Research ArticleBioengineering

Cell-selective arrhythmia ablation for photomodulation of heart rhythm

See allHide authors and affiliations

Science Translational Medicine  28 Oct 2015:
Vol. 7, Issue 311, pp. 311ra172
DOI: 10.1126/scitranslmed.aab3665

For heart cells only

Abnormal heartbeats, called arrhythmias, can be stopped by photoablation, but the use of light energy to terminate malfunctioning cardiomyocytes runs the risk of damaging the other dozen or so cell types in the heart. To be more specific in photoablation, Avula and colleagues devised a heart cell–targeted photosensitizer, which could be delivered systemically. Laser light was then used to ablate only cardiomyocytes while leaving the surrounding fibroblasts and other cells intact. The approach was tested in vivo in rodents and in sheep and rat hearts ex vivo, demonstrating that the technology is indeed able to avoid fibroblasts and block electrical conduction, returning the heart to its normal rhythm.


Heart disease, a leading cause of death in the developed world, is overwhelmingly correlated with arrhythmias, where heart muscle cells, myocytes, beat abnormally. Cardiac arrhythmias are usually managed by electric shock intervention, antiarrhythmic drugs, surgery, and/or catheter ablation. Despite recent improvements in techniques, ablation procedures are still limited by the risk of complications from unwanted cellular damage, caused by the nonspecific delivery of ablative energy to all heart cell types. We describe an engineered nanoparticle containing a cardiac-targeting peptide (CTP) and a photosensitizer, chlorin e6 (Ce6), for specific delivery to myocytes. Specificity was confirmed in vitro using adult rat heart cell and human stem cell–derived cardiomyocyte and fibroblast cocultures. In vivo, the CTP-Ce6 nanoparticles were injected intravenously into rats and, upon laser illumination of the heart, induced localized, myocyte-specific ablation with 85% efficiency, restoring sinus rhythm without collateral damage to other cell types in the heart, such as fibroblasts. In both sheep and rat hearts ex vivo, upon perfusion of CTP-Ce6 particles, laser illumination led to the formation of a complete electrical block at the ablated region and restored the physiological rhythm of the heart. This nano-based, cell-targeted approach could improve ablative technologies for patients with arrhythmias by reducing currently encountered complications.


Disturbances of the myocyte-generated electrical impulse may result in localized self-perpetuating arrhythmia sources. Many arrhythmias are associated with morbidity and/or deleterious effects such as heart failure, stroke, or sudden death (1). Arrhythmia management generally includes either antiarrhythmic drugs and/or interventional procedures such as surgery or catheter ablation. In the last few decades, catheter ablation has emerged as a common procedure used in the hospital setting to treat arrhythmias. Ablation technologies deliver energy locally to cardiac sites that harbor electrical sources of arrhythmia. Commonly used energy delivery modes are radiofrequency, cryoenergy, light amplification by stimulated emission of radiation (LASER), or ultrasound. The heart is a complex organ consisting of about 8 to 10 cell types, with cardiac myocytes accounting for only one-third of the total number of cardiac cells (2); the rest of the cells include fibroblasts and other connective tissue cells; nerve cells; smooth muscle and endothelial cells of the coronary vasculature and endocardium; and mast cells, stem cells, and pericytes (2, 3). Among these, only cardiac myocytes are the excitable cells that form an electrical syncytium and enable heart contraction and relaxation. Understandably, the resultant cellular damage from ablative therapies is nonspecific and, although the myocytes receive ablative energy, all bystander cells are also damaged. This creates the risk of serious complications, such as atrioesophageal fistula, pulmonary vein (PV) stenosis, and coronary artery injury (4, 5).

Herein, we demonstrate in vivo the feasibility of a myocyte-selective ablation through the use of photodynamic therapy (PDT). PDT is a clinically approved localized phototherapy that has been tested for several medical conditions, especially cancer (6). During PDT, the photosensitizer—excited by illumination from an appropriate wavelength light source (for example, a laser)—produces cytotoxic reactive oxygen species (ROS), resulting in cellular damage (6). PDT in the human heart is an emerging approach, currently limited to the treatment of coronary arterial disease. A molecular photosensitizer locates to the atheromatous plaque, and endovascular light illumination leads to plaque removal (7, 8). We previously used 50- to 60-nm methylene blue–loaded polyacrylamide nanoparticles as a PDT photosensitizer to treat cancer cells in vitro (9). By conjugating these nanoparticles to a cardiac-targeting peptide (CTP), we further demonstrated its selectivity to myocytes over fibroblasts in vitro (10).

However, in vivo use of polyacrylamide-based nanoparticles was limited because of the size of the cardiac capillary vessels’ fenestrations/pores (range, 6 to 25 nm) (11, 12). Thus, a new nanoparticle matrix was developed to pass through the cardiac capillary vessels in vivo. Here, we describe a myocyte-targeting nanoparticle conjugated to the photosensitizer chlorin e6 (Ce6) that can reach the heart in vivo. As a photosensitizer, the Ce6 molecule is a natural product that has high optical absorption in the near infrared, a wavelength that penetrates deeply into living tissues (13). The CTP-Ce6 nanoparticles enabled selective ablation of cardiomyocytes, leaving bystander cells untouched while terminating abnormal electrical activity both in vivo in rat and ex vivo in sheep hearts.


