Abstract

Circadian clocks are intracellular molecular mechanisms that allow the cell to anticipate the time of day. We have previously reported that the intact rat heart expresses the major components of the circadian clock, of which its rhythmic expression in vivo is consistent with the operation of a fully functional clock mechanism. The present study exposes oscillations of circadian clock genes [brain and arylhydrocarbon receptor nuclear translocator-like protein 1 (bmal1), reverse strand of the c-erbaα gene (rev-erbaα), period 2 (per2), albumin D-element binding protein (dbp)] for isolated adult rat cardiomyocytes in culture. Acute (2 h) and/or chronic (continuous) treatment of cardiomyocytes with FCS (50% and 2.5%, respectively) results in rhythmic expression of circadian clock genes with periodicities of 20–24 h. In contrast, cardiomyocytes cultured in the absence of serum exhibit dramatically dampened oscillations in bmal1 and dbp only. Zeitgebers (timekeepers) are factors that influence the timing of the circadian clock. Glucose, which has been previously shown to reactivate circadian clock gene oscillations in fibroblasts, has no effect on the expression of circadian clock genes in adult rat cardiomyocytes, either in the absence or presence of serum. Exposure of adult rat cardiomyocytes to the sympathetic neurotransmitter norephinephrine (10 μM) for 2 h reinitiates rhythmic expression of circadian clock genes in a serum-independent manner. Oscillations in circadian clock genes were associated with 24-h oscillations in the metabolic genes pyruvate dehydrogenase kinase 4 (pdk4) and uncoupling protein 3 (ucp3). In conclusion, these data suggest that the circadian clock operates within the myocytes of the heart and that this molecular mechanism persists under standard cell culture conditions (i.e., 2.5% serum). Furthermore, our data suggest that norepinephrine, unlike glucose, influences the timing of the circadian clock within the heart and that the circadian clock may be a novel mechanism regulating myocardial metabolism.

  • heart
  • metabolism
  • neurohumoral
  • rat
  • zeitgebers

circadian clocks are controlled by a set of genes that generate self-sustained positive and negative transcriptional feedback loops with a free-running period of 24 h (7, 11, 40). This molecular mechanism is intrinsic to the cell, persisting in cultured cells such as fibroblasts, vascular smooth muscle cells, and various cell lines (3, 13, 20, 23). Circadian clocks confer the selective advantage of anticipation, conditioning the cell to changes in its environment, such that it can respond rapidly to a specific extracellular signal at an appropriate time of the day. For example, our laboratory (30) has recently hypothesized that the circadian clock within cardiac and skeletal muscle allows the myocytes to anticipate diurnal variations in circulating fatty acids through modulation of fatty acid oxidative capacity at a transcriptional level. However, the existence of an intracellular circadian clock mechanism that is intrinsic to either the cardiomyocyte or skeletal myocyte has not yet been established.

Zeitgebers, or timekeepers, are factors that influence the timing of the circadian clock. The primary zeitgeber for the central (or master) clock located within the suprachiasmatic nucleus is light, whereas peripheral clocks (i.e., those distinct from the central clock) are influenced by neurohumoral factors (6, 7). Through the use of a primary cell culture system, Nonaka et al. (23) have reported that angiotensin II is a zeitgeber for the circadian clock within vascular smooth muscle cells. To date, no zeitgebers have been identified for the circadian clock within the cardiomyocyte.

Approximately 10% of all rodent myocardial genes exhibit significant diurnal variations in expression (31). These genes encode proteins involved in transcription, translation, protein turnover, signal transduction, ion homeostasis, and metabolism, as well as the clock components themselves (19, 30, 31, 39, 42). However, diurnal variations in myocardial gene expression in vivo could be due to diurnal variations in multiple neurohumoral influences and/or the circadian clock intrinsic to the cardiomyocyte. Identification of those genes specifically regulated by the circadian clock within the cardiomyocyte will undoubtedly provide insight regarding the role(s) of this molecular mechanism in the heart. Mouse models of disrupted circadian clock components (e.g., the clock gene) have proved useful in expanding our knowledge regarding the intricacies of this molecular mechanism (36). However, because the circadian clock is impaired in all cells of clock mutant mice, diurnal variations in neurohumoral factors are abnormal, preventing dissociation of intracellular (i.e., circadian clock) versus extracellular (i.e., neurohumoral) influences on diurnal variations in myocardial gene expression. An alternative strategy is therefore to remove any neurohumoral influence through the use of cell culture studies. Because of its intrinsic nature, only the intracellular circadian clock will mediate 24-h gene expression oscillations in cultured cells. This strategy has been successfully utilized for the identification of multiple tissue-specific circadian clock-regulated genes, such as insulin in pancreatic β-cells and peroxisome proliferator-activated receptor-α (pparα) in the liver (1, 24).

