BMAL1 Regulates the Daily Timing of Colitis

The animal studies outlined were reviewed and approved by animal care regulatory boards at McMaster University (#AUPP 19-02-09), or the University of Windsor (#AUPP 20-17). Bmal1^+/- mice (Jackson Laboratories, Bar Harbor, B6.129-Arntltm1Bra/J stock #009100) and Per2::Luciferase mice (Jackson Laboratories, Bar Harbor, B6.129S6-Per2tm1Jt/J stock #006852) were housed in a barrier facility under a 12-hour light/12-hour dark (LD) cycle (with lights on at 7am/Zeitgeber Time 0, and light off at 7pm/Zeitgeber Time 12). Litter mates were used in experiments and were co-housed prior to testing. Normal chow was administered ad libitum.

Mice aged 12-14 weeks were housed individually for 1 week before beginning the experiments to acclimate to the new cage. Bedding taken from their previous co-housed cages was sprinkled into the new individual cages in order to further control for microbiome effects. Untreated mice were collected following the same 1 week acclimation period (Figures 1, 4, 8). A 4% (w/v) solution of DSS (MW: 36-50 KDa, MP Biomedicals, Solon, Ohio) dissolved in normal drinking water was administered ad libitum to the mice over the course of 6 days (Figures 2–4, 6–9). Water consumption of normal drinking water and DSS water was measured twice a day (ZT0 and ZT12) and no differences were observed between controls and mutants in the amount of DSS-water consumed (data not shown). This same colitis assay was repeated in McMaster University using a 5% (w/v) solution of DSS in normal drinking water over 5 days (data shown in Figure 5).

After the 6-day DSS treatment, the DSS solution was removed and replaced with normal drinking water. Mice were tested 10 days after DSS removal (Figures 3, 8, 9). Weight and disease activity (see below) was monitored daily. If weight loss exceeded 25% over the duration of 3 days, or the mice appeared lethargic, it was taken as a sign of extreme sickness and the experiment was terminated and the mice humanely euthanized (data is shown in Figure 3A).

Mice were euthanized using CO₂ followed by cervical dislocation at various intervals over 24 hours (Zetigeber Times: ZT0, ZT4, ZT8, ZT12, ZT16, ZT20 as indicated in text). Colon tissue was removed and flushed with ice cold phosphate buffer saline (PBS). The tissue was cut longitudinally to open the lumen, and was then rolled up from the proximal to distal end. The dissected colon tissue was fixed in 4% paraformaldehyde (PFA) (Electron Microscopy Sciences, Hatfield, PA) diluted in PBS for 2 hours. Segments of the proximal, medial and distal colon were collected for gene expression analysis and placed in RNAlater buffer (Qiagen, Germantown, MD) at 4°C overnight (O/N), then stored at -80°C.

Disease Activity Index (DAI) score was monitored during DSS treatment to evaluate the progression of colitis symptoms. DAI score consisted of daily measurements of weight loss compared to initial weight, stool consistency and bleeding. Each disease criteria is assigned a score based on the severity of the symptom (weight loss 0:<1%, 1: 1–5%, 2: 5–10%, 3: 11-14%, 4:>15%, stool consistency 0: normal, 2: loose stools, 4: diarrhea, stool blood 0: negative, 2: positive, rectal bleeding 4: if present) (Kim et al., 2012).

Fixed colon tissues were washed 3 times in PBS to remove any PFA before being submerged in 30% (w/v) sucrose in PBS for 24 hours at 4°C. The tissue was transferred into Tissue-Tek O.C.T compound (Sakura, Flemingweg, Netherlands) and left at 4°C for 1 hour before being frozen at -80°C. The frozen tissue was sectioned at 10μm using a Leica CM 1950 Cryostat (Leica Biosystems). Sectioned tissues were washed 3 times in room temperature PBS for 5 minutes each. This was followed by antigen retrieval and another 3 washes of PBS for 5 minutes as needed. Tissue morphology was assessed using Hematoxylin & Eosin (Electron Microscopy Sciences, Hatfield, PA), and goblet cells were stained using Alcian Blue (Electron Microscopy Sciences, Hatfield, PA). H&E stained tissues were blindly scored for histopathology based on the degree of epithelial damage pertaining to crypt architecture (normal, 0 - severe crypt distortion with loss of entire crypts, 3), degree of inflammatory cell infiltration (normal, 0 - dense inflammatory infiltrate, 3), muscle thickening (base of crypt sits on the muscularis mucosae, 0 - marked muscle thickening present, 3), goblet cell depletion (absent, 0- present, 1) and crypt abscess (absent, 0- present, 1) (Cooper et al., 1993). DAB staining was performed using HRP/DAB (ABC) Detection IHC Kit as per manufacturers protocol (ab64261, Abcam). Antibodies and staining protocol used are listed in Supplementary Table 1.