Design and characterization of cardiac-targeting Ce6-PEG

Among available biocompatible materials, we chose the U.S. Food and Drug Administration–approved poly(ethylene glycol) (PEG) material with low toxicity, enhanced plasma circulation time, reduced protein fouling, and chemical versatility (14). We adapted a star-shaped eight-arm PEG supramolecule to enable the attachment of a chlorophyll-related dye molecule, Ce6, and multiple moieties of CTP to a single PEG molecule, thus achieving both therapeutic and targeting functionalities (15). The CTP, a 12–amino acid targeting moiety (APWHLSSQYSRT), provided cell-targeting specificity for myocytes (10, 16). Our synthetic strategy for preparing these multifunctional (targeting and therapy) nanoparticles is shown in Fig. 1A. As presented, free Ce6 was conjugated onto the eight-arm PEG, forming covalent chemical bonds to prevent the leaching out of the Ce6 from the nanoparticle matrix. Next, the CTP (modified with cysteine “C” attached) was covalently attached to the Ce6-PEG, using a bifunctional PEG as a cross-linker.

Fig. 1. CTP-Ce6-PEG nanoparticle synthesis and in vitro myocyte-specific uptake and photoablation.

(A) Schematic of nanoparticle preparation. Primary amine group on the eight-arm PEG is first conjugated to carboxylic acid group on Ce6 by carbodiimide cross-linking chemistry. Then, N-hydroxysuccinimide–PEG-maleimide was used to cross-link CTP- and Ce6-conjugated eight-arm PEG. (B) Coculture of primary adult rat ventricular myocytes (rod-shaped) and adult rat cardiac fibroblasts (irregularly shaped) were incubated with CTP-Ce6-PEG for 2 hours. calcein AM identified all live cells (green), and Ce6 fluorescence is seen in purple. Scale bars, 100 μm. (C and D) Targeted PDT in vitro in a rat myocyte–fibroblast coculture that had been incubated with CTP-Ce6-PEG for 2 hours and then exposed to laser illumination (405 nm, 7 to 10 mW) for 10 to 15 min. Uptake of PI, a dead cell indicator (red), or progressive loss of calcein AM, a live cell indicator (green), was imaged (C) and quantified (D). Images are representative of four experiments. Data are representative single cells followed over time after illumination.

Synthesis details and characterization of the CTP-Ce6-PEG are shown in fig. S1. The loaded Ce6-PEG showed three shifted absorption peaks, including 660 nm—a red shift compared to free Ce6. The Ce6 content of 2.2 (±0.1) mg in 100 mg of Ce6-PEG was calculated on the basis of a (free) Ce6 absorbance calibration curve using the area of peak (621 to 670 nm) (17), or roughly 1 or 2 molecules of Ce6 per eight-arm PEG. Ce6 excitation and fluorescence emission spectra appeared at 655 and 660 nm (λmax), whereas the Ce6-PEG peaks were shifted to 660 and 664 nm (λmax), respectively (fig. S1A). The fluorescence emission of Ce6-PEG shows a higher intensity than free Ce6, for a similar Ce6 concentration. After CTP conjugation, the intensity of the absorbance and fluorescence spectra of CTP-Ce6-PEG decreased compared to Ce6-PEG (fig. S1A). The number of conjugated CTP units was 2.9 (±0.3) per one CTP-Ce6-PEG nanoparticle. On the basis of correlation with the molecular weight, the hydrodynamic size of the CTP-Ce6-PEG was determined to be 6.7 (±0.9) nm, which should be small enough to penetrate the capillary vessels’ fenestrations/pores (range, 6 to 25 nm) (11, 12). However, the CTP-Ce6-PEG, due to its star shape, could have a smaller hydrodynamic size than linear proteins of the same molecular weight (14).

To determine the amount of ROS produced from the Ce6-PEG, the “k value” was calculated using first-order decay kinetics of the oxidative quenching of anthracene-9,10-dipropionic acid (ADPA) (fig. S1B). For Ce6-PEG (0.1 mg/ml) in phosphate-buffered saline (pH 7.4), the k value was found to be 2.99 × 10−4. The k value for the same amount of free Ce6 in the test solution of Ce6-PEG was 2.12 × 10−4. This increase is presumably due to the bulky eight-arm PEG, which creates distance between the fluorescent dye moieties and could thus reduce the self-quenching rate of Ce6 and, in turn, increase the fluorescence yield and ROS production. About 4.5-fold more ROS was produced from the Ce6-PEG nanoparticle than from the same total nanoparticle mass of our previously reported nanoparticles (9, 10).

Myocyte-specific ablation in vitro

We first tested the cell selectivity of the CTP-Ce6-PEG in vitro in cocultures of isolated adult rat ventricular myocytes and fibroblasts. Cells in the coculture were bathed in a medium containing targeted CTP-Ce6-PEG nanoparticles (0.1 mg/ml) [corresponding to free Ce6 (0.0016 mg/ml)]. As shown in Fig. 1B, Ce6 fluorescence was seen in myocytes, but not in fibroblasts, indicating specific uptake. Upon illumination with a 405-nm laser beam (7 to 10 mW, 10 to 15 min, ~5 J), myocytes rapidly changed from a rod-like shape to a random, shrunken shape (Fig. 1C). Propidium iodide (PI) uptake (red fluorescence) progressively increased, whereas the calcein AM staining (green fluorescence) vanished. In contrast, the fibroblasts showed no such changes (Fig. 1D). In a control experiment, myocytes and fibroblasts underwent PDT after incubation with free Ce6 (nontargeted, no CTP). Both cell types shrunk and showed nuclear PI uptake while simultaneously losing calcein fluorescence (fig. S2). This demonstrates that our targeted CTP-Ce6-PEG nanoparticles enabled PDT-induced cell death with a high degree of cardiomyocyte selectivity.

