Dysregulated miR-34a–SIRT1–Acetyl p65 Axis Is a Potential Mediator of Immune Activation in the Colon during Chronic Simian Immunodeficiency Virus Infection of Rhesus Macaques

Regardless of the route of transmission, mucosal tissues, particularly the gastrointestinal (GI) tract, are targeted by HIV/SIV, leading to rapid, severe, and sustained depletion of CD4⁺ T cells in HIV-infected individuals and SIV-infected rhesus macaques (1–5). As disease progresses, GI complications such as anorexia, weight loss, and diarrhea become frequent and are being reported in patients despite the extensive use of highly active antiretroviral therapy (6). Histologically, GI disease is characterized by infiltration of the lamina propria by T cells, plasma cells, macrophages, and morphologic changes such as villus blunting and crypt hyperplasia. An emerging feature of HIV/SIV pathogenesis is the markedly elevated levels of microbial translocation that occur in the later stages of infection (7, 8). This phenomenon has been proposed to play a key role in driving localized and systemic immune activation, which is a well-recognized correlate of HIV/SIV disease progression. The mechanisms leading to increased microbial translocation in AIDS patients remain largely unknown. However, the “leaky gut syndrome” is a preferred hypothesis, wherein luminal bacteria and/or their products enter the intestinal lamina propria through a disrupted epithelial barrier and pass via the portal blood into the systemic circulation.

Viral replication and CD4⁺ T cell depletion in the lamina propria leukocyte (LPL) compartment is associated with elevated expression of proinflammatory genes and reduced expression of genes involved in maintenance of epithelial barrier, repair, and digestive and metabolic functions (9–12). Furthermore, focused longitudinal examination of individual mucosal compartments has revealed deeper insights into the molecular pathological events occurring in the intestinal LPL and epithelial compartments (13, 14). Whereas inflammation and immune activation–related genes showed marked changes in the LPL compartment, genes regulating enterocyte maturation, differentiation, and epithelial barrier function such as Wnt-TCF7L2, Notch signaling proteins, adherens junction, hemidesmosomes, and desmosomes were found to be significantly dysregulated in the epithelial compartment following SIV infection (13, 14). Overall, these studies demonstrated considerable alterations in enterocyte structure and function that could facilitate microbial translocation.

Although multiple mechanisms involving transcription factors, chromatin modifications, and others such as histone modifications are known to regulate gene expression, one important mechanism mediated by small regulatory RNAs called microRNAs (miRNAs) has gained a lot of attention recently (15). miRNAs are ∼21–23 nt in length and have been described to impact almost all cellular processes by repressing gene expression at the posttranscriptional level (15). A growing body of evidence indicates that HIV infection is characterized by dysregulated miRNA expression (16), including direct targeting and crippling of the miRNA biosynthesis machinery by HIV (17). Recent studies performed in SIV-infected rhesus macaques also demonstrated dysregulated miRNA expression in plasma (18), brain (19), and monocyte derived macrophages (20). We recently reported altered miRNA expression in the intestine during acute SIV infection (21). More specifically, we identified miR-190b to be significantly upregulated as early as 7 d after SIV infection, and its expression remained elevated throughout SIV infection. Additional studies also suggested that miR-190b could influence disease pathogenesis by directly binding to the 3′ untranslated region (UTR) and regulating the expression of MMTR6, a phosphatidylinositol 1,3-bisphosphate previously shown to inhibit T cell and macrophage activation. In the present study, we performed miRNA profiling in colon during chronic SIV infection and detected marked changes in the expression of miRNAs linked to inflammation, cell cycle arrest, and senescence. Among these, the expression of miR-34a, a miRNA shown to regulate apoptosis, cell cycle control, and senescence (22–24), was markedly increased in both colonic epithelial and LPL compartments. Furthermore, increased miR-34a expression was associated with reduced expression of SIRT1, a longevity-inducing anti-inflammatory, anti-stress, and anti-aging class III histone deacetylase in the colonic epithelium and LPLs. Most noticeably, reduced SIRT1 expression in the LPL compartment was accompanied by enhanced expression of acetylated p65 (lysine 310), suggesting hyperactivation of the NF-κB signaling pathway. Furthermore, enhanced constitutive expression of acetylated p65 (lysine 310) was localized to lamina propria plasma cells. The results of our study indicate that the miR-34a–SIRT1–acetyl p65 axis represents a potential molecular pathological mechanism that could drive chronic hyperactivation of the intestinal B cell system, leading to their dysfunction.

