Chronic Ethanol consumption modulates growth factor release, mucosal cytokine production and microRNA expression in nonhuman primates

Ethics Statement

This study was performed in strict accordance with the recommendations detailed in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health, the Office of Animal Welfare and the United States Department of Agriculture. All animal work was approved by the ONPRC Institutional Animal Care and Use Committee (IACUC). All efforts were made to minimize animal suffering and all procedures performed with qualified veterinary staff present.

Ethanol self-administration model

We used a schedule-induced polydipsia approach described previously (Grant et al., 2008a, Grant et al., 2008b) to establish self-administration of 4% (w/v) ethanol in rhesus macaques. One male and one female cohort (n = 9 females, aged 5–6; n=12 males, aged 5–6) were used in this study. The average daily alcohol dosage (avg g/kg) for each animal ranged from 3.8–5.6 for females (n=6, with 3 controls) and 1.9–3.3 for males (n=8, with 4 controls), as outlined in Table 1.

Tissue collection and processing

All tissues were collected following terminal anesthesia (8 mg/kg ketamine followed by i.v. pentobarbital). Peripheral blood mononuclear cells (PBMC) were isolated by centrifugation over histopaque (Sigma, St Louis, MO) as per manufacturer’s protocol. Mesenteric and tracheobronchial lymph nodes, jejunum, duodenum, ileum and descending colon were collected in RPMI supplemented with FBS, streptomycin/ penicillin, and L-glutamine. RPMI was supplemented with 10% FBS for the collection of lymphoid tissues and 3% FBS for the collection of intestinal tissues. Lymph node lymphocytes were isolated by mechanical disruption and filtration of lymph nodes with a 70µm cell strainer (BD Pharmingen, San Diego, CA). Intestinal lymphocytes were isolated using DTT/EDTA digest (intra-epithelial cells) and collagenase IX/DNAase digest (lamina propria lymphocytes) and subsequently purified over Percoll (GE Healthcare Biosciences, Pittsburgh, PA) gradient as described previously (Veazey et al., 1997). Lavage of the right lung was with PBS was also performed at necropsy, with bronchial alveolar lavage (BAL) cells filtered, pelleted and resuspended in RPMI supplemented with 10% FBS. Cells were cryopreserved in Fetalplex™ Animal Serum Complex (Gemini Bio-Products, West Sacremento, CA)/DMSO.

Flow cytometry

For flow cytometric analysis of cytokine production by mucosal and lymph node lymphocytes, 1–2 × 10⁶ cells were stimulated for 6hr at 37 °C in supplemented RPMI (3% FBS) only or in the presence of phorbol myristate acetate/ionomycin. Brelfedin A was added for the last 5hr of the incubation to all wells. Following incubation, cells were surface stained for CD3, CD4 (eBioscience, San Diego, CA) and CD8 (Beckman Coulter, Brea, CA) and with AQUA Live/Dead stain (Invitrogen, San Diego, CA). Cells were subsequently fixed, permeabilized and stained intracellularly as described (Asquith et al., 201)) for TNFα, IFNγ (eBioscience) IL-2, and IL-17A (BioLegend). All flow cytometry samples were acquired with LSRII instrument (Beckton Dickinson) and analyzed using FlowJo software (TreeStar, Ashland, OR). Frequencies of cytokine-producing cells represent the stimulated sample value minus that of its unstimulated (media alone) control.

Cytokine, chemokine and growth factor analysis

PBMC culture supernatant samples were analyzed using Invitrogen Cytokine Monkey Magnetic 28-plex panel or G-CSF singleplex (Life Technologies, Grand Island, NY) per the manufacturer’s instructions. Samples were run in triplicates. Values below the limit of detection were designated as ND, or not detected.

RNA isolation, cDNA synthesis and analysis of microRNA and transcription factor expression

Total RNA was isolated from 2 × 10⁶ PBMC cells using Trizol method (Invitrogen, Grand Island, NY) as described previously (Asquith et al., 2012). Both PBMC and colon RNA was precipitated in ispropanol overnight at −20°C prior to final extraction. cDNA was synthesized using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) in conjunction with Taqman® microRNA 5x RT primer kits specific for each microRNA. MicroRNA expression was determined by quantitative Real Time PCR using Taqman ® microRNA assays (Applied Biosystems) and a StepOnePlus instrument (Life Technologies, Grand Island, NY). MicroRNA expression levels were normalized to control U6 miRNA expression for each sample using ΔCt calculations (Pfaffl, 2001).

