A previous study from our laboratory has shown that a single targeted heavy ion irradiation (THIR; 15 Gy) to rabbit hearts increases connexin43 (Cx43) expression for 2 wk in association with an improvement of conduction, a decrease of the spatial inhomogeneity of repolarization, and a reduction of vulnerability to ventricular arrhythmias after myocardial infarction. This study investigated the time- and dose-dependent effects of THIR (5–15 Gy) on Cx43 expression in normal rabbit hearts (n = 45). Five rabbits without THIR were used as controls. A significant upregulation of Cx43 protein and mRNA in the ventricular myocardium was recognized by immunohistochemistry, Western blotting, and real-time PCR from 2 wk up to 1 yr after a single THIR at 15 Gy. THIR ≥10 Gy caused a significant dose-dependent increase of Cx43 protein and mRNA 2 wk after THIR. Anterior, lateral, and posterior free wall of the left ventricle, interventricular septum, and right ventricular free wall were affected similarly by THIR in terms of Cx43 upregulation. The radiation-induced increase of immunolabeled Cx43 was observed not only at the intercalated disk region but also at the lateral surface of ventricular myocytes. The increase of immunoreactive Cx43 protein was predominant in the membrane fraction insoluble in Triton X-100, that is the Cx43 in the sarcolemma. In vivo examinations of the rabbits 1 yr after THIR (15 Gy) revealed no significant changes in ECGs and echocardiograms (left ventricular dimensions, contractility, and diastolic function), indicating no apparent late radiation injury. A single application of THIR causes upregulation and altered cellular distribution of Cx43 in the ventricles lasting for at least 1 yr. This long-lasting remodeling effect on gap junctions may open the pathway to novel therapy against life threatening ventricular arrhythmias in structural heart disease.
- gap junctions
- long-term effects
- radiation injury
sudden cardiac death resulting from ventricular tachyarrhythmias is a major clinical problem in most industrialized countries (24). It has been demonstrated by large-scale clinical trials that implantable cardioverter-defribrillator (ICD) therapy is superior over any pharmacological therapy to prevent life-threatening ventricular arrhythmias (5, 20). The usefulness of ICD currently available is, however, limited by a number of drawbacks. Those include myocardial damage causing shock-induced arrhythmia, hemodynamic dysfunction, and psychological disorders decreasing the quality of life of the recipient (9, 18, 30, 32). Increasing economical burden to the society is also a serious problem of ICD therapy. Development of a fundamentally new antiarrhythmic modality is, therefore, a matter of great concern not only to cardiologists but also to society.
Gap junctions are intercellular channels that allow the exchange of ions, second-messenger molecules, and other metabolites between adjacent cells. In the heart, coordinated action potential propagation depends on intercellular current flow passing through gap junctions. The principal gap junctional protein expressed in the ventricle is connexin43 (Cx43) (11). It has been reported in many experimental and clinical studies that alterations in Cx43 organization, expression, and phosphorylation are involved in the creation of substrates for ventricular arrhythmias under a variety of pathological conditions such as ischemia, hypertrophy, and cardiomyopathy (17, 28). In a previous report, our laboratory demonstrated that targeted heavy ion irradiation (THIR; at 15 Gy) to rabbit hearts caused upregulation of Cx43 in the ventricles 2 wk thereafter (1). After myocardial infarction (MI), this Cx43 upregulation was associated with a reduction of ventricular vulnerability and a decrease of spatial heterogeneity of repolarization (1). These observations suggest that THIR could have an antiarrhythmic potential through an improvement of electrical coupling. The time dependence of Cx43 expression by THIR and the minimum irradiation energy for the Cx43 expression remain to be investigated.
In the present study, we investigated time- and dose dependence of Cx43 upregulation by THIR in normal rabbits. Possible late adverse effects of THIR were also examined. The results have revealed that a single application of THIR (≥10 Gy) causes Cx43 upregulation for at least 1 yr without serious tissue injury.
Animal handling followed the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1996), and all procedures were approved by the Animal Experimentation Ethics Committee of Tokai University.
Animal model and heavy ion radiation.
