Effects of resveratrol on in vitro circadian clock gene expression in young and older human adipose-derived progenitor cells

APC donor characteristics

We assessed in vitro rhythmic clock gene expression in APCs that were cultured from previously collected adipose tissue biopsy specimens that were harvested from the abdominal depots of young (N=5) and older (N=5) female participants. Clinical characteristics of the APC donors are shown in Table 1. With the exception of age (P ≤ 0.0001) and systolic blood pressure (P = 0.027), there were no significant differences in race, height, weight, BMI, waist-hip ratio, DBP, fasting glucose, HbA_1C, total cholesterol, HDL, LDL, LDL/HDL ratio, VLDL, triglycerides and non-HDL cholesterol between groups.

In vitro rhythmic clock gene expression in APCs from young and older participants

To assess the rhythmic pattern of core clock gene expression in APCs derived from young and older female participants, we measured mRNA levels of core clock genes including BMAL1, CLOCK, PER1/2, CRY1/2. We also measured mRNA levels of the nuclear receptors NR1D1, NR1D2, and RORα which are part of a secondary regulatory loop contributing to the core clock, as well as the ubiquitous clock output gene, DBP. For these assessments, APCs were synchronized with 30% FBS for 2h. After 12h, APCs were then harvested every 6h for 48 h and mRNA levels were assessed (Figure 1). From these time points, we tested rhythmic profiles for significant circadian rhythmicity, which determines goodness of fit to a cosine wave, using ARSER analysis [24].

In young-APCs, circadian rhythmicity was detected in core clock components BMAL1, PER1, PER2, CRY2, NR1D1 and NR1D2 and the clock output gene DBP (all P<0.05). In older-APCs, circadian rhythmicity was detected in these same components, BMAL1, PER1, PER2, CRY2, NR1D1 and NR1D2 as well as DBP (all P<0.05). Also, rhythmic parameters were compared between young vs. older-APCs. These parameters included rhythmic period (duration of one complete rhythmic cycle), phase (location in time of rhythmic peak) and amplitude (half the distance from the rhythmic peak to the trough) [24]. Older-APCs were characterized by PER2 and NR1D2 phase delay, and longer period length for DPB expression, relative to young (Table 2; all P<0.05).

We also compared rhythmic expression levels between young- APCs vs. older- APCs using linear mixed model. If a significant interaction was detected, this was followed by post hoc comparisons to determine differences in mRNA expression at specific timepoints within the rhythmic profile. Overall, CLOCK gene expression was significantly higher in older-APCs vs. young-APCs all at time points, except t = 18. PER2 gene expression was significantly higher in older-APCs vs. young-APCs at t = 42 and t = 48, with a trend towards significance at t = 24. RORα gene expression was higher in older- APCs vs. young- APCs at t = 30.

RSV increases SIRT1 mRNA in APCs

In order to study the effects of RSV on APCs, we first compared SIRT1 mRNA expression profiles between young vs. older-APCs at baseline (e.g. no vehicle control) using linear mixed model followed by post hoc comparisons. Surprisingly, SIRT1 expression was significantly higher in older-APCs at t = 6, t = 12, t = 24, t = 30, t = 42, and t = 48 compared to young-APCs (Figure 2A; P<0.05).

As an indicator of RSV treatment, we tested the effects of RSV on SIRT1 mRNA expression profiles in young and older-APCs using the following two conditions: 100μM RSV vs. vehicle control. 100μM RSV for 4h has been previously shown to increase clock gene expression in fibroblasts derived from young and older humans [20], which supports our rationale for using this concentration. To examine if the duration and concentration of RSV affects SIRT1 expression, we compared SIRT1 mRNA levels between RSV vs. vehicle control at t = 0. Although SIRT1 was increased in older-APCs relative to young-APCs at baseline, RSV treatment further increased SIRT1 expression in both young (by 313.2%; P<0.0001) and older APCs (by 283.9%; P<0.0001; Figure 2B).