Our laboratory (43, 44) has previously characterized rhythmic expression of circadian clock genes in the intact rat heart. The purpose of the present study was to 1) expose the intrinsic nature of the circadian clock within the cardiomyocyte, 2) investigate whether glucose and/or norepinephrine influence the timing of this molecular mechanism, and 3) investigate whether the metabolic genes pyruvate dehydrogenase kinase 4 (pdk4) and uncoupling protein 3 (ucp3) are regulated by the circadian clock within the heart.

MATERIALS AND METHODS

Animals.

Male Wistar rats (Harlan; 200 g initial weight) were housed at the Animal Care Center of the University of Texas Health Science Center at Houston. All animal experiments were approved by the Institutional Animal Care and Use Committee. All animals were housed under controlled conditions (23 ± 1°C; 12-h light/12-h dark cycle) and received standard laboratory chow and water ad libitum. Ten days before they were euthanized, animals were housed in a separate environment-controlled room, in which a strict 12-hour light/12-hour dark cycle regime was enforced [lights on at 7:00 AM; zeitgeber time (ZT) 0]. For isolated cardiomyocyte studies, hearts were excised at ∼11:00 AM (i.e., ZT 4). For intact rat heart studies, animals were euthanized at 3-h intervals, beginning at 7:00 AM (i.e., ZT 0); to ensure reproducibility of gene expression cycles, at least two animals were euthanized at each time point on at least three separate days.

Adult rat cardiomyocyte isolation and culture.

Adult cardiomyocytes were isolated from male Wistar rats and cultured on laminin-coated plates, essentially as described previously (4, 29). Cardiomyocytes were cultured in glutamine-free DMEM (Sigma) supplemented with (in mM) 1 pyruvate, 4 NaHCO3, 8.6 HEPES (pH 7.3), 5 creatine, 2 l-carnitine, and 5 taurine, and 10 μM cytosine β-d-arabinofuranoside, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Unless otherwise specified, glucose was present at a concentration of 5 mM. The culture medium was also supplemented with a defined concentration of FCS. For the sake of consistency, all experiments followed the same general protocol (as depicted in Fig. 1), with variation in equilibration, challenge, wash, and postchallenge media supplementation (as specified for individual experiments), as well as the duration of the experiment. When only one set of conditions was investigated, cardiomyocytes were cultured up to 60-h postchallenge initiation. In contrast, when two groups were investigated (i.e., control vs. experimental), cardiomyocytes remained in culture only for a maximum of 36 h postchallenge initiation. This limitation on the number of time points was driven by the need for adequate cell numbers in each group, thereby enabling sufficient RNA isolation yields for subsequent gene expression analysis (see RNA extraction and quantitative RT-PCR). After the cells were plated to laminin for 2 h (4% FCS medium), cells were cultured overnight (equilibration medium with or without 2.5% FCS; to be specified). At time 0 (11:00 AM the following morning; ∼24 h postisolation) cardiomyocytes were challenged with fresh medium (challenge medium; supplements are specified for individual experiments). At time 2 h, cells were washed twice with wash medium (either in the absence or presence of 2.5% FCS; as specified), and subsequently cultured in postchallenge medium (either in the absence or presence of 2.5% FCS; as specified). All media were preequilibrated to 37°C in an atmosphere of 5% CO2 for at least 2 h before utilization with the use of 25-cm2 phenolic-style cap cell culture flasks (Corning). Furthermore, all media changes were performed on a 37°C preequilibrated heat pad within a sterile tissue culture cabinet, thereby minimizing temperature fluctuations. Cells were terminated at multiple time points (see Fig. 1 and individual experiments) and stored at −80°C before RNA extraction.

Fig. 1.

General cell culture protocol. After isolation, cells were cultured on laminin-coated plates for 2 h (4% FCS containing medium). Plated cells were allowed to equilibrate overnight (∼22 h) in either the absence or presence of 2.5% FCS (as specified). Cells were next challenged for 2 h by medium containing serum (2.5% or 50% FCS; as specified), glucose (5 mM or 15 mM), and/or norepinephrine (10 μM). After challenge, cells were rapidly washed twice with wash medium and subsequently cultured in postchallenge medium for up to 58 h. Supplementation for wash and postchallenge medium were specified for individual experiments.

RNA extraction and quantitative RT-PCR.

RNA extraction and quantitative RT-PCR was performed by using methods described previously (29, 30, 4244). Specific quantitative assays were designed from rat sequences available in GenBank for three core circadian clock genes [brain and arylhydrocarbon receptor nuclear translocator (arnt)-like protein 1 (bmal1), reverse strand of the c-erbaα gene (rev-erbaα), and period 2 (per2)], one circadian clock-regulated gene [albumin d-element binding protein (dbp)], two metabolic genes (pdk4 and ucp3), as well as the nuclear receptor pparα; primer and probe sequences for these assays have been published previously (42, 43). Gene expression data are represented as mRNA molecules (per ∼20 ng total RNA in the case of cultured cardiomyocytes or per 50 ng total RNA in the case of intact rat hearts). Because of the lower yield of total RNA for isolated cardiomyocytes versus intact rat hearts, accuracy in spectrophotometric determination of RNA concentration is reduced; typically 20 ng total RNA are used for each cardiomyocyte gene expression measurement.

