Experimental approval

All experiments were performed following the guidelines and approved protocols for animal care and use (approval number for animal use: 20–147) of the animal experiment committee at Kawasaki Medical School.

Masu and cherry salmon

Twenty-one juvenile fish, six months post fertilization (mpf), were purchased from Azuma Yougyojou Y.K. (Gunma, Japan). They were maintained for one week in 8-L freshwater tanks, with 10 fish per tank. A total of 30 masu salmon (O. masou landlocked type) 29–30 mpf (25 fish) and 34 mpf (5 fish) were purchased from Utanogawa Yamame Yougyojou (Yamaguchi, Japan), were maintained in 70-L freshwater tanks (5 or 10 fish per tank), and used in experiments within 1 week of their arrival. Eleven cherry salmon (O. masou sea-run type) at 29–30 mpf were purchased from Wakao-Suisan Co., Ltd. (Hyogo, Japan), maintained in 200-L seawater tanks with a salinity of 35 g/L (1.023 specific gravity) [73], at two or three fish per tank. For the cherry salmon, in vivo experiments were performed within two days of their transport. All fish tanks were maintained at 10°C with a 14-/10-h light/dark cycle, fed a commercial diet (Remix, Meito Suien Co., Ltd., Aichi, Japan, Cat# M-450). Fish were randomly selected for this study as sex could not be specified at the time of purchase.

Echocardiography and electrocardiography

Echocardiography in pulsed-wave Doppler mode and electrocardiography were recorded using an Aplio 300 system (Toshiba Medical System Corp., Tochigi Japan) with a 14-MHz transducer and three electrodes. Mild anesthetic induction by MS-222 (50 mg/L, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) [74] was used to treat five masu salmon and two cherry salmon until there was minimal gill movement. The anesthetized fish were placed upside down in the tank. The ultrasound transducer probe was placed on their heart positions, and the detection depth and brightness were adjusted for each fish (S2A Fig). The waveforms of atrioventricular inflows were observed immediately after ventricular outflow recordings because it is impossible to measure these flows simultaneously with a single probe. From the pulse-wave Doppler images, ventricular outflow times and velocities were calculated using Fiji software version 2.3.0 [75]. Intervals between the end of the ventricular outflow and the onset of the first atrioventricular inflow were calculated by subtracting the average time of ventricular outflow from each time from the peak of the R wave to the onset of atrioventricular inflow. The changes in the potential across the body surface by electrical excitation of the heart were measured by clipping electrodes to their pectoral and pelvic fins (S2A Fig). Aeration was performed during the measurements. After recording, all fish were euthanized using 250 mg/L MS-222 for organ sampling.

In vivo atrial and ventricular pressure measurements

Five masu salmon 34 mpf were anesthetized using 50 mg/L MS-222 and were placed upside down in the tank. A Single Transducer Set, DTX Plus DT-4812 (Argon Medical Devices, Inc., Plano, TX, USA), with needles was used to measure the pressure. Using an echocardiographic live image, a 25-G needle was inserted from the ventral side into the ventricle and a 27-G needle was inserted under the operculum into the atrium. The ventricular pressure tracing data were used to mediate the pressure recording interface (Power Lab 8/30, ADInstruments Pty Ltd., Dunedin, New Zealand) and were digitized at 200 Hz using LabChart 7 software (ADInstruments) and recorded. After recording, all fish were euthanized using 250 mg/L MS-222.

Ex vivo ventricular pressure–volume analysis

Four masu salmon and three cherry salmon were euthanized as described above, after which their hearts were removed. A catheter was inserted into the ventricles through the atrioventricular valve and fixed by ligation. Fluids in the ventricular lumen were washed using 10 U/mL heparin FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) and 10 mM 2,3-butanedione monoxime (BDM; FUJIFILM Wako Pure Chemical Corp.) in Ca²⁺- and Mg²⁺-free phosphate-buffered saline (PBS; Takara Bio Inc., Kusatsu, Japan). The aortic valve was then ligated. To measure the pressure–volume relationship, saline solution was injected into the ventricles using an infusion pump (Terumo Corp., Tokyo, Japan). The ventricular pressure data were recorded at 400 Hz with a multielectrode conductance catheter using the LabChart 7 software. The following exponential equation was fit to the EDPVR data: (1) where A, B, and C are constants that describe the ventricular exponential pressure-normalized volume property; constant A indicates the intercept on the pressure axis (A = -B), and C the index of the ventricular stiffness. V is the ventricular volume [4]. When comparing the ventricular stiffness in hearts of different sizes, V was normalized by the ventricular mass to calculate the stiffness per unit of ventricular mass and described by the following equation: (2) where v_t was the total volume of saline infused into the ventricle at time t, v₀ = 0, and m was the ventricular mass [15].

