Effect of combination of residual glucose concentration and subsequent increment by temporal glucose feeding on oscillation of clock gene Per2 expression

Abstract

With the aim of regulating clock gene expression to control cell activities in cell processing engineering, the effect of the combination of residual glucose concentration and subsequent increment by temporal glucose feeding on the oscillation of the expression of clock gene Per2 was investigated employing rat Mesenchymal stem cell (MSC)-like cells having Per2 promoter gene with a destabilized luciferase gene (Per2-dLuc). Two experiments with several initial glucose concentrations and different times of cultures (2 and 5 days) before temporal glucose feeding (0.9 g/L) were employed to realize various concentrations of residual glucose in the medium before the feeding. In these experiments, the lower residual glucose concentrations (0.002–0.02 g/L) before temporal glucose feeding tended to induce the larger amplitude of oscillation of Per2 expression than the higher ones (0.55–0.74 g/L). When the residual glucose concentration before glucose feeding was low (0.014–0.038 g/L), the higher temporal glucose concentration (0.23–0.9 g/L) feeding tended to induce the larger amplitude of oscillation of Per2 expression than the lower ones (0.012–0.023 g/L). Taken together, we found that the amplitude of oscillation of the expression of clock gene Per2 could be controlled by the combination of residual glucose concentration and glucose concentration of subsequent temporal feeding.

Introduction

In mammals, the time-keeping system is composed of a central clock in the superchiasmatic nucleus (SCN) that also coordinates the peripheral clocks in extensive peripheral tissues (Reppert and Weaver 2002; Yoo et al. 2004; Yamamoto et al. 2004). Even cultured fibroblasts have an active circadian clock (Fig. 5A), the molecular mechanism of which strikingly resembles that of the central clock in the SCN (Balsalobre et al. 1998; Yagita et al. 2001; Nagoshi et al. 2004, 2005; Welsh et al. 2004). Moreover, immortalized cultured cell lines such as Rat-1 (Balsalobre et al. 1998) and NIH3T3 (Akashi and Nishida 2000) fibroblasts have been found to produce a circadian rhythm and these cells thus provide a model system to dissect the molecular mechanisms underlying circadian oscillation.

Fig. 5.

Effect of added glucose concentration on amplification rate of oscillation. A Detrended oscillation of luminescence intensity after glucose addition at different concentrations of 0.9 (green), 0.23 (orange), 0.023 (yellow), 0.012 (blue), and 0 g/L (black, water) (n = 1). B Relationship between added glucose concentration and amplification rate (n = 2)

To generate the circadian rhythm in gene expression, mammalian clock genes form transcription-/translation-based negative feedback loops (Fig. 6) (Reppert and Weaver 2002; Lowrey and Takahashi 2004). In the main molecular loop, the bHLH-PAS transcription factors of CLOCK and BMAL1 dimerize to activate the transcription of Per and Cry through their bonding to E box element. Then, accumulated PER and CRY proteins negatively regulate the CLOCK-BMAL1-mediated transactivation of E-box-element-dependent Per and Cry transcription. The negative regulation of Per and Cry transcription leads to the reduction in the level of PER and CRY proteins. This event promotes the CLOCK-BMAL1 dimer to reactivate the transcription of Per and Cry. The turnover of these negative regulators, PER and CRY proteins, allows circadian oscillations of Per and Cry mRNA expression levels.

Fig. 6.

Proposed mechanism of oscillation of Per2 expression by added glucose. The main negative feedback loop of mammalian clock genes including Per2 (green area) and the proposed mechanism of oscillation of Per2 expression through Tieg1 activation by added glucose. Boxes, wavy lines, and ovals represent genes, mRNAs, and proteins, respectively. Dashed lines represent predicted relations. The proposed Tieg1-related mechanism was illustrated by the modification of Fig. 5D in the report by Hirota (Hirota et al. 2010)

Mammalian cell culture has a very important role in the manufacture of biopharmaceuticals, especially therapeutic antibodies, and regenerative medicine. Glucose is a main key element in a medium for cell culture because cells utilize glucose as a major carbon source and an energy source for their proliferation. Thus, glucose is not only an essential component of the initial medium but is also often added during the culture to prevent its depletion and maintain its concentration in culture at optimum levels.