Statistical analysis.

Periodicity analysis was performed in R stat analysis environment (http://lib.stat.cmu.edu/R/CRAN/) using nonlinear least-squares (nls) estimation. Before the fitting, all gene expression data were normalized to remove linear time trends and global shifts. For all experimental versus control data, a simple waveform was fit: Y(t) = A·cos (t/T·2π + P). For 60 h data, we included a decay term Y(t) = exp(-BtA cos (t/T + P), where Y is normalized expression value, t is time in hours, T is period (estimated by nls), P is phase shift (estimated by nls), and −B is decay term. A likelihood ratio test was performed to determine the significance of each fit for each gene in each experiment compared with the null hypothesis of no time pattern (a flat line). We call a periodic fit significant if 1) the likelihood ratio test P value was <0.05 and 2) the goodness of fit (R^2) was >0.2. For treatment versus control experiments, we also performed a comparative analysis on the coefficients resulting from the significant fits. In each case (each treatment), we performed a two-way ANOVA fitting a gene effect and an experiment effect for each of the outcomes: amplitude, period, and R^2. This analysis was restricted to the genes with significant periodic fits only (i.e., data without significant fits were excluded).

RESULTS

Serum influences circadian clock gene oscillations in adult rat cardiomyocytes.

The effect of a standard concentration of FCS (2.5%) on circadian clock gene expression in adult rat cardiomyocytes was investigated. After isolation, cardiomyocytes were divided into two groups, control and FCS. After plating was completed, control cardiomyocytes were cultured in the absence of FCS (i.e., with no FCS present in equilibration, challenge, wash, or postchallenge media). In contrast, 2.5% FCS was present in all media for FCS cardiomyocytes after plating. Figure 2 shows that culturing cardiomyocytes in medium supplemented with 2.5% FCS (i.e., FCS cardiomyocytes) results in high-amplitude gene expression oscillations for all four clock genes investigated, with periodicities between 20 and 24 h. In contrast, significant oscillations in these genes were either severely attenuated (bmal1, dbp; P < 0.05 for control vs. 2.5% FCS) or absent (rev-erbaα, per2) for cardiomyocytes cultured in the absence of FCS (i.e., control cardiomycytes; Fig. 2).

Fig. 2.

Gene expression of brain and arylhydrocarbon receptor nuclear translocator-like protein 1 (bmal1; A), reverse strand of the c-erbaα gene (rev-erbaα; B), period 2 (per2; C), and albumin D-element binding protein (dbp; D) for adult rat cardiomyocytes cultured in the absence (control) or presence of 2.5% FCS. In the latter case, 2.5% FCS was present in equilibration, challenge, wash, and postchallenge media. Values are means ± SE for three separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.01) oscillations are observed in cells cultured in the presence of 2.5% FCS for all four genes investigated; significant (P < 0.05) oscillations are observed in control cells for only bmal1 and dbp. FCS (2.5%) significantly (P < 0.05) increased oscillations in bmal1 and dbp compared with control cells.

It may be hypothesized that the severe attenuation of gene expression oscillations observed for control cardiomyocytes in Fig. 2 was due to irreversible impairment of the circadian clock under these particular culture conditions (i.e., absence of FCS). To rule out this possibility, cardiomyocytes were acutely challenged (2 h) with a high concentration of FCS (50%). This strategy, known as serum shock, has been utilized previously for the reactivation of the circadian clock in cultured cells (3). Thus, after isolation, cardiomyocytes were divided into two groups, control and shock. After plating was completed, control cardiomyocytes were cultured in the absence of FCS (i.e., with no FCS present in equilibration, challenge, wash, or postchallenge media). In contrast, 50% FCS was present in the challenge medium for shock cardiomyocytes (i.e., with no FCS present in the equilibration, wash, or postchallenge media). Figure 3 illustrates that an acute challenge with FCS is able to significantly reestablish rhythmic expression of all four circadian clock genes investigated.

Fig. 3.

Serum shock reactivates the circadian clock within adult rat cardiomyocytes. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) was measured for control and serum-shocked (shock) adult rat cardiomyocytes. FCS was omitted from equilibration, wash, and postchallenge media. In contrast, 50% FCS was present in challenge medium for shock cells. Serum was omitted from the challenge medium for control cells. Values are means ± SE for six separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.001) oscillations are observed in cells challenged with serum shock for all four genes investigated; significant (P < 0.001) oscillations are observed in control cells for only bmal1 and dbp. Serum shock significantly (P < 0.05) increased oscillations in bmal1 and dbp compared with control cells.