Tissue staining

Masu and cherry salmon hearts were fixed in 4% paraformaldehyde (PFA; Sigma-Aldrich, Merck) at 4°C for 48 h and embedded in paraffin blocks. The mounted glass slides with 5-μm-thick heart sections were immersed in Clear Plus (FALMA Corp., Tokyo, Japan) followed by 100%, 95%, and 70% ethanol solutions for deparaffinization, and finally rinsed with water.

Elastica van Gieson staining

The heart sections were immersed in Maeda’s resorcinol fuchsin stain solution (Muto Pure Chemicals Co., Ltd.) for 1 h and washed with 100% ethanol. Heart sections (5-μm thick) were mounted on glass slides and immersed in Weigert’s iron hematoxylin solution and washed with Milli-Q water. The specimens were then immersed in Picro–Sirius Red solution (ScyTek Laboratories, Inc., West Logan, UT, USA) for 15 min, air-dried, and rinsed with xylene. After staining, the tissue sections were dehydrated in 70%, 95%, and 100% ethanol solutions and ethanol/xylene mixed solution, immersed in xylene four times, and sealed using a mounting medium (Muto Pure Chemicals Co., Ltd., Tokyo, Japan). The sections were observed under a bright-field microscope BX53, (Olympus, Tokyo, Japan) using the cellSens software version 1.18 (Olympus) and BZ-X700 microscope (Keyence, Osaka, Japan). The collagen fiber content in the compact layer and cardiomyocyte density were measured using the Fiji software version 2.3.0 [75].

Hematoxylin and eosin staining

The heart sections were immersed in Hansen’s hematoxylin solution (Sigma-Aldrich, Merck) for 5 min, followed by washing with water. To define the nuclei, sections were immersed in 0.2% hydrochloric acid–70% ethanol solution for 1 min, and then immersed in eosin Y solution (Millipore, Merck) for 30 s. After staining, the tissue sections were dehydrated in 70%, 95%, and 100% ethanol solutions and ethanol/xylene mixed solution, immersed in xylene four times, and sealed using a mounting medium. The sections were observed under a bright-field microscope BX53, using the cellSens software version 1.18 and BZ-X700 microscope. The collagen fiber content in the compact layer and cardiomyocyte density were measured using the Fiji software version 2.3.0 [75].

Cardiomyocyte isolation and primary cultures

For isolation and primary cultures of the cardiomyocytes, we used a previously described protocol for zebrafish [76]. Two masu salmon and two cherry salmon 29–30 mpf were euthanized, their ventricles collected, and their lumens washed by injecting heparin buffer (10 U/mL heparin in PBS). The ventricles were dissected into 3-mm² columns and shaken at 800 rpm for 90–120 min at 30°C with 750 μL digestion buffer (12.5 μM CaCl₂, 5 mg/mL collagenase type II [Thermo Fisher Scientific Inc., Waltham, MA, USA], 5 mg/mL collagenase type IV [Thermo Fisher Scientific Inc.], 10 mM HEPES, 30 mM taurine, 5.5 mM glucose, and 10 mM BDM). Next, the cell suspension was filtered through a 100-μm nylon cell strainer, and 1 mL CaCl₂ buffer (12.5 μM CaCl₂, 10% fetal bovine serum [FBS; HyClone, Cytiva, Marlborough, MA, USA, Lot#15N353], 10 mM HEPES, 30 mM taurine, 5.5 mM glucose, and 10 mM BDM) was added and the cell suspension centrifuged at 200 × g for 5 min at 4°C. After discarding the supernatant, 1 mL of CaCl₂ buffer was added and the mixture was centrifuged at 200 × g for 5 min at 4°C. Isolated cells were resuspended using the culture media (Eagle’s minimum essential medium with 4.5 g/L glucose, 2 mM L-glutamine, 1 mM sodium pyruvate, and phenol red [FUJIFILM Wako Pure Chemical Corp.], 5 mM BDM, 5% FBS, and 100 U/mL penicillin-streptomycin [Thermo Fisher Scientific Inc.]) and seeded on a glass-bottom dish (Iwaki Co. Ltd., Tokyo, Japan) with 0.01% poly-L-lysine (Millipore, Merck) coating. Primary cells were cultured at 25°C under 5% CO₂. Twenty-four hours after incubation, the primary cells were observed using a BZ-X700 fluorescence microscope. The longitudinal length and area of the cardiomyocytes were measured using the Fiji software version 2.3.0 [75].