All experiments using rhesus macaques were approved by the Tulane University Institutional Animal Care and Use Committee (protocol no. 3574). The Tulane National Primate Research Center is an Association for Assessment and Accreditation of Laboratory Animal Care International–accredited facility (no. 000594). The National Institutes of Health Office of Laboratory Animal Welfare assurance number for the Tulane National Primate Research Center is A3071-01. All clinical procedures, including administration of anesthesia and analgesics, were carried out under the direction of a laboratory animal veterinarian. Animals were anesthetized with ketamine hydrochloride for blood collection procedures. Intestinal resections were performed by laboratory animal veterinarians. Animals were preanesthetized with ketamine hydrochloride, acepromazine, and glycopyrolate, intubated, and maintained on a mixture of isoflurane and oxygen. Buprenorphine was given intraoperatively and postoperatively for analgesia. All possible measures are taken to minimize discomfort of all the animals used in this study. Tulane University complies with National Institutes of Health policy on animal welfare, the Animal Welfare Act, and all other applicable federal, state, and local laws.

Colon tissues were collected from a total of 61 Indian-origin rhesus macaques, including 45 animals infected with pathogenic strains of SIV that use CCR5 in vivo and 16 uninfected macaques. All macaques were infected with SIVmac251. Animal IDs, duration of infection, route of infection, plasma and colonic viral loads, and intestinal histopathology in all SIV-infected animals are provided in Table I. For i.v. inoculation, a viral dose of 100 fifty percent tissue culture infective dose TCID₅₀ was used. However, for intravaginal and intrarectal inoculation, a higher viral dose of 500 and 1000 fifty percent tissue culture infective dose, respectively, was used. Twelve chronic SIV-infected animals (Table I) and four control animals (Table I) were used for TaqMan low-density array (TLDA) miRNA profiling. An additional 45 animals comprising 5 animals each at 7–10 and 13–14 d postinfection, 7 each at 21 and 90 d postinfection, 9 chronic SIV-infected, and 12 control macaques (Table I) were included for quantitative RT-PCR (qRT-PCR) confirmation, in situ hybridization, and immunofluorescence studies. Data regarding viral inoculum, duration of infection, and plasma and intestinal viral loads for the 24 animals at 7–10, 13–14, 21, and 90 d postinfection have been previously reported (21). In the control group, colon resection segments (∼5 cm long) were collected from three macaques (HF54-Pre, HR42-Pre, HD08-Pre) before SIV infection (Table I).

All tissues were collected at necropsy in RNAlater (Ambion, Austin, TX) for total RNA extraction and qRT-PCR. For histopathologic evaluation, colon tissues were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 7 μm, and stained with H&E for analysis. Sections were examined in a blinded fashion, and histological signs of inflammation were scored semiquantitatively on a scale of 0–3 as follows: 0, within normal limits; 1, mild; 2, moderate; 3, severe. Additionally, the presence of crypt dilatation/cryptitis, amyloidosis, and lymphoid hyperplasia were also recorded (Table I). Mild changes such as lymphoid hyperplasia were assigned a score of 1.

For TLDA studies ∼350 ng total RNA from intact colon tissue was first reverse transcribed, preamplified, and loaded onto TLDA cards and processed according to the manufacturer’s recommended protocol described previously (21).

Expression of miR-34a and SIRT1 mRNA was further investigated in both colonic epithelium and LPL compartments using qRT-PCR as previously described (21). For SIRT1 qRT-PCR assays, ∼1 μg total RNA was first reverse transcribed using the SuperScript. III first-strand synthesis system for RT-PCR kit following the manufacturer’s protocol. Each qRT-PCR reaction (20 μl) contained the following: 2× Power SYBR Green master mix without uracil-N-glycosylase (12.5 μl); SIRT1 forward, 5′-AAGATGA GCCTGATATTCCAGAGAGAGCTG-3′, reverse, 5′-GCCTCTTGATCATCTCCATCAGTCCCAAA-3′; GAPDH forward, 5′-CAAGAGAGGCATTCTCACCCTGAA-3′, reverse, 5′-TGGTGCCAGA TCTTCTCCATGTC-3′; and β-actin forward, 5′-CAACAGCCTCAAGATCGTCAGCAA-3′, reverse 5′-GAGTCCTTCCACGATACCAAAGTTGTC-3′ (200 nM); and cDNA (4 μl).

To determine the mucosal compartment contributing to miR-34a upregulation, we separated the intestinal epithelial cells from the underlying LPLs and fibrovascular stroma as described previously (21).