Samples were imaged using the Zeiss Axio Scan.Z1 slide scanner or Zeiss LSM900 Confocal (Zeiss, Toronto, Canada). Crypt length and area measurements were collected using the Zen Blue Software (Zeiss). Images were processed using Adobe Photoshop (San Jose, CA). Antibody-positive cells were quantified as the total number of labeled cells or nuclei/crypt, area (for measurements combining crypts and lesions), Villin positive epithelial cells, or CD45 positive blood cells. Ten crypts were sampled per animal for each colon region (proximal and distal) for quantification of ChgA, Nr1d1, CD45, pHH3, cleaved caspase-3 and Cldn1. Regions were averaged together for results shown if no significant differences were observed. The Villin/Nr1d1 and Cldn1 stained tissue were scanned using the Zeiss LSM900 Confocal while all remaining samples were imaged using the Axio Scan.Z1.

Per2::Luc mice 14-15 weeks of age were euthanized and sections of 0.5cm of the proximal colon were collected in Hanks’ balanced salt solution supplemented with HEPES, NaHCO3, and penicillin-streptomycin. Proximal colon from untreated animals (n=5) or animals treated for 6 days with 4% DSS (n=6) were cleaned of adipose tissue and cultured on 0.4 mm membrane inserts (Millipore) in sealed 35 mm Petri dishes with 1.5 mL of DMEM supplemented with HEPES, NaHCO3, penicillin-streptomycin, B27 and beetle luciferin (Gold Biotechnologies). Bioluminescence rhythms were measured for at least 4 days with a luminometer (Actimetrics Inc) in a light-tight incubator set to 37°C, period and amplitude were calculated with LumiCycle (Actimetrics Inc) (Kloehn et al., 2016; Carmona-Alcocer et al., 2018).

A bullet blender was used to homogenize tissue samples (Next Advance, Troy, NY). RNA was isolated using the RNEasy mini Purification Kit (Qiagen). Prior to production of complimentary DNA, RNA was purified from the DSS via LiCl precipitation as described in (Viennois et al., 2018). LiCl purification was performed on all samples regardless of treatment (DSS and untreated). Complementary DNA was produced using the iScript RT Supermix (Bio-Rad) according to the manufacturer’s protocol.

Transcriptional gene expression was analyzed using iTaq Universal SYBR Green Supermix (Bio-Rad) on a Viia7 real-time polymerase chain reaction plate reader (Thermo Fisher Scientific). Primer sequences were validated with an efficiency of approximately 80–110. The primers used are listed in Supplementary Table 2. Primer sequences were obtained from PrimerBank, Brooks et al. (2021), Kwon et al. (2019) and Yamazaki et al. (2011). All gene expression was normalized to Gapdh.

Using a whole tissue lysate of DSS treated (5 day, 5% w/v) colon tissue, levels of the pro-inflammatory cytokine TNF was determined by ELISA.

RNA in situ probe hybridization and detection was performed using RNAScope protocol according to manufacturer’s instructions (Advanced Cell Diagnostics/Bio-Techne). The probes used were RNAscope Probe- Mm-Lgr5 ACDBio Cat# 312171 and RNAscope Probe- Mm-Hopx-C2 ACDBio Cat# 405161-C2. RNAScope slides were imaged and quantified by counting visible RNA positive puncta and averaging puncta per crypt or puncta per 5000μm² in lesions.

Statistical analysis was performed using Prism 9 (Graphpad). Unpaired t-test, One-way ANOVA or Two-way ANOVA were used with Tukey’s multiple comparisons post-test; see Supplementary Table 3 for all details. Rhythmic profile was evaluated with Chronos-Fit program (P. Zuther, S. Gorbey and B. Lemmer, Chronos-Fit 1.06, 2009).

