min each with 0.1% (vol/vol) Tween 20 in PBS, and incubated with secondary antibodies in blocking buffer for 1 h. Images and the relative intensities of the immune reac-tive bands were captured using the Odyssey Infrared Imaging System (LI-COR Biotechnology). Expression levels of each sample were normalized to GAPDH levels. The primary and secondary antibodies used were Mitoprofile total OXPHOS human antibody cocktail (1:200 dilution), GAPDH (1:1000), COX 4 (1:1000), COX 6A (1:1000), SOD1 (1:1000), BCL2L1 (1:1000) and Alexa Fluor IRDye 680-conjugated goat antirabbit, IRDye 800-conjugated goat-anti-mouse (1:10,000 dilution), respectively. All antibody dilutions were made in blocking buffer.

Data Analysis

Gene expression differences and up-down fold regulation (∆∆Ct) were calculated using the online SABiosciences data analysis software. The p val-ues were calculated based on a Student’s t-test of the replicate (2−∆Ct) values for each gene in the control group and treatment groups, and p values < 0.05 were considered significant and used for the calculation of fold regulation. The fold-change (2−∆∆Ct) is the normalized gene expression (2−∆Ct) in the test sample divided by the normalized gene expression (2−∆Ct) in the control sample. Fold regulation values greater than 2 were considered as upregulation; fold regula-tion values less than −2 were considered as downregulation.

For all other experiments, statistical analysis was performed using two-way ANOVA. A p value < 0.01 was considered statistically significant.

RESULTS

Intra- and Interassay Variation of the OCR and ECAR of iPSC-Derived Cardiomyocytes

To ensure experimental consistency, we first determined the intra- and interassay variation of the OCR and ECAR rates of iPSC-derived cardiomyocytes. The intraassay (within plate) var-iation of the OCR was 13–17% and the average interassay (day-to-day) variation was 15.4% (Supplementary table 1A). The

intraassay variation of the ECAR was 11–14%, and the average interassay variation was 12.2% (Supplementary table 1B). The OCR and ECAR rates of iPSC-derived cardiomyocytes from the same batch were very consistent and reproducible measured on 5 separate days. Due to the consistency in the data, further experiments were conducted as two separate experiments.

ATP Levels of iPSC-Derived Cardiomyocytes

Mitochondria are the primary energy suppliers of the cell, providing energy in the form of ATP. Approximately 95% of ATP is produced during aerobic respiration by mitochondria, the rest being made during glycolysis in the cytosol. In order to establish appropriate culture conditions and time points for toxicity testing, iPSC-derived cardiomyocytes were thawed according to the manufacturer’s instructions and maintained in media with different carbon sources containing either HG, LG, LG + FA, GAL, or GAL + FA for up to 21 days. Cardiomyocytes grown in each of the five different media did not show obvi-ous morphological differences (Fig. 1), and ATP content nor-malized to cardiomyocyte number was approximately equal across all media conditions and showed a consistent intergroup increase over days 7, 14, and 21 in culture (Fig. 2). For experi-mental details, please refer to Materials and Methods section.

Bioenergetics of iPSC-Derived Cardiomyocytes

Mitochondria are essential for making almost 95% of the cellular energy supply in the form of ATP by OXPHOS. The relative contribution of OXPHOS and glycolysis to ATP production in iPSC-derived cardiomyocytes was determined by simultaneous measurement of dissolved oxygen in the media (expressed as OCR) and (ECAR), respectively, as detailed in Materials and Methods section. The processes underlying OCR were further investigated through modulating mitochondrial function and measuring the relative increases and decreases in oxygen consumption. OCR under basal conditions is due to OXPHOS where electron transport through the protein complexes I–IV is coupled to ATP synthesis (phosphorylation) catalyzed by the ATP synthase (complex V). Inhibition of the ATP synthase by oligomycin decreases OCR leaving only proton leak across the electron transport chain, cell surface oxygen consumption, peroxisomal oxygen consumption, and substrate oxidation (Abe et al., 2010). Mitochondrial and nonmitochondrial respiration can be further separated by addition of the complex I and III inhibitors rotenone and antimycin A, respectively, to stop proton leak. Conversely, maximal mitochondrial respiration can be investigated by the addition of carbonyl cyanide p-[trifluoromethoxy]-phenyl-hydrazone (FCCP). Addition of FCCP uncouples respiration and represents the mitochondrial capacity to respond to increased energy demands such as utilization of various substrates under conditions of cell stress. Sequential addition of oligomycin and FCCP illustrates the reserve and maximum respiratory capacities of the mitochondria. These bioenergetic parameters in iPSC-derived cardiomyocytes are shown in