Effects of resveratrol on in vitro rhythmic clock gene expression in APCs

Young-APCs: We tested for significant circadian rhythmicity, using ARSER analysis, on in vitro rhythmic profiles in RSV vs. control. After RSV treatment, CRY1 and CLOCK gained circadian rhythmicity, PER1 and CRY2 lost circadian rhythmicity, and BMAL1, PER1, NR1D1, NR1D2, and DBP rhythmic patterns were not changed (Figure 3). Regarding rhythmic parameters, RSV resulted in longer period length for DBP expression, increased NR1D1 amplitude and phase shift, and increased NR1D2 amplitude (Table 3; all P< 0.05).

We then compared rhythmic gene expression levels of these same conditions using linear mixed model followed by posthoc comparisons. RSV treatment increased mRNA expression of CLOCK (at t = 0, t = 12, t =18 and t = 48), CRY1 (at t = 12, t = 18), CRY2 (at t = 0), NR1D1 (at t = 6 and t = 18), NR1D2 (at t = 36), and RORα (at t=0).

Older-APCs: RSV treatment had the same effects on circadian rhythmicity of core clock components, as was observed in young-APCs. More specifically, after RSV treatment, CRY1 and CLOCK gained circadian rhythmicity, PER1 and CRY2 lost circadian rhythmicity, and BMAL1, PER1, NR1D1, NR1D2, and DBP rhythmic patterns were not changed (Figure 4). Regarding rhythmic parameters, RSV treatment resulted in PER1 phase shift, increased NR1D1 amplitude, and NR1D2 phase shift (Table 4; all P<0.05).

We then compared rhythmic gene expression levels of these same conditions using linear mixed model followed by posthoc comparisons. RSV treatment increased mRNA expression of BMAL1 (at t = 0 and t = 24), CLOCK (t = 12 and t =18), PER1 (at t = 0, t = 6, t = 12, t = 18, t = 24, and t = 48), CRY1 (at t = 0, t = 12, t = 18, and t = 30), CRY2 (t = 0 and t = 6), NR1D1 (at t = 6, t = 18, t = 24, t = 36, and t = 48), NR1D2 (at t = 0, t = 6, t = 12, t = 18, t = 24, t = 30, and t = 36), and RORα (at t = 0 and t = 36).

Participant donors

10 female participants completed the BioAge study at the Translational Research Institute, AdventHealth (AH TRI) in Orlando, Florida, The United States of America. Five participants were classified as older (aged 65-79) and five as young (aged 20-25). All included participants had a BMI ≤ 40 kg/m², were weight stable (±5 kg) for the last three months prior to the screening visit and were sedentary (<20 min of activity, 3x/week). Participants with a medical history of diabetes, coronary heart disease, stroke, chronic renal failure, chronic hepatic disease, autoimmune disease or gastrointestinal disease requiring medication, use of insulin or any other glucoregulatory drugs were excluded. Participants were not taking anti-inflammatory (within two days prior to biopsy) or antiplatelet medications (within seven days prior to biopsy), were not pregnant or less than 9 months postpartum. Other exclusion criteria included: drug or alcohol abuse (≥3 drinks per day) within the last five years, usage of oral contraceptives (<3 months on a stable regime) (with the exception of hormone replacement therapy), usage of tobacco or nicotine-containing products (within the past three months).

Adipose tissue biopsies

Adipose tissue biopsies of the abdominal subcutaneous region were performed under local anesthesia [35]. Adipose biopsies were taken in the overnight fasted state between 8 and 10 AM. The scWAT was digested with collagenase (type 1, 1 mg/2ml in HBSS) for 30-60 min at 37° C with shaking (100rpm), as described previously [36]. The mixture was passed through a 250-micron mesh, centrifuged at 500g for five minutes and floating adipocytes were discarded. Cell pellets were treated with erythrocyte lysis buffer (BioLegend, Cat: 420301) for five to 10 minutes at room temperature followed by centrifugation at 500g for five minutes. After discarding the supernatant, cells were resuspended in proliferation media (PM) (Gibco α-Minimum Essential Media (αMEM), 10% Fetal Bovine Serum (FBS) and 1% Penicillin-Streptomycin), then plated and cultured in an incubator at 37° C at 5% carbon dioxide. Media was replaced after four hours to remove blood cells. APCs continue to grow with PM changes every other day until they reached ~80% confluency, then they were frozen in freezing media (10% FBS, 10% DMSO in αMEM) and stored in dewars containing liquid nitrogen.

