Alcohol effects on colon epithelium are time dependent

Abstract

Background:

Alcohol intake increases the risk of developing colon cancer. Circadian disruption promotes alcohol’s effect on colon carcinogenesis through unknown mechanisms. Alcohol’s metabolites induce DNA damage, an early step in carcinogenesis. We assessed the effect of time of alcohol consumption on markers of tissue damage in the colonic epithelium.

Method:

Mice were treated by alcohol or PBS, at 4 h intervals for 3 days, and their colons were analyzed for (i) proliferation (Ki-67) and anti-apoptosis (Bcl-2) markers, (ii) DNA damage (γ-H2AX), and (iii) the major acetaldehyde (AcH)–DNA adduct, N²-ethylidene-dG. To model circadian disruption, mice were shifted once weekly for 12 h and then were sacrificed at 4h intervals. Samples of mice with a dysfunctional molecular Clock were analyzed. The dynamics of DNA damage-repair from AcH treatment as well as role of XPA in their repair were studied in vitro.

Results:

Proliferation and survival of colonic epithelium have daily rhythmicity. Alcohol induced colonic epithelium proliferation in a time-dependent manner, with a stronger effect during the light/rest period. Alcohol-associated DNA damage also occurred more when alcohol was given at light. Levels of DNA adduct didn’t vary by time, suggesting rather lower repair efficiency during the light vs. dark. XPA gene expression, a key excision repair gene, was time dependent, peaking at the beginning of the dark. XPA knockout colon epithelial cells were inefficient in repair of the DNA damage induced by alcohol’s metabolite.

Conclusions:

Time of day of alcohol intake may be an important determinant of colon tissue damage and carcinogenicity.

Introduction

Colorectal cancer (CRC) is the second leading cause of cancer related death in the United States(Siegel et al., 2017, Bhandari et al., 2017). The rate of CRC is rising among young adults, partly due to the growth in the occurrence of lifestyle habits that are associated with disease risk(Bhandari et al., 2017, Bishehsari et al., 2014). Alcohol intake, an increasingly common habit in the US(Dwyer-Lindgren et al., 2015), impairs the colonic mucosal barrier, and is a known risk factor for development of CRC(GBD 2016 Alcohol and Drug Use Collaborators, 2018; Rossi et al., 2018). While alcohol itself is not a carcinogen(Ketcham et al., 1963), metabolism of alcohol leads to formation of metabolites that can promote carcinogenesis through several mechanisms including induction of DNA damage(Bishehsari, 2016, Rossi et al., 2018). Acetaldehyde (AcH), the major metabolite of alcohol in the colon, directly affects DNA integrity by forming DNA adducts,(Rossi et al., 2018, Seitz et al., 1990, Seitz and Stickel, 2010) which if left unrepaired, are incorporated into the DNA during proliferation and can result in carcinogenesis(Fraenkel-Conrat and Singer, 1988, Balbo and Brooks, 2015). Alcohol consumption can increase the risk of epithelial carcinogenesis by increasing DNA damage due to adduct formation following alcohol intake or due to impaired repair efficiency. This is clinically supported by ample evidence from epidemiologic studies showing a dose response relationship between alcohol consumption and risk of developing CRC(Fedirko et al., 2011), as well as the observed effect of polymorphisms in the excision repair genes on alcohol-associated cancer risk(Abbasi et al., 2009, Li et al., 2014).

Accumulating evidence suggests that DNA adduct repair including the excision repair system is under circadian control(Yang et al., 2018). Excision repair, a key participant in the repair of alcohol-associated DNA damage(Balbo and Brooks, 2015, Brooks, 1997, Matsuda et al., 1998, Slyskova et al., 2011), shows a diurnal variation in several non-intestinal tissues (i.e., liver, brain and skin)(Kang et al., 2009, Kang et al., 2010, Gaddameedhi et al., 2011). The diurnal variation in the DNA repair has been linked to the variation in the epithelial damage from other adduct-forming agents, such as chemotherapeutics and UV radiation(Kang et al., 2010, Gaddameedhi et al., 2011). Soon after the DNA damage and formation of double stranded breaks, the phosphorylation of histone, H2AX (γ-H2AX) occurs, which is the first step in recruiting the DNA repair proteins(Kuo and Yang, 2008). The expression of XPA gene, a critical member of excision repair(Sugitani et al., 2016), affects the DNA repair proficiency(Gaddameedhi et al., 2011, Kang et al., 2010).