In cultured cells, it is known that total observable expression of a clock gene by cell population shows no oscillation (Fig. 1). This is explained that its expression by an individual cell has different phase of oscillation each other in cells. However, it is reported that glucose addition induces the oscillation of total expression of the clock gene. This induction was speculated to be caused by the resetting of all possible phases in the oscillations of individual cells in a population to the same phase (Fig. 1) (Nagoshi et al. 2004; Welsh et al. 2004).

Fig. 1.

Schematic diagram of the synchronization model for the observable oscillation of Per2 expression by glucose addition

Hirota et al. have reported that glucose addition to cell culture induced the oscillation of the expression of clock genes Per1, Per2, and Bmal1 in fibroblasts (Fig. 6) (Hirota et al. 2002). Moreover, they have shown that the transcription factor TIEG1 might be involved in this induction (Hirota et al. 2002). Thus, an increment in glucose concentration in culture may activate Tieg1 expression through carbohydrate response element (ChoRE) in Tieg1 (Hirota et al. 2010).

Tieg1 expression may lead to the suppression of Bmal1 expression and the subsequent down-regulation of Per2 expression, which may be an initiator of the normal oscillation of Per2 expression. If so, the change in glucose concentration before and after the temporal glucose addition during cell culture may affect the amplitude of oscillation.

With the aim of regulating clock gene expression to control cell activities in cell processing engineering, the effect of the combination of residual glucose concentration and subsequent increment by temporal glucose feeding on the oscillation of the expression of clock gene Per2 was investigated.

Materials and methods

Cells

Transgenic Wister rats carrying a bioluminescence reporter of Per2 expression were used, in which a destabilized luciferase (dLuc) reporter gene was expressed under the control of the mouse Per2 promoter (Per2-dLuc rats) (He et al. 2007). Per2-dLuc rats were reared in our animal quarters where environmental conditions were controlled (12-h light and 12-h dark with lights on at 06:00–18:00) (Nishide et al. 2014). Animals were cared for according to the “Guidelines for the Care and Use of Laboratory Animals in Hokkaido University” with permission No. 08-0250 from the Committee for Animal Experimentation in the Hokkaido University.

The rats were decapitated in the middle of the light period and their femurs were removed. Bone marrow cells were flushed from the femur into the medium using a 25-gauge needle. Bone marrow aspirates from different rats were pooled to provide a uniform inoculum. After cell debris was separated by sedimentation, the number of nucleated cells was determined using the Turk solution (Nacalai Tesque, Kyoto, Japan), a hemacytometer, and an inverted microscope.

Bone marrow aspirate was diluted with a growth medium (described in Media below), plated on a dish (35 mm, Corning, NY) to a concentration of 1.4 × 10⁶ nucleated cells/cm², and cultured at 37 °C in a humidified atmosphere containing 5% CO₂, during which the medium was changed to a flesh one on days 1, 2, and 6. On day 10, the cells were detached using trypsin–EDTA (Sigma, St. Louis, MO) and inoculated again to the same dish. When the culture reached subconfluence, the cells were detached using trypsin–EDTA, counted by the trypan blue dye-exclusion method, and stored in the medium containing 10% DMSO in liquid nitrogen.

Media

DMEM-LG (Gibco, Carlsbad, CA, USA) supplemented with 10% FCS (Invitrogen, Carlsbad, CA), 2500 U/L penicillin, and 2.5 mg/L streptomycin was used as the growth medium. The glucose concentration was 0.9 g/L, unless otherwise stated. d-luciferin potassium salt (100 μM, Wako, Osaka, Japan) was added to the growth medium before use for bioluminescence measurement.

Determination of glucose concentration

The concentration of glucose added to the culture and the residual glucose concentration in the culture were determined by the glucose oxidase method using Glucose CII-test Wako (FUJIFILM Wako Pure Chemicals, Osaka, Japan).