After the observation that serum maintains circadian clock gene oscillations in adult rat cardiomyocytes, we next decided to investigate the duration of these rhythms in vitro. Here, the culture conditions described for those experiments presented in Figs. 2 and 3 were combined, such that the equilibration, wash, and postchallenge media were all supplemented with 2.5% FCS, whereas the challenge medium was supplemented with 50% FCS. The data presented in Fig. 4 show that rhythmic expression of the three circadian clock components (bmal1, rev-erbaα, and per2) and the clock-regulated gene dbp persisted significantly for the duration of the experiment (i.e., 60 h), with 20- to 24-h oscillations (period lengths are 21.4, 20.6, 22.9, and 20.5 h for bmal1, rev-erbaα, per2, and dbp, respectively).

Fig. 4.

Duration of circadian clock gene oscillations in cardiomyocytes after serum shock. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) was measured for serum-shocked adult rat cardiomyocytes. FCS (2.5%) was present in equilibration, wash, and postchallenge media. In addition, 50% FCS was present in challenge medium. Values are means ± SE for six separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.001) oscillations are observed in cells challenged with serum shock for all four genes investigated. These oscillations persisted significantly for 60 h.

Comparable temporal relationship of clock gene oscillations in vitro and in vivo.

Although the data presented in Figs. 24 demonstrate an intrinsic circadian clock within adult rat cardiomyocytes, it is important to show that this mechanism operates in a manner similar to that observed in vivo. We therefore investigated rhythmic expression of the three core circadian clock components (bmal1, rev-erbaα, and per2), as well as the clock-regulated gene dbp, in the intact rat heart and compared the timing of these gene expression oscillations to those observed in vitro (Fig. 4). Figure 5 illustrates circadian clock gene oscillations in the intact rat heart. Expression of bmal1 peaks at ∼ZT 2, followed by rev-erbaα at ZT 7, dbp at ZT 11, and per2 at ZT 15. Table 1 compares the temporal relationship between these circadian clock components in the intact rat heart versus the isolated adult rat cardiomyocytes. Both in vivo and in vitro, the peak expression of the genes investigated follows a specific order: bmal1, rev-erbaα, dbp, and per2 (Table 1).

Fig. 5.

Diurnal variations in circadian clock gene expression in the intact rat heart. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) was measured in rat hearts isolated at 3-h intervals. Values are means ± SE for between 19 and 22 separate observations at each time point. Data are represented as number of mRNA molecules per 50 ng total RNA.

View this table:
Table 1.

Comparison of temporal gene expression in isolated adult rat cardiomyocytes versus the intact rat heart

Glucose does not influence circadian clock in adult rat cardiomyocytes.

We investigated whether glucose acts as a zeitgeber for the circadian clock within the cardiomyocyte by determining whether glucose either initiates or causes a phase-shift in circadian clock gene oscillations. To determine whether glucose can reestablish circadian clock gene oscillations in isolated cardiomyocytes, cells were divided into two groups, low (5 mM) and high (15 mM) glucose. Serum was omitted from all media after plating, thereby minimizing baseline circadian clock gene oscillations (as observed in Figs. 2 and 3). All cells were incubated in the presence of 5 mM glucose during the plating and equilibration phases. For low-glucose cells, 5 mM glucose was present in challenge, wash, and postchallenge media. In contrast, for high-glucose cells, 15 mM glucose was present in challenge, wash, and postchallenge media. Figure 6 shows that the elevation of the glucose concentration (from 5 mM to 15 mM) had no effect on rhythmic expression of the circadian clock genes investigated.

Fig. 6.

Influence of glucose on circadian clock gene expression in isolated adult rat cardiomyocytes. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) for adult rat cardiomyocytes cultured in low (5 mM) or high (15 mM) glucose. In the latter case, 15 mM glucose was present in challenge, wash, and postchallenge media. For both groups, FCS was omitted from all media after plating was completed. Values are means ± SE for three separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.01) oscillations are observed in cells cultured with either 5 mM or 15 mM glucose for bmal1 and dbp only.

After the lack of effect of glucose on the reestablishment of circadian clock gene oscillations in adult rat cardiomyocytes, we next investigated whether glucose influenced the timing of serum-sustained circadian clock gene rhythms. Thus cardiomyocytes were again divided into two groups, low (5 mM) and high (15 mM) glucose. In this set of experiments, FCS was present in all media (following plating) at a concentration of 2.5%. All cells were incubated in the presence of 5 mM glucose during the plating and equilibration phases. For low-glucose cells, 5 mM glucose was present in challenge, wash, and postchallenge media. In contrast, for high-glucose cells, 15 mM glucose was present in challenge, wash, and postchallenge media. Figure 7 shows that elevation of the glucose concentration has no effect on FCS-sustained circadian clock gene oscillations.

Fig. 7.