PDT in human induced pluripotent stem cell–derived cardiomyocytes

We tested the cell selectivity efficiency of CTP-Ce6-PEG in a coculture of human induced pluripotent stem cell (iPSC)–derived cardiomyocytes (hiPSC-CMs) and human fibroblasts. The experimental protocol was similar to that of the rat in vitro experiments (CTP-Ce6-PEG, 2-hour incubation; 405-nm laser, 7 to 10 mW, 7 to 10 min). Unlike adult rat cardiac myocytes and fibroblasts, hiPSC-CMs and human fibroblasts are not distinguishable morphologically. Thus, in addition to calcein AM and PI, the calcium indicator Fluo-4 AM was added to the coculture to indicate the hiPSC-CMs, because cardiomyocytes, but not fibroblasts, show spontaneous calcium depolarization at the frequency of ≈1 Hz (fig. S3). After targeted PDT in the presence of CTP-Ce6-PEG, the hiPSC-CMs were ablated (PI uptake and disappearance of calcein AM) (fig. S3A). In contrast, the fibroblasts showed no such changes. After nontargeted PDT with free Ce6 (no CTP), both cell types showed PI uptake (fig. S3B). This experiment demonstrates that the targeted CTP-Ce6-PEG nanoparticles enabled PDT-induced cell death specific to human cardiomyocytes.

Myocyte-specific ablation in vivo in rats

Next, we evaluated the ability of the CTP-Ce6-PEG nanoparticles to achieve myocyte-specific ablation in vivo in rats. In one group of animals, we confirmed CTP-Ce6-PEG delivery and myocyte-specific affinity and tissue damage. One hour after injection of CTP-Ce6-PEG (1.6 mg of Ce6 per rat), hearts were isolated and immunostained for myocytes and fibroblasts. Ce6 was present in myocytes but absent from fibroblasts (Fig. 2A). There was a significant linear correlation between myocyte heavy chain fluorescence (in cardiomyocytes) and Ce6 (nanoparticle) fluorescence, whereas there was nearly no correlation between vimentin fluorescence (in fibroblasts) and Ce6 fluorescence (nanoparticle) (Fig. 2B), indicating that ~60% of the myocyte area and ~4% of the fibroblast area had the nanoparticle attached. Photoablation at the left ventricle was performed by applying a laser beam (30 mW, 671 nm) onto the left ventricular epicardium for 10 to 15 min. Before tissue samples were collected, rats received an injection of PI in the tail vein. There was a greater amount of dead cells at the illuminated ablation site, with PI intensity gradually declining toward the nonablated site (Fig. 2C).

Fig. 2. Myocyte-specific targeting of CTP-Ce6-PEG in vivo in rats.

(A) One hour after CTP-Ce6-PEG injection (1.6 mg of Ce6 per rat), hearts were isolated, fixed, and stained for myocytes [cardiac myosin heavy chain (MHC)] and fibroblasts (vimentin). DAPI (4′,6-diamidino-2-phenylindole) was used to stain nuclei. The Ce6 nanoparticle is the red fluorescence. Scale bar, 50 μm. (B) Colocalization of cardiac or fibroblast fluorescence and Ce6 nanoparticle fluorescence, with corresponding Pearson correlation coefficient analysis (right) presented as a boxplot (n = 3 animals, six images per animal). (C) Heart tissue areas that received PDT, including nonablated surrounding regions, stained for live (calcein AM, green) and dead (PI, red) cells. PI intensity profiles are below the representative images (n = 3 animals, six images per animal).

After PDT therapy, images and intensity profile analyses indicated that the PI stain was only visible in myocyte nuclei, whereas it was absent from fibroblasts found between cardiomyocytes (Fig. 3A), indicating that the photodynamic effect and the ROS-induced cellular membrane breech occurred exclusively in myocytes. By comparison, with free Ce6 (nontargeted, no CTP), both myocytes and fibroblasts were killed during PDT (Fig. 3A), which supports our in vitro data with free Ce6 (fig. S2).

Fig. 3. Targeted photoablation in vivo spares bystander fibroblasts and vascular cells.

(A) Myocyte-specific (CTP-Ce6-PEG) and nontargeted (Ce6 only) photoablation were performed in n = 3 rats each. The photoablation sites were imaged for PI fluorescence. Cardiac myosin heavy chain identified myocytes, whereas vimentin identified fibroblasts. DAPI was used to stain nuclei. High-magnification images and corresponding fluorescence intensity profiles (right) show nuclear localization of PI. (B) Coronary vessels are not affected by CTP-Ce6-PEG PDT. Myocyte-specific (CTP-Ce6-PEG) and nontargeted (Ce6 only) photoablation were performed in n = 3 rats each. Tissue sections with vessels were stained for myocytes (green), nuclei (DAPI), PI (red), fibroblasts, and vascular endothelial and smooth muscle cells (vimentin, yellow). Fluorescent images from one representative field of view are merged (top), DAPI only (middle), and PI only (bottom). (C) PI fluorescence intensity in cardiomyocyte and vascular cell nuclei after myocyte-targeted and nontargeted PDTs. Data are means ± SD (n = 3 animals, 9 to 13 fields of view, 29 to 48 nuclei). P values were determined by paired Student’s t test. (D) Quantification of the number of nuclei with PI uptake in myocytes, fibroblasts, and vascular endothelial cells after targeted and nontargeted PDTs. Data are means ± SD (n = 3 animals; 11 to 18 fields of view, 257 to 365 nuclei). P values were determined by paired Student’s t test.