It is not clear whether susceptibility to alcohol-induced tissue damage and carcinogenesis in the colonic epithelium exhibits a daily variation. To fill this knowledge gap, we analyzed the daily expression of proliferation and survival markers in the colonic epithelium with alcohol administered either during the light period or the dark period. The data showed that both cell proliferation and survival had a diurnal variation in the colon. Epithelial proliferation was increased following alcohol intake, and this effect was markedly exacerbated when alcohol was consumed during the light period, suggesting a time of day-dependent effect. Moreover, alcohol-induced epithelial proliferation was associated with a sustained DNA damage when consumed during the light period. To explain the variation in the DNA damage, we found no daily pattern in the amount of the Ach-related DNA adduct in the colon, but did observe a diurnal pattern in XPA gene expression, which suggested a less efficient excision repair in the colon during the light period. The diurnal pattern of xpa expression in the colon was affected by both environmental and genetic models of circadian disorganization. In vitro, colon cells deficient in XPA had less capacity to repair Ach-induced DNA damage, confirming the participation of XPA in repairing DNA damage due to alcohol metabolism. In conclusion, time of day of alcohol intake may be an important determinant in the tissue damage in colon.

Methods

Animal Experiments

Alcohol Binge.

All experiments were reviewed and approved by the Institutional Animal Care and Use Committee at Rush University Medical Center (RUMC). Male C57BL/6J (6–8 wk old) mice (The Jackson Laboratory, Bar Harbor, ME) were housed individually, and maintained on a 12 hour light:12 hour dark cycle with standard rodent chow available ad libitum (Teklad Envigo no. 2018; Teklad, Madison, WI). Mice were randomly assigned to receive either alcohol (6 g/kg/day) or alcohol control (i.e., PBS) treatment, following the NIAAA model of alcohol binge(Bertola et al., 2013). Zeitgeber time (ZT) 0 and ZT12 are the times that the lights turn on and off, respectively. All procedures conducted during the dark phase (ZT12 to ZT0) were performed under red light conditions. Treatment to each mouse group was delivered by gavage (alcohol or PBS), at the same time of the day, at 4 h intervals (n=5/group per ZT), for three consecutive days before sacrifice. The tissue harvest was conducted 4h after the third and final binge. In this protocol, the serum alcohol level at the sacrifice was not affected by time, with the mean serum alcohol level per ZT not exceeding 1mM(Voigt et al., 2018).

Environmental Circadian Rhythm Disruption.

Male C57BL/6J (6–8 wk old) mice were either maintained on a constant 12h light:12h dark cycle (controls) or once weekly 12 h phase shift in the LD cycle for 22 weeks (i.e., circadian disrupted or shifted). We have previously used this light shifting protocol to promote central circadian rhythm disruption which results in intestinal barrier pathology and dysfunction(Summa et al., 2013). Mice were sacrificed at 4h intervals (ZT0, ZT4, ZT8, ZT12, ZT16, ZT20, n=5/ZT) and colon tissues were immediately harvested and dissected for subsequent histopathological assessment.

Genetic Circadian Rhythm Disruption.

The Clock mutant mouse has a dysfunctional molecular circadian clock and circadian rhythm disruption(Summa et al., 2013). To study the effect of the disruption of the circadian clock on the XPA gene expression, we studied the tissue from young adult male Clock mutant(Summa et al., 2013) versus wild type littermate controls, both maintained on a constant 12:12 LD cycle for 10 weeks, collected at two ZTs. The Clock mutation acts as a dominant negative, effectively disrupting the molecular circadian clock.

Cell Culture and Transfection

Human Caco-2 cells (ATCC CRL-2102) were grown to confluence and circadian rhythms in the cells were synchronized by the addition of 0.1 uM dexamethasone for 1h, after which media was changed(Hida et al., 2013). The time that dexamethasone was removed was designated ZT0 for timed experiments (i.e., RNA extraction). To assess the dynamics of DNA damage due to acetaldehyde (AcH) treatment, dexamethasone synchronized cells were treated with 3mM of AcH at ZT0, and then cells were fixed at several time points (i.e., 0, 0.5, 1, 2, 4, 6 hours post treatment) and stained for H2AX phosphorylation (γH2AX), a highly sensitive marker in monitoring DNA damages induced by DNA crosslinking agents(Clingen et al., 2008). We estimated the repair efficiency by monitoring resolution of γH2AX(Kuo and Yang, 2008).

To estimate the repair efficiency of AcH-associated DNA damage at different time points, dexamethasone synchronized cells were treated with AcH at 4 time points post synchronization (at 6 h intervals); cells were then fixed 4h post-treatment and stained for nuclear H2AX measurement. Each experiment was conducted three times in triplicate.