HUMAN-INDUCED PLURIPOTENT STEM CELL–DERIVED CARDIOMYOCYTES 119


http://toxsci.oxfordjournals.org/lookup/suppl/doi:10.1093/toxsci/kfs233/-/DC1


FIG. 1. Fluorescent microscope images of iPSC-derived cardiomyocytes grown in (A) HG, (B) LG, (C) LG +FA, (D) galactose, and (E) GAL + FA media on day 7 postplating.

FIG. 2. ATP levels were measured and normalized to their respective cell numbers in iPSC-derived cardiomyocytes grown in HG, LG, LG + FA, GAL, and GAL + FA on day 7, 14, and 21 postplating. Each data point represents the mean ± SEM of 10 wells, n = 2 separate experiments. There was no statistically sig-nificant difference between media conditions at all three time points.

120 RANA ET AL.


Figure 3. Detailed data for the mitochondrial bioenergetics profiles at days 7, 14, and 21 can be found in Supplemental tables 2–4, whereas Figures 3–6 only show representative examples of detailed bioenergetics parameters measured.

Basal Respiration and Glycolytic Rates of iPSC-Derived Cardiomyocytes

OCR and ECAR rates were measured in iPSC-derived car-diomyocytes grown in HG, LG, LG + FA, GAL, or GAL + FA at days 7, 14, and 21 postthaw. The OCR/ECAR ratio was determined to assess the relative contribution of glycolysis and mitochondrial respiration to energy generation (Fig. 4). A higher OCR/ECAR ratio indicates that energy is mainly gen-erated through mitochondrial OXPHOS, whereas a lower OCR/

ECAR ratio indicates that energy is predominantly generated by glycolysis. The OCR/ECAR ratios of iPSC-derived cardiomyo-cytes grown in HG, LG, and LG + FA were significantly lower compared with iPSC-derived cardiomyocytes grown in GAL and GAL + FA (Fig. 4). Furthermore, the OCR/ECAR ratio increased over time in iPSC-derived cardiomyocytes grown in media containing GAL and GAL + FA, suggesting that under these conditions iPSC-derived cardiomyocytes increased their mitochondrial function relative to glycolysis.

Reserve Capacity and Maximum Respiratory Capacity

To investigate a causal effect between energy substrate and mitochondrial reserve or maximum respiratory capacity, these two parameters were measured from cardiomyocytes grown

FIG. 3. This figure is a typical workflow example when utilizing the seahorse technology platform, and it shows the measurements that can be made using such platform. After the baseline OCR is established over a period of time (in this case four measurements over 25 min), mitochondrial modulators can be injected to establish a variety of bioenergetic parameters, such as the uncoupling efficiency, the proton leak, and reserve and maximum reserve capacity as well as respira-tion by other than mitochondrial source. In this particular example, the percentage change in OCR after injection of mitochondrial modulators in iPSC-derived cardiomyocytes growing in galactose-containing media on day 7 postplating on XF 96 flux analyzer is shown. After four baseline OCR measurements, oligomycin (3µM), rotenone (5µM), antimycin (3µM), and FCCP (5µM) are injected sequentially, and OCR measurements were recorded after each injection. To assess the mitochondrial reserve and maximum capacity, OCR was measured after injection (A) of 3µM oligomycin and injection (D) of 5µM FCCP. All the data acquired from bioenergetics measurement are provided in Supplementary tables 1–5.






on each of the five test media. The reserve capacity (Fig. 5A) was significantly higher in cardiomyocytes grown in GAL and GAL + FA for 7 and 21 days than in cells grown in HG, LG, or LG + FA (p < 0.01). In addition, the reserve capacity of the LG media group was significantly higher on day 21 than on day 7. Similar observations were made for the maximum respiratory capacity (Fig. 5B). Cardiomyocytes grown in GAL and GAL + FA had significantly higher values than cells grown in HG, LG, or LG + FA for 7 and 21 days. In addition, just as for the reserve capacity, the maximum respiratory capacity in the low-glucose media group was significantly higher on day 21 than on day 7.

