Daily rhythms control tight junction proteins in the mouse large intestine, affecting gut leakiness and risk of colitis

Female 5- to 6-week-old ICR mice purchased from Japan SLC (Tokyo, Japan), ICR mPer2^m/m and Clock^Δ19/Δ19 mice were bred under specific pathogen-free conditions. All mice were housed in 12-hours light/12-hours dark conditions (L/D cycles; the light was turned on at 6∶00 AM, Zeitgeber time [ZT] 0, and the light was turned off at 6∶00 PM, ZT12) with ad libitum access to food and water, for at least 2 weeks. All animal experiments were approved by the institutional review board of the University of Yamanashi and the Committee for Animal Experimentation of the School of Science and Engineering at Waseda University.

A quantitative real-time PCR analysis using cDNA from the mouse tissue specimens was performed using the AB7300 real-time PCR system (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions, using primers and probes for mouse Occludin, Zonula occludens (Zo)-1, Claudin-1, 2, 3, 4, 7, Junctional adhesion molecule (JAM), Myeloperoxidase (Mpo), Heme oxygenase (Ho)-1, Cyclooxygenase (Cox)-2, Prostaglandin E synthetase (Pges)-1, Per1, 2, 3, Clock, Bmal1, Cry1, IL-1β, TNF-α and Glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Applied Biosystems) as previously described [11]. The ratio of each gene to that of GAPDH was calculated, and the relative expression levels are shown.

IECs were isolated as described previously [12]. Briefly, an inverted large intestine of mice was cut into four segments and the segments were transferred to a 50-ml tube containing 10 ml RPMI1640 medium/5% FCS/25mM Hepes. The tube was shaken at 37°C at 100 rpm for 45 min. After incubation, vortex the cell solution vigorously for 20s and pass through a 70 µm cell strainer. Subsequently, the cells were suspended in 30% (wt/vol) Percoll and centrifuged at 20°C at 400×g for 20 min. After centrifugation, cells at the top of the 30% Percoll were enriched with IECs.

IECs were lysed in buffer containing 0.1 M Tris-HCl (pH 7.5)/2% SDS/10% glycerol/5% 2-mercaptoethanol, boiled at 95°C for 5 min and centrifuged at 15000 rpm for 5 min and analyzed by reducing SDS-PAGE. Immunoblots were performed using antibodies against Claudin-1 (Abcam, Cambridge, UK, ab15098, 1∶100), Occludin (Applied Biosystems, 71–1500, 1∶250), Clock (Abcam, ab3517, 1∶200), Bmal1 (Abcam, ab3350, 1∶200) and β-actin (Santa Cruz, Texas, sc-1615, 1∶200). Quantitative analysis of Western blotting was done by using the Scion Image Software package and relative intensities of the target proteins to β-actin were shown.

CHiP was performed using SimpleChIP® Enzymatic Chromatin IP Kit (Agarose Beads) (Cell signaling, Tokyo, Japan, #9002). Isolated IECs or cultured cells (HCT116 and Caco-2) were cross-linked in 1% formaldehyde, followed by the CHiP assay according to the manufacturer’s protocol. Rabbit polyclonal antibodies to CLOCK (ab3517) and BMAL1 (ab3350) were from Abcam (Cambridge, UK). DNA was purified and Q-PCR was done using SYBRgreen reagent (QIAGEN, Hilden, Germany). Primers used were as follows: mouse Occludin promoter forward 5′-CTCCCATCCGAGTTTCAGGT-3′ and reverse 5′-GCTGTCGCCTAAGGAAAGAG-3′, mouse Claudin-1 promoter forward 5′-GTTTGCAGAGACCCCATCAC-3′ and reverse 5′-AGAAGCCAGGATGAAACCCA-3′, human Occludin promoter forward 5′-CTCCCATCCGAGTTTCAGGT-3′ and reverse, human Claudin-1 promoter forward 5′-CTCCCCGCCTTAACTTCCTC-3′ and reverse 5′-CAGGAAGGCGAGAATGAAGC-3′, 5′-GGAGTGTAGGTGTGGTGTGT-3′.