To knock out XPA, first the lentiviral transfer plasmid GenCRISPR gRNA Human-XPA crRNA6 pLentiCRISPR v2 (Human-XPA crRNA6) encoding human XPA was purchased from GenScript. psPAX2 is a 2nd generation lentiviral packaging vector and contains Gag, Pol, Rev, and Tat genes (Addgene 12260). pMD2.G is a VSVG expressing lentiviral envelope plasmid (Addgene – 12259). Lentiviral particles were prepared as previously described(Bishehsari et al., 2018). Caco-2 cells were plated in a 6-well plate to be least 40% confluent on the day of the infection. Medium without antibiotics, the concentrated viral solution and 8μg/mL of polybrene supplementation were added to each well and plate was incubated overnight in 5.1% CO₂ at 37°C. After 24h, cells were harvested and seeded on medium containing 2μg/mL of Puromycin to select infected cells. Single cell dilution was performed before western blot analysis of XPA levels. Caco-2 cells with empty virus were utilized as controls.

RNA and Protein Analyses in Cells:

RNA was extracted using the RNeasy Mini RNA Extraction Kit (Qiagen). cDNA was prepared using the high capacity cDNA reverse transcription kit from the manufacturer (Applied Biosystems, Foster City, CA). The real time PCR was performed on an Applied Biosystems 7900HT Fast apparatus using primers (IDT, Coralville, IA) and Fast Sybr green (Applied Biosystems). The quantitative analysis was calculated from the ΔΔCt values normalized against the β-actin used as a housekeeping gene. Primer sequences are listed in supplementary Table 1. Immunoblotting in Caco2 cells was done for XPA expression. Samples were loaded to SDS gel, transferred to nitrocellulose blotting Membrane (GE Healthcare life Sciences CAT#10600004), incubated overnight at 4°C with primary antibody (Santa Cruz - XPA Antibody (B-1): sc-28353), followed by incubation with a secondary antibody, and developed by ECL solutions.

For the Caco-2 immunofluorescent (IF) staining for H2AX, cells were fixed with 4% PFA and permeabilized with 1% Triton-X100, and incubated with the primary antibody [Anti-gamma H2A.X antibody (ab11174)] overnight, and then revealed by anti-rabbit secondary antibody conjugated to Alexa Flour 488, followed by DAPI staining. Cells were imaged at 63x using a Zeiss LSM 700 confocal microscope (Zeiss, Oberkochen, Germany). At least three slides per ZT were used. In average 50 nuclei per group were included for the quantification. Number of positive foci per number of nuclei were counted and scored for the comparisons.

Tissue Staining

Proximal colon tissue were used for immunohistochemistry (IHC) and IF. Briefly, slides were deparaffinized and rehydrated. Slides underwent antigen retrieval, followed by incubation with 3% H₂O₂ for 5 minutes, and blocking for an hour. Then slides were incubated with primary antibodies (Ki-67 (Monoclonal, SP6) and Bcl-2 (Monoclonal, MM09–6)), rinsed and incubated with appropriate horse-radish-peroxidase antibodies. Color was developed with the DAB kit (SK-100). The slides were then counterstained with Hematoxylin and bluing solution for light field microscopy. Three (40x magnification) images were taken from different fields of each slide. Images were quantified for Ki67 as fraction of Ki67-positive epithelial cells (the Ki67 labeling index) (direct link at http://bit.ly/jski67) using ImageJ software. Ki67 positive staining cells were counted in average 50 crypts per animal. For Bcl-2, percent of the stained area to total epithelial area corrected by number of crypts was used for quantification using BioPix software.

IF staining was done using Anti-gamma H2A.X (phosphor S139) (ab11174 – 1:200) and PCNA (PC10, Cat# 2586, Cell Signaling Technology) as the primary antibody, followed by incubation with an appropriate fluorochrome tagged antibody [Invitrogen Alexa Fluor 555 or 488 – 1:250] for 45 minutes. Sections were then stained with DAPI and mounted using Fluoromount Aqueous Mounting Medium (Sigma, Catalog # F4680). To detect apoptotic cells, terminal deoxynucleotidyl transferase (TdT) dUTP Nick-End Labeling (TUNEL) assay was used (In Situ Cell Death Detection Kit, Fluorescein cat #11684795910-Roche). Microscope images were taken using Carl Zeiss confocal microscope (LSM 700). Three to five (20x magnification) images were taken from each slide to represent the sample. Number of positive H2AX foci per total number of nuclei were counted and scored for quantification. PCNA positive nuclei and apoptotic cells were calculated per total crypt surface area in average 50 crypts per animal.