In situ hybridization for miR-34a was performed using locked nucleic acid (LNA)-modified DNA probes (Exiqon, Vedbaek, Denmark). Briefly, 7-μm formalin-fixed, paraffin-embedded tissue sections were first deparafinized, rehydrated in descending series of ethanol, and pretreated in a microwave with citrate buffer (Ag unmasking solution; Vector Laboratories, Burlingame, CA) for 20 min at high power according to the manufacturer’s instructions. Thereafter, sections were thoroughly washed, placed in a humidified chamber, and hybridized overnight at 50°C with 60 nM digoxigenin-labeled LNA-modified miR-34a DNA probes (5DigN/ACAACCAGCTAAGACACTGCCA/3Dig_N/, RNA melting temperature of 85°C) or scrambled probe (5DigN/GTGTAACACGTCTATACGCCCA/3Dig_N/, RNA melting temperature of 87°C) in 1× miRNA in situ hybridization buffer (Exiqon). After hybridization, slides were washed with 5× SSC (SSC buffer), 1× SSC, 0.2× SSC, and blocked with Dako protein-free blocker (Dako Laboratories) for 1 h. Fab fragments of an anti–digoxigenin Ab conjugated with alkaline phosphatase (Roche Diagnostics, Penzberg, Germany) (1 in 250) were used to detect digoxigenin-labeled probes. Positive signals were detected by incubating the sections with permanent red (Dako Laboratories) for 7 min. Controls included matched tissues hybridized with LNA-modified scrambled probes.

Immunofluorescence (IF) for detecting γH2A.X (9F3, 1 in 250) (Abcam), acetyl p65 (lysine 310; 1 in 100) (Genetex), SIRT1 (E104, 19A7AB4, 1F3) (Abcam), SIRT1 (Bioss), SIRT1 (LS-B8356, 1 in 62.5) (LifeSpan Biosciences), and p21 (12D1, 1 in 50) (Cell Signaling Technology) was done as described previously (10). Immunophenotyping of acetylated p65⁺ cells was done using CD68 (1 in 20) (Dako, Glostrup, Denmark), CD163 (1 in 50) (Serotec, Raleigh, NC), CD3 (1 in 20) (Dako), anti-rhesus IgA (1 in 250), IgG (1 in 250) (Both Abs from Research Diagnostics, Flanders, NJ), and appropriate Alexa Fluor–conjugated secondary Abs.

Quantitation of cells and regions of interest (ROI) labeled by LNA-modified miR-34a in situ hybridization probes was performed using Volocity 5.5 software (PerkinElmer, Waltham, MA) after capturing images on a Leica confocal microscope. Several ROI were hand drawn on the epithelial and LPL regions in the images from colon. The data were first graphed and then analyzed using the Mann–Whitney U test employing the Prism v5 software (GraphPad software). A p value <0.05 was considered statistically significant.

The 3′ UTR of the rhesus SIRT1 mRNA contains two predicted miR-34a binding sites (TargetScan 6.2), that is, site 1, 5′-UUCCACAAGUAUUAAACUGCCAA-3′ (nt 891–897), and site 2, 5′-CCAGCUAGGACCAUUACUGCCAG-3′ (nt 1434–1440) (Table II). Accordingly, a G-block dsDNA sequence (∼125 bases long) containing both miR-34a binding sites along with ∼30 bases flanking the 5′ and 3′ ends of each site was synthesized (Integrated DNA Technologies, Coralville, IA) for cloning into the pmirGLO dual luciferase vector (Promega, Madison, WI). A second G-block sequence with the exact same flanking UTR sequences but with both miR-34a binding sites (total 14 nt from both sites) deleted was also synthesized to serve as a negative control. Both oligonucleotide sequences were synthesized with a PmeI site on the 5′ and XbaI site on the 3′ end for directional cloning. The pmirGLO vector was first cut with PmeI and XbaI restriction enzymes, gel purified, and ligated with either the wild-type (WT) sequence containing miR-34a binding sites (SIRT1-WT) or the deleted (Del) sequence (SIRT1-Del). HEK293 cells were plated at a density of 5 × 10⁴ cells per well of a 96-well plate. At 60% confluence, cells were cotransfected with ∼100 ng SIRT1-WT or SIRT1-Del UTR miRNA luciferase reporter vector and 100 nM miR-34a or negative control mimic (5′-CCACAACCUCCUAGAAAGAGUAGA-3′) using the DharmaFECT Duo transfection reagent (Thermo Fisher Scientific). In separate wells, cells were also transfected with pmirGLO vector (Promega) as a normalization control. After 72 h, the Dual-Glo luciferase assay was performed according to the manufacturer’s recommended protocol using the BioTek H4 Synergy plate reader (BioTek, Winooski, VT). The normalized firefly/renilla ratio was calculated to determine the relative reporter activity. Experiments were performed in six replicates and repeated twice.