Immunostaining

Primary cardiomyocytes were fixed in 4% PFA at 4°C for 15 min. They were then permeabilized with 0.02% Triton X-100 in PBS for 15 min and blocked with 5% bovine serum albumin (Sigma-Aldrich, Merck) in PBS for 3 h. These cells were incubated with primary antibody, monoclonal anti-α-actinin (Sarcomeric) antibody produced in mouse (1:1000, Sigma-Aldrich, Merck, clone EA-53, Cat# A7811), at 4°C overnight, and subsequently stained with a secondary antibody, polyclonal goat anti-mouse IgG (H+L) cross-adsorbed secondary antibody, Alexa Fluor 488 (1:1000, Thermo Fisher Scientific Inc., Cat# A-11001), and Hoechst33342 (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan, Cat# H342) for 1 h at 25°C. The sarcomeric structures and the number of nuclei in the cardiomyocytes were then examined using an FV1000 confocal laser scanning microscope system mounted on IX81 (Olympus).

Sodium dodecyl sulfate-agarose gel electrophoresis and western blotting

The atria, ventricles, and bulbus arteriosus were separated from the hearts of the masu and cherry salmon 29–30 mpf. Each sample was homogenized using Polytron (Kinematica AG, Malters, Switzerland) in a sample buffer (8 M urea, 2 M thiourea, 3% sodium dodecyl sulfate [SDS], 75 mM dithiothreitol, 0.03% bromophenol blue, 0.05 M Tris-HCl [pH 6.8], and protease inhibitor cocktail [Thermo Fisher Scientific Inc.]), heated at 65°C for 10 min, centrifuged at 16,000 × g for 5 min at 4°C, and the supernatant was collected. Cleared lysates were loaded into wells of the SDS-agarose gels (1% SeaKem Gold Agarose [Lonza, Basel, Switzerland], 30% glycerol, and 1 × Tris/glycine/SDS [TG-SDS] buffer [Takara Bio Inc.]) and then electrophoresed in a running buffer (1 × TG-SDS buffer and 10 mM 2-mercaptoethanol) at 0.01 A for 90 min [77]. To avoid the gel from sliding off the gel plate during manipulation, the SDS-agarose gel was stacked and fixed on a 1-cm-high acrylamide plug gel (12% acrylamide, 10% glycerol, 0.5 M Tris-HCl [pH 9.0], 0.0015% N,N,N′,N′-tetramethylethylenediamine, and 0.056% ammonium persulfate). The patterns of the band peaks on the Coomassie Brilliant Blue (CBB)-stained gels were scanned using the Gel Analyzer plugin of Fiji software version 2.3.0.

For western blotting, the electrophoresed proteins were transferred from SDS-agarose gel to a nitrocellulose membrane (Bio-Rad Laboratories, Inc., Hercules, CA, USA) with a semi-dry western blot system (Pierce Power Blotter, Thermo Fisher Scientific Inc.) at 25 V for 10 min. After blocking the membrane with Tris buffer saline containing 0.05% Tween 20 and 5% skim milk, the Connectin proteins in the heart tissue were identified using a polyclonal primary antibody, the C-terminus of chicken Connectin (1:1000, Pc72C) [78], and polyclonal goat anti-rabbit immunoglobulin/HRP (1:1000, DAKO Agilent Technologies, Inc., Palo Alto, CA, USA, Cat# P044801-2) as a secondary antibody. Connectin blots were detected using the chemiluminescence reagent Western Lightning ECL Pro (PerkinElmer Co., Ltd., Waltham, MA, USA) and acquired using ImageQuant LAS 4000 (GE Healthcare, Chicago, IL, USA).

Transmission electron microscopy

To observe the sarcomere structures in the cardiomyocytes using a transmission electron microscope (TEM), the fish hearts at 29–30 mpf were washed with 10 U/mL heparin and 10 mM BDM in PBS and fixed with 4% PFA and 2.5% glutaraldehyde in PBS for 1 h at 25°C. The fixatives were thoroughly washed with PBS. The samples were then post-fixed in 1% osmium tetroxide in PBS and immersed in 50%, 60%, 70%, 80%, 90%, 95%, 99%, and 100% ethanol consecutively for dehydration. Next, the samples were passed through propylene oxide, embedded in epoxy resin, and polymerized in an incubator for 72 h at 60°C. Sections that were 500 nm thick were produced from the resin block using an ultramicrotome and collected on glass slides. After reaching the desired section, 70-nm ultrathin sections were cut and placed on a single-hole grid coated with Formvar film. The grids were air-dried and stained with uranyl acetate and lead citrate. Electron micrographs were recorded using TEM JEM-1400 (JEOL Ltd., Tokyo, Japan), operated at 80 kV.

Statistical analysis

The experimental results are presented as the mean ± standard deviation (SD); the number of samples used in each analysis is indicated in each figure legend. P-values were calculated using Student’s t-test or one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test and are indicated on graphs and text. The threshold for significant difference was set at a p-value of ≤ 0.05. GraphPad Prism8 software version 8.4.3 (GraphPad Software, San Diego, CA, USA) was used to conduct all statistical analyses.

