For our study, male healthy Kunming mice aged 2 (young), 12 (middle-aged), and 24 months (aged) were obtained from the Laboratory Animal Center of the Academy of Military Medical Sciences (Beijing, China) and received humane care in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH publication no. 85–23, revised in 1996). We chose these ages of mice based on previous study (13). The ages of 2, 12 and 24 months are equivalent to the human age of ~20, 40 and 70 years, respectively. The characteristics of the mice are shown in Table I.

Sinus cycle length (SCL) was determined by averaging 3 consecutive R-R intervals. P-wave duration (PWD) was measured by determining the earliest onset and latest offset of atrial deflection from 3 simultaneously recorded surface leads.

For the transesophageal electrophysiological study, a 2-French octapolar mouse electrophysiological catheter [8 0.5 mm circular electrodes; electrode pair spacing, 0.5 mm (Ciber Mouse; NuMed Inc., New York, NY)] was used for recording cardiac electrograms as well as for pacing the heart using consecutive electrode pairs. This catheter was inserted into the esophagus with a depth of 3–4 mm and unipolar recordings were obtained from each ring electrode. The threshold of pacing was examined by a 1 msec pulse width and a pacing rate 10 beats/min faster than normal. The electrode catheter was adaptively positioned to the site closest to the left atrium to ensure that constant atrial capture and correspondence to the minimum threshold (<1.5 mA) were attained. Sinus node recovery time (SNRT) was measured after a 30-sec pacing train with a basic cycle length (BCL) of 100 msec, a stimulus amplitude of 2-fold diastolic capture threshold and a stimulus duration of 1 msec. The SNRT was defined as the interval between the last stimulus in the pacing train and the onset of the first sinus return beat. Rate-corrected SNRT (CSNRT) was defined as the SCL subtracted from the SNRT.

Inducibility of atrial tachycardia and fibrillation (AT/F) was examined by applying 15-sec bursts (25 msec BCL, 2-fold diastolic capture threshold and 1 msec duration). This series of bursts was repeated 10 times. AF was defined as a period of rapid and fragmented atrial electrograms with irregular AV-nodal conduction and ventricular rhythm for at least 1 sec (14). AT was defined as rapid and regular rhythm lasting for at least 30 sec.

After in vivo electrophysiological analysis, the mice were sacrificed and prepared for histopathological analysis. The aortas were then perfused via the common iliac bifurcation with pre-cooled PBS for 5 min. The left and right auricles were quickly cut and snap-frozen in liquid nitrogen for 1 min and stored at −80°C. Sections (5 μm in thickness) of the paraffin-embedded tissue were stained with picrosirius red (Direct Red 80; Sigma Aldrich) (15). Briefly, tissue sections were rinsed in distilled water and incubated in saturated 0.1% picrosirius red for 90 min. The sections were then rinsed twice with 0.01 N HCl for 1 min, dehydrated with an ethanol gradient and prepared for collagen analysis. Images were captured by a digital charge-coupled device (CCD) camera (Pro 150ES; Pixera Corporation, Los Gatos, CA), and connected to a polarized microscope (E600POL; Nikon, Tokyo, Japan) equipped and modified with a commercially available circular polarizer (Kenko, Tokyo, Japan). The collagen volume fraction (CVF) was calculated as described previously (15).

To isolate high-quality mouse atrial myocytes, we adopted a Langendorff-perfused method based on modified intubation skills (16). In brief, mice were anesthetized using an intraperitoneal injection of 0.4% pentobarbital sodium solution at a dose of 50 mg/kg body weight. The inferior vena cava was exposed under direct vision following an abdominal median incision. While 4°C heparinized saline at a dose of 4–5 ml (100 U/ml) was administrated into the inferior vena cava, the abdominal aorta was dissected for bloodletting. Subsequently, the thoracic aorta, isolated through thymus dissection following a left thoracotomy between the second and fourth intercostal space, was slightly intubated by a self-made catheter (7 F paracentetic needle with blunt tip, ring-shaped notching 0.5 mm from the distal ending) at a position of 2 mm distance from the root of the aorta. A syringe with 3 ml saline was connected to the end of the catheter for excluding the internal air. The successful intubation was fixed to the aorta by sealing with 5–0 silk. The heart was then quickly excised and removed to the Langendorff apparatus for immediate retrograde perfusion. Atrial myocytes were prepared using enzymatic digestion and then separated from Kraft-Brühe (KB) solution by centrifuging at 600 rpm for 3 min. The cells were maintained at room temperature for the experiments that followed and used within 10 h.

