High-resolution reconstruction of the beating zebrafish heart

Journal name:
Nature Methods
Volume:
11,
Pages:
919–922
Year published:
DOI:
doi:10.1038/nmeth.3037
Received
Accepted
Published online

The heart′s continuous motion makes it difficult to capture high-resolution images of this organ in vivo. We developed tools based on high-speed selective plane illumination microscopy (SPIM), offering pristine views into the beating zebrafish heart. We captured three-dimensional cardiac dynamics with postacquisition synchronization of multiview movie stacks, obtained static high-resolution reconstructions by briefly stopping the heart with optogenetics and resolved nonperiodic phenomena by high-speed volume scanning with a liquid lens.

At a glance

Figures

  1. Cardiac imaging in SPIM.
    Figure 1: Cardiac imaging in SPIM.

    (a) Schematic drawing of a zebrafish embryo at the intersection of illumination and detection axes. (b) Schematic of movie-stack acquisition (one movie per plane). (c) Single-view synchronization via similarity. Top, schematic comparison of three movies. The cardiac cycles with minimal image difference are highlighted. Bottom, reconstructed heart of Tg(myl7:GFP) at 48 h.p.f. in front, side and top views, with heart anatomy schematic shown at lower right. A, atrium; V, ventricle. Scale bar, 30 μm. (df) Schematics of dual-view movie stacks with perpendicular recordings (d), optogenetic manipulation to stop the heart beat and acquire a z stack (e) and volume scanning with ETL (f).

  2. Dual-view synchronization improves reconstruction at adverse imaging angles.
    Figure 2: Dual-view synchronization improves reconstruction at adverse imaging angles.

    (a) Schematic drawing of two synthetic movie stacks of a computer-generated heart-tube model at a favorable imaging angle (left) and at an adverse imaging angle (right). (b) Single- and dual-view synchronized (sync) synthetic movie stacks. (c) Single- and dual-view synchronization of movie stacks from Tg(myl7:GFP) embryos at 30 h.p.f. Arrowheads indicate artifacts. Scale bars throughout figure, 30 μm.

  3. Reconstructed hearts at different developmental stages.
    Figure 3: Reconstructed hearts at different developmental stages.

    Embryos (Tg(myl7:DsRed, kdrl:GFP)) are positioned head up. Red, myocardium; cyan, vasculature. (a) Schematic drawing of a 30-h.p.f. embryo at 60°. (b) Synchronized movie stack in maximum projection. (c) Endocardium at various phases (with time indicated relatively by clock schematic) during the cardiac cycle. (d) Schematic drawing of a 48-h.p.f. embryo positioned ventrally. (e) 3D rendering of synchronized movie stack cut open in the atrium. (f) Stills of the beating heart in one plane of the atrium. Arrowheads mark proximity of myo- and endocardium. (g) Schematic drawing of endocardial folds (arrowheads). (h) Schematic drawing of a 72-h.p.f. embryo positioned ventrally. (i) 3D-rendered heart in three rotated views. (j) Schematic drawing of a 5-d.p.f. embryo at 30°. (k) Single plane of the heart at different phases in the cardiac cycle. Boxed area is shown enlarged in l. Arrowheads indicate the atrioventricular valve. (l) Trabeculae with arrows indicating variable thickness of the cardiac wall. (m) Bisected heart in a 3D rendering. Scale bars throughout figure, 30 μm.

  4. Optogenetics yields high-resolution images of the heart, and ETL-SPIM captures intracardiac blood flow.
    Figure 4: Optogenetics yields high-resolution images of the heart, and ETL-SPIM captures intracardiac blood flow.

    (a) Three frames from a synchronized movie stack of a 5-d.p.f. embryo (Tg(myl7:lifeact, myl7:Gal4, UAS:NpHR-mCherry)). (b) Same heart as in a, stopped via optogenetics and imaged with higher magnification and longer exposure time. Inset shows enlarged detail of sarcomeres (arrowheads). (c) Front and side views of a 48-h.p.f. embryo (Tg(myl7:GFP, gata1a:DsRed)) imaged with movie-stack synchronization (left) and ETL-SPIM (right), showing myocardium in cyan and red blood cells in red. Solid vertical lines indicate position of cross section. Scale bars throughout figure, 30 μm.