Influence of glucose on serum-sustained circadian clock gene oscillations in isolated adult rat cardiomyocytes. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) for adult rat cardiomyocytes cultured in low (5 mM) or high (15 mM) glucose. In the latter case, 15 mM glucose was present in challenge, wash, and postchallenge media. For both groups, 2.5% FCS was present in all media after plating was completed. Values are means ± SE for three separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.01) oscillations are observed in cells cultured with either 5 or 15 mM glucose for all four genes investigated.

Norepinephrine reactivates circadian clock within adult rat cardiomyocytes.

We next investigated whether the sympathetic neurotransmitter norepinephrine acts as a zeitgeber for the circadian clock within the cardiomyocyte. Similar to the studies presented in Fig. 3 for serum-shock experiments, we investigated whether an acute exposure (2 h) to norepinephrine (10 μM) could reactivate the circadian clock; thus cardiomyocytes were divided into two groups, control and norepinephrine. After plating was completed, all cells were cultured in the absence of serum (i.e., no FCS in the equilibration, challenge, wash, or postchallenge media). During the challenge period, cardiomyocytes in the norepinephrine group were cultured in the presence of 10 μM norepinephrine. This challenge with norepinephrine resulted in significant circadian clock gene oscillations in adult rat cardiomyocytes (Fig. 8).

Fig. 8.

Norepinephrine (NE) reactivates the circadian clock within adult rat cardiomyocytes. Gene expression of bmal1 (A), rev-erbaα (B), per2 (C), and dbp (D) was measured for control and NE-stimulated adult rat cardiomyocytes. FCS was omitted from equilibration, wash, and postchallenge media. NE (10 μM) was present in challenge medium for NE cells. NE was omitted from challenge medium for control cells. Values are means ± SE for six separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA. Significant (P < 0.001) oscillations are observed in cells cultured challenged with norepinephrine for all four genes investigated; no significant oscillations were observed for control cells.

Oscillation of metabolic genes in isolated adult rat cardiomyocytes.

We next investigated whether metabolic gene oscillations observed in the intact rat heart persist in adult rat cardiomyocytes cultured under conditions associated with serum-mediated establishment of the circadian clock. Expression of both pdk4 and ucp3 was therefore measured for those samples obtained from the experiments presented in Figs. 2 and 4. Continuous culture of cardiomyocytes with 2.5% FCS is sufficient to maintain rhythmic expression of pdk4 and ucp3 with periodicities of 20–24 h (Fig. 9, A and B). In contrast, cardiomyocytes cultured in the absence of serum exhibit dampened oscillations in these metabolic genes (Fig. 9, A and B). Figure 9, C and D, shows that serum shock, in combination with continuous 2.5% FCS exposure, maintains pdk4 and ucp3 oscillations for ∼36-h postshock. Table 1 compares the temporal relationship between these metabolic genes in the intact rat heart (Fig. 9, E and F) versus isolated adult rat cardiomyocytes (Fig. 9, C and D), relative to the circadian clock components (Figs. 4 and 5). Both in vivo and in vitro, the peak expression of the genes investigated follows a specific order: bmal1, rev-erbaα, dbp, per2, pdk4, and ucp3 (Table 1).

Fig. 9.

Metabolic gene expression oscillations in isolated adult rat cardiomyocytes and intact rat hearts. Gene expression of pdk4 and ucp3 was measured for control vs. 2.5% FCS continuous exposure (A and B) and 2.5% FCS continuous exposure plus serum shock (50% FCS for 2 h; C and D). Diurnal variations in pdk4 and ucp3 were also measured in the intact rat heart (E and F). Values are means ± SE for between 3 and 22 separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA for isolated cardiomyocytes and mRNA molecules per 50 ng total RNA for intact rat hearts. Significant (P < 0.001) oscillations are observed in cells cultured with serum (both acutely and chronically) for both pdk4 and ucp3.

Given that both pdk4 and ucp3 are known PPARα-regulated genes and that the pparα gene appears to be modulated directly by the circadian clock within the liver (24, 38, 41), we investigated whether this nuclear receptor exhibited oscillations in expression in adult rat cardiomycytes cultured under conditions associated with serum-mediated establishment of the circadian clock. Expression of pparα was therefore measured for those samples obtained from the experiment presented in Fig. 4. Figure 10 shows a lack of circadian oscillation in pparα expression for cardiomyocytes subjected to serum shock (in combination with continuous 2.5% FCS exposure).

Fig. 10.

Lack of circadian oscillations in pparα gene expression in isolated adult rat cardiomyocytes. Gene expression of pparα was measured for serum-shocked adult rat cardiomyocytes. FCS (2.5%) was present in equilibration, wash, and postchallenge media. In addition, 50% FCS was present in challenge medium. Values are means ± SE for six separate observations at each time point. Data are represented as number of mRNA molecules per ∼20 ng total RNA.