Myocytes were transferred into an experimental chamber (~3 ml) mounted on the stage of an inverted microscope (TE2000-S; Nikon) and allowed to adhere to the glass bottom of the chamber. Cells were perfused with extracellular solution. A ruptured-patch whole cell voltage clamp was used to measure transient outward potassium (I_to) and ultra-rapid delayed rectifier potassium (I_kur) currents, and membrane capacitance (Cm), while a current clamp was used for measuring a single cell action potential (AP). For I_to and I_kur measurements, the internal solution contained (in mmol/l) 110 K-aspartate, 20 KCl, 1 MgCl₂, 5 Mg-ATP, 10 HEPES, 10 EGTA and 5 Na₂-creatine phosphate, pH 7.3–7.4 (NaOH). To block Ca²⁺-activated chloride currents, 0.01 mmol/l niflumic acid was added to the pipette solution. The external solution contained (in mmol/l) 137 NaCl, 5.4 KCl, 1.8 CaCl₂, 1 MgCl₂, 0.33 NaH₂PO₄, 5 HEPES, 10 D-glucose and 4-aminopyridine, pH 7.3–7.4 (NaOH). Ca²⁺ channels were blocked with 0.05 mmol/l CdCl₂. Atropine of 1 μmol/l was used to prevent muscarinic receptor activation, and 5 μmol/l E-4031 was added to the external solution to block the rapid delayed rectifier current (I_kr). In addition, EGTA was used for chelating free Ca²⁺, thereby blocking I_to2 and I_Na-Ca. For AP measurements, the internal solution contained (in mmol/l) 110 K-aspartate, 20 KCl, 1.8 MgCl₂, 10 HEPES, 5 Mg-ATP, 0.05 EGTA and 5 Na₂-creatine phosphate (pH 7.3–7.4, KOH). The external solution contained (in mmol/l) 137 NaCl, 5.4 KCl, 1 MgCl₂, 1.8 CaCl₂, 10 D-glucose, 0.33 NaH₂PO₄ and 5 HEPES (pH 7.3–7.4, NaOH). Signals were filtered with 2.9 and 10 kHz Bessel filters and recorded by an Axopatch700 A amplifier with the Digidata1320A-pClamp 8.2 Data Acquisition System (Axon Instruments, Foster City, CA). All experiments were conducted at room temperature (25±1°C).

Values are presented as the means ± standard deviation. Surface ECG and esophageal electrogram were measured by 3 independent observers and compiled for statistical interpretation. Differences between groups were analyzed using a two-tailed Student’s t-test, analysis of variance (ANOVA), Kruskal-Wallis H test, Mann-Whitney U test or Chi-square test, where appropriate. A two-tailed P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed with SPSS software.

A total of 32 mice had the transesophageal rapid pacing procedure performed. No complications due to anesthesia or surgical preparation of the esophagus were observed in the young group and middle-aged group. Two out of 10 mice in the aged group died from either sinus bradycardia or complete atrioventricular block by rapid atrial pacing and had to be excluded from the analysis.

The threshold of pacing (0.95±0.38 mA) in the young group was almost compatible with the middle-aged group (1.36±0.44 mA) and the aged group (1.22±0.21 mA). There was no significant difference in SCL among the 3 groups [112.73±17.78 sec (young), 110.42±11.64 sec (middle-aged) and 117.34±23.73 sec (aged)]. PWD in the aged group had a significantly higher prolongation than the young and middle-aged group (Fig. 1B) (P<0.05). A typical representation of SNRT measurement is shown in Fig. 1A. SNRT and CSNRT at a S1S1 stimulation cycle length of 100 msec in the aged group had a significantly higher prolongation relative to the young and middle-aged group (P<0.05); however, there was no significant difference between the young and middle-aged group (Fig. 1C and D) (P>0.05). The inducibility of AT/F among the 3 groups is shown in Table II. Although the duration of each AF episode was not significantly different among the groups, there was an overall induced rate of AT/F that increased significantly with age (P<0.05). Vulnerability of the atrial myocardium was defined as when AT/F was successfully induced in the mouse model.

Collagen fibers were much thicker in the right atrium than the left atrium, and the CVF was higher in the right atrium among the 3 groups (Fig. 2A). Spearman rank correlation analysis showed that the CVF in the left atrium significantly correlated with age (P<0.01, r=0.592) (Fig. 2B). The CVF in the right atrium was higher in the aged and middle-aged groups than in the young group; however, but there was no significant difference between the aged and middle-aged groups. Spearman rank correlation analysis showed that the CVF in the right atrium correlated with age (r=0.326, P<0.01); however, this did not achieve the level of the left atrium (Fig. 2B).