  5. Fixation introduces artifacts.
    Supplementary Fig. 1: Fixation introduces artifacts.

    (a) Anterio-ventral view of the heart of transgenic zebrafish embryos (Tg(myl7:GFP)) before and (b) after fixation. (c) Merged images after (d) registration. (e) Schematic of merged outlines of the beating heart at various phases in the cardiac cycle and the fixed heart. (f-h) Three individual data sets with outlines of the contraction phases in red lines and cyan area representing the fixed heart. Scale bar 30 µm.

  6. SPIM setup.
    Supplementary Fig. 2: SPIM setup.

    (a) Configuration of our SPIM setup for cardiac imaging with movie stacks. Illumination optics only partly shown. (b) Additional hardware for alternative methods for cardiac imaging including rotational motor for dual-view recording (i), an orange beam for optogenetical manipulation (ii), an ETL and a scanning mirror for volume scanning (iii), an LED for prospective gating (iv) and an additional illumination arm and camera for dual-plane recording (v). Not to scale.

  7. Heartbeat variability.
    Supplementary Fig. 3: Heartbeat variability.

    (a) Irregularities within consecutive heart beats in the same heart (Tg(myl7:GFP)). (b) Section of the heart at which data in (a) was taken. Lines indicate sites for kymographs in (c) through the atrium (A) and the ventricle (V). (d) Overlay of consecutive periods in the same heart with arrowheads indicating different wall motion. Scale bar 30 µm.

  8. Quality of synchronization depends on frame rate and movie length.
    Supplementary Fig. 4: Quality of synchronization depends on frame rate and movie length.

    (a) The quality of synchronization as a function of the camera frame rate. (b) Side view of reconstructed heart walls (Tg(myl7:GFP)) at different frame rates. Arrowheads indicate artifacts and the red dashes indicate the myocardium. Scale bar 10 µm. (c) Image difference as a function of the movie length for high (400 fps) and low frame rate (67 fps) data. Error bars represent standard error of the mean. * p-value < 0.01.

  9. Alternative approaches for cardiac imaging.
    Supplementary Fig. 5: Alternative approaches for cardiac imaging.

    (a) Principle of prospective gating with an external signal triggering image acquisition in each plane of the heart. (b) Principle of dual plane recording with two parallel light sheets.

  10. Phase delay introduces errors in single-view synchronization.
    Supplementary Fig. 6: Phase delay introduces errors in single-view synchronization.

    (a) Synthetic 4D heart tube model at 30°. (b) Image difference and shifts in synchronized synthetic movie stacks at various imaging angles. (c) Shifts in single-view synchronization of two synthetic movie stacks, one along and one orthogonal to the contraction propagation axis (imaging angle 0 and 90°, respectively). (d) Shifts in synchronization of two real perpendicular movie stacks.

  11. Optogenetics offers improved contrast and dynamic range.
    Supplementary Fig. 7: Optogenetics offers improved contrast and dynamic range.

    (a) Outlines of the heart during cardiac contraction and when stopped with optogenetics. (b) Shape of the heart in repetitive optogenetic acquisitions. (c) Intensity profile at various exposure times. (d) Images of the heart (Tg(myl7:H2B-GFP)) taken with different exposure times.