Effect of Mitochondrial OXPHOS Modulators on the ATP Production

The contribution of OXPHOS to ATP production under the various growth substrates was determined by assessing the effects of the ATP synthase inhibition (oligomycin; Fig. 6A) or inhibition of complexes I, III, and V (rotenone, antimycin, and oligomycin, respectively; Fig. 6B) on the ATP levels of iPSC-derived cardiomyocytes growing in different substrate media

at days 7 and 21. As shown in Figure 6A, inhibition of ATP synthase significantly reduced ATP levels of iPSC-derived car-diomyocytes grown in GAL and GAL + FA, but did not reduce ATP levels in cells grown in HG, LG, or LG + FA. Similar results were observed for inhibition of complexes I, III, and V (Fig. 6B).

Metabolism-Related Gene Expression in iPSC-Derived Cardiomyocytes Growing in Different Substrate Media

Commercial and custom RT2 profiler PCR arrays were used to define transcriptional correlates of iPSC-derived cardiomyo-cyte behavior: potential genomic components of bioenergetic adaptability, protective processes, and differences between the in vitro model and native tissues were investigated by compar-ing the expression levels of genes involved in mitochondrial OXPHOS, glycolysis, fatty acid/cholesterol metabolism, apop-tosis, and oxidative stress in iPSC cardiomyocytes across HG, LG, LG + FA, GAL, or GAL + FA culture media as well as to the expression levels present in adult cardiac tissue. The fold change of these genes across the five media compared with adult tissue is represented by the heat map in Figure 7 and Supplementary table 5. The first observation was that the

FIG. 4. OCR (pmol/min)/ECAR (mili pH/min) of iPSC-derived cardiomyocytes grown in HG, LG, LG + FA, GAL, and GAL + FA. Each data point repre-sents the mean ± SEM of 14 wells, n = 2 separate experiments. *p < 0.001 compares iPSC-derived cardiomyocytes grown in HG media against LG, LG + FA, GAL, and GAL + FA media. #p < 0.001 compares iPSC cardiomyocytes on day 7 against day 21.

122 RANA ET AL.



examined genes did not show significantly different expres-sion levels as a function of culture media (Fig. 7). Secondly, compared with the adult human heart tissue, ATP synthase, cytochrome C oxidase, and NADH dehydrogenase genes were under expressed, whereas genes involved in cholesterol metabolism (PRKAG1 and PRKAG2) and protection against apoptotic and oxidative stress processes (BCL2L1 and SOD1, respectively) were equal or over expressed.

Protein Expression of iPSC-Derived Cardiomyocytes Grown in Different Substrate Media

Protein expression analysis was used to further define the mechanisms underlying the bioenergetic behavior of iPSC-derived cardiomyocytes. Levels of the mitochondrial complexes I (NDUFB8 CI-20), II (FeS CII-30), III (CIII-Core

2), IV (CIV-II), and V (ATP synthase alpha subunit); COX 4; and COX 6A were determined for cardiomyocytes grown in each of the five media (Fig. 8). Growth in GAL or GAL + FA resulted in higher levels of mitochondrial complexes I–V than the growth of iPSC-derived cardiomyocytes in HG, LG, or LG + FA media. COX 4 protein was equally expressed across all five media conditions, whereas COX 6A levels were higher in LG + FA, GAL, or GAL + FA media versus LG or HG media. In comparison to adult cardiac tissue, the levels of mitochondrial complexes I–V and COX 6A were significantly less in iPSC-derived cardiomyocytes, whereas the level of COX 4 protein was similar. There was no observable expression of SOD1 and BCL2L1 protein in human heart, whereas there was slight expression of these proteins in iPSC-derived cardiomyocytes growing in all the media types.

FIG. 5. (A) Mitochondrial reserve capacity and (B) maximum respiratory capacity of iPSC-derived cardiomyocytes grown in HG, LG, LG + FA, GAL, and GAL + FA. Each data point represents the mean ± SEM of 12 wells, n = 2 separate experiments. *p < 0.001 compares iPSC-derived cardiomyocytes grown in HG media against LG, LG + FA, GAL, and GAL + FA media. #p < 0.001 compares iPSC-derived cardiomyocytes on day 7 against day 21.



