Introduction

In mammals, Suprachiasmatic nucleus (SCN) contains the central clock that synchronizes physiology, behavior and metabolism to the external environmental cues (Zeitgebers) (Jagota 2012; Roenneberg and Merrow 2016). SCN regulates circadian rhythms by core clock genes viz. Clock, Bmal1, Periods, Cryptochromes, Rev-erbα, Rorα etc. whose expression is orchestrated at transcriptional and translational levels to establish compact feedback loops that eventually result in ~ 24 h periodicity (Jagota 2012; Takahashi 2017). Such circadian clock machineries are present in almost every other cell in mammals and are considered as peripheral clocks, which are capable of exhibiting tissue specific rhythms (Schibler et al. 2015) like immune responses, energy metabolism etc. but are under the regulation of central clock (Dibner et al. 2010). Misaligned circadian rhythms may lead to various physiological, metabolic and behavioral disorders (Hatori et al. 2017).

Peripheral clocks in immune cells (Arjona and Sarkar 2005) temporally gate immune responses. Various inflammatory genes such as Nfκb, Tnfα, Il6, toll like receptors (tlrs) etc. playing important role in inflammation and immune response are known to show rhythmic expression (Scheiermann et al. 2013). CLOCK was shown to elevate transcription of NFκB mediated immune responsive genes (Spengler et al. 2012), while PER1 and CRY1 were shown to inhibit the activation of NFκB (Sugimoto et al. 2014; Yang et al. 2015). Further, CLOCK/BMAL1 complex has been reported to bind to the promoter of Tlr9 to initiate its transcription (Silver et al. 2012). However, circadian regulation of immunity is cell and tissue specific. For example: recruitment of leukocytes to tissues is observed at zeitgeber time (ZT)-13 and to blood at ZT-5; Tlr9 shows peak expression at ZT-11 in macrophages, at ZT-19 in spleen, at ZT-15 to ZT-19 in B cells (Silver et al. 2012; Tognini et al. 2018). Moreover, it has been well-established that desynchronizing circadian clock can adversely modulate the immune responses (Curtis et al. 2014). But most of these unveiled interactions between clock and immunity were confined to immune cells.

It has been reported recently that in kidney Bmal1, Cry1, Cry2 play critical role in circadian regulation of blood pressure (Doi et al. 2010; Hara et al. 2017; Douma et al. 2018). Further, CLOCK/BMAL1 complex was shown to regulate Na+/H+ exchanger NHE3 (Saifur Rohman et al. 2005). PER1 has been reported to regulate expression of renal epithelial sodium channel αENaC (Gumz et al. 2009). Various sodium transporters, water channels, vasopressin receptors were also shown to be under circadian regulation (Stow and Gumz 2011). Aging, as a global process was reported to deregulate circadian pattern of blood pressure that may lead to adverse conditions (Hart and Charkoudian 2014). In addition, aging was reported to be associated with chronic inflammation (Bolignano et al. 2014; Xi et al. 2014). Further, with aging increased activity of NFκB had been attributed to chronic kidney diseases (CKD) (Chen et al. 2016). As there are increasing evidences on immune perturbations with aging, we wanted to understand age-associated crosstalk between circadian and immune systems in peripheral clock kidney.

Curcumin, an active constituent of turmeric is known for its pleiotropic properties like anti-inflammatory, anti-oxidant, anti-cancerous, anti-microbial, anti-aging etc. (Hewlings and Kalman 2017). We have reported recently differential chronobiotic property of curcumin on clock genes expression in aged rat SCN (Kukkemane and Jagota 2019). We here report chronobiotic role of curcumin on age-induced alterations of expression of clock and inflammatory genes in peripheral clock kidney.

Methodology

Animals

All the studies were done with male Wistar rats. The rats were individually housed in standard polypropylene cages and maintained at 23 ± 1 °C; relative humidity 55 ± 6%; with LD, 12:12 [lights on: 06:00 AM (Zeitgeber time (ZT)-0) and lights off: 6:00 PM (ZT-12)] for 2 weeks prior to experiments. Food and water was provided ad libitum. Cage changing was done at random intervals. Dim red light was used for handling the animals in the dark (Mattam and Jagota 2014). All the experiments were performed as per Institutional Animal Ethics (approval number: IAEC/UH/151/2016/05/AJ/P12/Rats Wistar/M-144 dated 16/06/2016).

Animals were divided into three age groups: Group A—3 months (m), Group B—12 m and Group C—24 m. Each group (n = 48) was subdivided into three groups (I) Control (C) (II) Vehicle treatment (VT) and (III) Curcumin treatment (CT) with n = 16 in each sub group.

Control: Group A (I), B (I) and C (I) animals (n = 48) did not receive any treatment.

Vehicle treatment: Group A (II), B (II) and C (II) animals (n = 48) were administered with 0.5% carboxy methyl cellulose (CMC) orally at ZT-11 for 15 days.

Curcumin treatment: 100 mg/ml w/v of curcumin (Sigma) was suspended in 0.5% CMC. Required amount of curcumin was mixed freshly with CMC and stirred for at least 30 min. For 15 days, Group A (III), B (III) and C (III) animals (n = 48) were administered with curcumin (300 mg/kg body weight) orally at ZT-11, since it showed differential restoratory properties in rat central clock SCN (Kukkemane and Jagota 2019).

Tissue preparation

Animals of group A (I, II, III), B (I, II, III) and C (I, II, III) were sacrificed at ZT-0, 6, 12 and 18 (n = 4 at each time point; n = 16 for four time points) on 16th day and kidneys were removed carefully and snap frozen in liquid N2. Tissues were stored at – 80 °C until further by use.