A human Occludin (–60 to +240) or Claudin-1 (–1010 to +360) promoter fragment containing E-box elements was amplified by PCR from genomic DNA of TIG120 cells and cloned into the pGL3 vector (Promega, Fitchburg, Wisconsin) upstream of firefly luciferase. Primer sequences used were as follows: human Occludin promoter forward 5′-CTCGAGAAAGTGCTGAGTGCCTGGAC-3′ and reverse 5′-AAGCTTTCGGTGACCAATTCACCTGA-3′, human Claudin-1 promoter forward 5′-CTCGAGGGAAACTACAGTCCCAGCGA-3′ and reverse 5′-AAGCTTGATGTTGTCGCCGGCATAGG-3′. These reporter constructs were transfected into HCT-116 or Caco-2 cells cultured in 96-well plates using Fugene6 (Promega) in the presence of pcDNA3, pcDNA3-Clock or pcDNA3-Bmal1 expression plasmid. pcDNA3-Clock and pcDNA3-Bmal1 expression plasmids were kindly gifted by Dr. Fuyuki Sato (Hirosaki University, Aomori, Japan). At 48 hours after transfection, the luciferase activity was measured using the luciferase reporter assay system (TOYO ink, Tokyo, Japan).

Colon permeability to Evans blue was measured in vivo. This is a well-validated method for assessing epithelial permeability [13]. In isoflurane-N₂O anesthesia, spontaneously breathing animals were placed in a supine position on a heating pad and a laparotomy was performed. A small polyethylene tube (G22) was inserted into the proximal colon ascendens (immediately adjacent to the cecum) and secured by a ligature. Via this tube the colon was gently flushed until all stool was rinsed out, and 1 ml of Evans blue 1.5% (Sigma-Aldrich Japan, Tokyo, Japan, E2129) in PBS was instilled into the colon and left in place for 15 min. Then the colon was rinsed with PBS, until the perianal washout was clear. Animals were euthanized, and the colon was rapidly taken out. It was rinsed again with several milliliters of PBS, followed by 1 ml of 6 mM N-acetylcysteine to eliminate dye sticking in the colonic mucus. The colon was opened and rinsed once more with PBS, and its length was recorded. The whole colon was placed in 2 ml N,N-dimethylformamide for 12 h to extract the Evans blue dye. The dye concentration in the supernatant was measured spectrophotometrically at 610 nm and given as extinction per gram colonic tissue.

Colon permeability was measured in isolated intestinal segments using FITC-dextran or Horseradish peroxidase (HRP) as described previously [14], [15]. Briefly, 6-cm segments of the large intestine were removed, rinsed with ice-cold saline, everted, filled with 700 µl PBS and ligated at both ends. The filled colon segments were incubated in PBS containing 1 mM FITC-dextran (average mol wt, 4000 Sigma-Aldrich Co. Missouri, USA) or 0.4 mg/ml HRP (type IV, 300 units/mg solid, Sigma-Aldrich Co.). The colon sacs were removed after 45 minutes and the contents (∼500 µL) of each sac was collected carefully using a 1-ml syringe. The amount of FITC-dextran that transverse the colon was quantified by fluorescence plate reader at 521 nm. The HRP activity in the contents of each sac was determined spectrophotometrically based on the rate of oxidation of pyrogallol.

Colitis was induced by feeding mice 5% DSS dissolved in drinking distilled water (mol wt 5000; Wako Pure Chemical Co., Osaka, Japan) for 7 days, as previously described [16].

The stool scores were calculated based on a method by Cooper et al. with some modifications [17]. Briefly, the scoring was based on a maximum score of 4. Stool consistency was graded as: 0 = firm, 1 = loose, 2 = diarrhea. Blood in the stool was also evaluated on a 0- to 2-point scale. 0 = no blood, 1 = occult blood, 2 = gross rectal bleeding.

The concentrations of tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6) in the mouse colon tissue were determined using the mouse TNF-α and IL-6 enzyme-linked immunosorbent assay (ELISA) kits (R&D Inc, Minneapolis, MN), respectively, according to the manufacturer’s instructions.

Recipient wild-type mice or mPer2^m/m mice were lethally irradiated with 1050 rads using a Co irradiator and then were injected intravenously 6 hours later with 1×10⁷ bone marrow cells derived from the femurs of the respective donors.

Values represent the mean ± S.D. The statistical analysis was performed using the unpaired Studentns t-test. The statistical analysis of time-of-day-dependent variations in mRNA and protein levels was performed using a one-way analysis of variance (ANOVA). A value of P<0.05 was considered to be significant.