Overall, the right atrium showed significantly higher fibrosis than the left atrium in all the groups, and the extent of fibrosis in the left atrium had a higher positive correlation with age relative to the right atrium.

In the current clamp mode, APs were elicited by square current pulses of 400–600 pA amplitude and 3 msec duration. A steady-state AP was considered as the last of a train of 20 at the same stimulation rate. The APD at 20, 50 and 90% repolarization (APD₂₀, APD₅₀, APD₉₀, respectively) has a tendency towards prolonging followed by shortening in regard to the age development, whereas the APD₉₀ in the left relative to the right atrium demonstrated a significant difference in the aged group (P<0.05) (Table III).

In the whole voltage clamp mode, single atrial myocyte Cm was elicited by a square pulse of depolarization. Mean left atrial Cm was 34.2±1.4 pF (young) (n=33), 39.3±1.9 pF (middle-aged) (n=33) and 46.7±2.4 pF (aged) (n=19), respectively, while mean right atrial Cm was 30.0±1.4 pF (n=33), 32.8±1.9 pF (n=33) and 40.1±2.6 pF (n=19), respectively. These results demonstrate that Cm of the left and right atrium increases with advancing age, and there is a stronger tendency towards changes in left atrial Cm with age relative to right atrial Cm (Fig. 3).

Fig. 4A shows voltage-dependent I_to elicited by voltage steps of 100 msec to between −40 and +50 from −50 mV. The holding potential of −50 mV was used to exclude possible I_Na and I_Ca.T contamination during current recording. We found that I_to was activated rapidly and achieved peak current within minutes after delivering testing potentials, followed by a rapid decrease in the steady state baseline current level. I_to increased with testing potentials in a typically voltage-dependent manner. When 4 mM 4-AP was added to the external solution, I_to was blocked by >80% (Fig. 4A).

As shown in Fig. 4B, I_to was activated at −30 mV and current density was not significantly different among the 3 groups. The current density of I_to in the aged group was larger than that in the young and middle-aged groups at −10 mV (P<0.05), and this was even more significant when the testing potential was increased. At a test potential of +50 mV, the current density was 7.8±0.9 pA/pF (n=6), 10.0±1.0 pA/pF (n=6) and 12.0±1.4 pA/pF (n=5) in the young, middle-aged and aged groups, respectively, indicating that current density increased with age (P<0.05).

Fig. 4C shows the activation curve of I_to by depolarization to +50 from −50 mV. With age, the voltage-dependent I_to activation curve slightly shifted in a negative direction. V_1/2 and K were 12.8±0.9 and 9.4±0.9 mV (n=6), 11.5±0.6 and 9.5±0.6 mV (n=6), and 10.9±0.8 and 9.2±0.7 mV (n=6) in the young, middle-aged and aged groups, respectively, indicating that there were no significant changes in V_1/2 and K with age (P>0.05).

The voltage dependence of the steady state inactivation relationship was investigated using a standard 2-pulse protocol: a 2,000 msec preconditioning pulse ranging from −80 to 0 mV followed by a 140 msec test pulse to +50 mV. V_1/2 and K were −49.7±2.1 and 8.1±1.9 mV (n=6), −48.8±1.6 and 7.3±1.5 mV (n=6), and −48.7±1.5 and 7.0±1.4 mV (n=6) in the young, middle-aged and aged groups, respectively, indicating that there were no significant changes in V_1/2 and K with age (P>0.05) (Fig. 4D).

I_kur were elicited by a 80-msec prepulse to +30 mV to inactivate I_to, followed by 140-msec test pulses between −40 and +50 mV after a 10-msec interval from −50 mV at 0.2 Hz. At a test potential of +50 mV, the current density was 4.1±0.7 pA/pF (n=6), 4.2±1.0 pA/pF (n=6) and 4.2±0.8 pA/pF (n=5) in the young, middle-aged and aged group, respectively, indicating that there was no significant difference among the groups (P>0.05) (Fig. 5B).

In the present study, we performed measurements of in vivo global electrical properties by the transesophageal atrial rapid pacing method and found the following: (i Compared with the other 2 groups, PWD, as well as the SNRT and CSNRT at a cycle length of 100 msec were longer in the aged group, indicating the slowing of intra-atrial conduction and sinus nodal dysfunction. ii) Although the duration of each AF episode was not significantly different among the groups, the total inducibility, represented as the key point for evaluating vulnerability to AT/F, was significantly increased with aging. As regards changes in collagen in the left and right atrium, we found that the right atrium showed significantly higher fibrosis relative to the left atrium in all the groups, whereas the extent of fibrosis in the left atrium had a higher positive correlation with age compared to the right atrium. Furthermore, from the cellular electrophysiological experiments, the results showed that age-related augmented I_to currents and no age-related changed I_kur currents, as well as the subsequent AP discrepancy in old age, contributed to the dispersion of repolarization, which further promoted AF.

Abnormal pulse initiation due to the sinoatrial node dysfunction is associated with advancing age (17). If the sinus node fails to fire, then there is a significant tendency towards manifestation of abnormal impulse initiation in atrial cells. In the present study, measurements of SNRT and CSNRT at a cycle length of 100 msec showed that aged mice had a longer SNRT than the young and middle-aged mice. Recently, Stiles et al (18) demonstrated that the prolongation of SNRT, regarded as sinus node dysfunction, is a key contributor which facilitates the progression of paroxysmal AF to permanent AF. Surprisingly, it has been previously found that although the SCL increases with age, the SNRT decreases with age (19), which is in contrast to our results. The likely mechanism for these observations is that conduction in the sinus node worsens with age and limits the ability of an extra-stimulus to overdrive and suppress sinus node automaticity. A previous study found enlargement of the sinus node, hypertrophy of sinus nodal cells, and remodeling of the extracellular matrix in aged rats (20). Therefore, they concluded that the age-dependent decrease in sinus nodal function is due to structural remodeling of the sinus node. Consequently, decreased automaticity and conductivity of the sinus node are associated with shortening of atrial refractoriness and increasing the atrial effective refractory period (ERP) dispersion, which in turn creates a substrate for AF initiation and perpetuation (21).

Abnormal pulse conduction in the aged atrium is a contributor to the maintenance of AF. Hence, it is important to comprehend the nature of both the electrical and structural remodeling that occurs with age. In the current study, surface ECG showed that PWD was significantly longer in the aged group than in the young group and had a prolongation tendency in the middle-aged group, which indicated that the intra-atrial pulse conduction time was prolonged with increasing age. It has been reported that the PWD significantly correlates with age in animals (6) and humans (22), and depends mainly on intra-atrial conduction time and atrial size. The slowing of intra-atrial conduction, one of the most important requirements for the initiation of re-entry, is more frequently observed in patients with paroxysmal AF than in normal subjects (23). Moreover, atrial dilatation increases the amount of atrial tissue that can accommodate multiple re-entry circuits (3). Therefore, P-wave prolongation and atrial dilatation possibly reflect age-related atrial remodeling that is favorable for AF; thus, aged mice, compared to young mice, possess a substrate that is more susceptible to AF.

It is well known that the profound structural remodeling that occurs with age affects the ability of a premature beat to penetrate the substrate and initiate propagation. In the present study, we observed a significant correlation between the degree of interstitial fibrosis and aging in the left and right atrium. In a previous study, Burkauskiene et al (24) demonstrated that the diameter and area of collagen fibers positively correlate with aging, whereas the amount of collagen fibers does not alter with age. This indicates that changes in the collagen network may be a progressively developing process. Another study reported that spatial distribution of the collagen network was significantly different in patients with ischemic cardiomyopathy compared with those without heart disease (25). Taken together, these findings, including ours, suggest that atrial interstitial fibrosis, which is increased with age and heart disease, possibly enhances the heterogeneous slowing of atrial conduction, which may be responsible for the age-related increase in AF vulnerability. In general, there is a close correlation between the onset of AF and structural remodeling in the left atrium (26). However, Nakajima et al (27) found that although similar levels of cardiomyocyte apoptosis were present in the right and left atria of MHC-TGFcys³³ser hearts, the extent of fibrosis was more pronounced in the right atrium than in the left atrium. They also revealed cardiomyocyte cell activity in left atrial cardiomyocytes, but not in right atrial cardiomyocytes by tritiated thymidine incorporation studies. These data support the notion that cardiomyocyte cell cycle induction can antagonize fibrosis in the myocardium. In agreement with previous results, the present study found that the collagen volume of the right atrium was higher than that in the left atrium in all the groups. Although we found that the volume of fibers increased with age in the left and right atrium, less fibrosis was present in the left atrium, which may be due to faster cell cycle activity in the left atrium.