We further analyzed whether blood vessel injury occurs after PDT by quantifying endothelial and vascular wall smooth muscle cell death, recognized by vimentin staining (18). In the animals injected with the targeted CTP-Ce6-PEG, none of the vascular cells died, but cardiomyocytes were ablated (Fig. 3B). In animals injected with free Ce6 (nontargeted, no CTP), however, about 91% of endothelial cells and vascular wall smooth muscle cells died (Fig. 3D), comparable to the percentage (90%) of myocytes. After targeted PDT, the PI fluorescence of cardiomyocytes’ nuclei was several orders of magnitude more intense than the bystander vascular cell nuclei (Fig. 3C). In contrast, after nontargeted PDT, both cell types had similar PI fluorescence intensity. Quantification of fluorescent nuclei confirmed that fibroblasts and vascular endothelial cells were unaffected by CTP-Ce6 nanoparticle targeted PDT (Fig. 3D).

Electrophysiological effects of targeted PDT in rat hearts in vivo and ex vivo

We next tested the electrophysiological effects of myocyte-specific photoablation in vivo. The left atrium of the rat heart was exposed after left lateral thoracotomy (Fig. 4A). One hour after injection of CTP-Ce6-PEG or a nontargeted particle, Ce6-PEG, a laser beam (30 mW, 671 nm, 2-mm diameter) was directed toward the left atrium for 10 to 15 min (Fig. 4A). With the goal of obtaining complete left atrial ablation, the laser energy delivered was ~15 J in the open-chest setup. With a bipolar electrode positioned in the illuminated region, the left atrial appendage (LAA) electrogram amplitude was continuously recorded.

Fig. 4. In vivo cardiomyocyte-specific photoablation.

(A) Rat open-chest experimental setup. After anesthesia and lateral thoracotomy, the rat heart was exposed and a bipolar electrode was positioned on the LAA. LV, left ventricle. (B) In vivo PDT. After injection of CTP-Ce6-PEG (n = 5 animals), Ce6 (n = 3 animals), or no injection (sham surgery, n = 5 animals), a laser beam (671 nm, 30 mW, 1-mm diameter) was directed toward the LAA. LAA electrograms were recorded before and during laser illumination. Double, black vertical bars indicate a 6-min break in the electrogram. One representative example in each group is shown.

There was a progressive decrease in the LAA electrogram amplitude during targeted photoablation (Fig. 4B). After about 8 min, the local bipolar signal amplitude decreased significantly, on average by 82.8 ± 7.3%, compared to the amplitude before PDT (n = 5; P < 0.05, paired Student’s t test). This large decrease indicates that the voltage generated by the myocytes—the main source of cardiac electrical impulse—was markedly diminished after photoablation. The remaining electrical activity likely corresponded to a “far-field” detection of the electrical activity of the right atrium. PDT after injecting free Ce6 (nontargeted, no CTP) also led to a significant decrease in LAA electrogram amplitude (95.7 ± 0.4% average decrease, n = 3; P < 0.01, paired Student’s t test) (Fig. 4B). Laser illumination in the absence of Ce6 nanoparticles (sham) did not change the amplitude significantly (9.08 ± 3.8% decrease, n = 5; P = 0.24, paired Student’s t test) (one representative example from each group is shown in Fig. 4B). Any decrease larger than 60 to 70% (before versus after PDT in the same heart) indicates substantial efficacy. These data indicate that both specific and nonspecific approaches are effective at dampening electrical activity.

In a different group of animals, we injected CTP-Ce6-PEG (1.6 mg of Ce6 per rat), and hearts were explanted and Langendorff-perfused ex vivo. Then, blebbistatin (myosin IIA–actin binding inhibitors that disable contraction of the heart while preserving electrical conduction) and DI-4-ANEPPS (a dye that fluoresces in response to changes in membrane electrical potential) were added to the perfusate to enable motion correction and voltage optical mapping, as described previously (19). In the presence of acetylcholine (0.4 to 0.8 μM), atrial arrhythmia was initiated. After 10 min of arrhythmia stabilization, a laser beam (671 nm, 30 mW, 200-μm diameter) was directed toward the junction between the posterior left atrium (PLA) and the LAA (Fig. 5, inset). The laser therapeutic energy in the Langendorff experiments was about 2 J. [A lower energy was used ex vivo compared with in vivo (open heart) studies because the objective of the Langendorff experiments was to evaluate the modulating impact of cell-specific ablation in a small region of the PLA after an atrial arrhythmia had been initiated.] After 10 (n = 3) or 60 (n = 1) s of laser illumination, arrhythmia converted to normal sinus rhythm in all four hearts (Fig. 5A). Control hearts (n = 4) were similarly tested, except that CTP-Ce6-PEG was not injected (sham). In these animals, repeated laser illumination did not lead to arrhythmia termination (Fig. 5B). Altogether, these experiments indicate that photoablation targeted and limited to cardiomyocytes may lead to arrhythmia termination in the rat heart.

Fig. 5. Electrophysiological effects of myocyte-specific PDT on rat hearts ex vivo.

(A and B) Inset: Schematic showing the site of laser illumination at the junction PLA-LAA. LAA bipolar electrogram recordings, before and after laser illumination (10 or 60 s), in the presence (A) or absence (B) of CTP-Ce6-PEG (1.6 mg of Ce6 per rat injected before heart extraction). Data are from eight individual animals.