The mRNA levels of Occludin, Zo-1, Claudin-1, -2, -3, -4, -7 and JAM-A, major components of TJ, were examined in the colon tissue samples obtained from wild-type mice and mice with the loss-of-function mutation of a key clock gene Period2 (Per2; mPer2^m/m) every 6 hours during the day to investigate the possible circadian variations [18]. We found that the mRNA levels of Occludin and Claudin-1, but not Claudin-2, -3, -4, -7, Zo-1, and JAM-A, in the colon showed a time of day-dependent variation in wild-type mice (Fig. 1A). Detailed kinetic analysis of mRNA levels of Occludin and Claudin-1 at every 4 hours during the day revealed that the mRNA levels showed a nadir at Zeitgeber time (ZT) 12 (18∶00 PM) (Fig. S1A). We also found that Per1, Per2, Cry1, and Bmal1 mRNA levels in the colon exhibited a time of day-dependent variation (Fig. S2) and the kinetics of Per1 and Per2 mRNA expression were the opposite phase to Occludin and Claudin-1 mRNA kinetics (Fig. 1A). Consistently, the expression levels of Occludin and Claudin-1 proteins in colonic IECs showed a similar time of day-dependent variation to the mRNA expressions (Fig. 1B and C). Detailed kinetic analysis of protein levels of Occludin and Claudin-1 at every 4 hours during the day revealed that the protein levels showed a nadir at ZT 16 (22∶00 PM) (Fig. S1B), suggesting that there was a time-lag between mRNA and protein expression changes of Occludin and Claudin-1. However, please note that the circadian changes of Occludin, but not Claudin-1, protein levels did not reach a statistical significance (p = 0.053) and showed only a trend of circadian oscillations in this analysis. In contrast to wild-type mice, mPer2^m/m mice showed constitutively high expression levels of Occludin and Claudin-1 mRNAs and proteins without oscillations (Fig. 1). These results suggested that Occludin and Claudin-1 expressions in colonic epithelium showed circadian oscillations, which relied on normal Per2 activity.

The Clock-Bmal1 heterodimer binds directly to the E-box element of the clock controlled genes (CCGs) [19], [20]. Occludin and Claudin-1 possess E-box elements in the promoter regions of the genes [21], [22]. To examine whether Clock and Bmal1 bind to the E-box elements of Occludin and Claudin-1 promoter regions, we performed CHiP assays by using IECs isolated from the colonic tissues of wild-type and mPer2^m/m mice every 12 hours during the day.

Clock and Bmal1 bound to the E-box elements of Occludin and Claudin-1 promoter regions in mouse colonic IECs (Fig. 2A). Interestingly, the ratio of binding showed a time of day-dependent variation in IECs from the colon of wild-type mice, with a high at day (ZT4: 10∶00 AM) and low at night (ZT16: 10∶00 PM) (Fig. 2A). However, such a variation was absent in mPer2^m/m mice (Fig. 2A). The bindings of Clock and Bmal1 to the E-box elements of Occludin and Claudin-1 promoter regions were also confirmed in human colonic epithelial cell lines HCT-116 and Caco-2 cells (Fig. 2B and Fig. S3A).

To investigate whether Clock and Bmal1 regulate transcriptional responses of Occludin and Claudin-1, we performed reporter assays using luciferase-reporter plasmids containing human Occludin (–60 to +240) or Claudin-1 (–1010 to +360) promoter fragments with E-box elements. Transfection of the Occludin- or Claudin-1 promoter reporter plasmids into HCT-116 cells showed ∼20-fold increases in luciferase activity compared with transfection of control pGL3-basic plasmid without containing the promoter elements (data not shown), suggesting that these plasmids were activated in HCT-116 cells. Overexpression of Clock or Bmal1 in HCT-116 cells further increased the luciferase activity of Occludin- or Claudin-1-reporter plasmids (Fig. 2C). Similar findings were also observed in Caco-2 cells (Fig. S3B). These results suggested that Clock and Bmal1 bound directly to the E-box elements of Occludin and Claudin-1 promoters and increased the transcriptional responses.