To measure transcription factor expression, cDNA was synthesized from RNA isolated as above, using High Capacity cDNA Reverse Transcription kit (Applied Biosystems). Gene expression was determined using custom Taqman ® primer/probe kits specific for Macaca mulatta cDNA sequences. Transcription factor expression levels were calculated relative to the housekeeping gene Macaca mulatta glutathione synthetase. Samples with low cDNA yield (MGSS cycle number <35 cycles) were excluded from analysis.

Identification of miRNA targets

MicroRNA targets were first analyzed using the TargetScan algorithm (release 6.2, June 2012), using Macaca mulatta reference sequences. Positive hits were then verified using a second algorithm to improve specificity and avoid false positives as has been described previously (Asirvatham et al., 2008). The second algorithm used was miRanda software (August 2010 release) available at www.microRNA.org. For each putative target gene analyzed by miRanda for miRNA target sites, the homo sapiens sequence was analyzed using a mirSVR threshold < = − 0.1. Some transcription factor target genes selected for this approach were selected bioinformatically. In this instance, transcription factors predicted to bind to the promoters for VEGF, G-CSF, EGF and MIF were identified using the Champion ChiP Transcription Factor Portal (Qiagen), which uses SABiosciences’ Text Mining Application and the UCSC Genome Browser (available at www.sabiosciences.com).

Transfection of miRNA mimics and antagomirs into PBMC

PBMC were cultured at 1–2 × 10⁶ cells per well in a 96-well plate with RPMI-1640 supplemented with 10% FBS and transfected with 60 nM of either mimics (miR-181b, miR-221, miRNA- neg ctrl) or antagomirs (miR- 181b, miR- 221 or miRNA- neg ctrl) (Thermo-scientific) using nucleofection technology (human T cell Nucleofector kit, program F1–115) in accordance with manufacturer's recommendations. After a 24 h incubation, cells were stimulated overnight with 100 ng/ml PMA and 500 ng/ml ionomycin and harvested 14 h later for western blot analysis. Each transfection experiment was done in triplicate.

Western Blot Analysis

Total protein extracts were prepared in Ripa lysis buffer. The protein concentrations were determined by Bradford assay (BioRad). Approximately 30–40 µg of lysate was separated on a 10% SDS-PAGE and transferred to polyvinylidene difluoride membrane (Millipore). The membranes were incubated in 5% nonfat milk powder diluted in TBST for 30 min at room temperature, and then probed with a human monoclonal anti-STAT3 (1:2000), anti-ARNT antibody (1:1000), anti-HGF (1:5000), anti-VEGF (1:5000) (cell signalling) in 1% milk diluted in 1xTBST overnight at 4°C. The membrane was washed three times with 1xTBST and incubated with horseradish peroxidase conjugated secondary antibody (goat antirabbit, 1∶5000) for 2 h at room temperature. The Membrane was then washed three times with 1xTBST and 1x TBS respectively. Immuno-complexes were detected with an enhanced chemiluminescence method using DURA kit (GE Healthcare). The same membranes were stripped and re-probed with anti–β-actin monoclonal antibody (1:2000, cell signaling). Images of autoradiography were acquired using a scanner EPSON Perfection 2580 Photo (EPSON) and quantified by Image J 1.34 Software (http://rsb.info.nih.gov/ij).

Statistics

Statistical analysis and graphing was conducted with GraphPad Prism software (GraphPad Software, Inc, La Jolla, CA). Correlation analyses were performed with Spearman rank correlation test. Analyses that compared drinkers vs controls were performed using Student’s t test or nonparametric Mann-Whitney U test appropriate to normality distribution.

The impact of chronic alcohol consumption on cytokine, chemokine and growth factor production by PBMC

To initially examine the impact of chronic alcohol consumption on immune function, we compared soluble factor production by polyclonally-stimulated PBMC from drinkers versus those of control animals. Phorbol myristate acetate/ionomycin stimulation was chosen for these initial experiments given their broad range of action and their robust effect on cytokine production by T cells. We used a 28-plex array to measure protein levels of 13 cytokines (IFNγ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-12, IL-15, IL-17, TNFα, IL-1ra, IL-10 and MIF), 6 growth factors (EGF, FGF, G-CSF, GM-CSF, HGF and VEGF) and 9 chemokines (MCP-1, MIP1α, MIP1β, RANTES, EOTAXIN, MDC, IL-8, MIG and I-TAC). There were no differences in cytokine (Fig 1A) or chemokine (Fig 1B) production by PBMC obtained from control versus ethanol-consuming animals. By contrast, hepatocyte growth factor (HGF), granulocyte-colony stimulating factor (G-CSF) and vascular-endothelial growth factor (VEGF) were produced at significantly lower levels in drinkers versus controls (Fig 1) Epithelial growth factor (EGF) also showed a trend towards decreased expression in drinkers (p = 0.076). Since females had higher BECs than males, we also compared cytokines, chemokines and growth factor production after PMA/ionomycin stimulation between these two groups. We noted higher expression IL-1b, IL-15, ITAC, EGF, and IL-8 expression in females compared to males (data not shown). However, it is unclear whether theses differences are attributed to ethanol, gender or both.