New Zealand white rabbits (n = 50) weighing 3.5–4.0 kg were used. Forty-five rabbits received targeted THIR. We used carbon-ion beams provided by the Heavy Ion Medical Accelerator in Chiba (HIMAC) at the National Institute of Radiological Sciences, Japan (31). In this experiment, carbon ions were accelerated to be 290 MeV/u and extracted from vacuum tube. Monoenergetic carbon beam was modified to produce a 6-cm spread-out Bragg-peak, the distal end of the beam was focused through the left anterior breast, and exposed to the anterolateral left ventricular (LV) free wall with irradiation dose of 5–15 Gy. The dose setting was based on a previous study from our laboratory (1). Five rabbits that did not undergo THIR were used as controls. The control animals were not age matched to cover the whole observation period after THIR (2 wk–12 mo).
Ventricular tissue sections (12-μm-thick slices) from the midmyocardial layer were fixed and embedded in paraffin. For single staining of Cx43, the sections were incubated with an anti-Cx43 mouse monoclonal antibody (Chemicon; 1:100) and were then treated with the secondary antibody (Alexa Fluor 488 conjugated anti-mouse IgG) (15). The immunolabeled sections were examined with a laser confocal microscope (LSM510; Version 2.02) and analyzed with CLSM macro program (Carl Zeiss). To quantify the immunopositive Cx43, 10 randomly selected fields (×50) of longitudinally sectioned tissue (2.2 × 2.2 mm) were analyzed to obtain averaged values. To separate cell types expressing Cx43, we visualized three different proteins: the gap junctional protein Cx43; the myocyte marker α-actin, which is expressed in sarcomeres of myocytes (21); and the intermediate filament protein vimentin, which is expressed in the cytoskeletal intermediate filament of fibroblasts and endothelial cells (6, 16). Fibroblasts can be discriminated from endothelial cells by their different localization in cardiac tissue (15, 21). For triple staining, the sections were incubated with the following primary antibodies: anti-Cx43 mouse monoclonal antibody (Chemicon; 1:100), anti-vimentin mouse monoclonal antibody (DakoCytomation; 1:80), and anti-α-actin in mouse monoclonal antibody (DakoCytomation; 1:50) (25). The sections were then treated with the following secondary antibodies: horseradish peroxidase conjugated anti-mouse Ig (Vector; 1:1) for Cx43 and alkaline phosphates conjugated rabbit anti-mouse Ig (DakoCytomation; 1:30) for vimentin and α-actin. Staining was performed one by one with blocking previous antibodies and enzymes by citric acid buffer. The labeled sections were examined by digital microscope (COOLSCOPE; Nikon).
Real-time PCR and Western blotting.
To quantify mRNA expression of Cx43 in ventricular tissue, we performed a real-time PCR assay (Perkin-Elmer ABI Prism7700). 18S mRNA was used as an internal control. Sequence of PCR primers and sequence-specific probes are shown in the online data supplement in a previous article from our laboratory (1).
The amount of Cx43 protein was evaluated by Western blotting. The density of the Cx43 bands was quantified by densitometry and normalized to α-tubulin as control.
For separate analysis of Cx43 forming gap junctions (the junctional fraction) and that elsewhere in the cell (the nonjunctional fraction), we performed the Triton X-100 (TX-100) extraction method as described by Bruce et al. (4). In brief, samples of frozen ventricular tissue were homogenized by buffer solution containing 25 mM Tris (pH 7.4), 150 mM NaCl, 1/250 mammalian tissue extract protease inhibitors (Sigma), and 1% TX-100. After incubation on ice, the samples were centrifuged (15,000 rpm, 30 min, 4°C). Nonjunctional proteins in the supernatant were collected. Junctional proteins in the pellet were resuspended in buffer solution containing 4M urea. The samples were sonicated, incubated, and centrifuged to collect junctional proteins.
Late radiation injury of the heart.
Possible late radiation injury to the heart, such as newly emerging myocardial damage or proarrhythmia, was evaluated by longitudinal examination by echocardiography (UCG) and electrocardiography (ECG) in five rabbits before, 2 wk, and 3, 6, and 12 mo after THIR at 15 Gy. An in vivo electrophysiological study (EPS) was performed 1 yr after the THIR.
UCG and ECG were performed under light anesthesia with pentobarbital (0.125–0.5 mg/kg iv). LV end-diastolic diameter (EDD), LV end-systolic diameter (ESD), and ejection fraction (EF) were measured using two-dimensional M-mode echo signals. Fractional shortening (FS) was expressed as (EDD − ESD)/EDD × 100. The atrial filling velocity-to-early diastolic ratio (A/E) and deceleration time in transmitral flow (DT) were measured by Doppler imaging to estimate the diastolic LV function. RR, PQ, QRS, and QT intervals were measured from the ECG. Corrected QT interval (QTc) was defined as the QT interval divided by the square root of the RR interval.