Videos

  1. Shape of the beating heart at various contraction phases.
    Video 1: Shape of the beating heart at various contraction phases.
    Image of a 48 h.p.f. Tg(myl7:GFP) embryo with segmented outlines (pink dotted lines) of the heart over a full cardiac cycle.
  2. Movie stack synchronization.
    Video 2: Movie stack synchronization.
    Three synchronized planes of a movie stack of a 48 h.p.f. Tg(myl7:GFP) embryo and the 3D reconstruction. Scale bar 30 μm.
  3. Single view synchronization yields almost isotropic resolution.
    Video 3: Single view synchronization yields almost isotropic resolution.
    Synchronized movie stack of a 48 h.p.f. Tg(myl7:GFP) embryo in maximum projected front view and a horizontal and vertical cut. Scale bar 30 μm.
  4. Dual view sync of synthetic movie stacks.
    Video 4: Dual view sync of synthetic movie stacks.
    Reconstructions from movie stacks of a synthetic heart tube model synchronized using single- or dual-view synchronization. Scale bar 30 μm.
  5. Dual view sync of real movie stacks.
    Video 5: Dual view sync of real movie stacks.
    Reconstructions from movie stacks of a 30 h.p.f. Tg(myl7:GFP) embryo synchronized using single- or dual-view synchronization. Scale bar 30 μm.
  6. Reconstruction of endo- and myocardium in a 30 h.p.f. embryo.
    Video 6: Reconstruction of endo- and myocardium in a 30 h.p.f. embryo.
    Maximum projections showing cardiac contraction in real speed and slow motion in a 30 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo. Scale bar 30 μm.
  7. Reconstruction of endo- and myocardium in a 48 h.p.f. embryo.
    Video 7: Reconstruction of endo- and myocardium in a 48 h.p.f. embryo.
    Volume rendering (left) and single slice (right) showing cardiac cycle in real speed and slow motion in a 48 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo. Scale bar 30 μm.
  8. Reconstruction of endo- and myocardium in a 72 h.p.f. embryo.
    Video 8: Reconstruction of endo- and myocardium in a 72 h.p.f. embryo.
    Volume rendering of three different views showing cardiac contractions in real speed and slow motion in a 72 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo.
  9. Reconstruction of endo- and myocardium in a 5 d.p.f. embryo.
    Video 9: Reconstruction of endo- and myocardium in a 5 d.p.f. embryo.
    Single slices in front and side view showing cardiac contractions in a 5 d.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo.
  10. Optogenetically stopped heart.
    Video 10: Optogenetically stopped heart.
    Volume rendering showing the optogenetically stopped heart of a 5 d.p.f. Tg(myl7:lifeactGFP, myl7:Gal4, UAS:NpHR-mCherry) embryo.
  11. 4D reconstruction of blood flow with single view synchronization.
    Video 11: 4D reconstruction of blood flow with single view synchronization.
    Front and side view of synchronized heart of a 2 d.p.f. Tg(gata1a:DsRed, myl7:GFP) embryo. Scale bar 30 μm.
  12. 4D reconstruction of blood flow with ETL-SPIM.
    Video 12: 4D reconstruction of blood flow with ETL-SPIM.
    Front and side view of synchronized heart of a 2 d.p.f. Tg(gata1a:DsRed, myl7:GFP) embryo. Scale bar 30 μm.
  13. Arrhythmic heart imaged with ETL-SPIM.
    Video 13: Arrhythmic heart imaged with ETL-SPIM.
    Maximum projection of heart in 55 h.p.f. Tg(myl7:GFP) embryo treated with terfenadine. Scale bar 30 μm.

References

  1. Bakkers, J. Cardiovasc. Res. 91, 279288 (2011).
  2. Chopra, S.S. et al. Circ. Res. 106, 13421350 (2010).
  3. Sehnert, A.J. et al. Nat. Genet. 31, 106110 (2002).
  4. Hove, J.R. et al. Nature 421, 172177 (2003).
  5. Auman, H.J., Coleman, H., Riley, H., Olale, F. & Tsai, H. PLoS Biol. 5, e53 (2007).
  6. Huisken, J., Swoger, J., Del Bene, F., Wittbrodt, J. & Stelzer, E.H.K. Science 305, 10071009 (2004).
  7. Huisken, J. & Stainier, D.Y.R. Opt. Lett. 32, 26082610 (2007).
  8. Kaufmann, A., Mickoleit, M., Weber, M. & Huisken, J. Development 139, 32423247 (2012).
  9. Scherz, P.J., Huisken, J., Sahai-Hernandez, P. & Stainier, D.Y.R. Development 135, 11791187 (2008).
  10. Liebling, M., Forouhar, A.S., Gharib, M., Fraser, S.E. & Dickinson, M.E. J. Biomed. Opt. 10, 054001 (2005).
  11. Staudt, D.W. et al. Development 141, 585593 (2014).
  12. Taylor, J.M. et al. J. Biomed. Opt. 16, 116021 (2011).
  13. Bhat, S., Larina, I., Larin, K., Dickinson, M.E. & Liebling, M. IEEE Trans. Med. Imaging 32, 578588 (2013).
  14. Verveer, P.J. et al. Nat. Methods 4, 311313 (2007).
  15. Preibisch, S., Saalfeld, S., Schindelin, J. & Tomancak, P. Nat. Methods 7, 418419 (2010).
  16. Barry, A. Anat. Rec. 102, 289298 (1948).
  17. Garita, B. et al. Am. J. Physiol. Heart Circ. Physiol. 300, H879H891 (2011).
  18. Liu, J. et al. Development 137, 38673875 (2010).
  19. Arrenberg, A.B., Stainier, D.Y.R., Baier, H. & Huisken, J. Science 330, 971974 (2010).
  20. Fahrbach, F.O., Voigt, F.F., Schmid, B., Helmchen, F. & Huisken, J. Opt. Express 21, 2101021026 (2013).
  21. Nüsslein-Volhard, C. & Dahm, R. Zebrafish (Oxford University Press, 2002).
  22. Traver, D. et al. Nat. Immunol. 4, 12381246 (2003).
  23. Jin, S.-W., Beis, D., Mitchell, T., Chen, J.-N. & Stainier, D.Y.R. Development 132, 51995209 (2005).
  24. Huang, C.J., Tu, C.T., Hsiao, C.D., Hsieh, F.J. & Tsai, H.-J. Dev. Dyn. 228, 3040 (2003).
  25. Chi, N.C. et al. Genes Dev. 22, 734739 (2008).
  26. Arrenberg, A.B., Del Bene, F. & Baier, H. Proc. Natl. Acad. Sci. USA 106, 1796817973 (2009).
  27. Langheinrich, U., Vacun, G. & Wagner, T. Toxicol. Appl. Pharmacol. 193, 370382 (2003).
  28. Schindelin, J. et al. Nat. Methods 9, 676682 (2012).