RNA extraction and cDNA synthesis

Whole tissues were grinded with liquid N2 and 50 mg (dry weight) was used for RNA isolation from each sample. RNA extraction was carried out using TRI reagent (Sigma) according to manufacturer’s protocol (Vinod and Jagota 2017). RNA was dissolved in 50 µl nuclease free water. Concentration and purity of extracted RNA were quantified by using Nano drop spectrophotometer (Thermo Fischer) (Chomczynski and Sacchi 2006). cDNA was synthesized using Bioline cDNA synthesis kit following manufacturer’s protocol. cDNA was diluted to 1:20 in RNase free water for the further studies (Mattam and Jagota 2014).

Quantitative reverse transcriptase PCR (qRT-PCR)

The expression of clock genes (rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα and rRorα) and immune genes (rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9) mRNA transcripts were quantified using qRT-PCR by the SYBR Green (Applied Biosystems, Foster, USA) detection method (Mattam and Jagota 2014). Details of the primers used are given in Table 1.

Table 1 Forward and reverse primers of various clock and immune genes
Full size table

Dissociation curves for all the genes studied showed a single peak (Supplementary Fig. 1) representing specific amplified target gene. Threshold cycle (Ct) values were obtained from the exponential phase of amplification plots. The relative quantitative expression of clock genes were obtained by normalizing each target gene expression with Cyclophilin A (ΔCt = target gene Ct—Cyclophilin A Ct) in each sample and used 2−ΔCt method for analysis (Mattam and Jagota 2014).

Data analysis

Statistical analysis: GraphPad Prism software was used for the data analysis. Multiple comparisons of four time points within each age group were analyzed by one way ANOVA followed by Post hoc Tukey’s test. Student’s t test was performed to compare the mean 24 h levels and daily pulse between respective vehicle treated groups and curcumin treated groups.

Pearson correlation analysis was performed using R-program (Kukkemane and Jagota 2019). Pair wise correlations were analyzed in light (ZT-0, 6, 12) and dark (ZT-12, 18, 24/0) phase separately among rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα, rRorα, rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9 genes in 3, 12 and 24 m vehicle treated (VT) and curcumin treated (CT) kidney samples.

Gene to gene network analysis: We used weighted correlation network analysis (WGCNA) data mining package in R program to understand the gene to gene network alterations with the aging and curcumin treatment and the network images were developed using ‘Cytoscape’ software.

Results

Aging differentially alters clock genes expression in kidney

In the present study, we have measured the mRNA expression of clock genes rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα, rRorα at four different time points ZT-0, 6, 12, 18 in three age groups 3, 12, 24 m controls (C) and VT animals. There was no change in expression of all the genes between controls and vehicle treated animals. In all the age groups studied, rBmal1 showed significant daily rhythms with peak at ZT-0 and trough at ZT-12. Mean 24 h levels did not show significant change among age groups studied. Daily pulse i.e. Max/Min ratio did not show significant change in 12 m, but showed a significant decrease (p < 0.05) in 24 m in comparison to 3 m rat kidney. rPer1 did not alter its expression pattern in all the age groups with maximum expression at ZT-12 and minimum expression at ZT-0. Mean 24 h levels were not significantly changed among all age groups. Daily pulse did not change in 12 m but showed significant increase (p < 0.05) in 24 m in comparison to 3 m rat kidney. rPer2 also expressed maximum at ZT-12 in all age groups but minimum at ZT-6 in 3 m and ZT-0 in both 12 and 24 m rat kidney. Mean 24 h levels did not show significant difference among age groups studied. Daily pulse was significantly increased (p < 0.05) in 12 and 24 m in comparison to 3 m animals (Figs. 1 and 2).

Fig. 1
figure 1

Effect of curcumin administration on daily rhythms of rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα and rRorα mRNA expression in 3, 12 and 24 m old rat kidney. Each value is mean ± SEM (n = 4), p < 0.05 and expressed as relative mRNA expression. pa < 0.05; pb < 0.05, pc < 0.05 and pd < 0.05 (where ‘a’, ‘b’, ‘c’ and ‘d’ refers to comparison with ZT-0, ZT-6, ZT-12 and ZT-18 respectively within the group)

Full size image
Fig. 2
figure 2

Effect of curcumin administration on (i) Mean 24 h levels and (ii) Daily pulse of rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα and rRorα expression in 3, 12 and 24 m old rat kidney. Each value is mean ± SEM (n = 4), p < 0.05 and expressed as mean relative gene expression. pp < 0.05 (where ‘p’ refers to comparison with age-matched vehicle treated group). pq < 0.05 (where ‘q’ refers to comparison with 3 m vehicle treated group)

Full size image

In 3 and 12 m animals, rCry1 showed maximum expression at ZT-18 and minimum at ZT-6. But in 24 m, maximum expression was at ZT-12 with phase advance of 6 h, minimum expression was at ZT-6. Mean 24 h levels did not show significant change among age groups studied. Daily pulse was decreased significantly (p < 0.05) in 12 and 24 m in comparison to 3 m animals. rCry2 showed maximum expression at ZT-12 in all age groups but minimum at ZT-6 in 3 m and at ZT-0 in 12 and 24 m animals. Mean 24 h levels were not significantly different among the age groups. Daily pulse showed significant increase (p < 0.05) in 12 and 24 m with respect to 3 m animals. rRev-erbα expressed maximum at ZT-6 and minimum at ZT-18 in all the age groups. Mean 24 h levels also did not show significant difference among age groups studied. Daily pulse increased significantly (p < 0.05) in 12 and 24 m in comparison to 3 m animals. rRorα showed maximum expression at ZT-18, minimum at ZT-6 in 3 m. Interestingly, in 12 m rhythmicity was abolished. In 24 m, maximum expression was observed at ZT-12 and minimum at ZT-0. Mean 24 h levels did not show significant change among age groups studied. Daily pulse showed significant decrease (p < 0.05) in 12 m but significantly increased (p < 0.05) in 24 m with respect to 3 m animals (Figs. 1 and 2).