Effect of Tyrosine Kinase Inhibitors (Antineoplastic Agents) on ATP Production and OCRs

The previous results suggested that maintaining cardiomyocytes in GAL or GAL + FA resulted in ATP production predominately through mitochondrial respiration, rather than glycolysis. This hypothesis was confirmed by observing that inhibition of complex I with rotenone reduced the OCR of iPSC-derived cardiomyocytes growing in all different substrate media, however cell viability was only depleted in cardiomyocytes growing in GAL or GAL + FA (Fig. 9). As ATP production through mitochondrial respiration most closely mimics the native condition, a selection of six anticancer drugs (sorafenib, sunitinib, dasatinib, imatinib, lapatinib, and nilotinib) were tested for their effects on oxygen consumption and ATP content in iPSC-derived cardiomyocytes grown in media containing GAL or GAL + FA at day 21 (Fig. 10). Sorafenib reduced OCR and ATP content of the cells in dose-dependent manner (Fig. 10A). Sunitinib had

moderate effect on OCR while drastically reducing ATP content (Fig. 10B). Imatinib and dasatinib had moderate effects on both OCR and ATP content; predominantly observed at 100µM (Figs. 10C and D, respectively). Nilotinib and lapatinib did not have any major effect on either OCR or ATP content of the cells (Figs. 10E and F, respectively). Similar effects were observed when iPSC-derived cardiomyocytes were maintained in GAL + FA medium. The effect of kinase inhibitors on OCRs and cell viability on iPSC-derived cardiomyocytes growing in GAL + FA can be found in Supplementary figure 1.

DISCUSSION

The data presented here highlight several important and novel findings regarding the use and suitability of iPSC-derived cardiomyocytes as an experimental model. Several metabolic, transcriptional, and protein-based investigations defined the

FIG. 6. (A) ATP levels measured in iPSC-derived cardiomyocytes after ATP synthase (complex V) inhibition with oligomycin and (B) complexes I, III, and V inhibition with rotenone, antimycin, and oligomycin, respectively. iPSC-derived cardiomyocytes were grown in HG, LG, LG + FA, GAL, and GAL + FA. Each data point represents the mean ± SEM of eight wells, n = 2 separate experiments. *p < 0.001 compares iPSC-derived cardiomyocytes grown in HG media against LG, LG + FA, GAL, and GAL + FA media.

124 RANA ET AL.



FIG. 7. Heat map of expression analysis of genes involved in the response to mitochondrial oxidative phosphorylation, glycolysis, fatty acid and cholesterol metabolism, oxidative stress, and apoptosis. Fold regulation was calculated using the −∆∆Ct method comparing over treatment to RNA prepared from human heart tissue. An increase (fold regulation) in gene expression is depicted in red color, whereas a decrease in gene expression is represented by blue color. No differences in expression are depicted in black.



overall bioenergetic profile, metabolic processes, and underly-ing mechanisms of iPSC-derived cardiomyocytes and examined their responses to therapeutic compound with mitochondrial liabilities. To our knowledge, this is the first study to exam-ine these endpoints and together and this information provides a timely dataset that will aid in developing and implementing more relevant in vitro models.

Off- and on-target cardiotoxicity is a primary cause of compound attrition during drug development, and postlaunch restrictions, and withdrawal (Force and Kolaja, 2011). Drug-induced adverse cardiac effects include ion channel block, arrhythmia, altered contractility, ischemia, and many more. Therefore, there is an increasing need for relevant cardiotoxicity assessment throughout the drug development process and

FIG. 8. Western blotting analysis of (A) mitochondrial complexes I–V and human heart tissue; (B) SOD1, BCL2L1, COX 6A, COX 4, and human heart tissue. GAPDH was used as a loading control. (C) Quantification of mitochondrial complexes I–V, SOD1, BCL2L1, COX 6A, COX 4, and human heart tissue. Results are presented as means ± SD, n = 3 separate experiments. Statistical significance changes (*p < 0.001) from adult human heart tissue are reported.

126 RANA ET AL.


FIG. 9. The percentage change in OCR and cell viability of iPSC-derived cardiomyocytes growing (at 21 days) in (A) HG, (B) LG, (C) GAL, (D) LG + FA, and (E) GAL + FA containing media after addition of rotenone. Each data point represents the mean ± SD, n = 3 separate experiments. *p < 0.001 compares cell viability values to the OCR at various drug concentrations.