Download references

Author information

Affiliations

  1. Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany.

    • Michaela Mickoleit,
    • Benjamin Schmid,
    • Michael Weber,
    • Florian O Fahrbach,
    • Sonja Hombach &
    • Jan Huisken
  2. Max Planck Institute for Heart and Lung Research, Bad Nauheim, Germany.

    • Sven Reischauer
  3. Department of Biochemistry and Biophysics, University of California, San Francisco, San Francisco, California, USA.

    • Sven Reischauer
  4. Present address: Institute of Biochemistry, Genetics and Microbiology, University of Regensburg, Regensburg, Germany.

    • Sonja Hombach

Contributions

M.M. and M.W. designed and built the SPIM setup. M.M. developed the single-view synchronization routine, performed and analyzed all experiments and visualized the data. B.S. developed the dual-view synchronization algorithm as well as the synthetic-heart-tube model and wrote the software to operate the SPIM setup. M.W. built the hardware for optogenetic manipulation. M.W. and M.M. performed the optogenetic experiments. F.O.F. designed, built, programmed and operated the ETL-SPIM setup. S.H. made the Tg(myl7:Gal4) line, and S.R. made the Tg(myl7:lifeactGFP) line. J.H. designed and supervised the project. M.M. and J.H. wrote the manuscript with contributions by M.W., B.S. and F.O.F.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Author details

Supplementary information

Supplementary Figures

  1. Supplementary Figure 1: Fixation introduces artifacts. (327 KB)

    (a) Anterio-ventral view of the heart of transgenic zebrafish embryos (Tg(myl7:GFP)) before and (b) after fixation. (c) Merged images after (d) registration. (e) Schematic of merged outlines of the beating heart at various phases in the cardiac cycle and the fixed heart. (f-h) Three individual data sets with outlines of the contraction phases in red lines and cyan area representing the fixed heart. Scale bar 30 µm.

  2. Supplementary Figure 2: SPIM setup. (234 KB)

    (a) Configuration of our SPIM setup for cardiac imaging with movie stacks. Illumination optics only partly shown. (b) Additional hardware for alternative methods for cardiac imaging including rotational motor for dual-view recording (i), an orange beam for optogenetical manipulation (ii), an ETL and a scanning mirror for volume scanning (iii), an LED for prospective gating (iv) and an additional illumination arm and camera for dual-plane recording (v). Not to scale.

  3. Supplementary Figure 3: Heartbeat variability. (234 KB)

    (a) Irregularities within consecutive heart beats in the same heart (Tg(myl7:GFP)). (b) Section of the heart at which data in (a) was taken. Lines indicate sites for kymographs in (c) through the atrium (A) and the ventricle (V). (d) Overlay of consecutive periods in the same heart with arrowheads indicating different wall motion. Scale bar 30 µm.

  4. Supplementary Figure 4: Quality of synchronization depends on frame rate and movie length. (282 KB)

    (a) The quality of synchronization as a function of the camera frame rate. (b) Side view of reconstructed heart walls (Tg(myl7:GFP)) at different frame rates. Arrowheads indicate artifacts and the red dashes indicate the myocardium. Scale bar 10 µm. (c) Image difference as a function of the movie length for high (400 fps) and low frame rate (67 fps) data. Error bars represent standard error of the mean. * p-value < 0.01.