The Bmal1^-/- mutant has been reported to have metabolic and age-related phenotypes (Kondratov and Antoch, 2007), but we have previously shown that the small intestinal epithelium is morphologically and cellularly normal in young adult mice (Stokes et al., 2017). In order to investigate pre-existing differences in the colon epithelium, Bmal1^+/+ control and Bmal1^-/- mutant mice were collected at 13-15 weeks of age, before age-related phenotypes are apparent, and their colon epithelium was compared. The loss of Bmal1 produced no obvious differences in tissue morphology (Figure 1A). The number of crypts in both proximal and distal regions of the colon, and average length of the crypts in these were indistinguishable between Bmal1^+/+ control and Bmal1^-/- mutant genotypes, indicating no overt differences in cell proliferation/production (Figures 1B, C). We tested for changes in muscle thickness in the Bmal1^+/+ control and Bmal1^-/- mutant colon, to determine whether inflammation is Bmal1-dependent under normal conditions, but this was equivalent in both genotypes (Figure 1D). To directly investigate cell proliferation, we stained tissue using antibody to Ki67, a proliferation marker that labels cells in S/G2/M phases; cell proliferation is equivalent between these genotypes (Figures 1E, F). Goblet cells produce a protective mucus in the intestine, and previous studies have shown that there are fewer goblet cells present in Per1/2 mutant mice which predisposes these to colitis (Pagel et al., 2017). However, the Bmal1^-/- mutant contains equal goblet cell numbers and the staining of mucin is similar to controls, suggesting no obvious mucus deficiency (Figures 1G, H). Enteroendocrine cells are another type of secretory cell which act as key sensors of metabolites (Karaki et al., 2006; Nøhr et al., 2013). We discovered a significant decrease in enteroendocrine cell presence, marked by Chromogranin A (O’Connor, 1983; Cetin et al., 1989) in Bmal1^-/- mutant colon tissue (Figures 1I, J). Overall, our data shows that the colon epithelium does not appear to require Bmal1 for epithelial cell production or tissue morphology in undamaged conditions but has a ~50% decrease in enteroendocrine cells.

Several reports have indicated that inflammation can lower circadian rhythms (Liu et al., 2017; Weintraub et al., 2020). We therefore tested to see whether circadian clock-driven rhythmic gene expression in the colon is present during colitis. Per2::Luciferase (Per2::Luc) mice, a protein reporter of the circadian clock that contains Luciferase fused to the Per2 gene (Yoo et al., 2004), were treated with DSS to induce colitis. Colon tissue was then dissected from mice and colon tissue explants were tested for luminescence over a period of 4 days. This revealed that Per2::Luc DSS-treated colons exhibit dampened rhythmic luminescence readings compared to untreated controls (Figure 2A). DSS-treated Per2::Luc mice that are rhythmic show the same circadian period (Figure 2B), however the amplitude of these rhythms is much lower (Figure 2C) and half of DSS-treated mice do not show any rhythmicity (Figures 2D, E). When present, clock gene rhythms persist for several days in both controls and DSS-treated Per2::Luc mice, indicating that a free-running clock is present in the colon. These data suggest that clock transcriptional output in mice with colitis is weaker than normal, consistent with previous findings that show RT-qPCR rhythms in the DSS-treated colon are weaker (Wang et al., 2018).

The colon contains epithelial cells, but also many non-epithelial cells in the supporting lamina propria, surrounding muscle, blood vessels, and immune system. We further tested for circadian rhythms in epithelial and immune system cells at the protein level using histology sections. Bmal1^+/+ control and Bmal1^-/- mutant mice were treated with DSS, and colon tissue was collected at three time points. We probed for Nr1d1 protein, the transcriptional target of Bmal1, to serve as a reporter of circadian clock output, and co-stained for Villin and/or CD45⁺ to distinguish the epithelial colon cells from blood-derived cells that are interspersed in this tissue (Pilling et al., 2009). The expression of Nr1d1 was evident in both Villin⁺ epithelial cells and CD45⁺ blood cells (Figures 2F–H). In control animals, the number of Nr1d1⁺ cells are increased in both cell populations at ZT20 with decreased expression earlier at ZT4 and ZT12 in both Villin⁺ epithelial cells (Figure 2G) and CD45⁺ blood cells (Figure 2I). However, this increase in Nr1d1 in controls is subtle, consistent with previous reports that have shown Nr1d1 expression is suppressed during colitis (Wang et al., 2018). Nr1d1 was noticeably decreased in Bmal1^-/- mutant Villin⁺ epithelial cells (Figure 2G) and CD45⁺ blood cells (Figure 2I), with no significant daily rhythms in expression present (Figures 2F–I). These data suggest that Bmal1-dependent circadian rhythms are present in both the epithelial and blood cells in the colon during colitis. Together our data confirm that clock rhythms occur during colitis, and suggest that their weaker amplitude is not a result of desynchrony between colon epithelial and immune system cells.