FIG. 10. The percentage change in OCRe and cell viability of iPSC-derived cardiomyocytes growing on galactose containing media at day 21 after addition of (A) sorafenib, (B) sunitinib, (C) imatinib, (D) dasatinib, (E) nilotinib, and (F) lapatinib. Each data point represents the mean ± SD, n = 3 separate experiments. *p < 0.01 compares cell viability values to the OCR at various drug concentrations.

128 RANA ET AL.


in particular during the early stages of drug discovery when chemical matter is available that can be subjected to SAR medicinal chemistry efforts. In recent years, it has become routine for most pharmaceutical companies to evaluate cellular toxicity in basic immortalized cell lines. However, it is nearly impossible to predict particular organ toxicity from such a reduced approach (Lin and Will, 2012). For example, assessment of potential cardiac electrophysiological effects is often performed using various immortalized cell lines over expressing a single ion channel, such as human-ether-a-go-go related gene potassium channels (Moller and Witchel, 2011). Although such models provide a robust tool for examining single-ion channels, they miss potential protein-protein interactions. As a consequence of the disadvantages inherent in current models, recent approaches to predicting arrhythmic effects have utilized cardiomyocytes derived from human iPSCs (Guo et al., 2011; Moretti et al., 2010). Similarly, more holistic cellular approaches have been taken with H9c2, HL-1, neonatal rat and mouse ventricular myocytes, and adult rat/human cardiomyocytes for assessing nonelectrophysiological cardiotoxicity assessment, and the advantages and disadvantages have been discussed recently (Force and Kolaja, 2011).

Predicting specific biochemical and/or metabolic cardiotoxicities has been hampered by non-native cell culture conditions where most immortalized cell lines are cultured in high concentrations of glucose in order to facilitate rapid growth and generate energy. Cells in this environment generate ATP primarily through glycolysis, are less dependent on mitochondrial oxidative phosphorylation, and thus have minimal utility in assessing metabolism-related cardiotoxicity as native cardiac myocytes primarily use fatty acids and are highly oxidative in their ATP production (Pedersen, 2009). Here, we evaluated iPSC-derived cardiomyocytes grown in cell culture across differing carbon substrates for extended durations as a suitable in vitro model with an appropriate bioenergetic profile.

ATP levels in iPSC-derived cardiomyocytes were monitored for 21 days in days in culture and found to increase over time (Fig. 2), which may be due to postdifferentiation development and maturation of the metabolic machinery. Although iPSC-derived cardiomyocytes are fully differentiated cardiomyocytes, they have been shown to exhibit certain characteristics of matu-ration over time in culture using both electrophysiological and genomic markers (Babiarz et al., 2012; Sartiani et al., 2007).

Carbon source in the media was shown to influence the relative levels of active mitochondrial oxidative phosphorylation and glycolysis, with media containing galactose (GAL) or galactose and fatty acids (GAL + FA) showing the highest level of OXPHOS and thus most closely recapitulating the in vivo condition (Fig. 4). Relative levels of OXPHOS activity also showed an increase over time in culture, with GAL + FA reaching a maximum at day 14. Interestingly, when both glucose and fatty acids were available as a substrate (LG + FA), iPSCs remained fairly glycolytic. Despite the availability of FA, iPSCs

appeared to have a preference for glucose under the conditions that we used. The dependence upon mitochondrial respiration as the primary source of energy in cardiomyocytes grown in GAL or GAL + FA was confirmed by a decrease in oxygen consumption upon exposure to the mitochondrial modulators rotenone, oligomycin, and antimycin (Fig. 6) as well as a rotenone mediated loss in cell viability (Fig. 9); in agreement with work by Marroquin et al. (2007) and Will et al. (2007), who also demonstrated that cells grown in media containing galactose were susceptible to mitochondrial modulators.