In vivo photoablation and 3-day follow-up in rats

We compared the postoperative outcome and persistence of ablative lesion in targeted versus nontargeted ablation in rats. Laser therapeutic energy delivered was 5 J (0.03-W laser, 571 nm, 3 to 5 min), and the laser guide was placed on the anterior wall of the left ventricle. After non–cardiomyocyte-specific PDT (half the photosensitizer dose, 0.8 mg of Ce6 per rat), histological analysis revealed extensive coronary vessel injury (vacuolation and disruption of the tunica media of coronary vessels) (Fig. 6A). Conversely, the cardiomyocyte-specific PDT hearts showed no vascular abnormalities (Fig. 6A). Adjacent tissue sections from the PDT site (1 to 2 mm from the ablated zone) in the same animals showed no vascular abnormalities (Fig. 6A), demonstrating the localized effect of PDT where the laser is illuminated. During the postoperative follow-up, all four rats in the noncardiomyocyte-specific PDT died; however, no animals (zero of six) died after cardiomyocyte-specific PDT (Fig. 6B). With half the photosensitizer dose (0.8 mg of Ce6 per rat) in nontargeted PDT, two of three animals survived (Fig. 6B). Furthermore, 3 days after in vivo photoablation, the electrical block remained complete in the targeted hearts (movie S1), whereas laser-only control (no Ce6) application did not result in any conduction alterations (Fig. 6C).

Fig. 6. In vivo photoablation and 3-day histology and survival in rats.

(A) Hematoxylin and eosin (H&E)–stained sections of the photoablation site 3 days after nontargeted (one-half dose of Ce6 only) and cardiomyocyte-targeted (CTP-Ce6-PEG, 1.6 mg of Ce6 per rat) PDTs. Vacuolation and disruption of the tunica media are indicated by arrowheads. Images are representative of n = 2 controls and six targeted animals. Adjacent (1 to 2 mm from the ablated zone) H&E-stained tissue sections from the same animals are shown for comparison. Scale bars, 100 μm. (B) Survival rates 3 days after in vivo PDT. Nontargeted animals received full (1.6 mg of Ce6) or half (0.8 mg of Ce6) dose per rat. Sham animals received only laser illumination and no photosensitizer. P values were determined by log-rank (Mantel-Cox) test. (C) Voltage movie snapshots of Langendorff-perfused rat hearts during pacing removed 3 days after in vivo PDT with CTP-Ce6-PEG or laser only (sham, no Ce6). (The full-dose, Ce6-only animals did not survive up to 3 days postoperatively (B); we therefore did not obtain optical mapping results for these animals.)

Depth of the ablative lesion

To evaluate the ablated lesion depth, PDT was performed in Langendorff-perfused rat hearts previously injected with CTP-Ce6-PEG (1.6 mg of Ce6 per rat, tail vein). Confocal microscopy analysis of PI stain showed that the ablation depth extended to about 2.37 ± 0.12 mm into the left ventricle wall (fig. S4B). In a separate Langendorff-perfused heart, only laser was shone, without a previous injection of CTP-Ce6-PEG. In this heart, there was no detectable PI stain in the myocardial wall (fig. S4C).

Ex vivo optical mapping of sheep heart before and after PDT

Sheep hearts were explanted and Langendorff-perfused. Blebbistatin and DI-4-ANEPPS were added to the perfusate, similar to the rat studies. Sequential voltage movie snapshots were taken from optical mapping recordings obtained before and after photoablation (Fig. 7A; movies S2 and S3, n = 4). The same hearts served as both laser-only control and PDT group because nanoparticles were perfused after collecting the laser-only data. Before perfusion of CTP-Ce6-PEG, laser illumination did not alter electrical conduction in the region of illumination (Fig. 7A). After perfusion with CTP-Ce6-PEG, laser illumination (5 min on right atrial free wall; 1 and 2 min on right ventricle) resulted in an area of complete electrical block (Fig. 7A). As a result, in both the right atrial free wall and the right ventricle, the pacing impulse split into separate wavelets or circumnavigated the region of block. Complete block of atrioventricular (AV) nodal conduction was also demonstrated after perfusion with CTP-Ce6-PEG followed by laser illumination (671 nm, 30 mW) of the AV nodal region, whereas in the laser-only control hearts the AV nodal conduction was not blocked (Fig. 7B). By comparing apparent conduction velocity average values in ablated regions with nonablated regions (movies recorded at a 250-ms pacing cycle length), we confirmed that the electrical block across the ablated zones was complete (Fig. 7C, laser power 0.03 W). These results indicate that after short illumination periods (1 to 5 min), myocyte-specific photoablation can result in regional electrical block. These lesions were achieved nonthermally without increase in the temperature at the site of illumination (Fig. 7C and fig. S5).

Fig. 7. Optical mapping of ex vivo sheep hearts after photoablation.

(A) Sequential voltage movie snapshots obtained before (“laser only,” control) and after PDT in ex vivo sheep hearts (n = 4, same hearts served as both control and PDT group, because nanoparticles were perfused after collecting the data with laser only). Locations of laser illumination site, pacing site, and optically imaged areas are delineated by the red dotted boxes for the right atria (RA) and ventricles (RV). Laser only, no Ce6: 671 nm, 30 mW, 5 min; CTP-Ce6-PEG: the laser illumination was 5 min on right atrial (RA) free wall and 1 and 2 min on right ventricle. (B) Upper panel: Electrogram recordings from the AV node before and after targeted PDT or laser only [n = 3, as one heart from (A) transitioned to ventricular fibrillation before starting AV nodal ablation]. Lower panel: Representative bipolar recordings at a smaller time scale before and after photoablation. (C) Apparent conduction velocity measured in both ablated and nonablated regions (pacing cycle length, 250 ms; laser power, 0.03 W). Thermal imaging of right atrial free wall during PDT shows no significant change in the temperature at the site of ablation. Data are means ± SD (n = 3 to 4). P value was determined by paired Student’s t test.


