Report

Optogenetic Control of Cardiac Function

See allHide authors and affiliations

Science  12 Nov 2010:
Vol. 330, Issue 6006, pp. 971-974
DOI: 10.1126/science.1195929

Abstract

The cardiac pacemaker controls the rhythmicity of heart contractions and can be substituted by a battery-operated device as a last resort. We created a genetically encoded, optically controlled pacemaker by expressing halorhodopsin and channelrhodopsin in zebrafish cardiomyocytes. Using patterned illumination in a selective plane illumination microscope, we located the pacemaker and simulated tachycardia, bradycardia, atrioventricular blocks, and cardiac arrest. The pacemaker converges to the sinoatrial region during development and comprises fewer than a dozen cells by the time the heart loops. Perturbation of the activity of these cells was entirely reversible, demonstrating the resilience of the endogenous pacemaker. Our studies combine optogenetics and light-sheet microscopy to reveal the emergence of organ function during development.

In mammals, the heart rate is controlled by a specialized group of cells in the sinoatrial (SA) node, which act as the primary pacemaker. Together with cells in the atrioventricular (AV) node (secondary pacemaker) and specialized cells in the ventricular walls, they constitute the cardiac conduction system (CCS) (1, 2). In nonmammalian vertebrates, a CCS with properties similar to those of the mammalian CCS has been inferred from imaging of voltage signals (3, 4) or Ca2+ transients (5); the origin of the conduction wave was located in a region apparently homologous to the SA node. However, the exact role of this presumptive pacemaker and the consequences of its inactivation remain unclear.

We combined optical tools and transgenic expression of light-gated ion channels (68) in zebrafish (913) to locate and control cardiac pacemaker cells. Halorhodopsin (NpHR) (6) and channelrhodopsin-2 H134R (ChR2) (6, 14) enable temporally precise (tens of milliseconds) and spatially confined (single cells) control. The light-gated pump NpHR-mCherry was expressed in zebrafish cardiomyocytes by means of the Gal4/UAS system (13, 15) (fig. S1A). The protein was locally activated with light patterns generated with a digital micromirror device (fig. S2), which enables synchronous, uniform illumination of arbitrary shapes, unlike laser scanning (16) or fiber optic (13) techniques. High-speed video recordings of the heart were generated with a multidirectional selective plane illumination microscope (mSPIM) (1719) (Fig. 1A and fig. S3).

Fig. 1

Automated mapping of the cardiac pacemaker using patterned illumination in an mSPIM. (A) Microscope layout. A sheet of laser light (488 nm or 561 nm) illuminates the embedded fish (S) through objective lens OL1 or OL2. A computer-generated pattern is reflected off the digital micromirror device (DMD) and imaged onto the sample (S) through the detection lens OL3. See fig. S3 for details. LED, light-emitting diode (transmission); CH, chamber; DC, dichroic mirror; TL, tube lens; F, filter; M, mirror; PL, photographic lens; L, light bulb; CCD, charge-coupled device (camera). (B) A 3-dpf zebrafish heart expressing NpHR-mCherry stopped beating when illuminated with orange light and recovered instantaneously afterward. A high-speed video of an optical section was obtained by low-intensity light-sheet illumination. The kymograph shows the motion of the heart wall along the highlighted line. V, ventricle; A, atrium. The animal was transgenic for Et(E1b:Gal4-VP16)s1101t, Tg(UAS:NpHR-mCherry)s1989t, and Tg(UAS:Kaede)s1999t. Dark coloration corresponds to high Kaede fluorescence (inverted image). See movie S1. (C) The whole heart was sequentially illuminated with overlapping squares (see movie S2). (D) False-color image of the observed atrial heart rate after illumination of 20 sampled areas (ventral view, with head pointing upward). In this 3-dpf fish, the pacemaker cells are located in the dorsal-right side of the inflow ring.

Illuminating the entire zebrafish heart at 3 days post-fertilization (dpf) with orange light instantaneously blocked contractions (Fig. 1B and movie S1), indicating that the strong hyperpolarization induced by the activated chloride pump NpHR perturbed the well-balanced interplay of ion channels in cardiomyocytes (fig. S1B) and prevented depolarizations. The heart recovered to its original beat rate instantaneously after cessation of illumination [in zebrafish, blood circulation is dispensable for survival up to 6 dpf (20)].