Measurement of alcohol metabolite

N²-Ethyl-dG was measured as previously described(Balbo et al., 2012). DNA was extracted from the proximal colon, adjacent to the tissue used for tissue staining. [¹⁵N₅] N²-ethyl-dG was added as internal standard to allow for absolute quantitation. DNA was enzymatically hidrolyzed in the presence of NaBH₃CN which was used to convert the major acetaldehyde–DNA adduct, N²-ethylidene-dG, to the more stable N²-ethyl-dG. Sample enrichment and purification were then carried out, after removal of a 10 μl aliquot for 2′-deoxyguanosine (dG) analysis via HPLC. The final samples were analyzed by LC-ESI-MS/MS. Samples from all animals were processed simultaneously. Buffer blanks containing internal standard were processed and analyzed to check the MS instrument baseline and for possible contamination. Calf thymus DNA (0.1 mg) with internal standard added was used as a positive control to determine inter-day precision and accuracy. Each set of samples was run together with one buffer blank and three positive controls.

Statistical Analysis

SPSS version 23 (SPSS, Inc., Chicago, IL, USA) was used for data analyses. Results are shown as mean ± standard error. Data were analyzed using either a one way repeated measures ANOVA, a two-way repeated measures ANOVA, or a t-test as indicated in the text. Post hoc Tukey was used to identify significant between groups or between time differences. Cosinor method was used to test circadian pattern of the gene expression over time (R. Refinetti · www.circadian.org)· P<0.05 is considered significant. GraphPad Prism was used to generate figures. All authors had access to the study data and had reviewed and approved the final manuscript.

Results

Colonic Epithelial Proliferation from Alcohol is Time Dependent

We monitored Ki67, a marker for cell proliferation, at six different ZTs (every 4h for 24h) in epithelial tissue with and without alcohol. We found a significant effect of time on Ki67 index (ANOVA; F=8, P<0.001) (Fig 1 A), consistent with the diurnal pattern in proliferation and DNA replication in the colonic epithelium(Stokes et al., 2017, Potten et al., 1977, Buchi et al., 1991). Epithelial proliferation was significantly altered by alcohol treatment (ANOVA; F=44.4, P<0.001). In addition, there was a significant time by alcohol interaction (ANOVA; F-interaction=3.5, P<0.004). Proliferative response to alcohol was enhanced during the light (ZT4–8) vs. dark (ZT16–20) phase (ANOVA; F-interaction=9.8, P=0.02). Enhanced alcohol effect at ZT4 was confirmed by another proliferation marker, PCNA (Fig 1 B, two-tailed p value=0.04).

Figure 1:

Effects of alcohol and time on the colonic epithelium proliferation and survival. Mice kept under an LD 12:12 cycle were treated by alcohol or PBS at the indicated zeitgeber times (ZT 0 = light on), colon was harvested 4h later, and the levels of the indicated proteins in the colonic epithelium were determined by immunohistochemistry (IHC) and immunofluorescence (IF) as indicated. X-axis shows time of binge. Light and dark boxes correspond to light and dark periods, respectively. Quantitative analysis of several markers in the mouse colon epithelium tissue for (A) nuclear Ki67 index at different ZTs (B) proliferating cell nuclear antigen (PCNA) in PBS vs. Alcohol treated animals at ZT4 (C) cytoplasmic Bcl-2 expression at different ZTs, and (D) apoptosis index assessed by TUNEL in PBS treated animals at two ZTs. (A&C) Representative image (40x) of Ki67 and Bcl-2 at ZT4 is shown for alcohol treated mouse; (B&D) Representative images (40x) for each staining group are shown. Error bars represent mean ± SE (n = 5 mice at each time point).

One possible cause of an increase in cell proliferation could be due to reduced cell death and increased cell survival (Cohen, 1998). We measured expression of the Bcl-2, an anti-apoptotic cell survival marker every 4 hours in the colonic epithelium in alcohol and control (PBS) treated mice. While Bcl-2 expression was time-dependent (ANOVA; F=13.5, P<0.001), there was no significant effect of alcohol on the Bcl-2 expression. TUNEL staining verified the differential apoptosis rates between the time points with high and low Bcl-2 expression (Fig 1D). Apoptosis rates was higher during the dark at ZT16 (i.e, the time of low expression of the anti-apoptosis Bcl-2) versus the light at ZT4 (i.e., the time of high Bcl-2 expression) (two-tailed p value=0.03). Alcohol did not significantly affect the apoptosis rates at either of those time points (Data not shown). Therefore, increased proliferation from alcohol, particularly during the light period, was likely not from the consequence of increased survival.