The bioenergetic parameters of mitochondrial reserve and maximum respiratory capacity also showed a similar depend-ence upon carbon source and duration in culture. Mitochondria play a very important role in mediating the cellular response to oxidants formed during acute and chronic cardiac dysfunc-tion (Pedersen, 2009). When cells are subjected to a variety of different stress elements, mitochondria are capable of draw-ing upon a “reserve capacity” which is available to serve the increased energy demands for maintenance of organ func-tion, cellular repair, or detoxification (Hill et al., 2009). iPSC-derived cardiomyocytes grown in GAL or GAL + FA media have a higher reserve and maximum respiratory capacity sug-gesting that these cardiomyocytes have an appropriate “reserve capacity” relative to cardiomyocytes grown in glucose-con-taining media (Fig. 5). Taken together, these data demonstrate that when cultured under appropriate media conditions and durations, iPSC-derived cardiomyocytes exhibit an aerobically poised metabolism that recapitulates the in vivo condition and provides a suitable model for interrogating several aspects of basal and stressed mitochondrial respiration either as a poten-tial therapeutic or toxic outcome.

iPSC-derived cardiomyocytes showed differing levels of OXPHOS and glycolytic activities as a function of the carbon source in the media (Fig. 4). The expression levels of key genes involved in these processes were lower than those of adult car-diac tissue and did not show significant differences across the different media (Fig. 7). However, iPSC-derived cardiomyo-cytes growing in GAL and GAL + FA did show higher levels of mitochondrial OXPHOS proteins compared with cells growing in rest of the media conditions (Fig. 8). These results suggest that in iPSC-derived cardiomyocytes bioenergetic adaptability occurs posttranscriptionally, and these cells provide a model whereby that adaptability can be modulated or monitored for therapeutic or toxic outcomes, respectively.

Compared with adult human heart, iPSC-derived cardio-myocytes showed increased gene and protein expression of the antiapoptotic protein BCL2L1 and SOD1 (Figs. 7 and 8). Overexpression of BCL2L1 has been reported in other stem cell–derived lines and may promote survival and growth in vitro (Amps et al., 2011; Ardehali et al., 2011), whereas SOD1 represents the first line of defense against oxidative stress and may promote survival (Landis and Tower, 2005; Parker et al., 2004). Together, these data suggest iPSC-derived cardiomyocytes in culture may have adaptive processes related to their in vitro environment.



The data demonstrate that iPSC-derived cardiomyocytes grown in GAL or GAL + FA media mature and increase their mitochondrial contribution to energy generation over time and are most reliant on OXPHOS between day 14 and day 21. We tested several kinase inhibitors on iPSC-derived cardiomyo-cytes growing on this bioenergetically favorable media at day 21. The cytotoxicity rank order of these kinase inhibitors was sorafenib > sunitinib > dasatinib > imatinib > lapatinib > niol-tinib, which is in agreement with published literature that dis-cusses these kinase inhibitors leading to some level of clinical cardiotoxicity (Force and Kolaja, 2011). Will et al. (2008) have also studied the effect of multitargeted tyrosine kinase inhibiti-ors sorafenib, sunitinib, and imatinib on mitochondrial function in isolated rat heart mitochondria and in H9c2 cells and showed evidence that these kinases caused mitochondrial liabilities.

In summary, we have characterized iPSC-derived cardio-myocytes growing on different carbon substrates growth over an extended period of time up to 21 days. We conclude that iPSC-derived cardiomyocytes growing in HG, LG, and LG + FA rely primarily on glycolysis for their ATP production, whereas iPSC-derived cardiomyocytes growing on GAL or GAL + FA primarily rely on mitochondrial OXPHOS for their ATP produc-tion. The contribution of OXPHOS to ATP production, as well as the appearance of “reserve capacity” components increases over time, reaching a peak between 7 and 14 days in culture with GAL + FA. Gene expression analysis confirmed that these although overall expression levels of mitochondrial OXPHOS genes are lower than in native tissue, these cardiomyocytes are capable of adjusting posttranscriptionally in real-time to differ-ent carbon sources. Moreover, iPSC-derived cardiomyocytes growing in GAL and GAL + FA were more susceptible to mito-chondrial toxicants suggesting that may be more proficient in identifying cardiotoxicants that will induce mitochondrial liabil-ities. These data taken together demonstrate that iPSC-derived cardiomyocytes can provide an aerobically poised cellular model that recapitulates native myocyte behavior suitable for cardio-myocyte-based screening.

SUPPLEMENTARY DATA

Supplementary data are available online at http://toxsci.oxfordjournals.org/.

ACkNOWLEDgMENTS

We would like to thank Dr William Pennie and Dr Sashi Nadanaciva for their editorial input.