The possibility of altered epithelial permeability in wild-type and mPer2^m/m mice was investigated since the TJ proteins play an important role in epithelial permeability through the control of the paracellular pathway for movement of intestinal fluid and macromolecules [1]–[3]. Colonic permeability to Evans blue in vivo and to FITC-dextran and HRP ex vivo showed a time-of-day dependent variation in wild-type mice, with a high at night (ZT16: 10∶00 PM) and low at day (ZT4: 10∶00 AM), which was the opposite phase to Occludin and Claudin-1 protein levels (Fig. 2and3). However, such a variation was absent in mPer2^m/m mice and, as expected, the colonic permeability was lower in mPer2^m/m mice than that in wild-type mice (Fig. 3). These results suggested that colonic permeability was under control of the circadian clock, which was associated with temporal alterations of colonic permeability.

We then investigated whether mPer2^m/m mice with reduced colonic permeability were more resistant to colonic injury induced by DSS, because intestinal permeability was associated with susceptibility to colitis [1]–[3], [23].

mPer2^m/m mice given 5% DSS in drinking water over a period of 7 days showed less body weight loss, bloody stool, and colon shrinkage than the wild-type mice. Reduced levels of TNF-α and IL-6 proteins in the colon were observed in the mPer2^m/m mice in comparison to wild-type mice (Fig. 4A–D). The mRNA expression levels of pro-inflammatory molecular markers, myeloperoxidase (Mpo), heme oxygenase (Ho)-1, cyclooxygenase (Cox)-2, and prostaglandin E synthase (Pges)-1 in the colon were also less in mPer2^m/m mice than those in wild-type mice (Fig. S4A). mPer2^m/m mice drank equivalent amounts of DSS-containing water to wild-type mice per day during the period of DSS challenge (day 0–7; Fig. S4B). These results suggested that mPer2^m/m mice were more resistant to DSS-induced colitis in comparison to wild-type mice.

Bone marrow chimeric mice in which the donors and/or the recipients were either wild-type or mPer2^m/m mice were generated to further investigate whether the Per2 mutation in the tissue resident cells (most likely colonic epithelium) was important for the protection against DSS-induced colitis. DSS challenge of chimeric mice lacking the Per2 mutation in hematopoietic cells induced pro-inflammatory markers similar to those observed in wild-type mice reconstituted with wild-type bone marrow cells (Fig. 4E–H, Fig. S5). In contrast, chimeric mice lacking Per2 in the non-hematopoietic cells were resistant to DSS-induced colitis similar to the mPer2^m/m mice reconstituted with mPer2^m/m bone marrow cells, as indicated by the changes of those parameters (Fig. 4E–H, Fig. S5). These results suggested that the Per2 mutation conferred protection against DSS-induced colitis when it occurred in non-hematopoietic cells (presumably the cells of the colonic epithelium).

To further ascertain that the circadian clock regulated a time of day-dependent variation of Occludin and Claudin-1expression in the colon, we examined the mRNA levels of Occludin and Claudin-1 in the colon tissue samples obtained from wild-type mice and mice with the loss-of-function mutation of another key clock gene Clock (Clock^Δ19/Δ19 mice) every 6 hours during the day. The mRNA levels of Occludin and Claudin-1 in the colon showed a time of day-dependent variation in wild-type mice whereas the variation was absent in Clock^Δ19/Δ19 mice (Fig. S6A). Interestingly, the mRNA levels of Occludin and Claudin-1 in the colon were relative low in Clock^Δ19/Δ19 mice compared with those in wild-type mice. Consistently, Clock^Δ19/Δ19 mice were more sensitive to the colonic injury induced by DSS than wild-type mice as judged by body weight changes, stool property, and Ho-1 and Pges-1 expression (Fig. S6B–D). These results also suggested that disruption of the circadian clock affected TJ protein expression and barrier function. The opposite phenotypes observed in Clock^Δ19/Δ19 mice to those in mPer2^m/m mice may be attributed to the opposite function of Clock and Per2 in the circadian clockwork; Clock acts as a positive limb whereas Per2 acts as a negative limb in the transcriptional-translational feedback loop in the molecular clock machinery.

Major proinflammatory cytokines IL-1β and TNF-α have been shown to regulate epithelial barrier function [2]. To investigate whether the circadian clock also affects IL-1β and TNF-α expression, we examined mRNA levels of IL-1β and TNF-α in the colon tissue samples obtained from wild-type mice and mPer2^m/m mice every 6 hours during the day. We found that IL-1β and TNF-α mRNA levels did not show circadian oscillations in wild-type mice and they were comparable between wild-type and mPer2^m/m mice (Fig. S7).