In EPS, vulnerability to ventricular tachycardia/ventricular fibrillation was tested by programmed stimulation under norepinephrine infusion (0.1 μg·kg−1·min−1 iv). A pair of bipolar electrodes was placed at the epicardial surface near the LV apex. Following five basic stimuli (S1) at a cycle length of 200 ms, three extra stimuli (S2–S4) at twice diastolic threshold were applied with progressively shorter coupling intervals (1). After the completion of EPS, all hearts were fixed and embedded in paraffin. The ventricular sections were stained with hematoxylin/eosin and azan to analyze any pathological changes.
Data are presented as means ± SE. Data sets containing multiple groups were analyzed by two-way layout ANOVA and Bonferroni type multiple comparisons. Differences were considered statistically significant at a value of P < 0.05.
Time-dependent expression of Cx43 after THIR.
We first examined time-dependent changes of Cx43 expression in LV tissue after THIR at 15 Gy. Immunohistochemistry, Western blotting, and RT-PCR were carried out 2 wk and 3, 6, and 12 mo after the THIR in five rabbits in each group. Five rabbits without THIR were used as controls. Figure 1A shows representative confocal images of Cx43 immunolabeling in longitudinal sections of LV anterior wall. In control, Cx43 formed clusters of punctuate immunofluorescence domains confined to well-organized intercalated disks running perpendicular the longitudinal axis. THIR resulted in a characteristic increase in immunopositive Cx43 not only at the intercalated disk regions but also at the lateral cell borders during the entire follow-up period for 1 yr. Figure 1B shows the proportion of total cell area occupied by Cx43 immunoreactive signals in the combined analysis of 25 rabbits (5 in each group). The radiation resulted in a significant increase in the immunopositive signals by 71–116% compared with control (P < 0.05). We estimated the proportion of Cx43 label at the lateral cell surface (LS) over the total label including both at LS and at the intercalated disk region (ID). The values (LS/ID + LS) after THIR (28.2 ± 10.9% at 2 wk, 30.2 ± 15.7% at 3 mo, 28.1 ± 11.6% at 6 mo, and 18.9 ± 14.2% at 12 mo) were all significantly larger than controls (9.3 ± 6.3%; P < 0.05). The Cx43 protein amount estimated by Western blotting and the Cx43 mRNA estimated by RT-PCR were also increased after THIR throughout the entire follow-up period from 2 wk to 12 mo by 37–55% and by 24–59%, respectively, compared with controls (P < 0.05; Fig. 1, C and D).
In cardiac tissue, there is a substantial amount of fibrous tissue (collagen and fibroblasts) in between myocytes. The THIR-induced Cx43 upregulation, in particular that observed along the lateral borders of cardiac myocytes, could reflect Cx43 upregulation in fibroblasts. To address this issue, we carried out triple staining for Cx43, vimentin, and α-actin. Figure 2 shows representative images of LV anterior wall sections. In controls, Cx43 (brown) was recognized predominantly at the intercalated disk region. Vimentin (blue), a marker of fibroblasts (6), was recognized in between the myocytes. In the section 2 wk after THIR (15 Gy), Cx43 was upregulated not only at the ID but also at the lateral borders of myocytes. However, the lateralized Cx43 was not colocalized with vimentin. They were clearly separated in both longitudinal and transverse sections (see magnified images). The Cx43-positive domains were always in contact with actin. Similar distinct localization of Cx43 and vimentin was observed in four other rabbit hearts 2 wk after THIR. This observation strongly suggests that the lateralized Cx43 is expressed in cardiomyocytes, but not in fibroblasts. We quantified immunoreactive vimentin in control (n = 5) and 2 wk after THIR (n = 5). The total area occupied by vimentin immunoreactive signals was not affected by THIR (11.2 ± 4.5% in control vs. 15.7 ± 5.1% after THIR; P = 0.10).
Dose-dependent expression of Cx43 after THIR.