  5. Supplementary Figure 5: Alternative approaches for cardiac imaging. (101 KB)

    (a) Principle of prospective gating with an external signal triggering image acquisition in each plane of the heart. (b) Principle of dual plane recording with two parallel light sheets.

  6. Supplementary Figure 6: Phase delay introduces errors in single-view synchronization. (283 KB)

    (a) Synthetic 4D heart tube model at 30°. (b) Image difference and shifts in synchronized synthetic movie stacks at various imaging angles. (c) Shifts in single-view synchronization of two synthetic movie stacks, one along and one orthogonal to the contraction propagation axis (imaging angle 0 and 90°, respectively). (d) Shifts in synchronization of two real perpendicular movie stacks.

  7. Supplementary Figure 7: Optogenetics offers improved contrast and dynamic range. (207 KB)

    (a) Outlines of the heart during cardiac contraction and when stopped with optogenetics. (b) Shape of the heart in repetitive optogenetic acquisitions. (c) Intensity profile at various exposure times. (d) Images of the heart (Tg(myl7:H2B-GFP)) taken with different exposure times.

Video

  1. Video 1: Shape of the beating heart at various contraction phases. (2.11 MB, Download)
    Image of a 48 h.p.f. Tg(myl7:GFP) embryo with segmented outlines (pink dotted lines) of the heart over a full cardiac cycle.
  2. Video 2: Movie stack synchronization. (2.24 MB, Download)
    Three synchronized planes of a movie stack of a 48 h.p.f. Tg(myl7:GFP) embryo and the 3D reconstruction. Scale bar 30 μm.
  3. Video 3: Single view synchronization yields almost isotropic resolution. (2.07 MB, Download)
    Synchronized movie stack of a 48 h.p.f. Tg(myl7:GFP) embryo in maximum projected front view and a horizontal and vertical cut. Scale bar 30 μm.
  4. Video 4: Dual view sync of synthetic movie stacks. (4.75 MB, Download)
    Reconstructions from movie stacks of a synthetic heart tube model synchronized using single- or dual-view synchronization. Scale bar 30 μm.
  5. Video 5: Dual view sync of real movie stacks. (7.01 MB, Download)
    Reconstructions from movie stacks of a 30 h.p.f. Tg(myl7:GFP) embryo synchronized using single- or dual-view synchronization. Scale bar 30 μm.
  6. Video 6: Reconstruction of endo- and myocardium in a 30 h.p.f. embryo. (8.7 MB, Download)
    Maximum projections showing cardiac contraction in real speed and slow motion in a 30 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo. Scale bar 30 μm.
  7. Video 7: Reconstruction of endo- and myocardium in a 48 h.p.f. embryo. (4.18 MB, Download)
    Volume rendering (left) and single slice (right) showing cardiac cycle in real speed and slow motion in a 48 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo. Scale bar 30 μm.
  8. Video 8: Reconstruction of endo- and myocardium in a 72 h.p.f. embryo. (5.2 MB, Download)
    Volume rendering of three different views showing cardiac contractions in real speed and slow motion in a 72 h.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo.
  9. Video 9: Reconstruction of endo- and myocardium in a 5 d.p.f. embryo. (9.07 MB, Download)
    Single slices in front and side view showing cardiac contractions in a 5 d.p.f. Tg(myl7:DsRed, kdrl:GFP) embryo.
  10. Video 10: Optogenetically stopped heart. (3.28 MB, Download)
    Volume rendering showing the optogenetically stopped heart of a 5 d.p.f. Tg(myl7:lifeactGFP, myl7:Gal4, UAS:NpHR-mCherry) embryo.
  11. Video 11: 4D reconstruction of blood flow with single view synchronization. (2.75 MB, Download)
    Front and side view of synchronized heart of a 2 d.p.f. Tg(gata1a:DsRed, myl7:GFP) embryo. Scale bar 30 μm.
  12. Video 12: 4D reconstruction of blood flow with ETL-SPIM. (3.11 MB, Download)
    Front and side view of synchronized heart of a 2 d.p.f. Tg(gata1a:DsRed, myl7:GFP) embryo. Scale bar 30 μm.
  13. Video 13: Arrhythmic heart imaged with ETL-SPIM. (2.47 MB, Download)
    Maximum projection of heart in 55 h.p.f. Tg(myl7:GFP) embryo treated with terfenadine. Scale bar 30 μm.

PDF files

  1. Supplementary Text and Figures (3 MB)

    Supplementary Figures 1–7

Additional data