To investigate the effects which a non-functional circadian clock has on IBD, Bmal1^+/+ controls and Bmal1^-/- mutants were treated with DSS, and their recovery tracked for an additional 10 days. We observed a drastic decrease in the recovery of Bmal1^-/- mutant mice, indeed nearly all Bmal1^-/- mutants are not able to tolerate colitis and had to be euthanized much earlier than controls (Figure 3A). This supports a critical role for Bmal1 in IBD severity. We next tested if the increased disease severity of Bmal1 mutants indicates that the colon epithelium of mice lacking a functional circadian clock is compromised. Colon tissue from Bmal1^+/+ controls and Bmal1^-/- mutants were analyzed after DSS treatment. Bmal1^-/- mutant mice exhibited a decrease in overall colon length (Figures 3B, C). This supports recent studies which have shown decreased colon length in Rorα intestinal epithelial cell conditionally knocked out mice following DSS treatment (Oh et al., 2019), and in a different Bmal1^-/- mutant mouse strain than used in our current study (Wang et al., 2018). Pronounced morphological differences in Bmal1^-/- mutant colon epithelia were readily apparent (Figure 3D). An increase in the size of lesions, regions of epithelium which are morphologically abnormal and contain no crypts or structure was observed in Bmal1^-/- mutants (Figure 3E). Muscle tissue thickness, indicative of inflammation, was also increased in Bmal1^-/- mutants (Figure 3F). Histopathological scores assessed in a double-blind analysis, confirmed that Bmal1^-/- mutant colon tissue exhibits increased muscle thickening (Figure 3G). Loss of architecture in Bmal1^-/- mutant colon was also noted (Figure 3H), consistent with the previously mentioned increase in lesion size present in Bmal1^-/- mutants. Additionally, immune cell infiltration was increased in Bmal1^-/- mutant colons (Figure 3I). Although no differences were seen in goblet cell depletion and crypt abscess scores (Supplemental Figure 1A, B), together these features lead to increased overall histology score in Bmal1^-/- mice with colitis (Figure 3J). These results demonstrate that loss of colon epithelial structure underlie the increased colitis observed in Bmal1 mutant mice.

The circadian clock has recently been shown to play a role in the rhythms of the microbial response, and the Antimicrobial Peptide (AMP) gene expression of both Reg3g and Lcn2 exhibits diurnal rhythms (Bellet et al., 2013; Brooks et al., 2021). We therefore tested the expression of AMPs in untreated mice and mice treated with DSS to induce colitis (all samples collected at ZT16, when Reg3g and Lcn2 have been shown to be higher). Both Lcn2 and Reg3g exhibited significantly increased expression in Bmal1^+/+ control colon tissue following induction of DSS-colitis (Figures 4A, B). However, Lcn2 and Reg3g expression in DSS-treated Bmal1^-/- mutant tissue remains low. No significant differences were detected between Bmal1^+/+ controls and Bmal1^-/- mutants in the expression of the S100a8 AMP, or the anti-microbial enzyme Lyz1 (Figures 4C, D), although, the expression of S100a8 increased during colitis in both Bmal1^+/+ controls and Bmal1^-/- mutants. These data suggest that certain mechanisms of anti-microbial activity are compromised during colitis in Bmal1^-/- mutants.

To further investigate the colon epithelial integrity in the absence of clock function, genes involved in barrier function were measured. Previous studies have indeed shown that genes encoding the tight junction proteins Claudin1 (Cldn1) and Occludin (Zo1) are expressed in a daily rhythm in wildtype mice (Oh-oka et al., 2014). Since increased barrier permeability plays an important role in IBD, we stained for Cldn1 in Bmal1^+/+ controls and Bmal1^-/- mutant tissue with colitis (Figure 4E). No differences in Cldn1 were apparent. We also tested the transcriptional expression of Cldn1, Zo1, Tight junction protein 1 (Tjp1) and Mucin 2 (Muc2), however, no differences were detected (Figures 4F–I). These data suggest that the Bmal1^-/- mutant does not have an obvious barrier phenotype.