DISCUSSION

The present study reports circadian clock gene oscillations in cultured isolated adult rat cardiomyocytes, thereby demonstrating the intrinsic nature of this molecular mechanism within the myocytes of the heart. Standard culture conditions utilized by cardiovascular researchers (i.e., DMEM containing 2.5% FCS) are sufficient to maintain circadian clock gene oscillations. Our data also suggest that norephinephrine acts as a zeitgeber for the circadian clock within the cardiomyocyte and that the metabolic genes pdk4 and ucp3 are novel circadian clock-regulated genes in the heart.

Characterization of circadian clock within cardiomyocytes.

Circadian clocks confer the selective advantage of anticipation, allowing a cell to anticipate changes in its environment over the course of the day (7). Much progress has been made in the identification of the components of the circadian clock in Drosophila, wherein several of the key proteins have been characterized (10, 11). Moreover, mammalian orthologs of circadian clock components are emerging (9, 32, 36, 46). Of these, circadian locomotor output cycles kaput (CLOCK) was the first identified by screening offspring for abnormalities in behavioral circadian rhythms after treatment of male mice with N-ethyl-N-nitrosourea (i.e., chemical mutagenesis) (36). The core mammalian clock machinery consists of at least seven distinct proteins: BMAL1, CLOCK, REV-ERBAα, PER1, PER2, cryptochrome (CRY) 1, and CRY2. BMAL1 (also known as MOP3) and CLOCK are basic helix-loop-helix/PER-ARNT-SIM transcription factors that form a transcriptionally active complex upon heterodimerization (9, 14). This heterodimer binds to cis-acting elements (E-boxes) within the promoter of various target genes, including rev-erbaα, per1, per2, per3, cry1, cry2, multiple clock-controlled output genes, as well as bmal1 itself (25, 28, 40).

Several studies have now reported rhythmic expression of circadian clock genes in the intact rodent heart (22, 27, 30, 43, 44). By definition, the circadian clock is intrinsic to the cell, and as such, should persist when cells are isolated and cultured. One could hypothesize that oscillations in circadian clock components observed in intact organs in vivo are the result of diurnal variations in one or more neurohumoral factors that influence expression of these genes. Therefore, to conclusively demonstrate that the cardiomyocyte possesses an intrinsic circadian clock mechanism, one must expose its cell autonomous nature through the use of cultured organs or cells. The latter is feasible for the heart.

The present study reports that oscillations in circadian clock components persist in isolated adult rat cardiomyocytes, when these cells are cultured in the presence of a standard concentration of FCS (2.5%; Figs. 2 and 7). Furthermore, acute exposure (2 h) of cardiomyocytes to a high concentration of FCS (50%; known as serum shock) augments rhythmic expression of circadian clock genes (Fig. 4). Oscillations in circadian clock components exhibited periodicities of ∼20 to 24 h and persisted for at least 60 h (Fig. 4). Furthermore, the temporal relationship for the peak level of expression among the various clock components in isolated cardiomyocytes is very similar to that observed in the intact rat heart (Table 1). Taken together, these observations expose the intrinsic nature of the circadian clock within the cardiomyocyte and suggest that one or more neurohumoral factors in serum are required for the initiation and maintenance of this molecular mechanism.

When adult rat cardiomyocytes are cultured in the absence of FCS, oscillations in circadian clock components appear to be severely diminished (Figs. 2, 3, 6, and 8). It could be hypothesized that the attenuation of circadian clock gene oscillations for cardiomyocytes cultured in the absence of serum is due to an irreversible impairment of this molecular mechanism. To test this hypothesis, cardiomyocytes previously cultured in the absence of serum were challenged with a serum shock (i.e., 50% FCS for 2 h). This serum shock resulted in the reestablishment of circadian clock gene oscillations (Fig. 3), suggesting that the circadian clock can be reactivated under these conditions. The observation that circadian clock gene oscillations can be reversibly attenuated in cardiomyocytes provides an ideal model system within which potential zeitgebers for the circadian clock within the heart can be investigated.

It could be hypothesized that gene expression rhythms observed in isolated cardiomyocytes are due to the preservation of in vivo circadian clock gene oscillations through the isolation and culturing protocol. Oscillations in bmal1 routinely peaked at ∼6 h postchallenge in our in vitro experiments. This is at the end of the subjective day (i.e., ZT 10 or 5 PM). However, at ZT 10, bmal1 expression approaches its trough value in the intact rat heart. The antiphase nature of circadian clock gene oscillations in isolated cardiomyocytes versus intact rat hearts exposes a genuine reestablishment of this molecular mechanism in vitro. Such a reestablishment may have occurred during the plating phase of our protocol when all cardiomyocytes are initially exposed to 4% FCS for 2 h. Indeed, the modest circadian clock gene oscillations observed for cardiomyocytes cultured in the absence of serum after plating (e.g., bmal1 and dbp) are likely due to residual clock gene oscillation reestablishment during the plating phase. The presence of 2.5% FCS therefore likely provides the appropriate conditions for maintained circadian clock gene oscillations. Dampening of circadian clock gene oscillations in the absence of serum may be due to inactivation of the clock mechanism and/or dyssynchronization among cell-autonomous clocks within individual cells in culture. In the latter case, summation of circadian clock gene oscillations when individual cells in culture are asynchronous would be perceived as a loss of rhythmicity. Welch et al. (37) and Nagoshi et al. (21) have both recently reported that the culturing of fibroblasts for prolonged periods of time leads to asynchrony among individual cell-autonomous clocks and that serum reestablishes synchrony, as opposed to the reactivation of the clock mechanism itself.