Inflammatory genes exhibit mid-age perturbations in kidney

The mRNA expression of various pro-inflammatory genes like rNfκb1, rTnfα, rIl6 as well as pattern recognition receptor genes rTlr4 and rTlr9 were studied in 3, 12 and 24 ms (m) control (C) and vehicle treated (VT) animals. No change was observed between controls and vehicle treated animals. rNfκb1 showed maximum expression at ZT-12 and minimum expression at ZT-6 in 3 m animals. In 12 m, maximum expression was at ZT-6 and minimum at ZT-18 with 6 h phase advance in comparison to 3 m animals. In 24 m, maximum expression was at ZT-12 and minimum expression at ZT-18. Mean 24 h levels showed significant increase (p < 0.05) with aging. Daily pulse was significantly decreased (p < 0.05) in 12 m but increased in 24 m in comparison to 3 m animals. rTnfα showed maximum expression at ZT-12 and minimum expression ZT-6 in 3 m animals. In 12 and 24 m, maximum expression was at ZT-6 i.e. 6 h phase advance with respect to 3 m, but minimum expression was at ZT-12 and ZT-0 respectively. Mean 24 h levels significantly increased (p < 0.05) in 12 and 24 m with respect to 3 m animals. Daily pulse also showed significant increase (p < 0.05) in 12 and 24 m in comparison to 3 m. rIl6 expressed maximum at ZT-12 and minimum at ZT-0 in 3 m animals. In 12 m, maximum expression was observed at ZT-6 with phase advance of 6 h and minimum at ZT-18. In 24 m, maximum and minimum expressions were at ZT-12 and ZT-0 respectively. Mean 24 h levels did not show significant change among age groups studied. Daily pulse showed significant decrease (p < 0.05) in 24 m in comparison to 3 m animals (Figs. 3 and 4).

Fig. 3
figure 3

Effect of curcumin administration on daily rhythms of rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9 mRNA expression in 3, 12 and 24 m old rat kidney. Each value is mean ± SEM (n = 4), p < 0.05 and expressed as relative mRNA expression. pa < 0.05; pb < 0.05, pc < 0.05 and pd < 0.05 (where ‘a’, ‘b’, ‘c’ and ‘d’ refers to comparison with ZT-0, ZT-6, ZT-12 and ZT-18 respectively within the group)

Full size image
Fig. 4
figure 4

Effect of curcumin administration on (i) Mean 24 h levels and (ii) Daily pulse of rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9 expression in 3, 12 and 24 m old rat kidney. Each value is mean ± SEM (n = 4), p < 0.05 and expressed as mean relative gene expression. pp < 0.05 (where ‘p’ refers to comparison with age-matched vehicle treated group). pq < 0.05 (where ‘q’ refers to comparison with 3 m vehicle treated group)

Full size image

rTlr4 showed rhythmic expression with maximum at ZT-12 and minimum at ZT-6 in 3 m animals. In 12 m, maximum expression showed phase advance of 6 h and minimum expression was observed at ZT-12. In 24 m, maximum expression was at ZT-12 and minimum expression at ZT-18. Mean 24 h levels were increased significantly (p < 0.05) in 12 and 24 m in comparison to 3 m animals. Daily pulse also showed significant increase (p < 0.05) in 12 and 24 m animals. rTlr9 showed maximum expression at ZT-12 and minimum at ZT-6 in 3 m animals. In 12 m, maximum expression was at ZT-6 with phase advance of 6 h and minimum at ZT-18. In 24 m, maximum expression was at ZT-12 and minimum expression at ZT-0. Mean 24 h levels showed significant increase (p < 0.05) in comparison to 3 m animals. Daily pulse did not show significant change among age groups studied (Figs. 3 and 4).

Correlation analysis of daily rhythms of clock genes in kidney: age-induced alterations

Pairwise correlations were analyzed among clock genes in light phase (LP) and dark phase (DP) of 3, 12 and 24 m animals (Fig. 5). In light phase of 3 m, a significant negative correlation between rBmal1 and rRev-erbα (p < 0.001) was observed. Also a negative correlation was observed between rBmal1 and rPer1 (p < 0.01). Within the rPer1,2 genes and within the rCry1,2 genes there was a significant positive correlation (p < 0.001). However, rPer1,2 showed significant positive correlation with rCry2 (p < 0.001). rRev-erbα showed negative correlation (p < 0.05) with rRorα. We also observed that rRorα showed positive correlation (p < 0.001) with rCry1,2. In dark phase of 3 m, negative correlation persisted between rBmal1 and rRev-erbα and the negative correlation between rBmal1 and rPer1,2 genes became more significant (p < 0.001). Positive correlation (p < 0.001) persisted between rPer1,2 genes but positive correlation between rCry1,2 was abolished. Positive correlation between rPer1,2 and rCry2 genes persisted. Negative correlation between rRorα and rRev-erbα was abolished. But positive correlation between rRorα and rCry1,2 genes persisted. Moreover, there was a positive correlation between rRev-erbα and rPer1,2 genes (p < 0.001).

Fig. 5
figure 5

Effect of curcumin administration on Pair wise correlations among rBmal1, rPer1, rPer2, rCry1, rCry2, rRev-erbα, rRorα, rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9 in light (ZT-0, 6, 12) and dark (ZT-12, 18, 24/0) phase of 3, 12 and 24 m old rat kidney (LP light phase, DP dark phase, VT vehicle treated, CT curcumin treated). Intensity of color and size of circle represents correlation coefficient values between the genes. Where, positive correlations are indicated by shades of blue, negative correlations by shades of red and white indicates no correlation. ‘*’, ‘**’, ‘***’ indicates statistically significant correlations (p < 0.05), (p < 0.01). (p < 0.001) respectively. (B1—rBmal1; P1—rPer1; P2—rPer2; C1—rCry1; C2—rCry2; Re—rRev-erbα; Ro—rRorα; N1—rNfκb1; Tnf—rTnfα; Il6—rIl6; T4—rTlr4; T9—rTlr9)