We observed that the sinoatrial region was more susceptible to NpHR manipulation than the working myocardium in the atrium and ventricle, most likely because the current densities across the cell membranes of CCS cells are smaller by a factor of >10 (21, 22). To locate the pacemakers in a nonbiased way, we sequentially illuminated small, overlapping regions with constant intensity (Fig. 1C and movie S2). Maps showing the heart rate for every illuminated area were generated (Fig. 1D and fig. S4) to identify regions in which NpHR activation induced cardiac arrest or ventricular arrhythmia. Pattern generation, data recording, and video analysis were automated and computer-controlled to enable fast and reproducible analysis of a large number of hearts (23).

To follow the maturation of the CCS over embryonic and larval development, we generated heart rate maps for five stages (1 dpf to 5 dpf; n = 3 to 15 hearts analyzed, mean 7.4 per stage) from different angles. In 1-dpf animals, the heart stopped beating when a large region at the venous pole of the heart was illuminated at medium light intensities (Fig. 2A). At higher light intensities, it was possible to block the heartbeat by illuminating different smaller patches of the venous pole (fig. S5A). This observation indicates that strong electrical coupling seemed to be present and suggests that at this stage the cells required to initiate the heartbeat (pacemaker cells) cover a large area at the venous pole.

Fig. 2

Localization of the areas sensitive to hyperpolarization at the inflow and atrioventricular canal (AVC) regions of 1- to 5-dpf hearts. (A to E) Schematic drawings of the pacemaker (red) and AVC (green) regions for the developmental stages indicated. SAR, sinoatrial region.

In 2-dpf animals, the pacemaker region was more confined to the sinoatrial ring (SAR; Fig. 2B and fig. S5B). Although no consistent bias was found, individual hearts were more easily arrested by illuminating the right part of the SAR (right-biased: 8 hearts, left-biased: 5 hearts). Illumination of large areas adjacent to the pacemaker region did not arrest the heart (movies S3 to S5), indicating that the silencing effect did not spread farther than one or two cell diameters outside the illuminated area.

The atrioventricular canal (AVC) is formed at 2 dpf (24). At this stage, AV blocks were induced with high illumination intensities at the AVC; from 3 to 5 dpf, lower light intensities were sufficient to block ventricular systoles. No bias was detected around the AVC perimeter, which suggests that all AVC cells contribute equally to the conduction of the beat from the atrium to the ventricle. It has been shown that upon ablation of the primary pacemaker, downstream pacemakers such as the AV node can take over to maintain heartbeat (25). In our experiments, beating was always blocked when the SAR was silenced; this result implies that pacemaker currents outside the SAR are not strong enough to drive the larval heart immediately after SAR silencing.

The pacemaker region became further defined at 3 dpf (Fig. 2C and fig. S5C). The cells required to initiate the heartbeat were confined to the dorsal right quadrant of the SAR in 10 of the 13 hearts investigated, as revealed by mapping the heart from different angles. A refinement of the pacemaker area during development has also been reported in other species, where it narrows to a small region as it migrates from the atria to the right sinus primordium (3). In amniotes, the pacemaker is generally situated on the right side, which matches our findings in zebrafish. Furthermore, the center of the region appeared to be located about 10 μm (one or two cell diameters) farther anterior than the neck of the SAR (Fig. 2D). Interestingly, the cells that make up the lips of the sinus adjacent to the SAR appear to have an intrinsic pacemaking capability and poor coupling to the SAR, because they continued to contract in the arrested heart (movie S6), and photosilencing their contractions did not arrest the heart. The number of cells required for beating varied between animals. The size of the illuminated area generally corresponded to 10 to 30 cells, but some hearts could be arrested by illuminating very few cells (3 cells, movies S7 and S8). At 4 and 5 dpf, the pacemaker cells remained at the same relative position (Fig. 2, D and E, and fig. S5, D and E).