These results show that both cell proliferation and survival of colonic epithelial cells has a diurnal rhythm, alcohol treatment increased the proliferation rate, while the cell death rate remained unaffected. Colonic epithelial proliferation in response to alcohol varies based on the time of alcohol intake, with an enhanced proliferative response when alcohol was consumed during the light period. Increased proliferation, especially when unopposed by cell death, could convert the alcohol-induced DNA damage to persistent mutations unless the damage is effectively repaired(O’Neill, 2000, Kiraly et al., 2015, Alekseev and Coin, 2015). Therefore, next we examined whether alcohol-induced DNA damage is time dependent.

Alcohol-Induced DNA damage in the Colonic Epithelium is Time Dependent

Acetaldehyde (AcH), the major alcohol metabolites in the colon(Bishehsari, 2016), and colon carcinogenesis(Seitz et al., 1990, Seitz and Stickel, 2007), binds to deoxynucleotides which could lead to formation of DNA crosslinks(Grafstrom et al., 1994, Balbo and Brooks, 2015). DNA crosslinks are recognized and repaired by DNA repair systems resulting in the resolution of the DNA damage(Matsuda et al., 1998, Wang et al., 2001). To examine the dynamic resolution of AcH-induced DNA damage, we used a colon epithelial cell line (i.e., Caco2 cells), and monitored the phosphorylation of histone variant H2AX (γ-H2AX) upon treatment with the alcohol metabolite AcH. The dose of AcH selected was the minimum concentration previously shown to create DNA damage in vitro(Grafstrom et al., 1994). Nuclear γ-H2AX formation is the most sensitive marker to examine the DNA damage and the subsequent repair of the DNA lesion(Sharma et al., 2012). There was a significant effect of time on the average number of nuclear γ-H2AX foci (ANOVA; F=4.3, P<0.007) (Fig 2A). We observed that γ-H2AX foci started to form as soon as 30 min post Ach treatment (0.47±0.16 at 30 min vs. 0.12 ±0.07 at baseline, p=0.04), and was maximal by one hour, after which time the number of the foci started dropping. By 4 hours, there was a significant decrease in foci number (1.6±0.34 at 4 hrs vs. 2.8 ±0.50 at 1 hr, p=0.04), indicative of active DNA damage repair by this time (Fig 2A).

Figure 2:

Effect of time on the DNA damage in response to alcohol and its metabolite in the colon epithelium. (A) Synchronized Caco-2 cells were treated by AcH. Nuclear γ-H2AX was quantified and monitored by time to estimate the dynamics of DNA damage and repair status in response to AcH. After exposure to AcH, γ-H2AX increased and reached maximum level by 1h and by 4h there was a significant reduction in γ-H2AX, suggestive of DNA damage repair. Error bars show mean ± SE (n=3 experiments). (B) Mice were treated by alcohol either during the light (ZT4 and ZT8) or during the dark (ZT12 and ZT16). Averaged values of γ-H2AX at ZT4 & ZT8 (light) vs. at ZT12 & ZT16 (dark) are compared. (*) indicates p<0.05. (C) Mice were treated by alcohol at the indicated ZTs and the N²-ethyl-dG adduct level was determined in the fresh colon. Bar graphs show mean ± SE (n = 3–5 mice at each ZT).

We then turned our attention to mouse colon tissue to examine whether residual DNA damage from alcohol consumption varies according to the time of alcohol intake. Mice received an alcohol binge either during the light period (ZT4 and ZT8), or the dark period (ZT12 and ZT16) and colon tissue was collected 4h later and stained for nuclear γ-H2AX to estimate the DNA damage. There was significantly greater residual DNA damage when alcohol was administered during the light period (ZT4-ZT8) compared to the dark period (ZT12-ZT16) (P=0.04) (Fig 2B).

Diurnal Pattern of Alcohol-induced DNA Damage Does Not Depend on AcH Metabolite Concentrations

Variations in alcohol metabolism could affect the production of alcohol metabolites and subsequent DNA adduct formation, modulating the alcohol effect on the tissue(Seitz and Stickel, 2010, Chiang et al., 2012). Therefore, time of day dependent effects of alcohol on epithelial proliferation and DNA damage could be due to diurnal variations in AcH production (and subsequent DNA adduct formation). In other words, significant differences in the sustained colonic DNA damage during the light period versus the dark period could be secondary to the varying quantity of DNA adducts that need to be repaired. DNA damage from AcH is mainly due to the AcH reaction with deoxyguanosine (dG) which results in N²-ethylidene-dG (measured as the stable reduced form N²-ethyl-dG) in large quantities(Seitz and Stickel, 2007).