Catheter ablation using radiofrequency or other thermal energy is the most commonly used approach to treat drug-resistant atrial fibrillation (AF). Typical ablation procedures consist of energy delivery in well-defined cardiac regions responsible for initiation and/or perpetuation of arrhythmias (20). These approaches are nonspecific, in the sense that the ablative energy is delivered both to the electrically active cardiac tissue and to bystander cardiac or noncardiac tissues. Collateral damage and complications, such as atrioesophageal fistula, PV stenosis, or coronary artery injury, are known side effects of a nonspecific approach (5, 21). Thus, there is great interest in targeting specific cells perpetuating the arrhythmia while preserving the integrity of surrounding, healthy tissues.

Recently, several studies have suggested that PDT—which is nonthermal, ROS-dependent ablation using a photosensitizing agent—may be safely and successfully implemented in rat, porcine, and canine hearts (2224). However, in these studies, a nontargeted photosensitizer was used and resulted in nonspecific ablation in all areas illuminated by the laser. Thus, nonspecific PDT may only provide incremental improvement over existing radiofrequency and thermal ablation methods. Cell-specific PDT has been proposed to improve cancer therapies (9) but has not been tested in the heart to target cardiomyocytes.

We therefore provide proof of concept in this study of performing cardiac cell–specific PDT in small and large animal hearts, including hearts undergoing arrhythmias. Our photosensitizing nanoparticle, CTP-Ce6-PEG, shown previously in vitro to selectively attach to cardiac myocytes (10), similarly demonstrated PDT-induced cell death of hiPSC-derived cardiomyocytes with high specificity toward heart cells in a coculture with fibroblasts in vitro. In vivo in rats, CTP-Ce6-PEG targeted myocytes but not the interspersed fibroblasts after intravenous injection. Laser illumination led to complete electrical block in rat and sheep hearts ex vivo and in vivo in rats without damage to surrounding tissues. Collectively, these results demonstrate the feasibility of cell-specific photoablation in hearts undergoing arrhythmias.

Laser penetration demonstrated in our study—2.4 mm (at 0.03 W, for 5 min)—could be compatible with human atrial muscle. The human left atrial myocardium thickness is on average 4.5 mm at the roof, 3.9 mm at the lateral wall, and 3.3 mm at the anterior wall (25). The myocardial thickness values of structures that represent a challenge to ablate transmurally, such as the mitral isthmus or the left lateral ridge, are 3.8 and 2.8 mm, respectively (26). Also, the PLA’s thickness ranges from 1.7 mm on its superior level to 4.3 mm on its inferior level (27). With a moderate increase in laser power or illumination duration, we anticipate that we will be able to generate transmural lesions in the thickest regions of the human atrium. Here, it is important to consider that the level of laser energy that we use in our experiments is smaller by a factor of ~200 than previously demonstrated energy levels necessary for “non-PDT, high-power, laser-only” ablation in humans (5.5 to 15 W) (28). Furthermore, experiments ex vivo in explanted human hearts, which were previously rejected for transplantation, have demonstrated that CTP conjugated with 6-carboxyfluorescein was internalized by human cardiomyocytes (29), suggesting that our CTP-based approach may be translated to specific uptake in human tissues.

Our study indicates that the myocyte-specific photoablation is a sound approach to terminating arrhythmias in the rat heart and has a strong translational potential to interface with the current clinical setup. Looking forward, CTP-Ce6-PEG will require a pharmaceutical-grade upscaling of the nanoparticle production. Also, laser transmission through a bloodless field in the heart is a practical limitation. In humans, laser guides and/or balloons will need to be adapted with the objective of reaching any region of the heart. We also acknowledge that the optimal parameters (photosensitizer and laser doses, approach) will need to be adjusted to maximize efficacy while, at the same time, decreasing complications. As an example, we did not investigate whether the absence of postablation mortality in our targeted ablation group in rats followed the sparing of the pericardial circulation and the limiting in pericardial effusion. These aspects will be best evaluated in large animal studies and/or pilot clinical studies. Finally, we used a model of acetylcholine-induced atrial arrhythmia. This model does not fully recapitulate the PV triggers for AF seen in human AF. Thus, results here are interpreted cautiously. Future studies are warranted to demonstrate the efficacy of this approach in large animal models of persistent arrhythmias, such as the ones available in sheep or pig, which are more similar in anatomy and electrophysiology to humans.

Cell-specific laser photoablation may improve both endocardial and epicardial ablation approaches. During cardiac surgery in particular, the performance of epicardial cell–specific ablation could simplify the isolation of arrhythmogenic regions. Also, our work sets the stage for the development of cell-specific alternatives to common cardiac ablation methods. For example, nerve terminals or Purkinje myocytes are known to play a critical role in arrhythmia initiation (30, 31). Still, with traditional methods, it is nearly impossible to target these structures without affecting other cardiac cells. Cell-specific technologies could be implemented in well-defined clinical scenarios in which Purkinje cells or cardiac neurons are targeted. Altogether, cell-specific photoablation could advantageously complement current ablation systems with the dual goal of decreasing procedure time and the risk of complications, and improving efficacy.