If circadian rhythms have a regulatory role in colitis, disease severity should vary depending on time of day. In order to monitor the progression of inflammatory bowel disease, a disease activity score was assessed based on the severity of individual weight loss, blood presence in stool, softening of stool or diarrhea, and rectal bleeding (Kim et al., 2012). Disease activity was tested daily at ZT0 and ZT8 to assess differences in colitis severity at different times over a period of five days. ZT0 is the early morning, the end of nocturnal mouse activity, and ZT8 is just past midday, the middle of the rest phase where regenerative processes are thought to occur. When scored at ZT0 Bmal1^+/+ controls show significantly lower colitis than Bmal1^-/- mutants particularly in stool consistency (diarrhea) (Figure 5A) and fecal blood (Figure 5B), leading to increased total disease severity (Figure 5C) despite having no differences in weight loss (Supplementary Figure 2A). These differences are not present later in the day at ZT8. Instead, disease scores for Bmal1^+/+ were increased to levels that were not significantly different from Bmal1^-/- mutants (Figures 5D–F and Supplementary Figure 2B). We then re-analyzed Bmal1^+/+ disease scores and noted that stool consistency (diarrhea) is more severe at ZT8 driving an overall higher total disease score if mice were examined later in the day (Figures 5G–I and Supplementary Figure 2A, B). Notably, Bmal1^-/- mutants remained at the same high levels at both times examined (no significant differences in Figures 5A–F). These results suggest that colitis exhibits time-dependent disease activity in Bmal1^+/+ circadian-competent animals, with increased severity during rest (ZT8). Rhythms in disease severity in Bmal1^-/- clock-deficient animals are not present and instead disease activity is consistently high at different times of day.

Daily changes in disease activity suggest underlying changes of cellular activity in the colon leading to increased inflammation. Thus, we investigated the pro-inflammatory cytokine expression in Bmal1^+/+ control and Bmal1^-/- mutant colon tissue following DSS-colitis. While no differences are present in IL-1β, and IL-6, an increase in TNF expression was detected in Bmal1^-/- mutant colon tissue at ZT4 (Figures 6A–C). Two additional cytokines, IL-10 and IL-22, did not show a difference between genotypes (data not shown). Given that the change in TNF expression occurs at the onset of rest (for mice), we collected Bmal1^+/+ control and Bmal1^-/- mutant colon tissue at ZT0 and measured TNF colonic protein concentrations. TNF protein is higher in the Bmal1^-/- mutant colon suggesting an increase in inflammation (Figure 6D).

Next, we examined histological samples and determined cellular infiltration histopathological scores over 24 hours (ZT0, 4, 8, 12, 16, and 20). Cellular infiltration in the colon, by score, does not vary over time in either Bmal1^+/+ controls or Bmal1^-/- mutants (Figure 6E), however Bmal1^-/- mutants have significantly higher cellular infiltration scores overall (Figure 3I). To further test this, we probed the same cohort of Bmal1^+/+ control and Bmal1^-/- mutant mice for immune system blood cells (marked by anti-CD45) (Pilling et al., 2009) which infiltrate tissue lesions during DSS-induced colitis (Figure 6F). We examined both lesioned regions and morphologically normal regions to determine if a difference was present. When the colonic lamina propria (LP) was examined, no differences in the daily timing of CD45⁺ cellular presence was evident in either Bmal1^+/+ and Bmal1^-/- mice (Figure 6G). Reanalyzing these data by averaging CD45⁺ number irrespective of time shows that an increase in CD45⁺ cellular infiltration is present in the Bmal1^-/- mutants (Figure 6H). This is consistent with the increased cellular infiltration histology score (compare Figure 6E and Figure 3G). We then looked specifically at lesioned tissue in the DSS-treated colons to test the immune cell response in areas of greater injury. A significant increase in CD45⁺ cells at ZT12, the time when mice normally become active, was observed in control Bmal1^+/+ animals relative to times of rest (Figure 6I). This daily change in blood cell numbers is not present in Bmal1^-/- mutant tissue who have low levels of CD45⁺ cellular infiltration at these same times. We conclude that controls show an increase of immune cell presence in DSS-damaged tissue that peaks at the same time as activity usually begins, but these changes are absent in circadian-deficient Bmal1^-/- mice that show elevated inflammation throughout the colon.