Bioluminescence measurement and glucose addition

Cells were inoculated to a 35 mm dish with the growth medium containing 10% FBS at a density of 1.0 × 10⁴ cells/cm² and cultivated for 3 days in a CO₂ incubator (5% CO₂, 37 °C) (MCO-18 M, Sanyo, Tokyo, Japan). Then, the medium was changed to the growth medium containing 10% FBS (0.9 g/L glucose), unless otherwise stated, and D-luciferin, the time at which was defined as t = 0 (Fig. 2). The oscillation of luminescence intensity starts immediately after this medium change (MC) at t = 0 (Fig. 2). The 35 mm dishes were incubated in a dish-type luminometer equipped with an incubator (5% CO₂, 37 °C) (Kronos Dio AB-2550, ATTO, Tokyo, Japan), and bioluminescence intensity was measured every 10 min to monitor Per2 expression during incubation. On day 5 or 2 after the MC, 10 μl of glucose solution was added to the culture to increase the glucose concentration to the desired level (0.012–0.9 g/L).

Fig. 2.

Schematic diagram of oscillation analysis. A representative detrended luminescence oscillation is shown. After luminescence oscillation that occurred upon medium change (MC) at t = 0 was progressively reduced, oscillation occurred again after glucose addition. A_2/1 is the amplification rate of oscillation amplitude. The amplitudes of the latest oscillation before the addition and the earliest oscillation after the addition were ℓ₁ and ℓ₂, respectively. The dotted line indicates the unavailable moving average immediately after MC and glucose addition

Data analysis

Bioluminescence intensity data were detrended to eliminate baseline drifting by subtracting the moving average of 24 h from the raw data. Detrended data were further smoothed by five-point moving averaging (Nishide et al. 2014). A representative time course of detrended luminescence intensity data is shown in Fig. 2. The first luminescence oscillation caused by MC at time 0 was observed. After the reduction in the amplitude of the first oscillation over time, the second oscillation was observed immediately after glucose addition. The amplification rate (A_2/1) of two oscillation amplitudes immediately before (ℓ₁) and after (ℓ₂) glucose addition was defined as,

$${\text{A}}_{2/1}={\ell }_2/{\ell }_1$$

The culture for luminescence intensity measurement was performed in duplicate, which showed the same tendency and the representative result is shown in Results.

Results

Effect of residual glucose concentration in culture before glucose addition on amplification rate of oscillation

To investigate the effect of the residual glucose concentration in culture before glucose addition on the amplification rate of oscillation, 0.9 g/L glucose was added to cultures having various residual glucose concentrations. To realize such conditions of cultures, two setups were employed in cultures using MSC-like cells.

In the first setup, initial glucose concentrations of four cultures at MC (t = 0) were set to 0.9, 1.35, 1.8, and 2.25 g/L, and the residual glucose concentrations after 5 days of MC were 0.002, 0.012, 0.55, and 0.74 g/L, respectively (Fig. 3A). In the second setup, two cultures were started at the same initial glucose concentration of 0.9 g/L at MC (t = 0). These two cultures were continued for 2 and 5 days, which showed the residual glucose concentrations of 0.69 and 0.02 g/L, respectively (Fig. 4A).

Fig. 3.

Effect of residual glucose concentration in parallel cultures on the amplification rate of oscillation. A Time course of glucose concentration in culture for 5 days. B Detrended oscillation of luminencence intensity before and after glucose or water addition. At 5 days after MC, glucose (0.9 g/L) or water was added to the cultures containing residual glucose at various concentrations of 0.002 (4, green), 0.012 (3, orange), 0.55 (2, yellow), and 0.74 g/L (1, blue) (n = 1). C Relationship between residual glucose concentration and amplification rate (n = 1)

Fig. 4.

Effect of residual glucose concentration on amplification rate of oscillation after glucose addition at different times. A Time course of glucose concentration in culture on days 2 (1, dotted green line) and 5 (2, dotted orange line). B Detrended oscillation of luminescence intensity before and after glucose or water addition. On days 2 and 5 after MC, glucose (0.9 g/L) or water was added to the cultures containing different residual glucose concentrations of 0.69 (green) and 0.02 g/L (orange) (n = 1). C Relationship between residual glucose concentration and amplification rate (n = 2)

In the first setup, 10 μL of glucose solution (0.9 g/L) or water was added to four cultures on day 5 (Fig. 3B). With the lower residual glucose concentration (Fig. 3A) before glucose addition on day 5, the amplitude of luminescence intensity and the amplification rate (A_2/1) after glucose addition were larger (Fig. 3B, C).