Study design

Our overall objective was to demonstrate the feasibility of selective cardiomyocyte ablation in the heart. First, we sought to show efficacy in vitro in both rat and human iPSC-CMs in cocultures with cardiac fibroblasts. We designed these coculture experiments to mimic the pro?ximity of cardiomyocytes and fibroblasts in the in vivo heart. The end point was cellular uptake of the dead cell stain PI. Control experiments were performed similarly using nontargeted “free” Ce6 (without CTP) to confirm that in this setting, the cell damage was unspecific. Then, in vivo ablation experiments were performed in open-chest rat preparations. These experiments were designed to examine whether cardiomyocyte-specific PDT can achieve complete or nearly complete electrical ablation in the left atrium. The end point was the observation of a large decrease in the electrogram amplitude in the region ablated, which is a commonly accepted criterion of ablation efficacy. Also, in vivo open-chest rat ablation experiments were used to ascertain cardiomyocyte targeting efficacy. We conducted a confocal microscopy analysis of cell-specific antibody colocalization, and these experiments were repeated with free Ce6 to illustrate unspecific cell damage, and with laser only to demonstrate that laser alone does not damage in vivo cells.

We further examined cell-specific ablation efficacy ex vivo in isolated Langendorff-perfused rat and sheep hearts. The goal was to demonstrate that an electrical block occurs rapidly after targeted PDT ablation but that this block does not occur after laser-only illumination. Such experiments were repeated with laser only (no CTP-Ce6-PEG injected) to show that laser alone does not result in electrical ablation. These experiments were not performed with free Ce6 because it is known that free photosensitizer PDT ablation produces an electrical block (2224). Therefore, the point of comparison for such experiments is CTP-Ce6-PEG versus laser only.

Last, in vivo cell-specific PDT experiments were conducted to assess postablation 3-day survival and ablation lesion stability, in other words, whether the electrical block generated by cardiomyocyte-specific PDT ablation is still present 3 days after ablation. The study end points therefore were the 3-day postoperative survival and the recording of an electrical block with optical mapping techniques. We compared these results with the ones obtained after control laser-only illumination or after free-Ce6 nontargeted ablation. The postoperative care and histological analysis were done in a blinded fashion. The sample size is set so as to reach the statistical significance (P value <0.05 and confidence interval set at 95%). The animals (rats and sheep) used for the study were normal and healthy with similar body weights. They were randomly assigned to the targeted ablation or control group.

Preparation of the cardiac-targeted photosensitizer nanoparticle (CTP-Ce6-PEG)

The synthesis and characterization of CTP-Ce6-PEG nanoparticle are described in the Supplementary Materials and Methods.

Isolation of adult rat myocytes and fibroblasts

Adult cardiomyocytes from normal adult male rats (200 to 300 g) were isolated as described in (32, 33) and in the Supplementary Materials and Methods. In vitro imaging was performed using Nikon A1R confocal microscope (Nikon Instruments Inc.).

Open-chest rat model, protocol, and analysis

All procedures were approved by the University of Michigan Committee on Use and Care of Animals and complied with National Institutes of Health guidelines. We used male Sprague-Dawley rats (Charles River Laboratories International Inc.) weighing 200 to 250 g that were housed under conditions of controlled temperature and 12-hour light-dark cycle. The animals were anesthetized with ketamine (60 mg/kg IP) and xylazine (6 mg/kg), placed in dorsal recumbency, and then intubated and ventilated throughout the procedure with 100% humidified oxygen at 90 strokes/min and 10 ml/kg tidal volume. An extra dose of anesthetic (one-third of the initial dose of ketamine) was administered as needed at 45 min after the first dose to prolong anesthesia. Electrocardiography (ECG) limb electrodes were positioned and connected to an amplifier to record ECG continuously.

With the animal in dorsal recumbency, the skin on the ventral thorax was shaved and disinfected and was then incised laterally to the sternum near the third intercostal space. After incising the intercostal muscle, the left atrium was exposed and held with retractors. Then, the animal was allowed to stabilize for 15 min. A bipolar electrode was positioned on the LAA, and the amplitude was recorded. Then, CTP-Ce6-PEG was injected into the tail vein. Sixty minutes after CTP-Ce6-PEG injection, the laser beam (30 mW, 671 nm, 1-mm diameter, 2- to 3-mm epicardium laser guide tip distance) was directed toward the LAA, and changes in LAA electrogram amplitude were recorded. The value of laser therapeutic energy in the open-chest setup, with the goal of obtaining a complete left atrial ablation, was about 15 J. Control experiments were done in a similar fashion but with the injection of free Ce6 (nontargeted).

To register the ablated and nonablated regions for subsequent histological analysis, we used a 1-mm-diameter laser guide positioned in the vicinity of the left ventricular free wall in the apex region. The ablated region was registered as the region positioned immediately under the light guide tip. Then, we collected a tissue sample that included the ablated region as well as the most distant myocardial regions all the way to the basis of the left ventricular wall. After PI staining, the transition region between ablated and nonablated zones was identified as the region with a large decrease in PI intensity. For LAA amplitude analysis, bipolar recordings were filtered to eliminate alternating current power noise (60-Hz infinite impulse response band stop). LAA electrogram voltage peak-to-peak amplitude was measured after filtering (AcqKnowledge software, Biopac Systems Inc.).

Rat survival studies

The survival experiments were performed to monitor the postoperative outcome and the persistence of the postablation electrical block after targeted ablation in rats. Our hypothesis was that targeted ablation yields a lesser amount of postablation complications while allowing for the generation of a stable electrical block. In three additional groups of rats, we compared (i) CTP-Ce6-PEG photoablation (1.6 mg of Ce6 per rat; laser energy, 5 J; 0.03 W); (ii) nonspecific ablation (1.6 mg of free Ce6 per rat; laser energy, 5 J; 0.03 W); (iii) nonspecific ablation, half dose of Ce6 (0.8 mg of free Ce6 per rat; laser energy, 5 J; 0.03 W); and (iv) sham ablation (only laser illumination, no photosensitizer). After left thoracotomy, heart was exposed and laser therapeutic energy delivered was 5 J (0.03 W, 571 nm, 3 to 5 min) by positioning the laser guide placed on the anterior wall of the left ventricle. The ablated region was registered as the region positioned immediately under the light guide tip. The nonablated region was registered as the region away from the laser guide tip. Rats were then monitored postoperatively for 3 days (signs of distress and mortality). Persistence of electrical block in targeted ablation after 3 days was confirmed by optical mapping. Mortality rates and histological H&E staining of left anterior ventricular samples were compared between groups in a blinded fashion.