Next we examined dose-dependent changes of Cx43 expression in LV tissue (anterior wall) 2 wk after THIR at three radiation doses (5, 10, and 15 Gy; n = 5 in each group); five rabbits without THIR were used as controls (Fig. 3). The increase of Cx43 expression identified by immunohistochemistry was only significant in the rabbits receiving 10–15 Gy. The increase amounted to 88–114% (P < 0.05), the response to 5 Gy being insignificant (Fig. 3, A and B). Similarly, Cx43 protein amounts estimated by Western blotting were significantly increased in the rabbits receiving 10–15 Gy (by 21–42%; P < 0.05), but not in those receiving 5 Gy (Fig. 3C). The expression of mRNA estimated by RT-PCR was significantly increased at all three tested doses (by 18–46%; P < 0.05; Fig. 3D).
Regional difference of Cx43 upregulation after THIR.
Figure 4 shows the experiments to assess regional differences of Cx43 upregulation by THIR in five rabbits 2 wk after 15 Gy irradiation (5 without THIR were used as controls). Tissue sections after THIR from the LV anterior wall, the LV lateral wall, the LV posterior wall, the interventricular septum, and the RV free wall were compared with the LV anterior wall of control rabbits. Immunohistochemistry revealed that THIR caused a similar increase of Cx43-positive area in these five regions (by 88–148% compared with controls; P < 0.05; Fig. 4, A and B). The Cx43 amounts estimated by Western blotting and Cx43 mRNA levels estimated by RT-PCR were also similarly increased by THIR in these five regions (by 28–48% and 44–61%, respectively, compared with controls; P < 0.05). These results suggest that the THIR-induced Cx43 upregulation is not restricted to the radiation target area (LV anterolateral wall).
Junctional and nonjunctional Cx43 proteins.
Cx43 proteins recognized by standard Western blotting are composed of junctional and nonjunctional components. The junctional component is incorporated into the sarcolemma to form docked connexons, whereas the nonjunctional component is Cx43 en route to degradation. We discriminated the two components by using the TX-100 extraction method (4). Figure 5A shows the effects of THIR at 15 Gy on the LV anterior wall 2 wk and 3, 6, and 12 mo after the irradiation. The THIR caused prominent increases of junctional Cx43 protein throughout the entire follow-up period. The average increase in five hearts in each group compared with control (n = 5) was 3.6- to 4.2-fold (P < 0.05). In contrast, the nonjunctional Cx43 protein was virtually unaffected by THIR. Figure 5B shows the effects of THIR at 15 Gy on the LV anterior wall, LV lateral wall, LV posterior wall, interventricular septum, and RV free wall 2 wk after the irradiation. THIR caused a significant increase of the junctional Cx43 in all regions by 2.5- to 2.9-fold compared with control (LV anterior wall; P < 0.05; n = 5) but had no significant effects on the nonjunctional Cx43. These observations suggest that most of the Cx43 upregulated by THIR resides in the sarcolemma and may be assembled into gap junctional plaques.
Late radiation injury.
Longitudinal examinations by UCG and ECG were carried out to evaluate possible late radiation injury of THIR at 15 Gy. Figure 6A shows representative UCG images obtained from the same rabbit before, 2 wk after, and 12 mo after THIR. There were no appreciable changes in LV dimensions, LV contractility, and LV diastolic function over the period of a year. Figure 6B summarizes the results obtained in five rabbits. Parameters of LV dimension (EDD, ESD), those of LV contractility (EF, FS), and those of LV diastolic function (A/E, DT) did not change significantly from baseline (before THIR) throughout the whole follow-up period from 2 wk to 12 mo after THIR. In ECGs, RR, PQ, QRS, QT, and QTc intervals did not change significantly from baseline throughout the whole follow-up period. In the EPS 12 mo after THIR, no VT/VFs were induced by programmed stimulation. LV and right ventricular tissue sections, obtained from the five rabbits at 12 mo after THIR, showed no substantial histopathological features of myocardial injury (e.g., inflammation, degeneration, or fibrosis) in standard microscopy.
The major observations in the present study are as follows. 1) THIR at 15 Gy causes a significant upregulation of Cx43 protein and mRNA in normal rabbit ventricular myocardium lasting up to 1 yr after a single irradiation. The THIR-induced Cx43 upregulation is dose dependent, and the minimum irradiation energy is around 10 Gy. 2) The THIR-induced Cx43 upregulation is not restricted to the LV anterolateral free wall (target region), but also observed in other regions of the both ventricles. 3) The increase of immunoreactive Cx43 is predominant in the lateral cell abutment. 4) Rabbits 1 yr after THIR at 15 Gy show no significant changes in ECG, UCG, and EPS, suggesting absence of late radiation injury.
Long-term Cx43 upregulation by THIR.