Total RNA samples from all SIV-infected animals were subjected to a quantitative real-time TaqMan two-step RT-PCR analyses to determine the viral load in colon samples. Briefly, primers and probes specific to the SIV long terminal repeat sequence were designed and used in the TaqMan real-time RT-PCR assay. Probes were conjugated with a fluorescent reporter dye (FAM) at the 5′ end and a quencher dye at the 3′ end. Fluorescence signal was detected using the ABI Prism 7900 HT Fast Real-Time PCR System (Life Technologies). Data were captured and analyzed with Sequence Detection software v2.4 (Life Technologies). Viral copy number was determined by plotting C_T values obtained from the colon and jejunum samples against a standard curve (y = −3.384x + 39.029) (r² = 0.998) generated with in vitro–transcribed RNA representing known viral copy numbers.

TLDA-SDS–run files from 12 SIV-infected and 4 control animals were loaded onto Relative Quantification Manager software v1.2.1 and analyzed using automatic baseline settings and manual threshold of 0.2. One (EL66) of the four control samples was selected as a calibrator and set to 1. The remaining 15 samples were expressed as an n-fold difference relative to the calibrator sample. The results from the Relative Quantification Manager analysis containing five columns (well, sample, detector, task, and C_T values) were saved as a tab-delimited text file, imported, and analyzed using the DataAssist v3.01 software (Life Technologies) employing the global normalization method, as this method has been proposed to be more sensitive and accurate for analyzing high-throughput TaqMan miRNA qRT-PCR than endogenous controls (25, 26). The C_T upper limit was set to 33, meaning that all miRNA detectors with a C_T value ≥33 were excluded. Multiple comparisons correction for TLDA miRNA profiling (simultaneously applied to all 768 miRNA targets on cards A and B) was performed using the Benjamini–Hochberg method. A p value <0.05 was considered significant. TLDA miRNA data were deposited with Gene Expression Omnibus (accession no. GSE60773, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE60773).

For miR-34a and SIRT1 qRT-PCR studies, one uninfected control macaque with the highest (for miR-34a) or lowest (for SIRT1) ΔΔC_T value served as the calibrator/reference and was assigned a value of 1. All differentially expressed miRNAs or mRNAs in SIV-infected and other control macaques are shown as an n-fold difference relative to this macaque. This approach was preferred over calculating miRNA/mRNA fold change using an average of all control animal ΔC_T values mainly to facilitate graphing the control samples so that variation within the control group can be displayed. Individual miR-34a and SIRT1 mRNA qRT-PCR data were analyzed by a nonparametric Wilcoxon rank sum test for independent samples using RealTime StatMiner software (Integromics). A Spearman nonparametric one-tailed correlation analysis was performed to determine the degree of association between colon inflammation scores/tissue viral loads and miR-34a fold expression. Firefly/renilla ratios were analyzed using one-way ANOVA, and post hoc comparisons were performed using a Tukey multiple comparisons test.

All chronic SIV-infected macaques obtained at necropsy had substantial viral loads in the colon that ranged from 0.04 × 10⁶ to 2075 × 10⁶ copies/mg total RNA with a median of 0.9 × 10⁶ copies/mg total RNA (Table I). Similar to colonic viral loads, plasma viral loads in chronic SIV-infected macaques ranged from 0.3 × 10⁶ to 5000 × 10⁶ copies/ml with a median of 15 × 10⁶ copies/ml. Plasma viral loads were not available for 6 animals (GH25, CP47, HD08, DE50, HN15, and R191).

All chronic SIV-infected animals with the exception of four (HF27, HB48, GK31, and HG58) progressed to AIDS. CD4⁺ T cell data were only available for eight chronic SIV-infected macaques and, as shown in Supplemental Fig. 1, all eight had significant mucosal CD4⁺ T cell depletion compared with uninfected control macaques. Histologic evaluation of H&E-stained sections of colon (Table I) from all SIV-infected macaques revealed the presence of mild to severe colitis including other intestinal lesions such as crypt abscess/cryptitis (Table I). In contrast to the colon, only 9 of the 21 chronic SIV-infected macaques had mild to moderate enteritis (Table I). Opportunistic pathogens such as Mycobacterium avium intracellulare were not detected in 20 of 21 chronic SIV-infected macaques.