Evaluation of the ventricular hemodynamics

At 29–30 mpf, the cherry salmon were found to be approximately 2.4 times bigger and 17 times heavier than the masu salmon (Fig 1A and 1B; Table 1). We also investigated their hemodynamics at this developmental stage. Pulsed-wave Doppler echocardiography and electrocardiography were simultaneously performed to observe their atrioventricular inflow in sinus rhythms. To set the pulsed-wave Doppler gate window downstream of the atrioventricular valve, a sagittal long-axis view including the atrium, ventricle, and bulbus arteriosus was acquired (S2B and S2C Fig; S1 and S2 Movies). Electrocardiography revealed regular P waves, QRS complexes, and T waves (S2D Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Recordings of the waveforms of the atrioventricular inflows in masu and cherry salmon using pulsed-wave echocardiography.

(A, B) Representative Oncorhynchus masou masou pictures: (A) masu salmon at 29 months post fertilization, (B) cherry salmon reared in seawater for five months after being reared in freshwater for 24 months. Scale bars = 10 cm. (C, D, E) Representative velocity waveforms of atrioventricular inflows in the masu (C) and cherry salmon of monophasic (D) and biphasic patterns (E). Blood passing the atrioventricular valve was observed using pulsed-wave Doppler echocardiography simultaneously with electrocardiography. Green arrowheads indicate P waves. Black left–right double arrows indicate R-R intervals. a: atrial systolic velocity, e/a: fusion velocity of atrial systolic and early diastolic inflows, e: early diastolic velocity. (F) Peak wave velocity. ○: masu salmon atrial systolic velocity (31.3 ± 6.4 cm/s, N = 37 peaks from five fish), ▲: cherry salmon early diastolic waves and atrial systolic waves (e/a) fusion velocity (36.7 ± 3.6 cm/s, N = 12 peaks from two fish), ■: cherry salmon early diastolic velocity (16.5 ± 2.6 cm/s, N = 15 peaks from two fish), ●: cherry salmon atrial systolic velocity (34.7 ± 4.2 cm/s, N = 15 peaks from two fish). Lines and error bars indicate the mean ± standard deviation. Minimal data sets for atrioventricular velocities are found in S1 File.

https://doi.org/10.1371/journal.pone.0267264.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Body size and heart mass measurements (means ± standard deviation) for Oncorhynchus masou masou.

https://doi.org/10.1371/journal.pone.0267264.t001

Pulsed-wave Doppler echocardiography revealed that the atrioventricular inflow patterns in the masu salmon showed sequential single forward flow waveforms that were synchronized with the P waves, indicating atrial contraction (Fig 1C). The heart rate of masu salmon as recorded with echocardiography was 109 ± 12 beats/min (bpm, N = 5 fish) (S1 Table). In masu salmon #5, the time from the R wave to the end of ventricular outflow was 326.8 ± 1.0 ms (N = 4 beats) (S3A Fig). When the interval times for the onset of ventricular systole (R-R interval times) took 624 ms (short) and 958 ms (long), the time between the end of ventricular outflow and the appearance of atrioventricular inflow waveforms was estimated to be 160 ms and 396 ms, respectively (S3 Fig). Regardless of the duration of the R-R interval, the atrioventricular inflow waveforms showed a monophasic pattern; furthermore, the atrial pressure in the ventricular diastole of the masu salmon was the same as the ventricular pressure for a while after isovolumic relaxation and became higher than the ventricular pressure in the atrial systole (S4 Fig). Hence, the monophasic inflow waveforms in masu salmon were presumed to indicate atrial systolic velocity.

Interestingly, the forward flow waveforms began to appear in the atrioventricular inflow patterns of the cherry salmon, even before the P waves were recorded (Fig 1D). When the R-R interval took > 0.77 s, echography recorded clear biphasic waveforms (Fig 1E, S3 Fig). The low forward waveforms appeared before the P wave, and the high forward waveforms appeared after the P wave. The former and the latter were, therefore, presumed to be comparable to the early ventricular diastolic and atrial systolic velocities, respectively. In addition, the sequential monophasic waves in cherry salmon were likely to be the fusion of these two velocities (early ventricular diastolic and atrial systolic fusion velocity). The mean ± SD of the early diastolic velocity peak in the cherry salmon was 16.5 ± 2.6 cm/s, which was lower than the atrial systolic velocity (p < 0.0001, Fig 1F). The ratio of the early diastolic velocity/atrial systolic velocity was 0.48 ± 0.06 in the cherry salmon. In cherry salmon #1, the interval time from the R wave to the end of ventricular outflow was 352 ± 1 ms (N = 4 beats). Whether the R-R interval times were 720 ms (short), 772 ms (intermediate), or 908 ms (long), the times from the end of ventricular outflow until the appearance of atrioventricular inflow waveforms remained almost the same, at 92–96 ms (S3 Fig). Under MS-222 anesthesia, the heart rates of cherry salmon were 72 ± 5 bpm (N = 2 fish) and lower than that of masu salmon (p = 0.016, S1 Table).