The THIR-induced Cx43 upregulation in ventricular myocardium lasted for a surprisingly long period (at least 1 yr). Precise mechanisms of this long-lasting effect remain unclear. Irradiation with heavy ions could leave persistent radioactive isotopes in the affected tissue. Those isotopes might provide the source for low level persistent secondary radiation. However, the half-life of carbon-11, a major isotopic product of carbon beam irradiation, is only ∼20 min, which is incompatible with the effect lasting up to 1 yr. The radiation dose from such isotopic products also is estimated to be 4 to 5 orders lower than the dose from the primary beam. The long-term effect could be related to a more potent effect of heavy ions compared with other beams in terms of double-strand breaks of the DNA chain. In experiments using human fibroblast cultures, Cx43 upregulation induced by α-particle radiation lasted longer than that induced by photon beam radiation (x-ray, γ-ray) (2). Heavy ions are known to have a higher tumor cell killing effect than α-particles (3).
Regional variation of Cx43 upregulation.
The THIR-induced Cx43 upregulation was not restricted to the LV anterolateral wall targeted by radiation but was also observed in other regions in the both ventricles. This uniform modification of Cx43 expression can be explained in part by a bystander effect or adaptive response of radiation (22). The bystander effect is a phenomenon that neighboring nonirradiated cells show biological response (e.g., apoptosis and the mutagenesis) like irradiated cells (10). A variety of mechanisms have been suggested to be involved in this phenomenon; those include gap junction-mediated intercellular communication, secreted soluble factors, oxidative metabolism, plasma membrane-bound lipid rafts, and calcium fluxes (13, 23). An inhibition of phosphorylation of ERK was shown to suppress the bystander response, suggesting an involvement of MAPK signaling (14). In the rabbit heart, unlike in humans, it was practically difficult to synchronize the heart motion to correspond to the radiation release, because of the smaller organ size and faster heart rate. This may have limited the accuracy of irradiation to the targeted region.
Lateralization of Cx43.
THIR resulted in an increase of Cx43 fluorescence signals not only in the intercalated disk regions but also at the lateral cell borders (lateralization). There is a substantial amount of fibroblasts in cardiac tissue. The THIR-induced Cx43 upregulation might reflect increased expression of the gap junction protein not only in cardiac myocytes but also in fibroblasts in between cardiac myocytes (6). To address this issue, we carried out triple staining for Cx43, vimentin, and α-actin. The results revealed that Cx43-positive domains were not colocalized with vimentin-positive domain (a marker of fibroblasts). They were clearly separated. Instead, the Cx43-positive domains were always in contact with actin. This suggests that the THIR-induced upregulation of Cx43 occurred predominantly in cardiomyocytes but not in fibroblasts.
In a previous article from our laboratory, in normal rabbit hearts, we showed that Cx43 is expressed in the lateral cell abutments after THIR (1). This was not associated with the expression of cadherin, a protein forming the fascia adherents. There are two fractions of Cx43 proteins in the cardiac myocytes: one incorporated into the sarcolemma as docked connexon (junctional component) and another en route to degradation in the cytoplasm (nonjunctional component) (4). We carried out Western blotting after TX-100 extraction to discriminate the two fractions (4). The results revealed that the increase of Cx43 protein amount after THIR was predominant in the TX-100 insoluble junctional component. Previous in vivo epicardial excitation mapping experiments in normal rabbit hearts (2 wk after THIR) from our laboratory showed that transversal, but not longitudinal, conduction velocity was increased significantly by the radiation, giving rise to a significant reduction of the anisotropic ratio (1). It was also shown in rabbits after MI that the reduced conduction velocity and the increased dispersion of activation-recovery interval (an index of action potential duration) were reversed by THIR (1). These facts strongly suggest that the lateralized Cx43 proteins increased after THIR may form functional gap junctions even when Cx43 is dissociated from cadherin.
The functional implication of enhanced gap junction lateralization would be variable in terms of arrhythmogenesis. In cardiac muscle with no substantial source and sink imbalance of excitatory currents, an improvement of electrical coupling may increase conduction velocity and decrease spatial heterogeneity of repolarization, thereby reducing the risk of microscopic anisotropic reentry (7, 19). In the setting of a serious source and sink imbalance, a different scenario may be pertinent. Action potential propagation from a narrow strand to a large mass of cardiac muscle is known to be compromised, because limited amount of excitatory current is provided by the small source to the large sink (7, 19). In such a case conduction critically depends on the extent of electrical coupling; the higher the coupling, the greater the inhibition (19). Therefore, against the intuition, the safety of conduction may be reduced by an increase of electrical coupling and vice-versa. In theory, reduction of Cx43 in the diseased heart might increase the safety of conduction, providing a protective mechanism against reentrant arrhythmias (27). Interference of this response by increasing functional gap junction might, in some situations, be deleterious.