The impact of chronic alcohol consumption on mucosal cytokine production

In light of the increased susceptibility of alcoholics to lung infection, we examined cytokine, growth factor and chemokine production by lung resident immune cells. However, by contrast to PBMC, bronchoalveolar lavage (BAL) cells collected from drinkers and controls exhibited equivalent production of all proteins examined upon polyclonal stimulation (Fig S1). Similarly, the frequency of IL-2+, TNFα+, IFNγ+ and IL-17+ CD4 and CD8 T cells in BAL and tracheobronchial lymph node were comparable between the two groups (Fig 2A and B). Moreover, drinkers and controls had a similar frequency of both central and effector memory CD4 and CD8 T cells amongst BAL cells (Fig 2C and D) as well as total CD4 and CD8 T cells illustrated by comparable CD4/CD8 T cell ratio in the BAL (Fig 2E).

We next examined the frequency of CD4 and CD8 T cells producing IL-2, TNFα, IFNγ and IL-17 in jejunum, duodenum, ileum and colon, as well as the intestinal draining mesenteric lymph node by intracellular cytokine staining (Fig 3, S2, and S3). Both intra-epithelial and lamina propria lymphocytes were analyzed since they represent distinct subpopulations of the intestinal immune axis. Cytokine responses were typically more variable among drinkers, with more ‘low responders’ detected in this group. We detected a lower frequency of TNF-expressing CD8 T cells within the duodenal lamina propria of ethanol consuming animals (Fig 3A). In light of several recent reports that IL-17 and IFNγ are often co-expressed by mucosal T cells with both protective and pathogenic effector functions, we also enumerated the frequency of IFNγ/IL-17 co-expressing T cells. Interestingly, chronic alcohol exposure significantly reduced the frequency of these cells within CD4+ intraepithelial T cells of the jejunum (Fig 3C) and CD8+ T cells of the ileal lamina propria (Fig 3D). No differences in IL-2, IFNγ or IL-17 production by CD4 or CD8 T cells was observed between ethanol-consuming and control animals (Fig S3).

Since cytokine responses were particularly variable amongst drinkers, we next performed correlation analyses to establish dose-dependent effects of ethanol consumption on tissue cytokine production. We observed a negative association in the frequency of colonic CD4 T cells producing multiple cytokines as a function of increasing ethanol dose (Fig 4). This trend was significant for IL-2+, TNFa+, IFNγ+ and IFNγ+/IL-17+cells. We also observed a trend towards negative association with IL-17+ cells, albeit non-significant. CD8+ T cells did not show a significant association between increasing alcohol exposure and cytokine production. Thus, CD4+ T cells of the colon appeared exquisitely susceptible to immune modulation by ethanol. Moreover, the ratio of CD4:CD8 T cells was equivalent between controls and drinkers in all gut tissues examined (Fig S4), further indicating that this reduction was at the functional level.

Chronic alcohol consumption modulates microRNA expression

MicroRNAs are ∼ 22 nucleotide long endogenous RNAs that target mRNAs for translational repression or degradation (Tomankova et al., 2011). MicroRNAs are increasingly appreciated as potent regulators of immune function, and reports indicate that alcohol may modulate microRNA expression (Miranda et al., 2010). We therefore compared the expression of a panel of microRNAs highly expressed in lymphoid cells (Felli et al., 2005, Xiao et al., 2008, Li et al., 2007, Lu et al., 2010, Rouas et al., 2009, Ma et al., 2011, Lu et al., 2009, Rodriguez et al., 2007) by quantitative real-time PCR in both PBMC and colon. This panel included microRNAs miR-181a-5p, miR-146a-5p, miR-155-5p, miR-221-3p, miR-17-3p, miR-17-5p, miR-21-5p and miR29a.