APC cell culture procedures

APCs frozen cryovials were quickly thawed by directly placing them in a 37° C water bath, then centrifuged at 500g for five min to remove freezing media. APCs were plated in a T75 flask with 10mL of PM. The PM was replaced every two or three days until cells were ~80% confluent. Once confluent, cells were lifted by trypsin and cell number was determined by using trypan blue exclusion and an automatic cell counter (Countess II FL). The APCs were plated in a 12-well plate with αMEM and 5% FBS at a density of 50,000 cells per well. All APCs were passage 3-5 at the time of the experiment. In vitro synchronization for APCs was initiated 24 hours after plating.

In vitro synchronization

After 24h, cells were 70% confluent. Subsequently, cells were synchronized by adding media containing 30% FBS for 2h, as previously performed [8]. Subsequently, the αMEM containing with 2% FBS was added for 12h. Cells were harvested every 6h for a period of 48h.

Resveratrol treatment

RSV (3,5,40-trihydroxystilbene, purity ≥99%) was purchased from Sigma-Aldrich (R5010). 100mM RSV stock solution in 100% ethanol was made. APCs were synchronized as described above for 2h. Cells were treated with 100μM RSV or ethanol (vehicle) for 12h, and then cells were harvested every 6h for a period of 48h [7, 20].

RNA isolation and real-time quantitative reverse transcriptase-PCR

To isolate RNA from APCs, we used the RNeasy Lipid Tissue Mini kit (Qiagen, Cat: NC9307831). 200μL of RLT buffer per well was used. 2 wells of 12 well plates were combined per treatment. Harvested cells were immediately frozen at -80° C. RNA was isolated according to the protocol, and quality and quantity were assessed by using the Nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The PCR reagents were 2.5μL of 4X TaqMan Fast Virus Master Mix (Applied Biosystems, Cat: 4444434), 0.5μL of 20X primer and 10ng of RNA for a total reaction of 10μL. The initial step of PCR was five minutes at 50° C of reverse transcription, followed by a 20-second hold at 95° C for inactivation/initial denaturation. Followed by 40 cycles of PCR amplification for 15 seconds at 95° C and 60 seconds at 60° C. PCR reactions were performed in triplicate in 384-wells plates. The housekeeping gene peptidylprolyl isomerase A (PPIA) (Hs_01565699_g1) was used to normalize core clock gene expressions calculated by 2(^–ΔCt) method. PPIA was previously identified as a suitable reference gene in scWAT of mice fed during the resting or active phase, thus, this gene is optimal for studies of clock genes [37]. Pre-designed TaqMan gene expression assays from Applied Biosystems were: CLOCK (Hs00231857_m1), BMAL1 (Hs00154147_m1), PER1 (Hs00242988_m1), PER2 (Hs00256143_m1), CRY1 (Hs00172734_m1), CRY2 (Hs00323654_m1) and SIRT1 (Hs01009006_m1). The resulting transcript abundance was measured by a Real-Time Polymerase Chain Reaction (RT-PCR) (Applied Biosystems, Waltham, MA, USA).

Statistics

Statistical analyses were performed using SAS (9.4) or Prism (v9.3.0, GraphPad Software Inc.). Results were presented as mean ± standard error of the mean (SEM) unless stated otherwise. Two-tailed P-values ≤ 0.05 were considered statistically significant. The Shapiro-Wilk W-test was used to test the normality of the data. For between group comparisons a two-sample t-test or a Fisher’s exact test was performed. To analyze the differences between the age groups over time and the differences between the control and resveratrol treated group over time for the young and older participants, a linear mixed model for repeated measures over time was used. The final mixed model included group, time and their interactions. The rhythmicity of gene expression was analyzed using ARSER method of R MetaCycle package [38].