Fresh frozen proximal colon tissue, adjacent to the tissue used for staining above, were used to measure N²-ethyl-dG adduct levels. No significant time of day effects were observed in tissue N²-ethyl-dG levels (Fig 2C). In the absence of the significant time variation in the DNA adduct, we hypothesized that diurnal pattern of the DNA damage may be due to variations in the repair mechanism itself. Previous studies have shown that nucleotide excision repair (NER) is involved in the AcH-induced DNA damage(Balbo and Brooks, 2015, Brooks, 1997, Matsuda et al., 1998, Slyskova et al., 2011). We then studied whether time of day variations in DNA damage could be related to differential XPA gene expression, a major member of the NER pathway.

XPA Gene Expression in the Colon Epithelium is Time Dependent

XPA is a critical member of NER(Sugitani et al., 2016), and its variation affects the risk of developing digestive cancers(Sugimura et al., 2006, Hall et al., 2007). Previous studies have shown that the transcription of XPA is under circadian control in a tissue-specific manner(Kang et al., 2009, Gaddameedhi et al., 2011). Diurnal XPA pattern is closely associated with the oscillation in the rate of the DNA repair(Gaddameedhi et al., 2011, Kang et al., 2010). To explore whether XPA has diurnal expression we analyzed its transcript expression every 6h over a 24h period, in Caco-2 epithelial cells that were synchronized with dexamethasone. XPA transcript showed a circadian pattern (cosine, F(2,5)=7, p=0.03), as well as a significant time of day effect (One-way ANOVA: F=17.57, p=0.04) (Fig 3A), with XPA expression increasing at ZT12 (compared to ZT0 and ZT6, p=0.02, and p=0.04, post-hoc Tukey). While another NER member, ERCC1 transcript, showed a similar diurnal pattern, but did not have a significant time of day effect (One-way ANOVA: F=1.06, p=0.41) (Fig 3B).

Figure 3:

Diurnal pattern of DNA damage and repair in the colon epithelium. (A-C) Dexamethasone synchronized Caco-2 cells were harvested every 6h for gene expression analysis of the indicated genes. (D) Dexamethasone synchronized Caco-2 cells were treated by AcH at 4 time points as indicated. Nuclear γ-H2AX was quantified 4h later to estimate DNA damage. Error bars show mean ± SE (n=3 experiments). (E) Correlation between the average XPA gene expression in each ZT and the residual nuclear H2Ax foci in that ZT.

It has been previously reported that time-dependent variation in XPA levels is coupled with the circadian Clock, and is negatively associated with the expression of cryptochrome 1 (cry1), the primary repressor of clock-controlled genes(Gaddameedhi et al., 2011). To further explore whether XPA gene expression in the colon epithelial is coupled with the circadian clock, we monitored cry1 expression in CaCo-2 cells over time (Fig 3C). There was a significant inverse relation between expression of xpa and cry1 (Pearson’s R: −0.7, p=0.01); the nadir of cry1 was at ZT6, after which its expression increased (compared to ZT12, p=0.02, post-hoc Tukey), suggesting that the low point of cry1 (i.e., ZT6) occurred just before the high expression of xpa expression (i.e., ZT12).

Next we aimed to examine if the xpa daily variation is associated with differential repair efficiency, assessed by the amount of sustained DNA damage post AcH in the intestinal epithelial cell line. Following dexamethasone synchronization, Caco-2 cells were treated with AcH at different times (ZTs), and after 4h the nuclear H2Ax foci were counted. The amount of DNA damage showed a significant time of day effect (ANOVA; F=5.6, P<0.01) (Fig 3D). There was a significant inverse correlation between the average xpa expression in each ZT and the nuclear H2Ax foci after 4 h (Pearson’s R: −0.99, p=0.002, Fig 3E), with the least residual DNA damage from AcH corresponding to the time of the maximum level of xpa in Caco-2 cells at ZT12. The inverse association of the XPA gene expression levels with the amount of DNA damages suggests a possible differential repair capacity of the AcH-induced DNA damages by time in colon cells.

To directly examine the involvement of XPA in repairing AcH-induced DNA damage, XPA was stably knocked out (KO) in Caco-2 cells as shown by two independent transfections (XPA-KO1and XPA-KO2) (Fig 4A). Cells were treated by AcH, and DNA damage was assessed 4h later by nuclear H2Ax foci quantification. Caco-2 cells with XPA-KO had significantly higher residual DNA damage (XPA-KO1: 5.4 ± 0.50 and XPA-KO2: 6.2 ± 0.55) compared to Caco2 cells with functional XPA (2.1 ± 0.33; p<0.001) (Fig 4B). This supports that XPA is important for AcH-induced DNA repair.

Figure 4:

XPA participates in repair of DNA damage from alcohol’s metabolite. (A) XPA was knocked out in the Caco2 cells. Immunoblotting for XPA and housekeeping proteins from two independent transfections (XPA-KO1and XPA-KO2) are shown. Bar graph shows minimal XPA expression in XPA-KO cells versus controls. (B) Control Caco-2 cells as well as XPA-KO1 and XPA-KO2 Caco-2 cells were treated with acetaldehyde (AcH) and DNA damage was assessed 4h later by nuclear H2Ax foci quantification. Representative images from XPA-KO1 are shown on the top. Bar graphs show comparison of average γ-H2AX foci per nuclei cell type. (***) indicates p<0.001.

To examine whether variation in the xpa expression may explain the diurnal pattern of DNA repair in the colon tissue in vivo, we analyzed expression of xpa in the mouse colon across 24h. XPA expression was time dependent (ANOVA; F=2.9, P=0.02) (Fig 5A), suggesting that xpa expression varies across 24h, with its peak at the beginning of the dark phase in mice. This xpa expression did not fit a cosine curve. Using CircaDB database, we found that xpa has a significant circadian rhythm in the mouse colon (JTK_Cycle, p=0.0056)(Pizarro et al., 2013).

Figure 5:

XPA gene expression in the mouse colon is circadian coupled. (A) Mice were maintained on a constant 12h light:12h dark cycle (solid line represents non-shifted mice) or once weekly 12 h phase shift in the LD cycle (dotted line represents shifted mice) for 22 weeks, after which they were sacrificed at 4h intervals and colon tissues were used for the XPA gene expression analysis. Error bars represent mean ± SE (n = 5 mice at each time point) (B) Bar graphs for the XPA gene expression at ZT12 and ZT4 in Clock mutant mice versus wild type littermate controls and (C) The fold differences in the colonic XPA gene expression at ZT12 and ZT4 in Clock mutant mice versus wild type littermate controls. Error bars represent mean ± SE (n = 4 mice at each ZT). (*) indicates p<0.05.

To examine whether colonic xpa expression is circadian coupled, we measured xpa expression in environmental mode of circadian disorganization, using chronic shifts of the light/dark (LD) cycle. We have previously shown that this model affects intestinal pathologies(Summa et al., 2013, Voigt et al., 2014). Comparing colonic xpa expression between control and circadian disrupted mice revealed a significant time of day by circadian disruption interaction (ANOVA: F-interaction=2.90, p=0.03) (Fig 5A). Taken together these data suggest that circadian disruption alters the time of day-dependent expression of xpa in the colon (Fig 5A). To further explore whether time-dependent expression of xpa in the colon is coupled with the circadian clock, we compared the xpa expression in the colon of Clock mutant mice to their littermate wild-type between two ZTs. While wild type animals showed a significant differential xpa expression between the examined time points, the differential expression was dampened in the Clock mutant animals (Fig 5B,C), (p=0.04).

Discussion

Our findings show, for the first time, that alcohol-induced epithelial damage in the colon is time dependent. Alcohol intake during the light (i.e., rest) period compared to the dark (i.e., active) period, resulted in greater cell proliferation which coincided with decreased DNA repair and increased DNA damage. This is clinically relevant given the observed interaction of alcohol and circadian disruption in intestinal pathologies including barrier dysfunction and colon cancer(Bishehsari, 2016, Bishehsari F, 2017).

Circadian clocks keep intrinsic rhythms in the cells that endow fitness benefit to the host by gaining capacities to adapt to changes in the environment(Mohawk et al., 2012). Disruption of circadian rhythms has been linked with various chronic diseases of the digestive system including colon cancer (Bishehsari et al., 2016). Epidemiological studies have shown that chronic disruption of the circadian clock (i.e., light/dark shifting) could increase CRC risk(Schernhammer et al., 2003, Wang et al., 2015). More recently we observed that LD shift could promote CRC in mice fed alcohol(Bishehsari F, 2017). However the mechanism of how circadian disruption promotes colon carcinogenesis remains unclear.

More than 40% of protein-coding gene transcripts are clock controlled(Zhang et al., 2014). In the colonic epithelium, the circadian clock maintains a regular circadian rhythm, entrained by central signals (light/dark cycle) and/or time of food consumption(Sladek et al., 2007, Bishehsari et al., 2016). Indeed, a large body of evidence shows that cell proliferation, cell death, and DNA repair, all critical for epithelial carcinogenesis, are under circadian control.(Uchida et al., 2010) In the colon, genes that are involved in cell proliferation and death were found to have rhythmic expression as well(Hoogerwerf et al., 2008). Data from the current study confirm these diurnal variations and expands to show their implications in the effect of alcohol in colon.

Similar to previous studies(Stokes et al., 2017, Potten et al., 1977), we found that markers of cell proliferation and survival are both time dependent in the mouse colon. Mice are nocturnal, and light and dark period corresponds to rest and active phase. Alcohol increased the cell proliferation in the mouse colon, particularly when given during the light (i.e., rest) period. The anti-apoptotic BCL-2 expression although showed a significant time variation, was not significantly affected by alcohol. This pattern was confirmed by the direct assessment of the apoptosis rates. Earlier reports showed a higher BCL-2 expression in the peripheral tissue, suggestive of a survival tone during the light phase(Granda et al., 2005). The alcohol-induced enhancement of epithelial proliferation, if left unopposed by anti-survival signaling, may lead to processes culminating in mucosal damage and carcinogenesis(O’Neill, 2000).

Alcohol-induced proliferation was coupled with more residual DNA damage in our study. The repair of DNA damage from alcohol and its metabolite (AcH) was time dependent both in vivo and in vitro. We found more DNA damage (i.e., H2AX), when cells were exposed to alcohol during the light (i.e., rest) period suggesting loss of repair capacity in the colon epithelium. This is clinically relevant as synergism between DNA damage and replication is the critical step in mutation accumulation and cancer formation(Ames et al., 1993, Huang et al., 2016).

The time variation observed in DNA repair in the mouse colon, could at least partly be explained by the time of day dependent expression of XPA in colon. The importance of XPA in cancer development is clinically supported by the association between NER gene polymorphisms (including XPA) with the risk of developing several epithelial cancers(Lu et al., 2014, Zhang et al., 2018, Hall et al., 2007, Kiyohara and Yoshimasu, 2007) including colon carcinogenesis(Ho et al., 2018, Feng et al., 2018).

Prior studies have shown that XPA expression is rhythmic in several tissues of C57BL/6 mice(Kang et al., 2009, Kang et al., 2010, Gaddameedhi et al., 2011). Indeed, in the current study the highest XPA gene expression coincided with the lowest alcohol-induced DNA damage. In a murine model of skin cancer, it has been shown that excision repair has diurnal pattern with the most carcinogenic effect of the UV occurring at the time when the expression of XPA was low, and that XPA is controlled by the circadian clock (Gaddameedhi et al., 2011). In the mouse colon, we found that the diurnal pattern of XPA transcript changes by circadian disruption due to the shift in LD cycle, and that its oscillation is lost in the Clock mutant mice. These findings suggest that XPA gene expression and DNA repair could be circadian coupled in colon. Implications of the effects of circadian rhythm disruption on the colonic xpa expression are clinically relevant. Disruption of circadian rhythms may affect colon mucosal integrity and carcinogenesis by altering critical mechanisms regulating cell proliferation, apoptosis, and/or repair. In this regard, the time of alcohol consumption may impact CRC risk due to time of day differences in cell proliferation, apoptosis, and repair.

Our study has a number of limitations. Here we examined the daily variation of AcH-associated DNA adducts in the colon. We could not exclude diurnal variation in the level of other types of etheno-DNA adducts that form due to reactive oxygen species (ROS)(Rossi et al., 2018, Linhart et al., 2014). However, the microsomal ethanol–oxidizing system, that is the main contributor to ROS formation from alcohol, becomes primarily activated only in heavy chronic alcoholics. It should be noted that microsomal ethanol–oxidizing system (CYP2E1) does not typically participate in alcohol metabolism in lower moderate dosing of alcohol(Bishehsari, 2016), which is also associated with an increased risk of CRC(GBD 2016 Alcohol and Drug Use Collaborators, 2018). Secondly, while our data suggests diurnal variation in the XPA gene expression in the colon, we cannot exclude daily variation of other types of the DNA repair systems in the colon. We examined XPA, part of NER, as a participant in DNA repair associated with alcohol metabolism(Matsuda et al., 1998). XPA may also participate in double-strand repair(Fadda, 2016, Camenisch and Nageli, 2008). Here we confirmed that XPA participates in the repair of the DNA damages from AcH.

In conclusion, alcohol-induced epithelial damage in the colon appears to be time dependent. An increase in epithelial proliferation rate due to alcohol during the light period is coupled with a decrease in the DNA repair capacity in the colonic epithelium. Low gene expression of XPA is associated with low repair capacity in colon. Our data suggests that colonic XPA gene expression is circadian coupled. This could partly explain interaction of circadian disruption with alcohol in intestinal pathologies(Bishehsari, 2016, Summa et al., 2013), particularly in promoting colon cancer(Bishehsari F, 2017).

Abbreviations:

CRC

colon cancer

NER

Nucleotide excision repair

XPA

xeroderma pigmentosum, complementation group A

ERCC1

Excision Repair Cross-Complementation Group 1