We performed global miRNA profiling of colon tissue from 12 chronically SIV-infected and 4 uninfected control macaques using the TLDA cards to determine whether miRNA expression was altered in the GI tract/immune system during chronic SIV infection. After applying multiple comparisons correction, 10 miRNAs (3 upregulated and 7 downregulated) were found to be statistically significant (adjusted p < 0.05) and differentially expressed following analysis using DataAssist software version 3.01 (Table III).

Among the three upregulated miRNAs (cards A and B), the expression of miR-223* and miR-34a was >2-fold, and miR-424 showed an ∼1.4-fold increase in colon of chronic SIV-infected macaques (Table III). The expression of seven miRNAs decreased from 1.2- to 1.8-fold (Table III). At least 13 other miRNAs that comprised three previously well-characterized inflammation-associated miRNAs [let-7b (p = 0.0578), miR-146b-5p (p = 0.0719), and miR-21 (p = 0.0825)] showed a tendency to be statistically significant (p < 0.1) (Table III). When multiple comparisons correction was not applied, 68 (31 upregulated and 37 downregulated) miRNAs were found to be differentially expressed (Supplemental Table I). Among the upregulated miRNAs was miR-190b (3.2-fold increase), which we recently reported to be significantly upregulated throughout the course of SIV infection primarily in response to viral infection/replication and not immune/inflammatory response accompanying viral replication (21). These findings show that chronic SIV infection is characterized by marked changes in the expression of multiple inflammation-related miRNAs in the colon and that the profile is dominated by downregulated miRNAs.

Among the differentially expressed miRNAs, we decided to focus on miR-34a because it is a highly conserved miRNA previously linked to inflammation, apoptosis, and cellular senescence/cell cycle arrest (22–24). Further, miR-34a has been directly linked to “inflammaging,” an aging-related state characterized by systemic chronic inflammation proposed to promote premature aging of HIV-infected patients (27). Furthermore, miR-34a has also been associated with the induction of cellular senescence in normal colonic epithelial cells in response to treatment with irinotecan, a topoisomerase I–inhibiting anticancer drug (28). Finally, miR-34a was the most abundant (lowest C_T) among the 10 differentially expressed miRNAs whose expression was significantly upregulated in the colon of chronic SIV-infected macaques (Table III).

Because miR-34a expression was significantly elevated in intact colons of chronic SIV-infected macaques when using the TLDA cards, we next quantified its expression using a miR-34a–specific qRT-PCR assay separately in the colonic epithelial and LPL compartments to identify the mucosal compartment contributing to its upregulation. The protocol used for intestinal cell isolation in the present study yields epithelial cells with ∼85–90% purity with minimal contamination with intraepithelial lymphocytes (29). Similar purity has also been obtained for LPLs with minimum contamination with epithelial cells (29). The LPLs isolated consist mostly of lymphocytes (50–60%) but also contain substantial numbers of plasma cells and macrophages and smaller numbers of neutrophils, dendritic cells, and eosinophils. Interestingly, miR-34a expression was significantly elevated in both epithelium (∼3.8-fold increase; p = 0.0057) (Fig. 1A) and LPL compartments (∼2.7-fold increase; p = 0.0051) (Fig. 1B).

We then sought to determine the timing of miR-34a upregulation during the course of SIV infection by quantifying miR-34a expression in colon samples collected from 24 additional macaques at 7–10 (n = 5), 13–14 (n = 5), 21 (n = 7), and 90 (n = 7) d postinfection. CD4⁺ T cell status and plasma and intestinal viral loads for all 24 animals have been previously reported (21). As shown in Fig. 1C, no significant increase in miR-34a expression was observed at any of the time points compared with uninfected controls. This suggests that miR-34a upregulation is tied more tightly with inflammation, which is minimal during acute infection than viral load, which is very high acutely, particularly at 10–14 (peak viral load) d post infection. Additionally, miR-34a fold change in both colonic epithelium and LPL compartments showed very good positive correlation with tissue viral loads and inflammation scores (Fig. 2). Interestingly, the correlation was higher in the LPL (Fig. 2C, 2D) compared with the epithelial compartment (Fig. 2A, 2B). These data suggested that miR-34a upregulation had a strong association with inflammation and that both epithelial (Fig. 1A) and LPL (Fig. 1B) compartments contribute to its elevated expression detected in intact colon tissue (Table III).

Overall, upregulation of the inflammation-associated miR-34a in both epithelial and LPL compartments suggests that it could potentially exert distinct pathophysiologic effects in both mucosal compartments through epigenetic regulation of gene expression.