Full size image

In light phase of 12 m, correlation between rBmal1 and rRev-erbα was abolished. However, negative correlation between rBmal1 and rPer1 genes persisted (p < 0.001). Also, positive correlation between rPer1,2 genes persisted (p < 0.001), whereas correlation between rCry1,2 genes was abolished. Interestingly, a significant positive correlation appeared between rPer1,2 and rCry1 and positive correlation with rCry2 (p < 0.001) persisted. Negative correlation between rRorα and rRev-erbα was abolished but positive correlation between rRorα and rCry2 persisted (p < 0.001). In dark phase of 12 m, there was no correlation between rBmal1 and rRev-erbα, but negative correlation between rBmal1 and rPer1 genes persisted (p < 0.001). Positive correlation between rPer1,2 genes persisted (p < 0.001) but between rCry1,2 genes correlation was abolished. Positive correlation persisted between rPer1,2 and rCry2 (p < 0.001). A positive correlation appeared between rRorα and rRev-erbα (p < 0.01). rRorα also showed positive correlation with rCry2 (p < 0.001).

In light phase of 24 m, negative correlation between rBmal1 and rRev-erbα persisted (p < 0.001). Significant negative correlation persisted between rBmal1 and rPer1 genes (p < 0.001). Significant positive correlation persisted within and in between rPer1,2 and rCry1,2 genes (p < 0.001). rRorα did not show correlation with rRev-erbα, but showed positive correlation with rCry1,2 (p < 0.001). In dark phase of 24 m, correlations between clock genes were significantly affected where rBmal1 showed significant negative correlation (p < 0.001) with all the other clock genes. A significant positive correlation (p < 0.001) appeared among all the clock genes.

Correlation analysis of daily rhythms of immune genes in kidney: age-induced alterations

Pairwise correlation among immune genes in light phase (LP) and dark phase (DP) were analyzed in 3, 12 and 24 m animals (Fig. 5). In light phase of 3 m, rNfκb1 showed significant positive correlation (p < 0.001; p < 0.01) with all other immune genes except rIl6. In dark phase of 3 m, rNfκb1 showed positive correlation (p < 0.001) with all the immune genes except rTlr4. In light phase of 12 m, rNfκb1 showed significant positive correlation (p < 0.001) with all other immune genes. In dark phase of 12 m, rNfκb1 showed negative correlation with rTnfα and rTlr4 (p < 0.001; p < 0.01) but significant positive correlation with rTlr9 (p < 0.001) persisted and correlation with rIl6 was abolished. In light phase of 24 m, positive correlation of rNfκb1 with all the immune genes (p < 0.001; p < 0.05) persisted. In dark phase of 24 m, rNfκb1 showed a significant positive correlation (p < 0.001) with all the immune genes.

Correlation analysis between immune genes and clock genes in kidney

Pairwise correlation between clock and immune genes in light phase (LP) and dark phase (DP) were analyzed in 3, 12 and 24 m animals (Fig. 5). In light phase of 3 m, rNfκb1 and rTnfα showed significant positive correlation with rRorα and rCry1,2 (p < 0.001; p < 0.01) and negative correlation with rRev-erbα (p < 0.05). rIl6 showed significant positive correlation (p < 0.001) with rPer1,2 genes. rTlr4 and rTlr9 showed significant positive correlation with rRorα and rCry1,2 genes (p < 0.001; p < 0.01; p < 0.05). rTlr9 also showed positive correlation with rPer1,2 genes. In dark phase of 3 m, rNfκb1, rTnfα, rIl6, rTlr4 and rTlr9 showed significant positive correlation with rRev-erbα and rPer1,2 genes (p < 0.001; p < 0.01; p < 0.05).

In light phase of 12 m, rNfκb1 and rTnfα changed to negative correlation (p < 0.001) with rCry1, but significant positive correlation (p < 0.001) with rRev-erbα was appeared. rIl6, rTlr4, rTlr9 changed to negative correlation with rCry1 and rPer1,2 genes (p < 0.001; p < 0.01; p < 0.05). In dark phase of 12 m, rTnfα showed positive correlation with rCry1 (p < 0.01). rIl6 and rTlr4 showed negative correlation with rPer1,2 genes (p < 0.01; p < 0.05). rTlr4 changed to negative correlation (p < 0.001) with rRev-erbα, but positive correlation between rRev-erbα and rTlr9 persisted (p < 0.05).

In light phase of 24 m, positive correlation between rNfκb1 and rCry1,2 genes persisted (p < 0.001). Positive correlation of rIl6, rTlr4 and rTlr9 with rPer1,2 and rCry1,2 genes persisted (p < 0.001; p < 0.05). In dark phase of 24 m, all immune genes showed significantly altered correlations with all the clock genes where significant positive correlation appeared with all clock genes but showed negative correlation with rBmal1 (p < 0.001).

Gene to gene network interactions altered with aging in kidney

WGCNA analysis demonstrated in 3 m, rCry2 and rRorα showed interactions with rTlr4; rPer1 showed interaction with rIl6 and rTlr9. In 12 m, rRev-erbα showed interactions with all immune genes except rTlr4. In 24 m, interaction between rRev-erbα and rTnfα persisted; rPer1,2 showed interaction with all immune genes except rTnfα, whereas, rBmal1 showed interaction with rNfκb1 and rIl6 (Fig. 6).