Zeitgebers for circadian clock within cardiomyocytes.

Zeitgebers, or timekeepers, are factors that influence the timing of the circadian clock. The primary zeitgeber for the central clock located within the suprachiasmatic nucleus is light, whereas neurohumoral factors influence peripheral circadian clocks (6, 7). The physiological zeitgeber(s) for the circadian clock within the cardiomyocyte is (are) currently unknown. The present study clearly shows that, by definition, serum can be considered as a zeitgeber for the circadian clock within the cardiomyocyte. This is true also for various cell types in culture, in which serum shock has been shown to reactivate the clock mechanism (3). Serum contains numerous potential factors (e.g., cortisol, retinoic acid, and angiotensin II) that have been shown to act as zeitgebers for the circadian clock within other (i.e., noncardiomyocyte) cell types (2, 20, 23). In the present study, we have investigated whether glucose and/or norepinephrine act as physiological zeitgebers for the circadian clock within the heart.

Diurnal variations in circulating glucose levels are well documented. In addition to transient increases after a carbohydrate-rich meal, circulating glucose levels increase just before awakening of mammals (5, 17, 33). This phenomenon, the so-called dawn phenomenon, appears to be under the control of the central circadian clock (18). When investigating putative zeitgebers for rat-1 fibroblasts, Hirota et al. (13) observed that the addition of glucose to the culture medium induced rhythmic expression of several circadian clock components. Although no mechanism was identified, these data strongly suggest that glucose is a potential zeitgeber for peripheral circadian clocks. In addition, our laboratory (44) has previously reported that uncontrolled streptozotocin-induced diabetes mellitus causes a phase shift in the circadian clock within the heart. We therefore postulated that glucose may act as a physiological zeitgeber for the circadian clock within the cardiomyocyte. However, the data reported in the present study are not consistent with this hypothesis; elevation of the concentration of glucose in the culture medium (from 5 mM to 15 mM, representing the extremes of physiological diurnal variations in blood glucose levels) has no effect on circadian clock gene oscillations, either in the absence (Fig. 6) or presence (Fig. 7) of serum. Whether chronic hyperglycemia in vivo contributes toward the phase advance observed during diabetes through glucotoxicity mechanisms cannot be ruled out. Indeed, the present study deliberately avoided supramaximal glucose concentrations (i.e., >20 mM) that are associated with glucotoxicity in cardiomyocytes (8).

Diurnal variations in the activity of the sympathetic nervous system are well documented and likely comprise the major driving force for diurnal variations in multiple cardiovascular parameters, such as blood pressure, heart rate, and cardiac output (26, 34, 35, 45). Norepinephrine is a major sympathetic neurotransmitter. Given the dramatic diurnal variations in sympathetic activity (and plasma norepinephrine levels) and the strong influence of this stimulus on cardiovascular physiology and pathophysiology, we hypothesized that norepinephrine may act as a zeitgeber for the circadian clock within the cardiomyocyte. Our data show (Fig. 8) that acute (2 h) exposure of cardiomyocytes to norepinephrine (10 μM), thereby mimicking a burst of sympathetic activity before arousal of the organism, induced oscillations in the three circadian clock components investigated (bmal1, rev-erbaα, and per2), as well as the circadian clock-regulated gene dbp, consistent with our hypothesis that this neurotransmitter acts as a significant zeitgeber for the circadian clock within the heart. As cardiomyocytes were cultured in the absence of serum in this experiment, norepinephrine-induced clock gene oscillations are not unexpectedly short lived.

Metabolic genes as a novel output from the circadian clock within the cardiomyocyte.