In our study, we compared the inducibility of atrial AT/F to evaluate vulnerability to arrhythmias. Our results showed that the inducibility of atrial AT/F significantly increased with age, indicating that aging is a key contributor to AF episodes. In addition, inducible AF was also observed in the young and middle-aged group, in contrast to the negative results presented in the study by Guzadhur et al (28). A possible reason is that the susceptibility of the mouse atrium to the induction of AF appears to be increased when transesophageal pacing instead of isolated pacing (28) or transvenous pacing (29) is performed. It is possible that relative to isolated or transvenous stimulation, unipolar transesophageal atrial stimulation affects a larger area of the atrium, which may involve the vagal nerve. This may result in a more global inhomogeneous and abnormally delayed activation of the atrium, which is believed to be the basic mechanism of induction of atrial arrhythmia by premature depolarization. Moreover, the atrial stretch is known to enhance atrial vulnerability and may be augmented in this setting of transesophageal stimulation. Of note, the present study showed that PWD was only prolonged in aged group relative to other 2 groups, not reconciling the high incidence of tachyarrhythmia in the middle-aged and aged group. The discrepancy may be due to more electrophysiological remodeling in the middle-aged group, while structural remodeling was predominant in the aged group.

AP changes with age are affected by different types of inward and outward ion currents. I_to and I_kur, which are the major ion currents involved in the repolarization process in mice, contribute to age-related AP changes and AF initiation and maintenance. In the present study, a Ca²⁺-independent transient outward current (I_to1) was recorded by the addition of CdCl₂ to the external solution to block the L-type Ca²⁺ current and the addition of a Ca²⁺ chelating agent (EGTA) to the internal solution. Our results showed that I_to was suppressed by >80% after the addition of 4-AP. Dun et al (30) found that I_to current densities of canine right atria were significantly increased in the aged group compared with the adult group, whereas inactivation slowing and steady state inactivation curves showed a trend to shift toward depolarization. Mansourati et al (31) also found that I_to was decreased in adult patients with heart diseases; however, the results did not reflect the normal electrophysiological properties of humans due to the existence of pathological factors. We found that I_to current densities increased with age, which is in agreement with the results presented in the study by Crumb et al (32), whereas the activation and inactivation properties of I_to were not different among the 3 groups. As such, the augmentation of I_to may have contributed to the shortening of atrial APs in the aged group. Following the inactivation of I_to1, a slowly inactivating current remains, termed I_sus or I_kur. There is a general consensus that the K_V 1.5 subunit is largely responsible for I_kur in human atrial myocytes, similar to other species, such as the mouse and dog. Since this subunit is much less abundant in the ventricle than in the atrial myocardium, it appears to be quite specific to atrial myocytes. I_kur densities are significantly increased in aged canine right atrial cells compared with adult cells (30). However, KN-93, a specific CaMKII inhibitor, has been shown to dramatically inhibit I_kur in aged right atrial cells compared with adult right atrial cells, indicating that increased I_kur in aged RA cells is most likely due to CaMKII upregulation (33). To the best of our knowledge, no age-related changes in I_kur have thus far been observed in human atrial myocytes. In alignment with human atrial data, the present study using mice also demonstrated that there were no age-related changes in I_kur among the different groups, which may be due to stability after birth.

The abnormal shape of the AP contour derived from a single isolated cell is conducive to AF. Anyukhovsky et al (34) found that APD₉₀ averaged across all regions was significantly longer in the aged compared with adult tissues. Notably, the range of APD₉₀ values was wider in the aged vs. the adult group as a result of an increase in duration, occurring mainly in right atrial fibers (34). However, clear-cut evidence of age-related changes in APs in normal human atria is lacking, as no age-related changes in APD and ERP have been obtained (35). In the present study, we found that APD₉₀ in the left and right atrium was prolonged with age, followed by a decrease in the aged group, particularly in the left atrium. In agreement with our report, Brorson et al (36) used 4 age groups to examine right atrial ERPs and monophasic APs in healthy male humans in vivo. They found that the middle-aged group showed a significantly longer right atrial ERP than the other groups. However, there was no progressive trend in APD with age. Thus, these results do not support the traditional view that APDs are prolonged with age. In general, AP is prolonged followed by shortening with age may due to a decrease in the calcium current in aged atrial tissue (6).

The transesophageal stimulation technique used in the current study includes difficulties in atrial electrogram detection during AF, and therefore, the evaluation of the atrial AF heart rate is not always possible. Additionally, due to the limitations of experimental devices and technologies, measurements of atrial ERP and intra-atrial conduction velocity were not conducted for further insight into the electrophysiological mechanisms. Finally, it should be acknowledged that the pathophysiological characteristics of AF in mice may be different from those in humans.

In conclusion, the results from our study demonstrate that aging causes structural remodeling, as well as integrative and cellular electrophysiological changes, facilitating increased dispersion of repolarization and re-entry formation. This in turn creates a substrate for AF initiation and perpetuation. Furthermore, there may be an association between structural and electrical remodeling, which can exert vicious cycle effects on the development of AF.