Analysis of ventricular stiffness

To investigate ventricular stiffness, we performed the EDPVR analysis ex vivo. The interventricular pressure of masu salmon reached 20 mmHg in response to injection with 0.27–0.31 mL saline (S4A Fig), while that of cherry salmon reached ~20 mmHg in response to injection with 1.13–1.91 mL saline, and the smaller ventricles tended to show steeper EDPVR curves (S4 Fig). However, the normalized EDPVR curves of the cherry salmon shifted to the left were steeper compared to those of the masu salmon (Fig 2A). The indices of ventricular stiffness were 1.04 ± 0.19 in masu salmon and 2.21 ± 0.74 in cherry salmon, and the ventricles of cherry salmon were significantly stiffer than those of masu salmon based on the measurement of ventricular stiffness per unit myocardial mass (p = 0.02, Fig 2B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Relative ventricular stiffness of masu and cherry salmon.

(A) The normalized end-diastolic pressure–volume relationship (EDPVR) at 30 mpf for masu and cherry salmon. The horizontal axis shows the ventricular volume normalized by the mass of the masu (N = 4) and cherry salmon ventricles (N = 3). The vertical axis shows the ventricular pressure. Ventricular mass is shown in S5B Fig. (B) Scores for ventricular stiffness in the masu and cherry salmon. The results of the EDPVR (A) were appended in Eq (1) to obtain the exponential C. Black horizontal lines in (B) indicate the means. ○: masu salmon, ●: cherry salmon.

https://doi.org/10.1371/journal.pone.0267264.g002

Morphological and histological analysis

At 29–30 mpf, the ventricles of both fish displayed a pyramidal shape (Fig 3A and 3B). The heart mass of cherry salmon was approximately 11 times greater than that of masu salmon, although the heart mass relative to body weight was significantly greater for masu salmon than for cherry salmon (p = 0.006, Table 1). The juvenile ventricles at 6 mpf also exhibited pyramidal shapes (S1D and S1E Fig). The Elastica van Gieson-stained and hematoxylin and eosin-stained sections of masu and cherry salmon hearts (Fig 3C–3J and S6 Fig) showed that their ventricles had at least two layers of myocardium, i.e., a compact layer and a spongy layer (Fig 3C, 3D, 3G and 3H). The compact layer of cherry salmon had more coronary vessels compared with that in masu salmon (Fig 3D and 3H). Further, this layer was thicker in cherry salmon than that in masu salmon (Fig 3D and 3H), and the ratios of the compact layer area to the total ventricular area on sagittal heart slices were significantly different between masu (29.0% ± 5.1%) and cherry salmon (39.6% ± 2.5%; p = 0.039, Fig 3K). The cell density in the compact layer was 6297 ± 762 and 6734 ± 1021 nuclei/mm² in masu and cherry salmon, respectively, and there was no significant difference between masu and cherry salmon (p = 0.318, Fig 3L).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Histological analysis of masu and cherry salmon hearts.

(A, B) Representative images of masu (A) and cherry (B) salmon hearts 29 mpf. At: atrium, V: ventricle, Ba: bulbus arteriosus. (C–J) Elastica van Gieson staining images of the sagittal sections of the hearts of masu and cherry salmon 29 mpf. (C, G) Images of the whole hearts. (D, H) Magnified images of the ventricles. Black double arrows indicate the thickness of the compact layer. Black arrowheads indicate coronary vessels. (E, I) Higher magnification images of the compact layers. (F, J) Higher magnification images of the spongy layers. Collagen fibers: bright red; cytoplasm: yellow; elastic fibers: purple-black; nuclei: dark black. Cl: compact layer, Sl: spongy layer. Scale bars = 1 cm in (A) and (B); 5 mm in (C) and (G); 1 mm in (D) and (H); 100 μm in (E), (I), (F), and (J). (K) Percentage compact layer in the sagittal ventricle sections. (L) The number of nuclei per yellow area in the compact layer. (M) The collagen fiber area percentage in the compact layer. Bar graphs and error bars indicate the means and standard deviations, respectively. ○: masu salmon, ●: cherry salmon. Minimal data sets of (K–M) are found in S3 File.

https://doi.org/10.1371/journal.pone.0267264.g003

The extracellular matrix is one of the parameters influencing tissue compliance. For example, elastin is an extracellular matrix that provides tissues with high levels of elasticity [79]. Bulbus arteriosus is an elastin-rich tissue [80] and stains purple–black with Elastica van Gieson staining (Fig 3C and 3G). The collagen fibers were tinted an intense red with Elastica van Gieson staining. Investigation of the heart sections revealed that the compact and spongy layers included collagen fibers (Fig 3E, 3F, 3I and 3J). The epicardium, the area around the coronary vessels, and the contact areas of the compact and spongy layers were found to be rich in collagen fibers (Fig 3D and 3H). The collagen fibers occupied 1.5% ± 0.9% of the compact layer in the ventricles from masu salmon and 5.8% ± 2.7% of the compact layer in the ventricles from cherry salmon, and the latter exhibited a significant accumulation of collagen fibers (p = 0.0003, Fig 3M).