It has been shown in a previous study from our laboratory using a rabbit model of subendothelial MI, which was created by intracoronary microsphere injection, that THIR decreased the vulnerability to VT/VF in association with decrease of spatial heterogeneity of repolarization (1). However, further experimental studies will be required to extrapolate the observation to other proarrhythmic pathological conditions.
Late radiation injury.
In radiation therapy of cancer patients, renal and liver dysfunction often develops 6–12 mo after the radiation (8). Newly developed cancers or leukemia may also occur several years later (8). Although the heart is relatively resistant to late radiation injury, MI was reported to develop after radiation therapy in patients with Hodgkin's disease (29) and in those with breast cancer (12). A possible beneficial effect of THIR might be offset by an adverse effect on the coronary vasculature. In the present study, significant complications in the heart were not observed as far as hemodynamic and electrophysiological parameters are concerned, even 1 yr after irradiation. Moreover, no pathological changes were induced in cardiac muscle. Nevertheless, longer follow-up experimental studies in the heart as well as in other organs will be required to evaluate the safety of THIR to the heart.
First, in the present study, we did not examine the functional consequence of Cx43 upregulation in rabbit hearts lasting at least 1 yr after a single THIR at ≥10 Gy. The issue was discussed based on previous in vivo epicardial potential mapping experiments in rabbits 2 wk after a single THIR at 15 Gy from our laboratory (1). More extensive electrophysiological studies using isolated hearts will be required to confirm the functional consequence. Experiments using animal models of diseased heart (e.g., ischemia, infarction, hypertrophy, or heart failure) will also be required to evaluate the potential usefulness of THIR as a novel antiarrhythmic procedure. Second, we did not use age-matched control animals in the experiments to see the time-dependent effect of THIR (2 wk to 12 mo after the irradiation). To the best of our knowledge, there are no reports showing substantial changes of Cx43 expression in the heart during aging for 12 mo in adult rabbits. Third, we used the midmyocardial layer of ventricular tissue for the immunohistochemistry and did not examine potential changes of transmural Cx3 expression. This does not seem to invalidate the present results, since it was confirmed in our pilot experiments that THIR at 15 Gy caused almost uniform upregulation of Cx43 through the whole thickness of LV free wall 2 wk after the radiation (unpublished data). Fourth, there was no appreciable regional difference in the Cx43 upregulation in the heart despite the beam targeting to the LV anterolateral wall. We cannot rule out inaccuracy of the targeting because of small animal size. Finally, we cannot rule out potential serious adverse effects of THIR (e.g., carcinogenesis), which might appear later than 1 yr after irradiation.
A single application of THIR to rabbit hearts caused upregulation and altered cellular distribution of Cx43 in the ventricles for at least 1 yr. This long-lasting remodeling effect on gap junctions may open the pathway to novel therapy against life-threatening ventricular arrhythmias in structural disease.
This work was supported by Grants-in-Aid for Scientific Research (B) 17390236, Grant-in-Aid for Young Scientists (B) 20790555 from Japanese Society for the Promotion of Scientists, Tokai University School of Medicine Research Aid, and The Vehicle Racing Commemorative Foundation.
No conflicts of interest are declared by the author(s).
We are grateful to many colleagues for technical support: Sachie Tanaka, Noboru Kawabe, Katsuko Naito, Hideaki Hasegawa, Yoshiro Shinozaki, and Jobu Itoho at Department of Teaching and Research Support Center in Tokai University; Mayumi Hojo at Nagoya University; Maki Asano, Takeshi Murakami, and Kumie Nojima of the HIMAC Cooperative Research Project; Masaya Sakai, Norihiko Mishima, and Satoshi Yamazaki at Fukuda Denshi; Daisuke Araki and Michinari Kaneko at Unique Medical; Daisuke Nakata at Life Line; and Nobue Kumaki at Department of Pathology in Tokai University.
↵* M. Amino and K. Yoshioka contributed equally to this work.
- Copyright © 2010 the American Physiological Society