Recent studies show that miR-34a transcription is elevated in response to irradiation or oxidative stress (inflammation)–induced DNA damage (30, 31). Consequently, we verified the presence of dsDNA damage response (DDR) in the colon by determining γH2A.X expression. Both dsDNA breaks and telomere uncapping can induce a DDR, resulting in the activation and recruitment of ATM/ATR proteins to the site of damage (32), which then phosphorylates histone H2A.X proteins on Ser¹³⁹ to facilitate the focal assembly of checkpoint and DNA repair proteins to the DNA damage site (32). Consistent with these findings, enhanced phospho-γH2A.X (Ser¹³⁹) expression was detected in colonic crypt epithelial cells of chronic SIV-infected macaques (Fig. 3A, 3B), an indicator of active DDR signaling. More strikingly, epithelial cells present along the entire circumference of the crypts showed intense nuclear staining for phospho-γH2A.X (open circles in Fig. 3A, 3B). In contrast, phospho-γH2A.X staining was either extremely weak or completely absent in the crypt epithelial cells of the control macaque (Fig. 3C). Additionally, elevated phospho-γH2A.X expression was also detected in LPLs (Fig. 3A–C).

Interestingly, persistent DDR manifested by phospho-γH2A.X accumulation at DNA damage foci is also considered an initiating signal for cellular senescence/aging (33). In this context, elevated miR-34a expression in renal and endothelial cells was well demonstrated to promote cellular senescence (23, 24). Accordingly, we used a combination of phosho-γH2A.X and the proliferation marker Ki67 (32) where senescent cells should stain negative for Ki67 and positive for phospho-γH2A.X. Surprisingly, most of the colonic crypt epithelial cells of chronic SIV-infected macaques strongly expressed both phospho-γH2A.X (Fig. 3A, 3B) and Ki67 (Fig. 3D–F), suggesting that these cells, although showing clear evidence of DDR, did not undergo senescence and retained the capacity to proliferate. Additionally, the absence of cell cycle arrest/cellular senescence was further confirmed based on the expression pattern of cyclin-dependent kinase inhibitor p21, a cell cycle arrest/cellular senescence–promoting protein that unlike Ki67 stained differentiated enterocytes near the lumen but not the crypt epithelial cells (Supplemental Fig. 2). Quantitative image analysis showed significantly increased expression of phospho-γH2A.X in both colonic epithelial (p < 0.0001) and LPL (p = 0.0034) compartments (Fig. 4A). Interestingly, between the two mucosal compartments, phospho-γH2A.X expression in the epithelium showed better statistical significance compared with the LPL compartment (Fig. 4A). Collectively, the increased expression of phospho-γH2A.X expression in colonic crypt epithelial cells of chronic SIV-infected macaques provides compelling evidence of oxidative stress-mediated DDR secondary to inflammation and represents a potential mechanism driving miR-34a upregulation (34, 35). Nevertheless, the positive Ki67 staining suggests that unlike inflammatory bowel disease (IBD) patients (36), colonic crypt epithelial cells in chronic SIV-infected macaques have not been subjected to oxidative stress that is long enough (persistent DDR) to induce changes consistent with cellular senescence.

To better understand the biological significance of miR-34a upregulation, we focused on SIRT1, a predicted and validated target of miR-34a (34). SIRT1 is a NAD⁺-dependent class III histone/protein deacetylase known to deacetylate transcription factors, enzymes, DNA binding histone proteins, and such, and in doing so regulates critical cellular processes such as inflammation, DNA repair, cell senescence/aging, cellular metabolism, and stress response (35). Furthermore, decreased expression of SIRT1 has been detected in several chronic inflammatory conditions such as chronic obstructive pulmonary disease, atherosclerosis, obesity, and dextran sulfate sodium (DSS)–induced colitis (35, 37). Most importantly, SIRT1 inhibits cellular senescence/aging and also functions as a negative regulator of inflammation and immune activation by deacetylating the p65 subunit of NF-κB at lysine 310 (35, 37).

The 3′ UTR of the rhesus macaque SIRT1 mRNA has two predicted miR-34a binding sites (38) (Table II). Although human SIRT1 mRNA has been validated as a direct target of miR-34a (34), it is not known whether mml-miR-34a can bind the 3′ UTR of rhesus macaque SIRT1 mRNA and repress its expression. Because mRNA destabilization has been proposed as a predominant mechanism by which miRNAs repress gene expression (39), we next quantified SIRT1 mRNA separately in colonic epithelium and LPLs. Interestingly, SIRT1 mRNA expression significantly decreased by ∼2.4- and ∼1.8-fold in colonic epithelium (p = 0.0118) and LPLs (p = 0.015), respectively (Fig. 4B, 4C). Furthermore, cotransfection of SIRT1-WT UTR vector with miR-34a mimic resulted in significant reduction (∼30%) in firefly/renilla ratios (p < 0.05)). In contrast, cotransfection of SIRT1-Del UTR vector with miR-34a restored the firefly/renilla ratios to the level observed with unmanipulated pmirGLO vector (Fig. 4D). Interestingly, cotransfection of SIRT1-WT UTR vector with a negative control mimic did not reduce firefly/renilla ratios compared with wells transfected with miR-34a mimic (p < 0.05) (Fig. 4D). These results clearly show that mml-miR-34a can physically interact with the 3′ UTR of rhesus macaque SIRT1 mRNA and regulate its expression.