To investigate whether cell hypertrophy is involved in heart growth, the cardiomyocytes of masu and cherry salmon were isolated, and their morphologies were analyzed. Most of the isolated cells that were attached to the dish showed spindle or rectangular shapes (Fig 4A and 4B). These cells were stained in a band pattern using the cardiomyocyte marker α-actinin 2 (S7A–S7D Fig). Smaller and round-shaped cells were negative for α-actinin 2 (S7E and S7F Fig). The longitudinal length and area of the cultured primary cardiomyocytes were not significantly different between masu (56 ± 19 μm and 559 ± 239 μm²) and cherry salmon (58 ± 20 μm and 518 ± 197 μm²; p = 0.7 and p = 0.4; Fig 4C and 4D, respectively). In these primary culture experiments, we found one binucleated cardiomyocyte in cherry salmon, and the rest of the cardiomyocytes were mononucleated (S8 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Morphology and sizes of the isolated cardiomyocytes.

(A, B) Representative bright-field images of the primary cultured cardiomyocytes of masu and cherry salmon 29 mpf. Scale bars = 20 μm. (C) Length of the long axis of the isolated cardiomyocytes; masu salmon (56 ± 19 μm, N = 55), cherry salmon (58 ± 20 μm, N = 50). (D) Area of the isolated cardiomyocytes; masu salmon (559 ± 239 μm², N = 55), cherry salmon (518 ± 197 μm², N = 50). Lines and error bars indicate the mean ± standard deviation. ○: masu salmon, ●: cherry salmon. Actual measurements of cardiomyocyte morphology are shown in the S4 File.

https://doi.org/10.1371/journal.pone.0267264.g004

Connectin expression patterns

The expression patterns of connectin isoforms affect ventricular diastolic function [30, 35–37]; these patterns are regulated by thyroid hormone in mammals [81]. In O. masou, increased secretion of this hormone induces downstream migratory behavior and smoltification [69, 72]. From these assumptions, we hypothesized that masu and cherry salmon exhibit different expression patterns for cardiac Connectin isoforms. To investigate the cardiac Connectin expression patterns, SDS-agarose gel electrophoresis was performed (Fig 5A). Band pattern analysis of the CBB-stained gel revealed the presence of two peaks in the α-connectin (T1) zone, indicating expression of intact Connectin in the ventricles (Fig 5B). Furthermore, western blotting with the anti-Connectin antibody [78] confirmed that these bands corresponded to Connectin (S9 Fig). These results suggest that O. masou at 29–30 mpf expressed Connectin (~3,700 kDa) in their ventricles and that there was no difference in the expression pattern of Connectin between masu and cherry salmon.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Molecular weights of the Connectin isoforms and sarcomere lengths in the hearts of masu and cherry salmon.

(A) A representative CBB-stained gel image. Lanes 2–7 show the molecular weights of Connectin isoforms in the atrium, ventricle, and bulbus arteriosus of masu salmon (Ma) and cherry salmon (Ch). The following samples were used as molecular weight standards: a major Connectin isoform in zebrafish heart (Zf, 3500 kDa) in lane 8; Connectin N2A isoform from the skeletal muscles of mice (M1, 3700 kDa) in lanes 1 and 10; the Connectin N2B isoform expressed in the left ventricle of mice (M2, 3000 kDa) in lane 9. α-connectin (T1) and β-connectin (T2) indicate an intact Connectin and its degraded product, respectively. MHC indicates Myosin heavy chain (molecular weight ~220 kDa each), which was used as a loading control. (B) Electropherogram of lanes 2–7 of the CBB-stained gel in (A). Black arrowheads indicate the band peaks for the ventricles of masu and cherry salmon in the α-connectin zone. (C–F) Representative transmission electron microscopy images of the cross-sections and longitudinal sections in the ventricular sarcomeres of masu and cherry salmon 29 months post fertilization. Insets of (C, E) showing higher magnification images of the cross-sections of myofibril bundles. Scale bars = 500 nm in (C) and (E) and 1 μm in (D) and (F). (G) The ventricular sarcomere lengths of the z-line to z-line in the ventricles of masu salmon (2.06 ± 0.08 μm, N = 42) and cherry salmon (2.05 ± 0.11 μm, N = 90). A black double arrow indicates sarcomere length. Lines and error bars indicate the means ± standard deviations. ○: masu salmon, ●: cherry salmon. A raw data image of CBB-stained gel is found in S1 Raw images. Minimal data sets for (G) are found in S5 File.