Fig. 6
figure 6

WGCNA analysis between clock and immune gene clusters: a clock and immune gene clusters (b) effect aging on gene to gene network in 3, 12 and 24 m old rat kidney (upper Panel) and effect of curcumin administration (lower Panel). Color of node indicates no. of interactions (highest—purple; intermediate–sion and least—yellow). Color of edge indicates the strength of interaction (strongest—purple; sion—intermediate and weakest—yellow). Green and red dots indicate clock and immune genes respectively

Full size image

Effect of curcumin administration on daily rhythms of clock genes expression in kidney

With curcumin treatment, rBmal1 showed maximum expression at ZT-18 i.e. 6 h phase advance in 3 m animals with respect to 3 m vehicle group and minimum expression was at ZT-6. In 12 and 24 m, maximum expression persisted at ZT-0 and minimum at ZT-12 with curcumin treatment in comparison to age-matched vehicle groups. Curcumin administration reduced mean 24 h levels significantly (p < 0.05) in 3 and 12 m in comparison to age-matched vehicle groups, but did not show any change in 24 m. Daily pulse showed significant decrease in 3 m, did not vary in 12 m, but restored in 24 m animals with significant increase (p < 0.05) in comparison to age-matched vehicle group. rPer1 showed maximum expression at ZT-6 in 3 m CT group with phase advance of 6 h in comparison to 3 m vehicle group, minimum expression was at ZT-0. In both 12 and 24 m, curcumin treatment showed similar expression as in age-matched vehicle groups. Mean 24 h levels were significantly decreased (p < 0.05) in 3 and 12 m with respect to age-matched vehicle groups, but in 24 m CT the decrease was not significant. Daily pulse showed significant increase (p < 0.05) in 3 and 12 m with curcumin treatment in comparison to age-matched vehicle groups. In 24 m CT animals, significant decrease (p < 0.05) was observed in comparison to 24 m vehicle group. In 3 m, curcumin treatment did not alter rPer2 maximum expression but minimum expression was shifted from ZT-6 to ZT-0. In 12 and 24 m, curcumin administration did not change the expression pattern of rPer2 with respect to age-matched vehicle groups. There was no significant change in mean 24 h levels of rPer2 in 3, 12 and 24 m in comparison to age-matched vehicle groups. Daily pulse was significantly increased in 3 m but decreased in 12 and 24 m CT (p < 0.05) with respect to age-matched vehicle groups (Figs. 1 and 2).

In 3 m, maximum expression of rCry1 persisted at ZT-18 but minimum expression was observed at ZT-0 with curcumin treatment. In 12 m, curcumin did not alter daily rhythm in comparison to vehicle group. In 24 m, curcumin restored maximum expression at ZT-18 in comparison to 3 m vehicle group. With curcumin administration mean 24 h levels were significantly decreased (p < 0.05) in all age groups with respect to age-matched vehicle groups. Daily pulse was also reduced significantly (p < 0.05) in all age groups. rCry2 showed maximum expression at ZT-6 and minimum expression at ZT-0 in 3 m CT group. However, in 12 and 24 m, expression patterns were similar to the age-matched vehicle groups. Mean 24 h levels were significantly decreased (p < 0.05) in all age groups with curcumin administration. Daily pulse was increased in 3 m, unaltered in 12 m and decreased in 24 m (p < 0.05) upon curcumin treatment. In case of rRev-erbα, curcumin treatment did not change daily rhythm pattern in all the age groups with respect to age-matched vehicle groups. Mean 24 h levels showed significant decrease in 3 m but did not show significant change in 12 and 24 m group in comparison to age-matched vehicle groups. Curcumin treatment significantly decreased (p < 0.05) daily pulse in 3 and 12 m with respect to age-matched vehicle groups. However, curcumin restored daily pulse in 12 m in comparison to 3 m vehicle group. But in 24 m, daily pulse increased (p < 0.05) in comparison to 24 m vehicle group. rRorα showed maximum expression at ZT-6 and minimum expression at ZT-0 in 3 m CT group. In 12 m, curcumin restored rhythmicity with maximum expression at ZT-12 and minimum expression at ZT-6. In 24 m, maximum expression persisted at ZT-12 but minimum expression was at ZT-18 with curcumin administration. Mean 24 h levels were significantly decreased (p < 0.05) with curcumin in all age groups with respect to age-matched vehicle groups. Curcumin decreased daily pulse in 3 and 24 m (p < 0.05), however increased (p < 0.05) in 12 with respect to age-matched vehicle groups (Figs. 1 and 2).

Effect of curcumin administration on daily rhythms of inflammatory genes expression in kidney

Curcumin administration resulted in alleviation of expression of inflammatory genes in kidney. Curcumin administration phase delayed rNfκb1 about 6 h with maximum expression at ZT-18 and minimum expression at ZT-0 in 3 m. In 12 m, rNfκb1 showed restoration in comparison with 3 m vehicle group with maximum expression at ZT-12 and minimum expression at ZT-6. In 24 m, rNfκb1 was phase advanced by 6 h with maximum expression at ZT-6 and minimum expression at ZT-18 with respect to 24 m vehicle group. Being an anti-oxidant curcumin significantly reduced mean 24 h levels in all age groups (p < 0.05). Daily pulse did not show significant change in 3 and 12 m with respect to their age-matched vehicle groups. However, in 24 m, curcumin decreased (p < 0.05) and restored daily pulse with respect to 3 m vehicle group. rTnfα showed phase delay of 6 h with maximum expression at ZT-18, minimum expression at ZT-6 in 3 m CT. In 12 m, maximum expression was observed at ZT-0 which is 6 h phase advance with respect to 12 m vehicle group and minimum expression was at ZT-6. In 24 m, expression pattern was not changed with respect to 24 m vehicle group. Curcumin significantly reduced (p < 0.05) the mean 24 h levels in all the age groups. But in 24 m, curcumin restored mean 24 h levels in comparison with 3 m vehicle group. In 3 and 12 m, daily pulse was reduced significantly (p < 0.05) in comparison to age-matched vehicle group. In 24 m, daily pulse remained unaltered with respect to 24 m vehicle group. rIl6 showed 6 h phase advance with maximum expression at ZT-6 and minimum expression at ZT-12 in 3 m CT group. In 12 m, curcumin treatment did not change expression pattern with respect to 12 m vehicle group. But in 24 m, maximum expression was observed at ZT-0 which is 12 h phase advance in comparison to 24 m vehicle group and minimum expression at ZT-18. Curcumin administration did not alter mean 24 h levels in all age groups. Daily pulse remained unaltered in 3 m, significantly increased in 12 m (p < 0.05), and significantly decreased in 24 m (p < 0.05) with cucumin administration (Figs. 3 and 4).