REFERENCES

Abe, Y., Sakairi, T., Kajiyama, H., Shrivastav, S., Beeson, C., and Kopp, J. B. (2010). Bioenergetic characterization of mouse podocytes. Am. J. Physiol. Cell Physiol. 299, C464–C476.

Allard, M. F., Schönekess, B. O., Henning, S. L., English, D. R., and Lopaschuk, G. D. (1994). Contribution of oxidative metabolism and glycolysis to ATP production in hypertrophied hearts. Am. J. Physiol. 267(2 Pt 2), H742–H750.

Amps, K., Andrews, P. W., Anyfantis, G., Armstrong, L., Avery, S., Baharvand, H., Baker, J., Baker, D., Munoz, M. B., Beil, S., et al. (2011). Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat. Biotechnol. 29, 1132–1144.

Anson, B. D., Kolaja, K. L., and Kamp, T. J. (2011). Opportunities for use of human iPS cells in predictive toxicology. Clin. Pharmacol. Ther. 89, 754–758.

Ardehali, R., Inlay, M. A., Ali, S. R., Tang, C., Drukker, M., and Weissman, I. L. (2011). Overexpression of BCL2 enhances survival of human embryonic stem cells during stress and obviates the requirement for serum factors. Proc. Natl. Acad. Sci. U.S.A. 108, 3282–3287.

Babiarz, J. E., Ravon, M., Sridhar, S., Ravindran, P., Swanson, B., Bitter, H., Weiser, T., Chiao, E., Certa, U., and Kolaja, K. L. (2012). Determination of the human cardiomyocyte mRNA and miRNA differentiation network by fine-scale profiling. Stem Cells Dev. 21, 1956–1965.

Brand, M. D., and Nicholls, D. G. (2011). Assessing mitochondrial dysfunction in cells. Biochem. J. 435, 297–312.

Chapin, R. E., and Stedman, D. B. (2009). Endless possibilities: Stem cells and the vision for toxicology testing in the 21st century. Toxicol. Sci. 112, 17–22.

Cohen, J. D., Babiarz, J. E., Abrams, R. M., Guo, L., Kameoka, S., Chiao, E., Taunton, J., and Kolaja, K. L. (2011). Use of human stem cell derived car-diomyocytes to examine sunitinib mediated cardiotoxicity and electrophysi-ological alterations. Toxicol. Appl. Pharmacol. 257, 74–83.

Davila, J. C., Cezar, G. G., Thiede, M., Strom, S., Miki, T., and Trosko, J. (2004). Use and application of stem cells in toxicology. Toxicol. Sci. 79, 214–223.

Dykens, J. A., and Will, Y. (2007). The significance of mitochondrial toxicity testing in drug development. Drug Discov. Today 12, 777–785.

Force, T., and Kolaja, K. L. (2011). Cardiotoxicity of kinase inhibitors: The prediction and translation of preclinical models to clinical outcomes. Nat. Rev. Drug Discov. 10, 111–126.

Gohil, V. M., Sheth, S. A., Nilsson, R., Wojtovich, A. P., Lee, J. H., Perocchi, F., Chen, W., Clish, C. B., Ayata, C., Brookes, P. S., et al. (2010). Nutrient-sensitized screening for drugs that shift energy metabolism from mitochon-drial respiration to glycolysis. Nat. Biotechnol. 28, 249–255.

Guo, L., Abrams, R. M., Babiarz, J. E., Cohen, J. D., Kameoka, S., Sanders, M. J., Chiao, E., and Kolaja, K. L. (2011). Estimating the risk of drug-induced proarrhythmia using human induced pluripotent stem cell-derived cardio-myocytes. Toxicol. Sci. 123, 281–289.

Hill, B. G., Dranka, B. P., Zou, L., Chatham, J. C., and Darley-Usmar, V. M. (2009). Importance of the bioenergetic reserve capacity in response to car-diomyocyte stress induced by 4-hydroxynonenal. Biochem. J. 424, 99–107.

Kakinuma, Y., Miyauchi, T., Suzuki, T., Yuki, K., Murakoshi, N., Goto, K., and Yamaguchi, I. (2002). Enhancement of glycolysis in cardiomyocytes elevates endothelin-1 expression through the transcriptional factor hypoxia-inducible factor-1 alpha. Clin. Sci. 103(Suppl. 48), 210S–214S.

Landis, G. N., and Tower, J. (2005). Superoxide dismutase evolution and life span regulation. Mech. Ageing Dev. 126, 365–379.