We next performed IF on formalin-fixed paraffin-embedded colon sections using anti-SIRT1–specific Abs to check whether changes in mRNA expression were paralleled at the protein level. After evaluating five anti-human SIRT1-specific Abs, one cross-reacted with the rhesus macaque for IF studies. As shown in Fig. 5, SIRT1 protein expression was detected in the colonic epithelium and LPLs of chronic SIV-infected (Fig. 5A, 5B) and control macaques (Fig. 5D, 5E). Furthermore, SIRT1 protein localized to both cytoplasm and nuclei of epithelial and LPLs (Fig. 5C). Consistent with decreased SIRT1 mRNA expression, image analysis confirmed significantly (p < 0.05) reduced SIRT1 protein expression in the epithelial and LPL compartments of both chronic SIV-infected compared with uninfected control macaques (Fig. 5F). Based on the results of the luciferase assay (Fig. 3C), decreased SIRT1 mRNA and protein expression in the colonic epithelium and LPL may be directly attributed to miR-34a binding and inhibition of target gene expression.

Because SIRT1 is known to deacetylate the p65 subunit of NF-κB on lysine 310 (35), we hypothesized that miR-34a–mediated downregulation of SIRT1 will result in constitutive expression of the transcriptionally active form of p65 (acetylated on lysine 310) in the colon. In agreement with our hypothesis, IF studies revealed intense expression of acetylated p65 exclusively in the LPL compartment of chronic SIV-infected macaques (Fig. 6A, 6B). The inset image in Fig. 6A and 6B shows a magnified view of acetylated p65⁺ cells in the colonic lamina propria. This Ab detects p65 only when acetylated on lysine 310 and has been used previously for immunohistochemistry studies (40). Additionally, the Ab recognizing the nonacetylated peptide has been removed by chromatography using nonacetylated peptide sequences. In contrast to chronic SIV-infected macaques, occasional acetylated p65 (lysine 310)⁺ cells were detected in the colonic lamina propria of the control macaque (Fig. 6C). We next immunophenotyped acetyl p65⁺ cells using markers specific to macrophages (CD68, CD163) and T lymphocytes (CD3). Surprisingly, none of the markers colocalized with acetylated p65⁺, suggesting that neither macrophages nor T lymphocytes were contributing to the elevated acetylated p65 expression (Fig. 6D, 6F). Finally, using anti-rhesus IgA and IgG Abs we successfully identified Ab-producing plasma cells as the primary cell type contributing to enhanced acetylated p65 (lysine 310) expression (Fig. 7A, 7B). IgA⁺/acetylated p65⁺ and IgG⁺/acetylated p65⁺ cells were rare in the control macaque (Fig. 7C, 7D). Furthermore, image analysis revealed significantly increased numbers of IgA⁺/acetylated p65⁺ and IgG⁺/acetylated p65⁺ cells in the colonic lamina propria of chronic SIV-infected macaques compared with uninfected control macaques (Fig. 8A–D). Image analysis provided additional evidence that plasma cells in the colonic lamina propria of chronic SIV-infected macaques carry a higher percentage of IgG⁺/acetylated p65⁺ cells compared with control macaques (Fig. 8B, 8D). In contrast, there were very few IgG⁺/acetylated p65⁺ cells in the three control macaques (Fig. 8B, 8D). Similar to the colon, IgA⁺/acetylated p65⁺ cells were also detected in the jejunum but only in macaques that showed histological signs of enteritis (HD08) (Table I) (data not shown). IgA⁺/acetylated p65⁺ cells were extremely rare in the jejunum of a chronically SIV-infected macaque (GA19) that did not show histological evidence of enteritis (Table I) (data not shown). Although we could not detect acetylated p65 in T cells and macrophages, it is possible that the transcriptionally active protein could be present at levels below the detection limit of IF.