In 3 m curcumin treated group, rTlr4 showed maximum expression at ZT-18 with phase delay of 6 h in comparison to 3 m vehicle group, and minimum expression was observed at ZT-0. In 12 m, maximum expression was observed at ZT-0 with the phase advance of 6 h and minimum was observed at ZT-6. In 24 m, maximum expression was observed at ZT-6 with phase advance of 6 h and minimum expression was observed at ZT-18. Curcumin administration did not change mean 24 h levels in 3 m, but significantly reduced in 12 and 24 m (p < 0.05) with respect to age-matched vehicle groups. Interestingly, in 24 m, curcumin restored the mean levels in comparison to 3 m vehicle group. Daily pulse did not alter in 3 m and 24 m, but significantly reduced in 12 m (p < 0.05) with respect to age-matched vehicle groups. With curcumin treatment rTlr9 showed rhythmic expression with maximum at ZT-18 i.e. 6 h phase delay with respect to 3 m vehicle group and minimum at ZT-6. In 12 m, maximum expression was observed at ZT-18 with phase delay of 12 h and minimum expression at ZT-6. In 24 m, maximum expression was observed at ZT-6 with phase advance of 6 h and minimum expression was observed at ZT-18. With curcumin treatment, mean 24 h levels were significantly increased in 3 m (p < 0.05), remained unaltered in 12 m and significantly decreased in 24 m (p < 0.05) with respect to their age-matched vehicle groups. Daily pulse was unaltered in 3 and 24 m, significantly reduced in 12 m (p < 0.05) with curcumin treatment (Figs. 3 and 4).

Correlation analysis of clock genes with curcumin administration in kidney

In light phase of 3 m CT group, negative correlation of rBmal1 with rPer1 genes and rRev-erbα (p < 0.001) persisted. Within and between rPer1,2 and rCry1,2 genes positive correlation persisted (p < 0.001; p < 0.01). rRorα changed to positive correlation with rRev-erbα (p < 0.001) but positive correlation with rCry2 (p < 0.001) persisted. In dark phase of 3 m CT group, negative correlation of rBmal1 with rRev-erbα (p < 0.001) persisted. Positive correlation persisted between rPer1,2 and rCry1,2 genes (p < 0.001). rRorα showed insignificant negative correlation with rRev-erbα but significant positive correlation (p < 0.001) with Cry1,2 genes persisted (Fig. 5).

In light phase of 12 m CT group, negative correlation of rBmal1 with rRev-erbα was restored and negative correlation with rPer1 gene (p < 0.001; p < 0.01) persisted. Positive correlation between rPer1,2 genes persisted (p < 0.001). Curcumin administration abolished correlation between rCry1,2 genes. rRorα showed significant positive correlation with rCry2 (p < 0.05) but not with rRev-erbα and rCry1. In dark phase of 12 m CT group, negative correlation of rBmal1 with rRev-erbα was restored and negative correlation with rPer1 gene (p < 0.001) persisted. Positive correlation between rPer1,2 genes (p < 0.001) persisted. Interestingly, rCry1 showed negative correlation with rCry2 (p < 0.05). rRorα showed positive correlation with rCry2, rRev-erbα (p < 0.001) and negative correlation with rCry1 (p < 0.01) (Fig. 5).

In light phase of 24 m CT group, negative correlation of rBmal1 with rPer1 genes and rRev-erbα (p < 0.001) persisted. Positive correlation within the rPer1,2 and rCry1,2 genes persisted (p < 0.001; p < 0.05). In dark phase of 24 m CT group, negative correlation of rBmal1 with rPer1 and rRev-erbα (p < 0.001) persisted. Positive correlation within and between rPer1,2 and rCry1,2 genes persisted (p < 0.001; p < 0.01) (Fig. 5).

Correlation analysis of immune genes in kidney upon curcumin administration

In light phase of 3 m CT group, curcumin administration abolished correlation between rNfκb1 and rTnfα, but showed significant negative correlation between rTnfα and Il6 (p < 0.001). Positive correlation of rNfκb1 with Tlr9 and Tlr4 (p < 0.001) persisted. In dark phase of 3 m CT group, positive correlation of rNfκb1 with rTnfα, rTlr9 and rTlr4 (p < 0.001; p < 0.05) persisted. But rNfκb1 changed to negative correlation with rIl6 (p < 0.001).

In light phase of 12 m CT group, positive correlation of rNfκb1 with rTnfα, rTlr4 and rTlr9 (p < 0.001) persisted. Significant negative correlation was observed between rTnfα and rIl6 (p < 0.001). Significant negative correlation was observed between rIl6 and rTlr4,9 (p < 0.001). In dark phase of 12 m CT group, curcumin treatment resulted in abolition of correlation of rNfκb1 with rTnfα and rTlr4. However, positive correlation was restored between rNfκb1 and rIl6 (p < 0.05); rTnfα and rIl6 (p < 0.001). Interestingly, negative correlation appeared between rNfκb1 and rTlr9 (p < 0.05). Positive correlation between rTnfα and rIl6 (p < 0.001) was restored, and positive correlation between rTlr4 and rTlr9 (p < 0.001) was restored.

In light phase of 24 m CT group, positive correlation of rNfκb1 with rTnfα and rTlr4,9 (p < 0.001; p < 0.01) persisted. Curcumin administration resulted in abolition of correlation between rNfκb1 and rIl6, and resulted in restoration in comparison to LP of 3 m vehicle group. Correlation between rIl6 and rTlr4,9 abolished. In dark phase of 24 m CT group significant correlation of rNfκb1 with rTnfα abolished, but positive correlation with rTlr4,9 (p < 0.001; p < 0.01) persisted. Positive correlation of rIl6 with rTlr4,9 (p < 0.01; p < 0.05) persisted.