In the second setup, 10 μL of glucose solution (0.9 g/L) or water was added to the culture on days 2 and 5 (Fig. 4B). When the residual glucose concentration before glucose addition was lower on day 5 (0.02 g/L) than that on day 2 (0.69 g/L) (Fig. 4A), the amplitude of luminescence intensity and the amplification rate (A_2/1) after glucose addition were larger on day 5 (Fig. 4B, C).

In these two setups, it was found that the lower residual glucose concentrations (0.002–0.02 g/L) before temporal glucose feeding tended to induce the larger amplification rate of oscillation of Per2 expression than the higher ones (0.55–0.74 g/L).

Effect of concentration of added glucose on amplification rate of oscillation

To study the effect of the concentration of added glucose on the amplification rate of oscillation, glucose was added at various concentrations (0.012–0.9 g/L) to the culture at day 5 (Fig. 5). After 3 days of cultivation of MSC-like cells, the first oscillation of luminescence intensity was caused by MC at t = 0 (Fig. 5A). On day 5 after MC, the first oscillation was damped to a low level and the residual glucose concentrations were considerably low (0.014 – 0.038 g/L) (data not shown). Then, 10 μL of glucose solutions of various glucose concentrations were added to the culture to increase the glucose concentration of the culture in respective increment of 0, 0.012, 0.023, 0.23 and 0.9 g/L (Fig. 5A).

All glucose additions other than water addition clearly induced the oscillation of luminescence intensity (Fig. 5A). In Fig. 5A, the change of period length and phase shift of oscillation after glucose addition at different glucose concentration was analyzed. The added glucose concentration (0.012–0.9 g/L) did not affected on period length (25.5–28.3 h) after glucose addition. Additionally, the first, the second and the third peaks after glucose addition at various concentrations (0.012–0.9 g/L) respectively showed similar time (data not shown), indicating no phase shift occurred among different glucose concentrations.

As shown in Fig. 5B, the amplification rates of oscillation amplitude (A_2/1) at the added glucose concentration of 0, 0.012, 0.023, 0.23, and 0.9 g/L were 1.2, 5.8, 5.3, 8.4, and 12.1, respectively. Thus, there is an apparent positive correlation between the added glucose concentration and the amplification rate of oscillation caused by glucose addition.

Discussion

The amplitude of oscillation of Per2 expression induced by MC at t = 0 was rapidly decreased over several days (Figs. 3B, 4B, and 5A). As shown by the synchronization model mentioned in Introduction (Fig. 1), it has been reported that this reflects a loss of synchrony among individual cell oscillators (Nagoshi et al. 2004; Welsh et al. 2004). Briefly, although the phases of Per2 oscillations in most individual cells were synchronized upon MC at t = 0, the synchrony was progressively lost due to slightly different circadian cycles among cells. This is why the observable oscillation in a cell population was damped over time. The oscillation of Per2 expression was induced again by glucose addition after the damping (Figs. 3B, 4B, and 5A). This reinduction may be explained by the synchronization model in Fig. 1, in which the phases of Per2 oscillations of cells are aligned to the same phase (Nagoshi et al. 2004; Welsh et al. 2004).

In these reinduction of Per2 expression (Figs. 3B, 4B, and 5A), the change of osmotic pressure by glucose addition was possibly supposed to elicit the oscillation of Per2 expression. For example, 0.9 g/L of glucose addition increase the osmotic pressure in culture medium by 5 mOsm/L. However, we have confirmed that NaCl addition equivalent to the osmolarity increase by glucose addition has no effect on the bioluminescence rhythm (data not shown).