Langendorff perfusion and optical mapping methods

Rat and sheep hearts were excised and Langendorff-perfused as described previously (34).

Rats. Hearts were excised and Langendorff-perfused, and then, the voltage sensitive dye DI-4-ANEPPS (10 mg/ml, Life Technologies) and the excitation contraction uncoupler blebbistatin (10 μM) were injected into the perfusate. A single charge-coupled device camera (SciMeasure Analytical Systems Inc.) was directed toward the atrial appendage and the ventricles. The PV antrum (the PV ostial region) is typically targeted during an AF ablation procedure with the goal of obtaining electrical disconnection between the PV myocardium and the remainder of the atrial muscle. A 1-W, 532-nm green laser was used as excitation light, and movies were recorded (red filter, 2 s, 1000 frames/s). The value of laser therapeutic energy in the Langendorff experiments was about 2 J.

Movies were filtered and voltage intensity maps were obtained and color-coded as previously described (35) using in-house developed algorithms (PV-Wave). Because ANEPPS dyes may be toxic to tissues after illumination, optical mapping recordings were only obtained after PDT-induced arrhythmia termination (except for one rat heart, top panel in Fig. 5A) in which we recorded one 2-s movie about 15 min before termination. Control hearts, in which AF did not terminate after PDT (Fig. 5B), were also injected with DI4-ANEPPS.

Sheep. Optical movies were obtained as described for rats. After obtaining baseline pacing movies, CTP targeted nanoparticles were perfused. After 15 min, we delivered a total therapeutic energy of up to 10 J (30 mW, 1 to 5 min) to generate photoablation point lesions either on the right atria, the right ventricle epicardium, or the AV node region. The same hearts served both as laser-only control and as PDT. Nanoparticles were indeed perfused before PDT after the data obtained with laser-only illumination had been collected. We chose to perform PDT on the right atrial free wall because the right atrium is quite representative of the large variation in thickness present in the human atrium. We also evaluated our technology on the right ventricle because its thickness (up to ~10 mm) is representative of the largest atrial thickness values that may be encountered in the human atrium. Local temperature levels were monitored with a thermal imaging infrared camera (Fluke Ti32, Fluke Inc.) before, during, and after photoablation.

Human cardiomyocytes and fibroblast coculture

hiPSC-derived cardiomyocytes (iCell cardiac myocytes, Cellular Dynamics International) and human fibroblasts from a BJ cell line (Stemgent) were used for targeted and nontargeted PDT experiments. Details are presented in the Supplementary Materials and Methods.

Statistical analysis

All the statistics were performed using GraphPad Prism software, and data are presented as means ± SD. When comparing two data sets for significance, P values were calculated using paired Student’s t test with two-tailed distribution and confidence interval set at 95%. Postoperative survival curves (Fig. 6B) between targeted and nontargeted groups were compared using log-rank (Mantel-Cox) test. Statistical significance was set as P < 0.05. Nanoparticle colocalization (Fig. 2B) was quantified using Pearson correlation coefficient analysis with a built-in plug-in of confocal microscope. All the statistics data were reviewed by an independent statistician.



Fig. S1. Characteristics of CTP-Ce6-PEG nanoparticle.

Fig. S2. Nontargeted control experiments in vitro.

Fig. S3. Targeted and nontargeted PDT in a coculture of human cardiac myocytes and fibroblasts.

Fig. S4. Photoablated lesion depth analysis.

Fig. S5. Thermal imaging during PDT ablation.

Movie S1. Optical mapping of a rat heart 3 days after photoablation.

Movie S2. Optical mapping of the sheep right atrial free wall.

Movie S3. Optical mapping of the sheep right ventricle.

References (3641)


  1. Acknowledgments: We thank E. Rothman (Center for Statistical Consultation and Research, University of Michigan) for statistical consultation and guidance; M. Hoenerhoff (In-vivo Animal Core Pathology, University of Michigan) for interpreting the data from the postoperative rat histology tissue sections; R. Tang, K. Campbell, C. Willis, G. Guerrero-Serna, and G. Kim for their helpful suggestions; and J. Jalife and H. Valdivia for their support. Funding: Supported by National Heart, Lung, and Blood Institute grant R21HL111876 to J.K. and R.K. Author contributions: U.M.R.A., H.K.Y., C.H.L., and K.K. performed in vitro rat experiments. U.M.R.A., J.K., and H.K.Y. performed in vivo and ex vivo rat experiments. U.M.R.A., J.K., Y.T., R.J.R., and S.R.E. performed ex vivo sheep experiments. U.M.R.A., J.K., and R.K. designed the in vitro and in vivo rat experiments. U.M.R.A., J.K., O.B., and F.M. designed the ex vivo sheep experiments. U.M.R.A., J.K., H.K.Y., and R.K. wrote the manuscript. T.H. provided hiPSC-derived cardiomyocytes and fibroblasts, designed in vitro iPSC experiments, and interpreted the data. Competing interests: One patent (J.K., R.K., U.M.R.A., and H.K.Y.) has been filed concerning this work: Publication no. WO2013163187A1, International Application no. PCT/US2013/037807. Data and materials availability: All data and materials are available in this article.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article