https://doi.org/10.1371/journal.pone.0267264.g005

The sarcomere length of the ventricular myocardium did not vary significantly between the masu (2.06 ± 0.08 μm) and cherry salmon (2.05 ± 0.11 μm; p = 0.8, Fig 5C–5G). Similarly, the sarcomere length of the atrial myocardium also did not vary significantly between the masu (2.13 ± 0.16 μm) and cherry salmon (2.12 ± 0.11 μm; p = 0.7, S10A–S10E Fig). Moreover, microfilaments and collagen fibers were observed in the bulbus arteriosus of both the masu and cherry salmon (S10F–S10H Fig) [82].

Differing ventricular diastolic hemodynamics

Pulsed-wave Doppler echocardiography revealed that landlocked masu salmon and sea-run cherry salmon have different ventricular diastolic properties despite being the same age and species. In masu salmon, monophasic atrioventricular inflow waveforms were observed immediately after the P wave in the electrocardiograph (Fig 1C), and therefore, these inflows may be attributed to atrial contraction. Simultaneous measurement of atrial and ventricular pressures showed that the atrial-ventricular pressure gradient occurred during atrial systole, but not the early diastolic phase between completion of ventricular ejection and atrial systole (S4 Fig). Our results suggested that there was no ventricular suction of blood from the atrium in masu salmon. Conversely, cherry salmon mainly displayed two types of atrioventricular inflow patterns depending on the heart rate: (1) early ventricular diastolic and atrial systolic fusion and (2) biphasic waveform (Fig 1D and 1E, S4 Fig). The inflows observed before detecting the P wave indicated passive ventricular filling, suggesting that the ventricle in cherry salmon acquired the ability to suction blood from the atrium. However, the atrial systolic velocity was greater than the early diastolic velocity (Fig 1F, early diastolic velocity/atrial systolic velocity ratio = 0.48 ± 0.06); it is clear, therefore, that atrial contraction plays a dominant role in ventricular filling in cherry salmon, at least under anesthesia.

A negative correlation has been reported between body weight and resting heart rate in both mammals and birds [83, 84], but this rule does not always apply to fish [85]. However, in our study, at 29–30 mpf, cherry salmon was over 10 times heavier than masu salmon (Table 1), and the heart rate of masu salmon was higher than that of cherry salmon (S1 Table, p = 0.0164). A previous hemodynamic study of rainbow trout suggested that the heart rate is associated with the appearance of monophasic or biphasic atrioventricular inflow waveforms [86]. Pulsed-wave Doppler echocardiography of rainbow trout recorded biphasic waveforms in specimens with a low heart rate (44–69 bpm) [86]. In cherry salmon, monophasic atrioventricular inflow switched to biphasic when the R-R interval times extended, but the time duration between the end of ventricular ejection and the onset of new inflow was constant at 92−96 ms, independent of the R-R interval (S4 Fig). In comparison, the lower-heart-rate cherry salmon had longer ventricular diastole, and ventricular filling occurred before atrial contraction began. No biphasic inflow waveform was observed, in masu salmon, even though the time duration from ventricular ejection to the onset of atrioventricular inflow increased. Further, since juvenile Atlantic salmon of 15 g body weight showed biphasic waveforms [87], it is possible that our data derived from pulsed-wave Doppler echocardiography are species specific. Elastic reaction forces, muscle contraction, and differences in chamber size between the atria and ventricles have been considered as candidates for the driving forces of ventricular filling, but are not yet sufficiently understood [88, 89]. Our study revealed that the same species adapted to different environments show different ventricular filling systems during diastole. Thus, histological and material changes in the ventricles of cherry salmon contribute to generating the early diastolic ventricular inflow (Figs 2 and 3).

Relative ventricular stiffness indices

The mechanical properties of ventricles of different sizes cannot simply be evaluated by comparing the individual mechanical parameters [4]. To assess ventricular contraction, end-systolic pressure–volume analysis has been carried out by normalizing ventricular volume per myocardial mass with respect to its contraction indexes [90]. EDPVR analysis, based on a time-varying elastance model, is an important method to assess ventricular stiffness [4]; however, its application in comparing the stiffness of ventricles of different sizes is difficult. In a smaller ventricle, the ratio of the pressure increase relative to the volume increase is greater than for a larger one. Thus, the EDPVR curve for a smaller heart would be steeper than that for a larger ventricle, even with the same actual stiffness. To evaluate the relative ventricular stiffness of masu and cherry salmon, we calculated the stiffness per unit of myocardial mass by normalizing the ventricular end-diastolic volume by the ventricular mass [15]. The index of ventricular stiffness per unit myocardial mass of cherry salmon, calculated from the normalized EDPVR curve (Fig 2A), was significantly higher than that of masu salmon (Fig 2B), thus suggesting that the ventricle of cherry salmon has a relatively reduced ventricular diastolic compliance.