Our laboratory (29, 30, 42) has shown previously that the intact rat heart exhibits dramatic diurnal variations in multiple metabolic genes. The most striking rhythms were observed for PPARα-regulated genes, such as pdk4 and ucp3. The circadian clock directly regulates pparα expression in the liver (24). Furthermore, the circadian clock component REV-ERBAα attenuates PPARα-mediated transcription (16). Consistent with their antagonist function, rhythms in pparα and rev-erbaα expression are antiphase in the intact rat heart (30). These observations lead us to hypothesize that the circadian clock within the cardiomyocyte mediates diurnal variations in myocardial pdk4 and ucp3 expression through PPARα and/or REV-ERBAα (30). Consistent with this hypothesis, we find that the reestablishment of circadian clock gene oscillations within isolated adult rat cardiomyocytes is associated with the induction of pdk4 and ucp3 rhythmic expression (Fig. 9). Rhythms in pdk4 and ucp3 mRNA were in similar phases to one another and were antiphase with respect to rev-erbaα mRNA (Figs. 9, C and D, and 4B, respectively). In contrast, significant oscillations in pparα mRNA were not observed, suggesting a potential greater role for rev-erbaα as a mediator of pdk4 and ucp3 rhythmicity (Fig. 10). As for the circadian clock components, rhythmic expression of pdk4 and ucp3 in isolated cardiomyocytes was temporally similar to that observed in the intact rat heart (Table 1). These data strongly support the hypothesis that the circadian clock within the heart regulates myocardial metabolic gene expression.

Unlike oscillations of circadian clock components in cardiomyocytes after serum shock, rhythmic expression of pdk4 and ucp3 mRNA was observed for only the first 36 h, after which time expression of these metabolic genes declined steadily. Although the exact mechanism responsible for suppression of these rhythms is currently unknown, several possibilities exist. Decreased amplitude of clock gene rhythms (i.e., peak minus preceding trough) as the culture continues likely affects downstream clock-controlled genes (e.g., pdk4 and ucp3) before the clock mechanism itself. In addition, given the decreased energetic demands on the cardiomyocyte in vitro compared with the intact heart in vivo, it is not surprising that the expression of genes involved in ATP generation declines. Furthermore, adult cardiomyocytes are known to become increasingly fetal-like as culture duration continues (i.e., dedifferentiation). PPARα-regulated genes, such as pdk4 and ucp3, are expressed at low levels in the fetal heart.

Implications of this study.

The data reported within this study have multiple implications at both the bedside and benchside levels. Several studies have reported that diurnal variations in cardiovascular parameters are controlled by an unidentified intrinsic mechanism in humans (12, 15). Two major possibilities therefore exist: 1) the central circadian clock may regulate diurnal variations in sympathetic activity and/or other neurohumoral influences, which in turn modulate cardiovascular function; and/or 2) the circadian clock within the cardiomyocyte may influence responsiveness of the heart to diurnal variations in its environment (e.g., sympathetic activity), thereby influencing diurnal variations in cardiovascular function. The latter hypothesis becomes tenable only after our characterization of the intrinsic nature of the circadian clock within the cardiomyocyte. Through the use of this in vitro system, multiple putative zeitgebers for the circadian clock within the heart can be investigated, which in turn will improve our understanding of not only cardiovascular physiology but also pathophysiology. Indeed, given the identification of norepinephrine as a zeitgeber for the circadian clock within the cardiomyocyte, it could be hypothesized that this molecular mechanism will be abnormal under conditions associated with chronic alterations in sympathetic activity, such as hypertension, diabetes, obesity, shift work, and sleep apnea. Indeed, our laboratory (43, 44) has found alterations in the circadian clock of the heart in rodent models of diabetes and hypertrophy.

Many investigators utilize the cell culture system in an attempt to maintain a controlled set of conditions. Our data clearly show that the presence of a standard concentration of serum within the culture medium (i.e., 2.5% FCS) is sufficient to maintain a fully functional circadian clock. This becomes problematic when the process under investigation is regulated in some manner by the circadian clock within the cardiomyocyte. For example, DBP is a circadian clock-regulated transcription factor of unknown function in the heart. When the circadian clock is active, dbp oscillates. If an investigator were to challenge cells with a factor that reactivates or augments the circadian clock within the cardiomyocytes, then one may falsely conclude that that factor induces or represses dbp expression, depending upon the length of time after the challenge is initiated (see Figs. 2 and 8, for serum and norepinephrine, respectively). It is therefore important for an investigator to establish that the circadian clock is not active or synchronized between the cultured cells (e.g., culture in the absence of serum for prolonged periods of time) and/or to ensure that the intervention does not modulate the circadian clock, which in turn may influence the conclusions drawn.

In conclusion, the present study has exposed the intrinsic nature of the circadian clock within the cardiomyocyte. Culturing adult rat cardiomyocytes in the presence of a standard concentration of FCS (2.5%) is sufficient to maintain circadian clock gene oscillations. Our data suggest that norepinephrine, unlike glucose, is a zeitgeber for the circadian clock within the cardiomyocyte. Furthermore, we suggest that the metabolic genes pdk4 and ucp3 are novel circadian clock-regulated genes in the heart.

GRANTS

This work was supported by the American Heart Association Texas Affiliate Beginning Grant-In-Aid Award 0365028Y and by National Heart, Lung, and Blood Institute Grant HL-074259–01. E. D. Abel is an established investigator of the American Heart Association.

Acknowledgments

We thank Dr. William C. Stanley for constructive comments before submission.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract