Effect of Alcohol on miR-212 Expression in Intestinal Epithelial Cells and Its Potential Role in Alcoholic Liver Disease

Authors

  • Yueming Tang,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Ali Banan,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Christopher B. Forsyth,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Jeremy Z. Fields,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Cynthia K. Lau,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Li Juan Zhang,

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author
  • Ali Keshavarzian

    1. From the Division of Digestive Disease and Nutrition, Department of Internal Medicine (YT, AB, CBF, JZF, CKL, LJZ, AK); Department of Pharmacology (AB, AK); and Department of Physiology (AK), Rush University Medical Center, Chicago, Illinois.
    Search for more papers by this author

  • No conflicts of interest exist.

Reprint requests: Yueming Tang, MD PhD, Assistant Professor of Medicine, Section of Gastroenterology, Department of Internal Medicine, Rush University Medical Center, 1725 W. Harrison (Suite 206), Chicago, IL 60612; Fax: 312-942-5664; E-mail: yueming_tang@rush.edu

Abstract

Background and Aims:  Alcohol-induced gut leakiness is a key factor in alcoholic liver disease (ALD); it allows endotoxin to enter the circulation and initiate liver damage. Zonula occludens 1 (ZO-1) protein is a major component of tight junctions that regulates intestinal permeability. microRNAs (miRNAs) are recently discovered regulatory molecules that inhibit expression of their target genes. The aims of our study were: (i) to investigate the effect of alcohol on miRNA-212 (miR-212) and on expression of its predicted target gene, ZO-1, (ii) to study the potential role of miR-212 in the pathophysiology of ALD in man.

Methods:  Using a TaqMan miRNA assay system, we measured miR-212 expression levels in colon biopsy samples from patients with ALD and in Caco-2 cells (a human intestinal epithelial cell line) treated with or without EtOH. We measured ZO-1 protein levels using western blots. ZO-1 mRNA was assayed using real-time PCR. Intestinal barrier integrity was measured using fluorescein sulfonic acid clearance and immunofluorescent staining for ZO-1.

Results:  Ethanol increased miR-212 expression, decreased ZO-1 protein levels, disrupted tight junctions, and increased the permeability of monolayers of Caco-2 cells. An miR-212 over-expression is correlated with hyperpermeability of the monolayer barrier. miR-212 levels were higher in colon biopsy samples in patients with ALD than in healthy controls; ZO-1 protein levels were lower.

Conclusion:  These data suggest a novel mechanism for alcohol-induced gut leakiness, one in which EtOH induces miR-212 over-expression which causes gut leakiness by down-regulating ZO-1 translation. This mechanism is a potential therapeutic target for leaky gut in patients with or at risk for ALD.

Alcoholic Liver Disease (ALD) is one of the most common and serious complications of heavy drinking, and heavy drinking is a major health problem in the United States, consuming 15% of total health care costs. ALD has high morbidity and mortality with no satisfactory therapy (Bergheim et al., 2005; Bode and Bode, 2005; Grant et al., 1988; O’Connor and Schottenfeld, 1998; Reuben, 2006). Our previous studies (Banan et al., 1999, 2000; Keshavarzian and Fields, 2000, 2003; Keshavarzian et al., 1994, 1999, 2001; Mathurin et al., 2000) showed that: (i) clinically significant liver disease occurs only in a subset (15–30%) of alcoholics (Bode and Bode, 2005; Grant et al., 1988), indicating that excessive ethanol (EtOH) consumption is necessary but not sufficient to induce liver injury, and therefore, one or more cofactors are required. (ii) Gut-derived endotoxin is a required cofactor because removal of gut flora with antibiotics prevents ALD in animal studies and because an endotoxin-initiated hepatic necroinflammatory cascade causes liver injury in ALD (Criado-Jimenez et al., 1995; Hunt and Goldin, 1992; McClain and Cohen, 1989). (iii) Gut leakiness appears to be the cause of the endotoxemia in ALD (Rao et al., 2004). We showed that intestinal barrier hyperpermeability occurs only in alcoholics with ALD and not in those alcoholics without liver disease (Keshavarzian et al., 1999). Taken together, these data strongly suggest that EtOH-induced intestinal barrier disruption is the key mechanism in endotoxemia in ALD. However, the mechanism by which EtOH disrupts intestinal barrier dysfunction is unclear. As gut leakiness is important for the development of ALD, knowledge of the underlying mechanisms should help us to understand better and manage patients with ALD.

The integrity of the intestinal barrier depends on both healthy epithelial cells and on an intact paracellular pathway, which appears to be the main route for permeation of macromolecules, such as endotoxin (Hollander, 1992). This pathway is a complex array of structures that includes tight junctions between gut epithelial cells. Tight junctions function as gates that regulate intestinal permeability (Clayburgh et al., 2004; Ivanov et al., 2005; Laukoetter et al., 2006; Turner, 2006). These dynamic tight junctions are highly regulated and are able to change their size under various physiological and pathological conditions (Madara, 1990).

A major protein of tight junctions is Zonula occludens 1 (ZO-1). Because changes in ZO-1 are associated with EtOH induced increases in intestinal permeability, we became interested in the regulation of ZO-1. In general, the functional and structural integrity of ZO-1 and other tight junction proteins can be influenced by the rate and accuracy of their protein synthesis. Novel regulators of protein synthesis are microRNAs (miRNAs), a class of small noncoding RNAs that are 18 to 25 nucleotides in length. miRNAs control gene expression by targeting mRNAs and triggering either translational repression or RNA degradation. These effects inhibit the synthesis of proteins coded by the miRNA’s target genes (Ambros, 2004; Carthew, 2006; Nakahara et al., 2005; Sontheimer and Carthew, 2005). Therefore, it is not surprising that miRNAs have been implicated in many cellular processes including cell proliferation, differentiation, apoptosis, and metabolism (Czech, 2006; Hammond, 2006; Zamore and Haley, 2005). Aberrant expression of miRNAs appears to be involved in many human diseases, including infection and cancer (Ambros, 2004; Esquela-Kerscher and Slack, 2006; Krutzfeldt et al., 2006; Kutay et al., 2006). Whether EtOH affects expression of miRNAs and their regulation of their target genes in intestinal epithelial cells is unclear.

One can surmise that EtOH might have such an effect. Each of the many known miRNAs has either documented or predicted target genes that could influence the synthesis of the proteins coded by the target genes. For example, the tight junction protein ZO-1 is one of miR-212’s predicted target genes according to the mature miR-212 sequence in an miRNA database (Blow et al., 2006; Griffiths-Jones et al., 2006), but no experiment has yet validated this conjecture. It is interesting that the chromosome location of miR-212 is in 17p13.3-D fragile region that EtOH may affect (Krek et al., 2005; Rajewsky, 2006).

This putative effect of EtOH on miR-212 and ZO-1 could be important because it is well known that ZO-1 plays an important role in the regulation of intestinal permeability (Sawada et al., 2003). However, roles for miRNAs in regulation of ZO-1 synthesis, in regulation of the intestinal barrier, and in alcohol-induced gut leakiness in patients with ALD remain to be established.

Our hypothesis was that EtOH induces intestinal barrier hyperpermeability by increasing miR-212 expression, which then down-regulates expression of its target gene, ZO-1. miR-212 inhibits ZO-1 protein synthesis and disrupts ZO-1’s organizational network. The aim of the present study was to test this hypothesis by examining miR-212 and its target gene, ZO-1’s expression in (Caco-2 cells) treated with or without EtOH, and in colon biopsy samples from healthy controls and patients with ALD. We found that miR-212 expression is increased in ALD patient and that EtOH-induced over-expression of miR-212 and down-regulation of its target gene (decreased ZO-1 tight junction protein) results in intestinal hyperpermeability in patients with ALD.

Material and methods

Human Subjects’ Involvement and Characteristics

We collected endoscopic mucosal biopsies from the sigmoid colon (20 cm from the anus) of alcoholic patients with a drinking history ≥10 years (the minimum time thought to be required for development of liver disease) and with mild to moderate liver disease. We also collected endoscopic mucosal biopsies from the sigmoid colon of healthy control subjects. Controls and patients with ALD had similar demographic characteristics (age, gender, and ethnicity). Endoscopic samples were obtained during an unprepared, limited sigmoidoscopy and were snap-frozen in liquid nitrogen. All procedures were carried out after written informed consent was obtained. All alcoholic patients fulfilled National Institute on Alcohol Abuse and Alcoholism (O’Connor and Schottenfeld, 1998) & Diagnostic and Statistical Manual of Mental Disorders criteria (Ball et al., 1997) for alcoholism. All enrolled patients with ALD had either clinical evidence of liver disease (hepatomegaly, splenomegaly, and esophageal varices), or had serum bilirubin and alanine amino transferase or aspartate amino transferase greater than 1.5X normal levels. As gut leakiness is required for the development of clinically important liver disease, liver disease was clinically rather histologically defined. Clinical characteristics of study subjects are shown in Table.1.

Table 1.   Characteristics of Patients With Alcoholic Liver Disease
GroupsAge (year)Duration of EtOH intake (year)Amount of EtOH intake (g/d) Bilirubin (mg/dl) AST (U/l) ALT (U/l) Albumin (g/dl)
  1. ALD, alcoholic liver disease; AST, aspartate amino transferase test (normal <40 U/l); ALT, alanine amino transferase test (normal <50 U/l).

  2. *< 0.05 compared to control group.

Controls56 ± 2.90.5 ± 0.118 ± 620 ± 54.06 ± 0.08
ALD57 ± 2.815.3 ± 1.4172 ± 18.32.1 ± 0.3*132 ± 18*62 ± 12*3.25 ± 0.11*

The study was approved by the Institutional Review Board of Rush University. Written consent was obtained from each participant before enrollment in the study.

Cell Culture

Caco-2 cells were obtained from American Type Culture Collection (Manassas, VA) at passage 15. This cell line was chosen because it forms monolayers of cells that morphologically resemble intestinal epithelial cells with defined apical brush border and a highly organized tight junction network upon differentiation (Banan et al., 1999, 2000, 2005b; Forsyth et al., 2007). Cells were maintained at 37°C in complete Dulbecco’s minimum essential medium (DMEM) in an atmosphere of 5% CO2 and 100% relative humidity. Cells grown for barrier function experiments were split at a ratio of 1:2 and seeded at a density of 200,000 cells/cm2 into 0.4 μM Biocoat collagen I cell culture inserts (0.3 cm2 growth surface; BD Biosciences Discovery Labware, Bedford, MA). Experiments were performed at least 7 days postconfluence. The media were changed every 2 days. The utility and characterization of this cell line have been previously reported (Banan et al., 1999, 2000, 2005b; Forsyth et al., 2007).

Real-Time Quantification of miRNAs by Stem-Loop RT-PCR

Total RNA from tissue samples and cells were isolated using mirVana RNA isolation kits (Ambion, Austin, TX) according to the manufacturer’s protocol. TaqMan microRNA assay kits with stem-loop RT primer (Applied Biosystems, Foster, CA) were used for detection of miR-212 and miR-145. Real-time PCR was performed on 7900HT Sequence Detection System (Applied Biosystems). The threshold cycle (CT) was defined as the fractional cycle number of fluorescence passes through the fixed threshold. RNU6B was used as an internal control. Experiments and data analysis were carried out according to the manufacturer’s protocol. Relative quantities of miRNA expression were calculated as in our previous studies (Clayburgh et al., 2005; Zhang et al., 2004).

Real-Time PCR for ZO-1 mRNA

Isolation of total RNA was carried out using the RNEasy Kit (Qiagen, Valencia, CA). RNA was quantified using a spectrometer. Reverse transcription of RNA to cDNA was carried out using the High Capacity cDNA RT Kit (Applied Biosystems). Real-time quantitative PCR was then performed using the resulting cDNA with the SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). We used 40 cycles of 95 to 60 and to 75°C (1 minute for each temperature) for PCR amplification. The reactions were performed in 96-well plates using a 7900HT Prism apparatus (Applied Biosystems). Primers for RT-PCR used in the present study were: β-actin forward primer: GCCAGGTCCAGACGCAGG; β-actin reverse primer: TGCTATCCAGGCTGTGCTA; ZO-1 forward primer: CAGAACCAAAGCCTGTGTATG; ZO-1 reverse primer: TTAGGTAGGACACCATCAGATGGA. Primers were designed with Oligoperfect Software and purchased from Invitrogen (Carlsbad, CA). Relative quantitative gene expression was calculated as follows (Clayburgh et al., 2005; Zhang et al., 2004): (i) CT value for each target gene was determined and corrected by subtracting the CT for β-actin (ΔCT) of each sample assayed (the CT value represents the cycle when the sequence detection application begins to detect increases in signal associated with exponential growth of the PCR product). (ii) Untreated controls or 0-time samples served as reference samples, and the ΔCT for all treated experimental samples were subtracted from the ΔCT for the control samples (ΔΔCT). (iii) Treated target gene mRNA abundance (A) relative to control was calculated by the formula A = 2−(ΔΔCT). All assays were performed in triplicate.

Transfection of miR212 Precursors

miR-212 precursors are small, partially double-stranded RNAs that mimic endogenous precursor miR-212 (Ambion). The design of miR-212 precursors ensures that the portion of the miRNA precursors identical to the mature miRNA is selected for uptake and activation by miRNA processing pathways. Transfections were performed by nuclear transfection using siPORTTMNeoFXTM transfection agents (Cat#4510; Ambion). Transfection conditions for Caco-2 cells were optimized according to the manufacturer’s protocol. Caco-2 cells were transfected with miR-212 precursors or control precursors at a final concentration of 100 nM. Cells were harvested at 72 hours after transfection for assaying the effect of the transfected miR-212 precursors.

Barrier Integrity and Permeability of Intestinal Cell Monolayers

The integrity of the monolayer barrier was assessed by a widely used validated technique that measures the apical to basolateral paracellular flux of fluorescent markers, such as fluorescein sulfonic acid (FSA, 200 μg/ml; 0.478 kDa) as we described (Banan et al., 1999, 2000, 2005b; Forsyth et al., 2007). In brief, fresh phenol-free DMEM (800 μl) was placed into the lower (basolateral) chamber and phenol-free DMEM (300 μl) containing probe (FSA) was placed in the upper (apical) chamber. Aliquots (50 μl) were obtained from the upper and lower chambers at 0-time and at subsequent time points and transferred into clear 96-well plates (clear bottom; Costar, Cambridge, MA). Fluorescent signals from samples were quantified using a fluorescence multiplate reader (FL 600; Bio-Tek Instruments, Winooski, VT). Excitation and emission spectra for FSA were 485 and 530 nm, respectively. Clearance (CL) was calculated using the following formula: CL (nl/h/cm2) = Fab/([FSA]a × S), where Fab is the apical to basolateral flux of FSA (light units/h), [FSA]a is the concentration at baseline (light units/nl), and S is the surface area (0.3 cm2). Simultaneous controls were performed with each experiment.

Western Blots for ZO-1 Protein

Western blotting was performed using the Bio-Rad mini-Protein system (Bio-Rad, Hercules, CA) as previously described (Banan et al., 2005b; Forsyth et al., 2007). Protein was transferred to nitrocellulose and blocked with 5% bovine serum albumin for 1 h at 4°C and then incubated at 4°C overnight with primary Ab, anti-ZO-1 (Zymed, South San Francisco, CA). Blots were washed, incubated with horseradish peroxidase-conjugated 2Ab for 1 hour at 4°C, washed, developed with enhanced chemiluminescence solution (Amersham, GE Healthcare, Piscataway, NJ), exposed to Fuji film (Fuji, Tokyo, Japan), and finally scanned for analysis with Image J software (NIH, Bethesda, MD; Banan et al., 2005b; Forsyth et al., 2007).

Immunofluorescent Staining of ZO-1 and Tight Junction Morphology

Cells from monolayers were fixed in cytoskeletal stabilization buffer and then postfixed in 95% ethanol at −20°C. Cells were subsequently processed for incubation with a primary antibody [monoclonal mouse anti-ZO-1(Zymed), 1:200 dilutions] for 1 hour at 37°C and then incubated with a secondary antibody (fluorescein isothiocyanate-conjugated goat anti-mouse; Sigma-Aldrich, St. Louis, MO; 1:50 dilution) for 1 hour at room temperature. Slides were washed 3X in d-phosphate-buffered saline and mounted in Aquamount (VWR International, West Chester, PA). Following staining, cells were observed with an Axiovert 100 microscope and Axiovision software (Carl Zeiss Inc., Thornwood, NY). The three-dimensional reconstruction of deconvoluted z-stacks was performed using Zeiss Axiovision software (Carl Zeiss, Inc., Thornwood, NY). Immunofluorescent microscopy analysis of Caco-2 monolayers has been previously described by us in detail (Banan et al., 2004, 2005a,b; Forsyth et al., 2007).

Statistical Analysis

Data are presented as mean ± SD. Statistical analysis comparing treatment groups was performed using analysis of variance followed by Student’s t-test. p-values <0.05 were deemed statistically significant. Data were analyzed using spss software for Windows (SPSS, Chicago, IL).

Results

miR-212 Is Highly Expressed in Intestinal Tissues

To determine if miR-212 is expressed in intestinal tissue, we examined miR-212 expression levels in 8 tissues from normal rat. We found that the highest expression level of miR-212 was in intestines including jejunum, ileum, and colon, though miR-212 was also expressed in spleen, heart, lung, kidney, and liver (Fig. 1).

Figure 1.

 miR-212 expression in different tissues. miR-212 expression was examined in different tissues from normal rats. miR-212 expression level in heart served as a reference (value was set at 1.0). RNU6B was used as an internal control. Data analysis is described in Materials and Methods. miR-212 was highly expressed in intestines including jejunum, ileum, and colon compared with heart (n = 4, p < 0.05), although it was also expressed in lung, kidney, liver, and spleen. Data are expressed as mean ± SD. *p < 0.05, **p < 0.01 compared to heart.

Effect of EtOH on miR-212 Expression in Caco-2 Cells

We could not find any published study about the effect of EtOH on miR-212 expression in intestinal epithelial cells. To determine if ethanol affects miRNA-212, and to further confirm that miR-212 is expressed in intestinal epithelial cells, we studied the effect of EtOH on miR-212 expression in a human intestinal epithelial cell line, Caco-2 cells. Caco-2 cells were treated with 0.1 to 1.0% EtOH for 3 hours and then harvested. We confirmed that miR-212 is highly expressed in these intestinal epithelial cells (Fig. 2A). We found that EtOH significantly increased miR-212 expression in Caco-2 cells in a concentration-dependent manner (p < 0.05, Fig. 2A). These data suggest that EtOH increases miR-212 expression in intestinal epithelial cells.

Figure 2.

 Effect of ethanol (EtOH) on miR-212 expression and monolayer integrity of colon epithelial cells (Caco-2). (A): Effect of EtOH on miR-212 expression. Caco-2 cells were treated with 0.1 to 1.0% EtOH for 3 hours. miR-212 was measured using a specific stem-loop primer RT and real-time PCR. One percent EtOH significantly increased miR-212 expression in Caco-2 cells (p < 0.05). (B): Time course for the effect of EtOH on monolayer integrity. Decreases in monolayer integrity are seen as increases in fluorescein sulfonic acid clearance, which is an indication of hyperpermeability of the monolayers. EtOH disrupted monolayer integrity in a concentration-dependent manner. (C): Correlation between miR-212 expression and monolayer integrity after 3 hours of exposure to EtOH (0.1 to 1%). (r = 0.92, p < 0.05). Data are from triplicate cells from 1 of 3 independent experiments and expressed as mean ± SD. *p < 0.05 compared to control (media alone, 0% EtOH).

Association of EtOH-Induced Intestinal Barrier Integrity and miR-212 Expression Levels

To evaluate mechanisms of EtOH-induced barrier disruption, we used an in vitro model, monolayers of Caco-2 cells. This is a practical and highly reproducible model for studying aspects of intestinal physiology (Banan et al., 1999, 2000, 2005b; Forsyth et al., 2007) and has been widely used to study mechanisms of intestinal barrier function. This in vitro system is also a relevant model for studying mechanisms of EtOH-induced intestinal barrier disruption (Banan et al., 1999, 2000, 2005b). We incubated Caco-2 monolayers with EtOH (0.1 to 1.0%) and found that EtOH significantly increases barrier permeability in a time- and concentration-dependent manner (p < 0.05, Fig. 2B). The EtOH-induced increase in monolayer barrier permeability was robustly and significantly correlated with EtOH-induced miR-212 expression after 3 hour of EtOH exposure (0.1 to 1.0%) (p < 0.05, Fig. 2C). The results suggest that miR-212 is involved in the regulation of intestinal barrier integrity.

Effect of EtOH on ZO-1, a Predicted Target Gene of miR-212

Caco-2 cells were treated with 1.0% EtOH for 3 hours and then harvested. Western blots showed that EtOH significantly decreased ZO-1 protein levels compared with controls (Fig. 3A). ZO-1 bands of western blots were scanned and assessed using Image J densitometry software. ZO-1 levels were normalized by β-actin levels. The data showed that 1% EtOH induced decreases in ZO-1 protein levels of 71% (p < 0.05, Fig. 3B). EtOH (1.0%) exposure did not change ZO-1 mRNA levels (Fig. 3C). These results indicate that EtOH-induced miR-212 over-expression may only inhibit ZO-1 protein levels, but not ZO-1 mRNA levels.

Figure 3.

 Effect of ethanol (EtOH) on the expression of Zonula occludens 1 (ZO-1), miR-212’s predicted target gene. (A): Effect of EtOH on ZO-1 protein levels. Western blot data show that ZO-1 protein levels in colon epithelial cells (Caco-2) treated with 1% EtOH for 24 hours were decreased compared with control. The blot was stripped and reprobed for β-actin as a protein loading control. The blot shown is a representative of 3 experiments. (B): Density of ZO-1 protein bands. ZO-1 bands of Western blots were scanned and assessed using Image J densitometry software. ZO-1 levels were normalized by β-actin levels. The data showed that 1% EtOH induced a decrease in ZO-1 protein levels by 71% (n = 6, p < 0.05). (C): Effect of EtOH on ZO-1 mRNA. Real-time PCR data showed that ZO-1 mRNA expression levels in Caco-2 cells treated with 1% EtOH for 24 hours were not significantly different compared with control. (D): miR-212 down-regulates ZO-1 protein. Caco-2 cells were transfected with either 100 nmol/L miR-212 precursors (miR-212) or control precursors. The cells were harvested at 72 hours after transfection. ZO-1 protein levels were measured by Western blots. ZO-1 protein levels were significantly decreased in Caco-2 cells transfected with miR-212 precursors compared with cells transfected with control constructs (p < 0.05). Data are from triplicate cells from one of 3 independent experiments and expressed as mean ± SD.

To confirm that increased miR-212 expression directly causes a decrease in ZO-1 levels, we transfected Caco-2 cells with miR-212 precursors and monitored ZO-1 levels using western blots. The results show that ZO-1 protein levels are significantly decreased in Caco-2 cells transfected with miR-212 precursors compared with cells transfected with control constructs (Fig. 3D, p < 0.05). These studies indicate that over-expression of miR-212 can decrease ZO-1 protein levels.

Effect of EtOH on Tight Junction Morphology (ZO-1 Distribution)

Tight junctions regulate the size of the spaces between adjacent intestinal epithelial cells, thereby regulating paracellular permeability. Tight junctions are maintained by the interactions of ZO-1 and other proteins that regulate epithelial permeability (Clayburgh et al., 2005; Ivanov et al., 2005; Laukoetter et al., 2006; Sawada et al., 2003). To further determine the role of ZO-1 in the effect of EtOH on tight junction morphology in intestinal epithelial cells, we treated Caco-2 monolayers with 0.1 to 2.5% EtOH for 24 hours. At the end of the incubation, trypan blue dye exclusion was used to assess cell viability. Incubating Caco-2 cells with 0.1 to 2.5% EtOH for 24 hours did not result in a significant reduction in cell viability. Immunofluorescent staining of ZO-1 shows that EtOH in a concentration-dependent manner caused disruption of the tight junctions of Caco-2 cells, including extensive disorganization, kinking, condensation, and beading of its normal ring structure (Fig. 4). To show the effect of EtOH on the three-dimensional structure of ZO-1 containing tight junctions, we reconstructed three-dimensional ZO-1 tight junctions by deconvoluting z-stack data. The three-dimensional pictures clearly showed the dramatic disruption (“zig-zag” pattern) of ZO-1 tight junctions caused by 0.5% EtOH (Fig. 5). These data show that EtOH decreases ZO-1 levels and disrupts tight junctions between intestinal epithelial cells.

Figure 4.

 Effect of ethanol (EtOH) on tight junction morphology of colon epithelial cells (Caco-2). To study the effect of EtOH on tight junction morphology, monolayers were treated with 0.1 to 2.5% EtOH for 24 hours. Immunofluorescent staining of Zonula occludens 1 protein [ZO-1; (fluorescein isothiocyanate-conjugated goat anti-mouse, green)] showed that EtOH in a concentration-dependent manner caused disruption of tight junctions in intestinal epithelial cells. The EtOH-induced displacement of ZO-1 tight junction protein including extensive disorganization, kinking, condensation, and beading of its normal ring structure are pointed out using red arrows. Representative images from 3 independent experiments are shown.

Figure 5.

 Effect of ethanol (EtOH) on the three-dimensional structure of Zonula occludens 1 protein (ZO-1) tight junctions. Colon epithelial cells (Caco-2) cultured on glass coverslips were treated with media alone (control, A and B) or 0.5% EtOH (C and D) for 24 hours. Then, ZO-1 was stained with anti-ZO-1 antibody and fluorescein isothiocyanate-conjugated goat anti-mouse secondary antibody. The cut view of ZO-1 reveals EtOH-induced disruption of tight junctions (A and C) which are similar to Fig. 4. To show the effect of EtOH on the three-dimensional structure of ZO-1 containing tight junctions, we reconstructed three-dimensional ZO-1 tight junctions by deconvoluting z-stack data from A and C (Materials and Methods). The three-dimensional pictures clearly showed the dramatic disruption (“zig-zag” pattern) of ZO-1 tight junctions caused by 0.5% EtOH (D). Representative images from 3 independent experiments are shown.

miR-212 Is Highly Expressed in Patients with ALD

To determine if our in vitro findings are relevant to ALD, we measured miR-212 expression levels in sigmoid colon biopsy samples from healthy controls and patients with ALD. miR-212 expression in patients with ALD was many fold higher than in healthy controls (p < 0.05, Fig. 6A). To confirm that EtOH-induced miR-212 expression is specific, we also measured miR-145 expression levels in colon biopsy samples in patients with ALD. miR-145 is one of the miRNAs expressed in human colon tissue (Cummins et al., 2006), but does not have ZO-1 as a target gene (Blow et al., 2006; Griffiths-Jones et al., 2006). Unlike miR-212, miR-145 expression levels in patients with ALD were not significantly different from healthy controls (Fig. 6A). This indicates that EtOH induction of the expression of miR-212 is specific. Western blot data showed that ZO-1 protein levels were decreased in patients with ALD (Fig. 6B). ZO-1 protein bands of western blots were scanned and assessed using Image J densitometry software. ZO-1 levels were normalized by β-actin levels. The data showed that ZO-1 protein levels in colon biopsy samples of patients with ALD are 59% lower than in controls (p < 0.05, Fig. 6C). ZO-1 mRNA levels in colon biopsy samples of patients with ALD were measured using real-time PCR (Materials and Methods). The data showed that ZO-1 mRNA expression levels in patients with ALD were not significantly different than in controls (Fig. 6D). These data are consistent with our in vitro data (Figs. 2A–C) and indicate that EtOH-induced miR-212 over-expression is specific in intestinal epithelial cells and may play an important role in the mechanism of intestinal barrier dysfunction in patients with ALD.

Figure 6.

 Expression of miR-212 and its target gene, Zonula occludens 1 (ZO-1), in colon biopsy samples from patients with alcoholic liver disease (ALD). (A): miR-212 expression was significantly higher in patients with ALD than in healthy controls (n = 6; p < 0.05). This difference was specific to miR-212 as the expression of miR-145, which does not have ZO-1 as its target gene, did not show a significant difference. (B): ZO-protein levels in patients with ALD. Western blots showed that ZO-1 protein levels in ALD are lower than in controls. The blot was stripped and reprobed for β-actin as a protein loading control. The blot shown is a representative of 3 experiments. (C): Density of ZO-1 protein bands. ZO-1 bands of Western blots were scanned and assessed using Image J densitometry software. ZO-1 levels were normalized by β-actin levels. The data showed that ZO-1 protein levels in ALD are lower than in controls by 59% (n = 6, p < 0.05). (D): ZO-1 mRNA levels in patients with ALD. Real-time PCR (Materials and Methods) data show that ZO-1 mRNA expression levels in patients with ALD were not significantly different than control. Data are expressed as mean ± SD. CON, Control; *p < 0.05 compared with controls.

Discussion

To the best of our knowledge, this study is the first to investigate the effect of EtOH on miRNAs in intestinal epithelial cells. Collectively, our results show that miR-212 is highly expressed both in intestinal epithelial cells and in colon biopsy tissues from patients with ALD. Our in vitro and human data show that EtOH increases miR-212 expression in intestinal epithelial cells, decreases ZO-1 protein levels, but does not affect ZO-1 mRNA. miR-212 disrupts integrity at the tight junction and disrupts barrier integrity of monolayers of intestinal cells. These data suggest a pathophysiologic role for miR-212 in gut leakiness in patients with ALD. In our current view, miR-212 down-regulates ZO-1 translation which disrupts the integrity of tight junctions and causes gut leakiness.

In health and in disease states, dynamic paracellular pathways are highly regulated by tight junctions and are able to change their size under various physiological and pathological conditions (Madara, 1990). Tight junctions function as gates that regulate intestinal permeability (Clayburgh et al., 2004; Laukoetter et al., 2006; Turner, 2006). Cytoplasmic plaque proteins, such as ZO-1, constitute the major part of tight junctions (Sawada et al., 2003). ZO-1 may work as a scaffold in tight junctions because it interacts with all integral membrane proteins. Thus, abnormalities in ZO-1 synthesis, structure or function can cause disruption of the intestinal barrier. Indeed, we recently showed disruption of tight junctions by oxidants, and alcohol resulted in increases in intestinal permeability (Banan et al., 1999, 2000, 2004, 2005b; Forsyth et al., 2007). After alcohol induces miR-212 over-expression, the excess miR-212 molecules may bind to ZO-1 mRNA which then inhibits ZO-1 translation.

To confirm a role for ZO-1 in EtOH-induced intestinal barrier dysfunction, one needs to silence or knock down ZO-1 in epithelial cells. Indeed, the direct effect of ZO-1 depletion on the integrity of tight junctions has been well studied by other researchers (Sawada et al., 2003; Sheth et al., 1997; Shin et al., 2006; Umeda et al., 2004). ZO-1 has been implicated in the development of tight junctions during mouse blastocyst formation (Sheth et al., 1997). To further enable analyses of the role of ZO-1 genes, a ZO-1 knockout cell line was generated by gene-targeting in EpH4, an established line of mouse epithelial cells (Umeda et al., 2004). The studies demonstrated that tight junction formation in the ZO-1 knockout cell line is delayed compared to wild-type cells, and that there is a change in the localization of ZO-1-bound protein (Umeda et al., 2004). The role of ZO-1 in the regulation of tight junctions in mammalian epithelial cells is very important for regulation of intercellular permeability (Sawada et al., 2003; Shin et al., 2006).

The regulation of miRNAs is one of several mechanisms for posttranscriptional modification of proteins. miRNAs are highly conserved regulatory molecules expressed in eukaryotic cells. They are short, noncoding RNAs that regulate gene expression by binding to target mRNAs, which leads to reduced protein synthesis and sometimes decreased steady-state mRNA levels (Carthew, 2006). Although hundreds of miRNAs have been identified, much less is known about their biological function (Ambros, 2004; Krutzfeldt and Stoffel, 2006; Zamore and Haley, 2005). There is evidence that miRNAs affect pathways fundamental to metabolic control in higher organisms, such as adipocyte and skeletal muscle differentiation. Also, some miRNAs are implicated in lipid, amino acid, and glucose homeostasis (Gauthier and Wollheim, 2006).

The selective expression of many miRNAs in some tissues but not others suggests that miRNAs might be involved in the determination or maintenance of specific cell functions. Several reports have suggested that dysregulation of miRNA-triggered cascades might lead to human diseases, including infection and cancer (Ambros, 2004; Krutzfeldt et al., 2006; Kutay et al., 2006). Thus, miRNA abnormalities may contribute to common metabolic diseases. At the same time, miRNA targeting may provide opportunities for the development of novel therapeutic agents (Krutzfeldt and Stoffel, 2006).

The role of miRNAs in regulation of intestinal function including intestinal permeability has not yet been reported. Our study shows that miR-212 is highly expressed in intestinal tissues and may play an important role in the regulation of the intestinal barrier. ZO-1 is predicted to be one of miR-212’s target genes in an established miRNA database (Blow et al., 2006; Griffiths-Jones et al., 2006). Our study provides for the first time, the experimental data to show that ZO-1 is one of miR-212’s target genes.

Ethanol-induced miR-212 expression inhibits ZO-1 translation, but not transcription. This is consistent with other reports that miRNAs reduce steady-state protein levels of their gene targets without affecting corresponding levels of mRNA in animals (Olsen and Ambros, 1999). If miRNAs indeed regulate translation, but not the stability of target mRNAs, it might explain, at least partially, why gene expression profiles based on mRNA analysis do not always correlate with protein expression data (Kern et al., 2003).

Our data suggest that miR-212 plays a fundamental role in the pathogenesis of gut leakiness. As recent studies have shown that miRNA expression can be inhibited in vivo using a pharmacologic approach (Czech, 2006), the potential for using molecular medicine to design new therapeutic weapons based on the modulation of miRNAs for treatment of gut leakiness seems quite feasible (Czech, 2006). But, further studies are needed to confirm our initial findings.

One of the important challenges in studying miRNAs in human biopsy samples are development of better methods. Current methods for detection and quantification of miRNAs are largely based on microRNA microarray data and Northern blotting (Lagos-Quintana et al., 2001). Although miRNA microarrays can provide global throughput of miRNA profiling, the method is relatively limited in terms of sensitivity and specificity (Krichevsky et al., 2003; Liu et al., 2004). Northern blotting while useful for detecting the presence of miRNAs in samples is slow, and requires more initial sample. Real-time PCR is the gold standard for the quantification of gene expression (Heid et al., 1996; Livak and Schmittgen, 2001). Microarray data have to be validated by real-time PCR. Another disadvantage of microarrays and northern blotting is that they need large amounts of RNA to profile miRNA expression compared with the TaqMan MicroRNA assay. It is thus difficult to use microarrays and Northern blotting to profile miRNAs in the small human samples typically obtained during endoscopic procedures, such as sigmoidoscopy.

Our data show that it is possible to use specific stem-loop primers combined with real-time PCR to measure miR-212 expression in human biopsy samples. This assay is specific for measuring mature miRNAs and discriminates among related miRNAs that differ by as little as one nucleotide (Chen et al., 2005). The high sensitivity, specificity, and precision of this method enables us to detect miR-212 expression profiling and can identify and monitor potential biomarkers specific to tissues or diseases (Chen et al., 2005). This technique needs only very small amounts of sample, amounts that are easily obtained during typical endoscopic procedures. Thus, it will provide an opportunity to study the role of miRNAs in a wide range of digestive diseases and to easily evaluate the effects of therapeutic interventions in the same subject because of the ease in obtaining serial samples during a limited, unprepared sigmoid-endoscopy which is a well-tolerated and safe procedure.

It is known that the human genome contains hundreds of miRNAs and that each miRNA can regulate a large number of mRNA targets. Binding of multiple miRNAs to one target could further increase the complexity of target predictions and validation (Krek et al., 2005). Although our study shows that EtOH affects miR-212 and ZO-1 expression in intestinal epithelial cells, many questions still remain to be answered. For example, further studies are needed to determine whether there are other miRNAs that target ZO-1 or to determine the signal pathways through which EtOH induces miR-212 over-expression and to identify the miR-212 binding site for ZO-1. It also remains to be determined whether there is any other target gene for miR-212 that can contribute to intestinal hyperpermeability.

Further studies are also needed to determine the potential role of miRNAs in the regulation of other tight junction proteins and adhesion junction proteins (e.g., occludin, claudin, β-catenin, cadherin, actin and tubulin) known to be critical to barrier function. The potential interactions between miRNAs and other signaling molecules involved in EtOH-induced intestinal barrier dysfunction (e.g., myosin light chain kinase, protein tyrosine phosphatase, PKC, to name a few) also require further investigation (Rao et al., 2004).

Finally, other miRNAs may also be involved in alcohol-induced gut hyperpermeability. For example, our group’s previous studies support a role for activation of NFκB, a transcription factor, in EtOH-induced inducible nitric oxide synthase (iNOS)-up-regulation and the consequent oxidative tissue damage that results in gut leakiness (Banan et al., 1999, 2000, 2001, 2004). A recent study demonstrated that NFkB-dependent induction of miR-146 expression regulates toll-like receptors and cytokine signaling through a negative feedback loop (Taganov et al., 2006). miR-375 is one of the most abundant pancreatic islet-specific miRNAs (Poy et al., 2004). miR-375-mediated changes in myotrophin levels are relevant to the effects of NFκB (Taganov et al., 2006). These studies suggest that NFκB plays an important role in miRNA expression. Further studies are needed to investigate the role of other miRNAs in alcohol-induced gut leakiness and also to determine whether and how alcohol-induced activation of NFκB or up-regulation of iNOS affects the expression of miRNAs.

In summary, we found that miR-212 plays an important role in EtOH-induced intestinal barrier dysfunction. EtOH-induced miR-212 expression results in disruption of intestinal paracellular tight junctions by down-regulation of the translation of ZO-1, one of its target genes. Understanding the biological function of miRNAs, the cellular mechanisms that regulate them and their roles in gut leakiness could lead to the development of novel therapeutic strategies for ALD.

Acknowledgments

This work was supported in part by a grant from the Department of Internal Medicine, Rush University Medical Center, and by National Institutes of Health Grants: NIAAA 013745 (to A. Keshavarzian) and NIDDK 60511 and NCCAM 01581 (to A. Banan).

Ancillary