The ventricular walls of the cherry salmon were thicker, and their coronary vascular network was well developed compared to those of the masu salmon (Fig 3D and 3H). Coronary circulation presents hemodynamic characteristics that are not observed in other organs: during ventricular systole, arterial blood flow is blocked because of the pressure associated with contraction, while during diastole, blood flows only after the percutaneous pressure is reduced. Therefore, excessive stretching of the coronary vessels associated with ventricular dilation increases vascular resistance and decreases blood supply to myocardium. Thus, the stiffening of the ventricle in cherry salmon might support the coronary vessel shapes and set the upper threshold for stroke volume.

Collagen accumulation in the thick compact layer of cherry salmon

In this study, histological analysis provided the causes of ventricular stiffness. The ventricle of cherry salmon had a thicker compact layer compared to that of masu salmon (Fig 3D, 3H and 3K). This is consistent with previous reports on other salmonids wherein the compact layer thickens with growth [56, 57]. This histological change indicates that the growth ratio of the compact and sponge layers was different during the growth process of the masu and cherry salmon. Hypertrophy of the ventricular wall in response to hypertrophic cardiomyopathy is one of the factors impairing the diastolic function of ventricles [91, 92]. Importantly, hypertrophy of the ventricular tissue is sometimes accompanied by an increase in cardiomyocyte size [93]. Mammalian cardiomyocytes become larger and maturate via multinucleation and polyploidization after birth [94, 95]. However, there were no differences in the morphology and nuclear number of each isolated cardiomyocyte between the masu and cherry salmon (Fig 4; S7 and S8 Figs). Therefore, O. masou hearts may show growth because of proliferation of cardiomyocytes or the differentiation of progenitor and stem cells instead of cardiomyocyte hypertrophy.

The compact layer of cherry salmon also contained more collagen fibers as compared with masu salmon (Fig 3E, 3I and 3M). Increased stiffness in the ventricle correlates with tissue fibrosis [66, 96], and is reportedly induced by collagen accumulation in the myocardium [97]. In a study involving rainbow trout, breeding at lower temperatures reduced ventricular compliance mediated by accumulating collagen fibers, but not because of hypertrophy of the compact layer [65]. Hence, the stiffening of the ventricle in cherry salmon is assumed to be strongly correlated to increasing collagen fiber content in the compact layer.

The bulbus arteriosus of teleost fishes contain abundant extracellular matrixes and smooth muscle cells [98–100]. Bulbus arteriosus of masu and cherry salmon appear to contain elastin, an elastic protein, stained purple-black by Elastica van Gieson staining (Fig 3C and 3G). They also had collagen fiber (S10F–S10H Fig). The superior extensibility of the vessel provided by these extracellular matrices contributes to maintaining the blood pressure of blood ejected from the ventricles [98–100].

Similar expression patterns of Connectin isoforms

The expression ratios of the Connectin isoforms affect cell passive tension [30, 35–37]. The mammalian heart expresses the lower-molecular-weight Connectin isoform N2B and the higher-molecular-weight Connectin isoform N2BA [31–33]. Rainbow trout express N2B-like and N2BA-like proteins in their hearts [101]. Multiple connectin splicing variants are reportedly expressed in zebrafish [102]. SDS-agarose gel electrophoresis showed the expression of two types of Connectins in the ventricles of the masu and cherry salmon (Fig 5A and 5B). Western blotting could not, however, discriminate between these Connectins because either of the isoforms might not have been detected by the antibody (S10 Fig). The similarity in the expression pattern of Connectins in the ventricles of cherry and masu salmon indicated that the difference in ventricular stiffness was not because of regulation by splicing of connectin molecules. There was no difference in sarcomere length under unloaded conditions as assessed by TEM (Fig 5C–5G). However, the correlation between the molecular sizes of Connectins and sarcomere length is controversial [103, 104]. The regulation of tensile strength by Connectin phosphorylation is a possible mechanism, other than the expression of its isoforms [27], but we did not analyze it.

We used O. masou of the same age but did not consider sex because it could not be specified at the time of purchase. Differences between male and female wild-type zebrafish strains at 9 mpf have been reported [105]. Furthermore, the amount of connective tissues in the ventricles of rainbow trout tends to differ based on their sex [106]. We used non-strained O. masou with non-uniform genetic backgrounds in this study, but a rigorous experimental system with a molecular basis is needed to provide definite answers to questions regarding the processes of heart development, morphogenesis, functions, and gene expression patterns.