The amplification rate of oscillation of Per2 expression increased with increasing concentration of glucose added (Fig. 5A). On the basis of the synchronization model mentioned above, this result suggests that Per2 expression was reset and restarted in a larger part of the cell population or that phases of Per2 expression in individual cells in a population shifted closer to the same phase, when the intracellular glucose concentration increased largely.

Although the oscillation of Per2 expression upon glucose addition was observed in cultures with quite low residual glucose concentrations (0.002–0.02 g/L) (Figs. 4C and 5C), it did not occur when the residual glucose concentrations (0.55–0.74 g/L) were high (Figs. 3C and 4C). This trend was confirmed in both the first (Fig. 3) and second (Fig. 4) setups.

In the second setup (Fig. 4), the amounts of metabolites in culture secreted by cells, such as lactate, ammonia, and amino acid, might be different between days 2 and 5. Therefore, we investigated the effect of the metabolites accumulated in culture on the amplitude of oscillation of Per2 expression in an experiment in which glucose was added to the cultures containing low and high concentrations of the metabolites. However, we found no effect of the metabolites on the amplitude of oscillation of Per2 expression (data not shown).

As mentioned above, it was found that the oscillation of Per2 expression upon glucose addition occurred when the residual glucose concentration was very low. It was speculated by Hirota et al. that the added glucose activates Tieg1 gene expression through the binding of the transcription factor carbohydrate responsive element-binding protein (ChREBP) to ChoRE in Tieg1 (Fig. 6) (Hirota et al. 2002, 2010). Our results may be discussed on the basis of this speculation, namely, if the residual glucose concentration in the medium is very low, the increase in intracellular glucose concentration by glucose feeding may promote the binding of ChREBP to ChoRE and the activation of Tieg1 expression. On the other hand, when the residual glucose concentration is very high, it may be difficult for fed glucose to stimulate the ChREBP binding owing to the still high intracellular glucose concentration. To verify this hypothesis, the change of Tieg1 expression with different concentration of added glucose should be determined in future studies.

Glucose addition might affect not only Tieg1 activation but also glucose metabolism in cells. Putker et al. have reported that the amplitude and phase of PER protein expression is sensitive to redox balance and oxidative pentose phosphate pathway (PPP) included in glucose metabolic pathways (Putker et al. 2018). In their experiment, they changed the initial glucose concentration in culture medium to monitor Per2 expression. On the other hand, we changed the added glucose concentration during culture to do it. Since acute increase of glucose concentration during culture in our experiment might also have an influence on redox flux and PPP, we should investigate the change of redox balance and PPP by glucose addition in future. Moreover, although Putker et al. did not investigate the effect of the residual glucose concentration, we found that the residual glucose concentration also related to the amplitude of Per2 expression.

The expression of some genes correlated to cell division, carbohydrate metabolism, and lipid metabolism was reported to oscillate in accordance with the oscillation of clock genes expression such as Per and Cry (Akhtar et al. 2002). Genome-wide gene expression profiles in the SCN and various peripheral tissues have recently shown that hundreds of genes are expressed under the control of circadian regulation systems (Panda et al. 2002; Storch et al. 2002; Ueda et al. 2002; Duez and Staels 2010).

For example, some genes concerning to cell differentiation induction of stem cells such as MSCs might also oscillate at their expression level influenced by clock genes during differentiation step. These results might suggest the possibility of the control of cell differentiation yield by clock gene expression. If the expression peak times of differentiation genes in all cells in a population were synchronized via clock genes by glucose addition, we could expect more effective cell differentiation.

For example about cell division, the molecular component of the circadian clockwork can directly regulate the expression of the cell-division-related gene wee1 in a regenerating liver of mice (Matsuo et al. 2003). Thus, there may be the possibility to control the cell proliferation through clock gene expression by timing or amount of glucose feeding.

Thus, glucose may be one of the key tools to control the cell growth and differentiation utilizing clock genes.

Consequently, this study would be valuable as the early step for the control of clock gene expression and its utilizing for the control of cell activities.

In conclusion, we found that the amplification rate of oscillation of clock gene Per2 expression might be controlled by the combination of residual glucose concentration and subsequent increment by temporal glucose feeding.