The zebrafish heart can be manipulated very precisely to rapidly switch between a healthy and a diseased state. We generated prominent cardiac pathologies such as AV blocks, tachycardia, and bradycardia. When illuminating the AVC (Fig. 3A), we induced different AV blocks by varying the light intensity. At high intensities, 3:1 or 4:1 AV blocks could be induced (Fig. 3, B and C); medium intensities caused 2:1 AV blocks (movie S9). Notably, the 2:1 AV block was stable over a range of intensities (55 to 95% maximal light intensity; see vertical lines in Fig. 3C) and for prolonged periods of time (50 s; Fig. 3C). The induction of different AV blocks was reproducible in many different hearts (Fig. 3D), although the required light intensity varied between hearts, probably because of the varying expression levels of NpHR. In some hearts, a complete block of ventricular beats could be produced with high light intensities (movie S9).

Fig. 3

Optically induced atrioventricular arrhythmia, tachycardia, and bradycardia. (A to D) Ventricular blocks of different severity are induced by illuminating the AVC with well-defined intensities. (A) Schematic of the spatial and temporal pattern of illumination. The atrioventricular ring in NpHR-expressing hearts was illuminated with slowly decreasing intensity. (B) Kymograph and (C) AV ratio for different illumination intensities. High intensities induced a 3:1 block, medium intensities a 2:1 block. (D) Summary for 10 different hearts. See movie S9. (E) Schematic of the region and temporal pattern of illumination. The sinoatrial ring in ChR2 expressing hearts was illuminated with short pulses with varying frequency. (F) The heart rate (brown bars) relative to the driving frequency (black dots) during the course of the experiment. The heart rate followed the driving frequency within the indicated dynamic range (between Min and Max). Error bars represent SD. See movie S10. A, atrium; V, ventricle; AVC, atrioventricular canal; SAR, sinoatrial ring.

Periodic photoactivation of the SAR with ChR2 (Fig. 3E) was used to control the heart rate in 4-dpf animals within the frequency range 2.7 to 4.7 Hz (Fig. 3F; heart rate before perturbation was 3.3 Hz). Above 4.7 Hz, atrial contractions started to skip pulses and became variable, suggesting that the heart rate cannot be sustained at such high frequency. Interestingly, our method enabled a reduction in heart rate, a modulation that was not reported in a recent study using a different laser-based stimulation technique (26).

In vertebrate hearts, the conduction direction is determined by the beating rates of primary (faster) and downstream (slower) pacemakers. Retrograde excitation has been reported in the zebrafish heart (27), and we were interested to know whether it could be forced to beat regularly in the retrograde direction. We applied light pulses to a ventricular region close to the bulbus to rhythmically activate ChR2 and found that cardiac conduction could be reversed (movie S11 and fig. S6) for at least 30 consecutive heartbeats.

Our results show that a surprisingly small number of pacemaker cells is indispensable for heartbeat initiation. This makes the embryonic heart very vulnerable, as no compensating mechanism seems to be in place on the time scales observed. Moreover, we have shown that this area can be optically targeted; our photostimulation methods allowed us to optically control heart rate, reverse cardiac conduction, and induce disease-like states in a reversible manner. This work opens a new avenue for controlling hemodynamic forces during studies on epigenetic factors of heart formation (28) and blood vessel development (29).

Supporting Online Material

www.sciencemag.org/cgi/content/full/330/6006/971/DC1

Materials and Methods

Figs. S1 to S7

Movies S1 to S11

References and Notes

  1. See supporting material on Science Online.
  2. We thank T. Mikawa, R. Shaw, and A. Müssigbrodt for feedback and comments on the manuscript. Supported by NIH grant HL54737 (D.Y.R.S.), the Packard Foundation (D.Y.R.S.), NIH grant R01 NS053358 (H.B.), a Sandler Opportunity Award, the Byers Award for Basic Science (H.B.), and the NIH Nanomedicine Development Center “Optical Control of Biological Functions” (H.B.). J.H. was supported by a Human Frontier Science Program (HFSP) cross-disciplinary fellowship. A.B.A. was supported by a Boehringer Ingelheim Fonds (B.I.F.) and a Krevans fellowship. The optogenetic idea, transgenic lines, and original observation originated in H.B.’s lab; follow-up experiments were carried out in D.Y.R.S.’s lab as an equal collaboration between A.B.A. and J.H. All authors worked on the manuscript, which was drafted by A.B.A. and J.H.
View Abstract

Navigate This Article