Effect of curcumin on pairwise correlation between inflammatory genes and clock genes in kidney

In light phase of 3 m CT group, positive correlation of rCry1,2 genes with rNfκb1 (p < 0.001; p < 0.05) persisted but abolished with rTnfα. Positive correlation of rTlr4 and rTlr9 with Cry1 (p < 0.001) persisted. Curcumin resulted in abolition of correlation between rIl6 and rPer1,2 genes. In dark phase of 3 m CT group, significant positive correlation of rPer1,2 genes with rNfκb1 (p < 0.001) persisted. Positive correlation of rPer1,2 with Tlr4 and Tlr9 persisted (p < 0.001; p < 0.01) (Fig. 5).

In light phase of 12 m CT group, rNfκb1, rTnfα and rTlr4,9 changed to positive correlation with rCry1 and rRev-erbα. Positive correlation of rTlr4 and rTlr9 with rPer1,2 genes abolished. In dark phase of 12 m CT group, rTlr9 showed significant negative correlation with rPer1,2 genes (p < 0.001). Correlation of rCry1 with rTnfα, rIl6 and rTlr4 abolished. Correlation between rRev-erbα and rTnfα abolished (Fig. 5).

In light phase of 24 m CT group, correlation between rTlr4,9 and rPer1,2 genes abolished. rIl6 showed negative correlation with rPer1,2 genes (p < 0.001). Correlation between rNfκb1 and rCry1 abolished. rIl6 and rTlr4,9 changed to negative correlation with rCry1. In dark phase of 24 m CT group, there was abolition of correlation between rTlr4.9 target gene C and rPer1,2 genes. rTnfα, rIl6 changed to negative correlation with rCry1. Correlation between rCry1 and rNfκb1 abolished with curcumin treatment (Fig. 5).

Curcumin differentially altered gene to gene network interactions in kidney

WGCNA analysis demonstrated in 3 m CT, rCry1 showed interactions with rNfκb1, rTlr4,9. In 12 m CT, interactions of rRev-erbα with rIl6 and rTlr9 persisted and rBmal1 lost the interactions with other clock genes. In 24 m CT, the interactions between clock and immune genes showed similarity with the interactions observed in 12 m VT, where rRev-erbα showed interactions with all immune genes except rIl6; rBmal1, rPer1,2, rCry2 showed interactions with each other (Fig. 6).

Discussion

In the present study, all the clock genes studied showed significant daily rhythms in kidney of 3 m old animals (Fig. 1). Elevated levels of rBmal1 at dark phase are corroborated with previous studies in different peripheral clocks across different species (Christiansen et al. 2016; Yang et al. 2016). This emphasizes the importance of well organised synchrony between the clocks for a better survival of an organism (Hatori et al. 2017). Interestingly, in 3 m rat SCN, rBmal1 maximum expression was seen at ZT-18 (Mattam and Jagota 2014) which is 6 h earlier to kidney, this further demonstrate the relation between master and slave clocks (Balsalobre 2002). rPer1 and rPer2 showed offset at ZT-12 (Fig. 1) and is in agreement with previous studies in different tissues and species (Pizarro et al. 2013; Yang et al. 2016). However, in SCN, only rPer2 showed peak expression at ZT-12 (Mattam and Jagota 2014). Expression pattern of rCry1, rCry2 and rRev-erbα (Fig. 1) also corroborates to the previous studies (Takeda et al. 2012; Yang et al. 2016; Astafev et al. 2017). In the present study, rPer1 did not show any change in expression with aging. Interestingly, mRNA levels of αENaC (alpha subunit epithelial Na+ channel), essential for regulation of salt and water reabsorption was reported to be under PER1 regulation (Gumz et al. 2009), did not vary with aging (Haloui et al. 2013). rBmal1, rPer2, rCry2 and rRev-erbα also did not show significant variations across the age groups studied (Fig. 1). However, in SCN rBmal1 and rPer1 showed variations as age progress (Mattam and Jagota 2014). This implies that SCN is more sensitive towards age-related attritions than peripheral clock kidney. However, in kidney only rRorα exhibited mid-age perturbations, where the rhythmicity was completely abolished. Interestingly, all the clock genes did not show significant variation in mean 24 h levels with aging (Fig. 2). On the other hand, SCN displayed an increase in the mean 24 h levels of rPer2, rCry1 and rCry2 in 12 m (Mattam and Jagota 2014) further signifying the sensitivity of central clock in mid-age.

rBmal1 showed significant negative correlation with rPer1 and rRev-erbα in both light phase and dark phase of 3 m (Fig. 5). These tightly regulated interactions are essential for the sustained metabolism in organisms (Solt et al. 2011). However, rBmal1 did not show any correlation with rRor α in LP and DP of 3 m (Fig. 5), which is in agreement with the minimal role of rRor α on circadian clock in kidney (Takeda et al. 2012). Positive correlation between rCry1 and rCry2 in LP did not vary with aging in kidney in present study (Fig. 5) but abolished in SCN (Mattam and Jagota 2014). Negative correlation between rBmal1 and rPer1 observed in kidney was not observed in SCN 3 m (Mattam and Jagota 2014). This emphasizes the fact that central and peripheral clocks exhibit several variations though similar transcriptional feedback loops are involved in both (Schibler et al. 2015).

Recently, we reported restoratory effect of curcumin administered at ZT-11 on rPer1, rPer2, rCry1 and rCry2 in 12 m and rPer1 in 24 m rat SCN (Kukkemane and Jagota 2019). Here we explored the chronobiotic properties of curcumin on peripheral clock kidney for the first time. Curcumin administration did not show significant changes in expression pattern of rBmal1 and rRev-erbα. However, rPer1, rCry1 and rCry2 were altered with curcumin treatment only in 3 m but remained unaltered in 12 and 24 m animals with respect to their age-matched vehicle groups. Further detailed study is required to understand the underlying mechanism. Interestingly, only rRor-α showed sensitivity towards curcumin in all age groups studied (Fig. 1).