The epithelial barrier is damaged during IBD and is required to ultimately resolve inflammation. We tested the epithelial response during DSS-treatment, to determine if Bmal1^-/- mutants have defects in regeneration. We focused on the non-ulcerated regions of the colon where crypt structures containing stem and precursor cells persist during colitis. Substantial proliferation differences (marked by Ki67) are evident in Bmal1^+/+ control mice based on time of day, while proliferation is consistently low in Bmal1^-/- mutants (Figures 7A, B). This contrasts with crypts prior to the induction of colitis, where there is no difference in proliferation (Figures 1E, F). We further tested proliferation using a cohort of Bmal1^+/+ control and Bmal1^-/- mutant mice treated with DSS, collected at 6 time points during the entire 24-hours. Diurnal rhythms in colon epithelial mitosis (marked by phosphorylated-Histone-H3; pHH3) are present in Bmal1^+/+ controls throughout the day, with a peak in mitosis taking place around ZT0 at the onset of rest (Figures 7C, D). These rhythms are completely disrupted and decreased in Bmal1^-/- mutants. Of note, we also tested for epithelial cell apoptosis in the same samples (marked by activated Caspase-3; Cas3) but no differences are apparent in a time- or genotype-dependent manner (Figures 7E, F). The Ki67 and pHH3 data suggest that daily proliferation rhythms during colitis are present in the epithelium of Bmal1^+/+ controls and are suppressed in circadian-deficient Bmal1^-/- mutants. These data indicate that the regenerative ability of the colon follows daily rhythms that are opposite to those observed in blood cell activity (Figure 6), peaking at the end of the active period rather than the onset of activity. Like colitis symptoms (Figure 5), these rhythms are not present in Bmal1^-/- mutant mice.

The colon epithelium is regenerated by stem cells and precursors that replenish cells lost during infection or injury (Yui et al., 2018). We probed for markers of these cells in the crypts of the colon epithelium prior to DSS-treatment, during DSS-induced colitis, and following a 10-day recovery period (Figure 8A). RNA probes were used to identify intestinal stem cells, marked by Lgr5, that renew the colon daily (Barker et al., 2007), as well as fetal-like precursors marked by Hopx, involved in the colitis response (Takeda et al., 2011; Wang et al., 2019). In the untreated colon, the expression of Hopx and Lgr5 is significantly weaker in the Bmal1^-/- mutant epithelium (Figures 8B–D). This is also apparent upon DSS-induced colitis where the expression of Hopx and Lgr5 remains higher in the morphologically intact proliferating crypts of Bmal1^+/+ controls at 6 days but is decreased in Bmal1^-/- mutant animals (Figures 8E–G). Of note, in the lesions present in the same DSS-injured colon tissues Hopx is decreased in Bmal1^-/- mutants (Figures 8H, I), but Lgr5 expression is reduced in both Bmal1^+/+ control and Bmal1^-/- mutant lesions (Figure 8J). These data are consistent with a recent study (Girish et al., 2021) which has shown Lgr5 is sensitive to DSS treatment. Finally, following the 10-day recovery period, Hopx expression is similar in both genotypes but Lgr5 expression remains significant reduced in Bmal1^-/- mutants (Figures 8K–M). At this stage of recovery, Bmal1^+/+ controls no longer have any DSS-induced lesions present (colitis is resolved), but Hopx and Lgr5 expression persists in the Bmal1^-/- mutant lesions (Figures 8N–P). Overall, these data suggest the expression of stem cell and precursor cell markers is reduced in Bmal1^-/- mutants at all stages pre- and post-colitis. We note that these timelines are generally consistent with previous findings (Harnack et al., 2019; Wang et al., 2019), but vary slightly by time perhaps due to the mouse genetic background and/or different levels of DSS used in our study.

To further assess the regenerative ability of mice without a functional circadian clock, we examined morphological differences between Bmal1^+/+ controls and Bmal1^-/- mutants following 10 days of recovery (Figure 9A). An increase in lesions is apparent in Bmal1^-/- colon tissue, who cannot resolve colitis. In contrast, nearly all lesions in control animals that were present (see Figure 3E) are now resolved and crypt architecture has returned to normal (Figure 9B). However, large lesions continue to be present in Bmal1^-/- mutants well into recovery (Figure 9B), and a corresponding decrease in the crypt density of Bmal1^-/- colons is evident (Figure 9C). In the non-lesioned parts of the colon, Bmal1^-/- mutants also exhibit an increased average crypt length reminiscent of the hyperplasia associated with IBD (Erben et al., 2014) (Figure 9D). An increase in average muscle thickness is also present in Bmal1^-/- colon tissue following 10 days of recovery (Figure 9E), similar to that noted during colitis (Figure 3F). These results indicate that the regenerative abilities of Bmal1^-/- mutants are significantly impaired.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.