Electrotonic coupling of excitable and nonexcitable cells in the heart revealed by optogenetics

Abstract

Electrophysiological studies of excitable organs usually focus on action potential (AP)-generating cells, whereas nonexcitable cells are generally considered as barriers to electrical conduction. Whether nonexcitable cells may modulate excitable cell function or even contribute to AP conduction via direct electrotonic coupling to AP-generating cells is unresolved in the heart: such coupling is present in vitro, but conclusive evidence in situ is lacking. We used genetically encoded voltage-sensitive fluorescent protein 2.3 (VSFP2.3) to monitor transmembrane potential in either myocytes or nonmyocytes of murine hearts. We confirm that VSFP2.3 allows measurement of cell type-specific electrical activity. We show that VSFP2.3, expressed solely in nonmyocytes, can report cardiomyocyte AP-like signals at the border of healed cryoinjuries. Using EM-based tomographic reconstruction, we further discovered tunneling nanotube connections between myocytes and nonmyocytes in cardiac scar border tissue. Our results provide direct electrophysiological evidence of heterocellular electrotonic coupling in native myocardium and identify tunneling nanotubes as a possible substrate for electrical cell coupling that may be in addition to previously discovered connexins at sites of myocyte-nonmyocyte contact in the heart. These findings call for reevaluation of cardiac nonmyocyte roles in electrical connectivity of the heterocellular heart.

Keywords: cardiac; electrophysiology; fibroblast; genetically-encoded voltage indicator; heterocellular coupling.

Fig. 1.

VSFP2.3 is highly expressed in myocytes of αMHC;VSFP2.3^+;+ mice. (A–D) YFP fluorescence in (A and B) whole hearts and (C and D) histological sections of LV tissue immunolabeled with anti-YFP antibody showing (A) high and uniform VSFP2.3 expression localized at (C) the myocyte sarcolemma in αMHC;VSFP2.3^+;+ mice, with (B and D) no discernible expression in αMHC;VSFP2.3^+;− mice. (E) Representative immunohistochemical staining of VSFP2.3 (anti-YFP antibody; green), cell nuclei (DAPI; blue), and myocytes (antisarcomeric α-actinin antibody; red) in a histological section of LV tissue from an αMHC;VSFP2.3^+;+ mouse, with extensive colabeling showing high specificity of myocyte targeting.

Fig. 2.

VSFP2.3 reports electrical activity of myocytes in αMHC;VSFP2.3^+;+ hearts. (A) CFP (blue) and YFP (green) fluorescence (F; expressed as percentage change ΔF/F_o) simultaneously collected from the LV epicardium of an αMHC;VSFP2.3^+;+ mouse heart showing electrical activity of myocytes, with signal enhancement by YFP/CFP ratiometry (black; expressed as percentage ΔR/R_o). (B) Simultaneously collected YFP (green) (Fig. S1A and Movie S1) and voltage-sensitive fluorescent dye di-4-ANBDQPQ (red) (Fig. S1B and Movie S2) fluorescence in αMHC;VSFP2.3^+;+ mouse heart and their normalized comparison (F_n), illustrating the known slower de- and repolarization kinetics of VSFP2.3 compared to di-4-ANBDQPQ signals but otherwise, good AP detection fidelity.

Fig. S1.

Selected frames of (A) Movie S1 and (B) Movie S2 showing simultaneously collected (A) VSFP2.3 YFP and (B) voltage-sensitive fluorescent dye di-4-ANBDQPQ signals collected in an αMHC;VSFP2.3^+;+ mouse heart. Comparison shows the difference in kinetics between the two signals. (Scale bar: 100 μm.)

Fig. 3.

VSFP2.3 is highly expressed in nonmyocytes but not myocytes of WT1;VSFP2.3^+;+ mice. (A–G) YFP fluorescence in histological sections of LV tissue showing increased VSFP2.3 expression as detected by anti-YFP antibody in interstitial nonmyocytes of (A) cryoinjured vs. (C) sham-operated WT1;VSFP2.3^+;+ mice, lack of expression in WT1;VSFP2.3^+;− mice (B) with and (D) without cryoinjury, (E) increase in expression from remote tissue to the scar border and into the scar in injured hearts, and (F and G) evenly distributed expression across the ventricular wall in sham hearts. Endo, subendocardium; Epi, subepicardium; Mid, mid-myocardium; NS, not significant. *P < 0.05; **P < 0.001. (H) Trichrome-stained longitudinal section of a cryoinjured heart showing the extent of scar tissue (collagen is stained bluish green, and myocytes are stained pink). RV, right ventricle. (I–N) Immunohistochemical staining of VSFP2.3 (green), cell nuclei (blue), and (I and J) myocytes (red) or (K–N) nonmyocytes (red) in histological sections of LV tissue from WT1;VSFP2.3^+;+ mice, with colabeling showing lack of VSFP2.3 in myocytes of (I) cryoinjured and (J) sham-operated animals (a rare example of a VSFP2.3-expressing myocyte is shown in Inset) but high nonmyocyte-specific expression in the remote tissue of (K) cryoinjured and (L) sham hearts, which increases (M) in the scar border and (N) into the scar. Data are presented as mean ± SEM; n ≥ 26 per group.

Fig. 4.

VSFP2.3 reports electrical activity of nonmyocytes at the scar border in WT1;VSFP2.3^+;+ hearts. YFP fluorescence (F; expressed as percentage change ΔF/F_o) collected from remote, central scar, scar border, and adjacent tissue (both toward and away from the scar center) in a WT1;VSFP2.3^+;+ mouse heart showing AP-like electrical activity in nonmyocytes at the scar border (indicating electrotonic signal transmission from myocytes to nonmyocytes) (Fig. S2 and Movie S3) and with lower amplitude in adjacent tissue (suggesting reduced passively conducted signal amplitude toward the scar center and reduced heterocellular electrotonic coupling away from the scar). No rhythmic polarizations were seen in central scar or remote tissue. [The number of fluorescence oscillations in border vs. adjacent tissue differed due to nonsimultaneous recordings.]

Fig. S2.

Selected frames of Movie S3 showing VSFP2.3 YFP signal collected at the scar border in a WT1;VSFP2.3^+;+ mouse heart. The patchy, locally heterogeneous signals suggest a lack of contribution from VSFP2.3-expressing myocytes. (Scale bar: 100 μm.)

Fig. 5.

Tunneling nanotubes connecting nonmyocytes to myocytes at the scar border. (A) Thick section (275-nm) transmission electron micrograph showing extensive finger-like membrane protrusions (examples indicated by white arrowheads) extending from a nonmyocyte to a myocyte in the scar border zone. (B and C) Serial ET slices (separated by 20 nm in the z axis) reconstructed in the plane of membrane protrusions indicated by the (B) blue and (C) red boxes in A, including (C, row 5) a segmentation and 3D rendering representing nonmyocyte sarcolemma (green), myocyte surface sarcolemma (blue), and myocyte basement membrane (yellow). In addition to nonmyocyte-derived nanotubes that either (B) extend toward the myocyte basement membrane or (C) penetrate that to contact the myocyte sarcolemma, (D) occasionally both cell types possess sarcolemmal membrane extensions toward one other (Movie S4). M, myocyte; nonM, nonmyocyte.

Fig. S3.

CFP fluorescence collected from scar border tissue in a WT1;VSFP2.3^+;+ mouse heart at the same location as the YFP signal shown in Fig. 4. No decrease in fluorescence indicating AP-like activity was detected because of the weak CFP signal, but also, no increase in fluorescence similar to that of YFP that would indicate a contribution of motion to the recorded signals was detected.