Time dependent immune responses are well documented in several immune cells; and there are remarkable evidences to show that these are cell and tissue specific circadian regulations (Curtis et al. 2014). NFκB1, an important regulatory transcription factor in inflammation, plays a central role in inducing transcription of Tnfα, Il6 and several other cytokines and also involved in apoptosis, cellular growth and differentiation (Hoesel and Schmid 2013). Several studies showed that these inflammatory genes in immune cells show diurnal expression in rodents (Keller et al. 2009; Cermakian et al. 2013; Curtis et al. 2014). Interestingly, in our study, we observed that rNfκb1, rTnf-α and rIl6 showing significant daily rhythms in 3 m kidney (Fig. 3) and corroborates to the previous reports in immune cells (Curtis et al. 2014). In addition, LPS induced phase shift of circadian rhythms in SCN were observed to be through TNFR1 receptors (Paladino et al. 2014). In this context, it would be of greater importance to understand role of Tnfα on circadian rhythms in kidney to address renal chrono-inflammatory aberrations. Tlr9 contains canonical E-boxes at its promoter site where CLOCK/BMAL1 complex can induce the expression, but its circadian rhythms are cell and tissue specific (Silver et al. 2012). In our study, we observed both rTlr9 and rTlr4 showing peak expression at ZT-12 (Fig. 3) which is similar to expression seen in inflammatory cells (Silver et al. 2012, 2018). Interestingly, all the immune genes studied showed phase advance of 6 h in 12 m but remained unaltered in 24 m in comparison to 3 m kidney (Fig. 3). Further studies are essential in understanding the mid-age perturbations in chrono-immune system.

Curcumin administration had profound effect on daily rhythms of all the immune genes. Interestingly, curcumin had similar chronomodulatory effects on rNfκb1 and rTnfα in all age groups (Fig. 3). This could be because of curcumin’s regulation on rTnfα through NFκB1. Several researchers demonstrated the role of elevated pro-inflammatory molecules like NFκB1 and TNFα in various renal disorders (Tilstra et al. 2011; Wang et al. 2017). Here we report that rNfκb1 and rTnfα expressions were significantly elevated with aging (Fig. 4) which corroborates to previous studies (Tilstra et al. 2011; Xi et al. 2014). Curcumin being an anti-inflammatory molecule reduced the expression of rNfκb1 and rTnfα in all age groups and restored rTnfα in 24 m kidney (Fig. 4). Further, we also report the gradual increase in transcription of rTlr4 and rTlr9 with aging which supports previous studies (Xi et al. 2014). Curcumin reduced the mean 24 h levels of rTlr4, which is in agreement with previous studies (Zhu et al. 2014) and restored in 24 m animals (Fig. 4). It has been reported that the anti-inflammatory action of curcumin could be through the activation of Nrf2 which was shown to attenuate inflammatory responses (Wardyn et al. 2015). Interestingly, in 3 m, curcumin significantly increased Tlr9 levels but remained unaltered in 12 and 24 m with respect to age-matched vehicle groups (Fig. 4).

Pairwise correlation analysis revealed the change of positive correlation of rNfκb1 with rTnfα and rTlr4 in DP of 12 m kidney (Fig. 5); this suggests the deregulated interactions between inflammatory genes with aging. rNfκb1 showed significant positive correlation with other immune genes in 24 m, this further suggests altered inflammatory status with aging. Curcumin showed significant alterations in correlations among the immune genes as it reduced the expression of several immune genes (Fig. 5).

We also correlated clock genes and immune genes in order to understand the possible interactions with each other. rPer1 showed positive correlation with rIl6 in LP and DP of 3 m animals (Fig. 5), whereas PER1 negatively regulates IL-6 expression in spinal astrocytes (Sugimoto et al. 2014). Tlr9 shows Per2 dependent circadian expression in macrophages (Silver et al. 2012), we also observed positive correlation between rPer2 and rTlr9 in LP and DP of 3 m animals (Fig. 5). REV-ERBα shows inhibitory action on TLR4 expression in human macrophages (Fontaine et al. 2008), but in our study we observed insignificant negative correlation between rRev-erbα and rTlr4 in LP of 3 m animals (Fig. 5). In LP and DP of 12 m, these correlations were altered (Fig. 5), suggesting desynchrony between immunity and circadian clock. CRY proteins were proposed to inhibit Il6 expression by blocking NFκB activity in fibroblasts and macrophages (Narasimamurthy et al. 2012). Interestingly, in our study we observed a positive correlation between rCry2 and rIl6 (Fig. 5). In another study, overexpression of CRY1 reduced the TLR4 expression in atherosclerosis mouse model (Yang et al. 2015). But, we observed a positive correlation between rCry1 and rTlr4 in LP of 3 m animals (Fig. 5). Though our study involves mRNA expression, the study at protein level may yield a better understanding on such interactions. In DP of 24 m, all the immune genes showed significant positive correlation with all clock genes except rBmal1, this provides significant basis for desynchronised clock and immune systems with aging (Fig. 5). WGCNA analysis between clock and immune gene clusters demonstrated that in young age clock and immune genes exhibit interactions in two different groups with rTlr4 and rPer1 being the hub genes and intensity of the interactions are medium to weak (Fig. 6). With aging the rRev-erbα showed maximum interactions with immune and with increased strength of interactions. In 24 m, interactions between clock and immune genes were increased with increased intensity of strength. This suggests that increased inflammatory status with aging might be because of subtle change in clock system. However, with curcumin treatment in 24 m, interactions between clock and immune genes showed similarity with 12 m VT, suggesting the potential of curcumin as chronobiotic to regulate both clock and immune system (Fig. 6).

The present study demonstrates that aging renders desynchronization between the expression of clock and immune genes in kidney. Curcumin administration resulted in differential restoration of immune gene expressions and their correlation with clock genes in aged kidney. Our results have given novel insights of curcumin as a chronobiotic on immune genes to further establish it as a potent drug against age associated chrono-immune attritions.