Lasser, K. E., Allen, P. D., Woolhandler, S. J., Himmelstein, D. U., Wolfe, S. M., and Bor, D. H. (2002). Timing of new black box warnings and withdraw-als for prescription medications. J. Am. Med. Assoc. 287, 2215–2220.

Lawrence, C. L., Pollard, C. E., Hammond, T. G., and Valentin, J. P. (2008). In vitro models of proarrhythmia. Br. J. Pharmacol. 154, 1516–1522.

Lexchin, J. (2005). Drug withdrawals from the Canadian market for safety rea-sons, 1963-2004. Can. Med. Assoc. J. 172, 765–767.

Lin, Z., and Will, Y. (2012). Evaluation of drugs with specific organ toxicities in organ-specific cell lines. Toxicol. Sci. 126, 114–127.

130 RANA ET AL.


http://toxsci.oxfordjournals.org/

Lopaschuk, G. D., and Jaswal, J. S. (2010). Energy metabolic phenotype of the cardiomyocyte during development, differentiation, and postnatal matura-tion. J. Cardiovasc. Pharmacol. 56, 130–140.

Ma, J., Guo, L., Fiene, S. J., Anson, B. D., Thomson, J. A., Kamp, T. J., Kolaja, K. L., Swanson, B. J., and January, C. T. (2011). High purity human-induced pluripotent stem cell-derived cardiomyocytes: Electrophysiological prop-erties of action potentials and ionic currents. Am. J. Physiol. Heart Circ. Physiol. 301, H2006–H2017.

Mandenius, C. F., Steel, D., Noor, F., Meyer, T., Heinzle, E., Asp, J., Arain, S., Kraushaar, U., Bremer, S., Class, R., et al. (2011). Cardiotoxicity testing using pluripotent stem cell-derived human cardiomyocytes and state-of-the-art bioanalytics: A review. J. Appl. Toxicol. 31, 191–205.

Marroquin, L. D., Hynes, J., Dykens, J. A., Jamieson, J. D., and Will, Y. (2007). Circumventing the Crabtree effect: Replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol. Sci. 97, 539–547.

Moller, C., and Witchel, H. (2011). Automated electrophysiology makes the pace for cardiac ion channel safety screening. Front. Pharmacol. 2, 73.

Moretti, A., Bellin, M., Welling, A., Jung, C. B., Lam, J. T., Bott-Flügel, L., Dorn, T., Goedel, A., Höhnke, C., Hofmann, F., et al. (2010). Patient-specific induced pluripotent stem-cell models for long-QT syndrome. N. Engl. J. Med. 363, 1397–1409.

Parker, J. D., Parker, K. M., Sohal, B. H., Sohal, R. S., and Keller, L. (2004). Decreased expression of Cu-Zn superoxide dismutase 1 in ants with extreme lifespan. Proc. Natl. Acad. Sci. U.S.A. 101, 3486–3489.

Pedersen, P. L. (2009). Mitochondrial matters of the heart: A plethora of regula-tory modes to maintain function for a long lifetime. J. Bioenerg. Biomembr. 41, 95–98.

Rossignol, R., Gilkerson, R., Aggeler, R., Yamagata, K., Remington, S. J., and Capaldi, R. A. (2004). Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. Cancer Res. 64, 985–993.

Sartiani, L., Bettiol, E., Stillitano, F., Mugelli, A., Cerbai, E., and Jaconi, M. E. (2007). Developmental changes in cardiomyocytes differentiated from human embryonic stem cells: A molecular and electrophysiological approach. Stem Cells 25, 1136–1144.

Wallace, D. C. (2005). A mitochondrial paradigm of metabolic and degenera-tive diseases, aging, and cancer: A dawn for evolutionary medicine. Annu. Rev. Genet. 39, 359–407.

Will, Y., Dykens, J. A., Nadanaciva, S., Hirakawa, B., Jamieson, J., Marroquin, L. D., Hynes, J., Patyna, S., and Jessen, B. A. (2008). Effect of the multitar-geted tyrosine kinase inhibitors imatinib, dasatinib, sunitinib, and sorafenib on mitochondrial function in isolated rat heart mitochondria and H9c2 cells. Toxicol. Sci. 106, 153–161.



characterization of human-induced pluripotent stem cell–derived

Documents