Strikingly, both PBMC and colon exhibited differential expression of specific microRNAs between drinkers and control animals (Fig 5). Within PBMC, the expression of both miR-181a and miR-221 was significantly upregulated in animals that chronically consumed ethanol with the rest showing a trend of equivalent or higher expression than controls (Fig 5A). Further analysis revealed increased expression of miRNA 221, miRNA 17-5p, and miRNA 21 in females compared to males following ethanol consumption (data not shown). However, as noted earlier, it is unclear at this point whether this difference could be attributed to increase in ethanol intake, gender or both. Notably, none of the microRNAs examined was downregulated in PBMC from drinkers. By contrast to PBMC, miR-181 and miR-221 had equivalent expression in colon sections from ethanol-consuming and controls (Fig 5B). Instead, microRNA miR-155 was significantly upregulated in colon from ethanol-exposed animals. Therefore, reduced growth factor and cytokine production following chronic alcohol consumption is associated with tissue-specific upregulation of microRNAs with potential immune modulatory activity.

Expression of transcription factors containing predicted miRNA target sites in drinkers vs controls

Since we observed an alcohol-dependent upregulation of microRNAs miR-181a and miR-221 in PBMC, we next examined whether cytokines found to be differentially expressed were potential microRNA targets. Using miRanda software, we used a bioinformatics approach to search target sequences for either miR181a or miR-221 in the sequence of growth factors downregulated in PBMC from ethanol-consuming animals (VEGF, G-CSF, HGF). Using a high threshold to identify potential binding sites (mirSVR threshold < = − 0.1) to identify potential binding sites, we found that none of the microRNAs examined were predicted to bind the aforementioned cytokine target sequences. We therefore tested the hypothesis that these microRNAs may bind transcription factors known or predicted to regulate expression of these genes. Interestingly, multiple transcription factors (summarized in Fig S5) were found to contain 3’-UTR sequences with target sites for miR221 and miR181a.

We therefore analyzed expression of some of the transcription factors with predicted miRNA target sites (cFos – miR181a and miR221; ARNT – miR221 and miR181a; STAT3 – miR181a; RUNX1 – miR221 and AhR – miR221) in PBMC from both drinkers and control animals (Fig 6A). Notably, chronic alcohol consumption led to significant downregulation of both ARNT and STAT3 mRNA expression. By contrast, expression of transcription factors cFos, RUNX1 and AhR mRNA in PBMC was comparable between both groups. We then compared the protein expression levels of STAT-3 and ARNT in PBMC isolated from drinking or control animals (Fig 6B). In line with the reduction in gene expression levels, protein levels of both STAT-3 and ARNT were reduced in ethanol-consuming animals (Fig 6B and C). These data suggest that chronic ethanol consumption led to a significant downregulation of transcription factors that regulate cytokine expression and are predicted targets for ethanol sensitive microRNAs.

miR-181 and miR-221 modulate STAT3, ARNT, HGF, VEGF and G-CSF protein expression

We next determined whether STAT3 and ARNT expression is modulated by these two microRNAs. To that end, we analyzed STAT3 and ARNT expression in PBMC following transient transfection of chemically-stabilized dsRNA oligomers that mimic the function of endogenous mature miRNA (mimics), or chemically-modified single-stranded antisense oligomers that inhibit miRNA function (antagomirs) (Fig 7). In PBMC transfected with miRNA mimics specific for miR-181, miR-221, miR-ctrl and then stimulated them with PMA and ionomycin we observed down-regulation of both STAT3 and ARNT protein levels (Fig 7A and C). As a corollary, transfection of the polyclonally stimulated PBMC with antagomirs induced up regulation of STAT-3 and ARNT protein levels (Fig 7B and C).

Since STAT3 and ARNT modulate HGF, VEGF and G-CSF expression a (Albasanz-Puig et al., 2012, Lin et al., 2012, Tacchini et al., 2003, Nguyen-Khac et al., 2006, Zhang et al., 2012, Zimmers et al., 2003), we analyzed HGF and VEGF levels by western blot in lysates of PBMC transfected with miRNA 181/221 and mimics or antagomirs and then stimulated with PMA/ionomycin (Fig 7D–F). Secretion of G-CSF into culture supernatant was carried using singleplex assay because we were not able to detect protein expression using the commercially available antibody (Fig 7G). Immunoblot analysis revealed that levels of VGEF and HGF were reduced in PBMC transfected with mimics (Fig 7D and F) while the opposite effect was observed with antagomirs (Fig 7E and F). Our analysis also shows that increasing levels of miR-181a and 221 through transfection of mimics led to decreased levels of G-CSF. Conversely, microRNA antagomirs results in increased secretion of G-CSF (Fig 7G).