Because acetylated p65 expression was detected only in plasma cells, we next determined whether these cells also expressed miR-34a to directly link expression of constitutive acetylated p65 to elevated miR-34a expression. As shown in Fig. 9A and 9B, after 7 min incubation with permanent red substrate (Dako), miR-34a expression was strongly detected in colonic epithelium and LPLs of both chronic SIV-infected macaques. In contrast, miR-34a staining was very weak in uninfected control macaques (Fig. 9D, 9E) examined in the same experiment. Furthermore, the section stained with the scrambled probe did not yield any signal (Fig. 9C). All images were captured using a Leica confocal microscope employing the same laser strength (red and green). These findings are strongly supported by quantitative image analysis showing significantly elevated miR-34a signal intensity in colonic epithelial and LPL compartments of chronic SIV-infected macaques compared with controls (Fig. 9F). Double labeling with miR-34a and acetylated p65 was attempted without success, as not all Abs work well on tissue sections previously subjected to in situ hybridization. The in situ hybridization results were in perfect agreement with the qRT-PCR data showing elevated miR-34a expression in colonic epithelium and LPLs of chronic SIV-infected macaques (Fig. 1A, 1B). Collectively, the intense miR-34a expression in IgA⁺ plasma cells (Fig. 9A, 9B) suggests that enhanced miR-34a expression may promote constitutive activation of acetylated p65 (lysine 310), possibly through SIRT1 downregulation.

GI inflammation, a major manifestation of HIV/SIV infection that is necessary for pathogen clearance, can, however, result in structural and functional damage when not properly regulated that can promote microbial translocation and systemic immune activation (1–4). In addition to proinflammatory cytokines and transcription factors, recent studies have proposed critical roles for miRNAs in controlling and managing certain aspects of the inflammatory process (41). Moreover, the recent finding that miRNAs primarily regulate gene expression by destabilizing and degrading mRNA targets (39) suggests that the previously reported alterations in intestinal gene expression profiles by us (13, 14) and others (11, 12) could be directly attributed to the regulatory functions of miRNAs. Therefore, the next logical step is to profile, identify, and characterize the specific miRNAs that exert these regulatory effects. To this end, we have reported miRNA dysregulation in colon during acute SIV infection (21). More specifically, we identified miR-190b to be significantly upregulated in both colon and jejunum at all stages of SIV infection primarily in response to viral replication. In the present study, we again focused on the colon, as it was found to be more severely impacted than the jejunum/small intestine during chronic SIV infection (Table I). Furthermore, and more importantly, the increased bacterial concentration in the colon (10¹² bacterial organisms/ml contents) (42) makes it a major source of bacteria/microbes contributing to microbial translocation during chronic HIV/SIV infection. In the present study, we report the identification of miR-34a, a miRNA previously linked to chronic inflammation, cell cycle arrest, and senescence (22–24, 27, 28), to be significantly dysregulated in both colonic epithelial and LPL compartments. Furthermore, our findings suggest that the miR-34a–SIRT1–acetylated p65 axis could potentially play a critical role in promoting immune cell activation, in particular Ab-producing plasma cells.

Among the 10 differentially expressed miRNAs, we further characterized miR-34a in the epithelium and LPL compartments to better understand its role in regulating GI structure and function. Interestingly, miR-34a showed significant upregulation in both colonic epithelium and LPL compartments of chronic SIV-infected macaques. Furthermore, miR-34a is upregulated in response to DNA damage that can be a result of oxidative stress induced by chronic inflammation leading to activation of the p53-mediated DDR (30, 31). Therefore, we next probed the expression of the DDR sensor protein γH2A.X and detected elevated nuclear expression of phospho-γH2A.X in the colonic crypt epithelial and LPL compartments. These findings demonstrate activation of DDR signaling and represent a potential mechanism triggering and driving miR-34a upregulation. Additionally, the viral protein (HIV/SIV Tat) was also shown to induce the expression of miR-34a (43, 44). Therefore, it is likely that Tat protein (direct effects of viral replication) and oxidative stress-induced DDR (indirect effects of viral replication) can separately or collectively upregulate miR-34a expression in the colonic epithelium and LPL compartments. The altered expression of miR-34a suggests that it could play a critical role in regulating epithelial cell cycle, survival, lamina propria immune cell activation, and inflammatory responses, all of which are dysregulated during chronic HIV/SIV infection.

Because miR-34a expression was markedly elevated in both epithelial and LPL compartments, we next investigated the functional relevance of its upregulation by examining the expression of situin1 (SIRT1), a validated target of miR-34a. SIRT1 is a NAD⁺-dependent class III histone deacetylase that regulates a wide array of cellular processes such as glucose homeostasis, energy metabolism, cell senescence/aging, apoptosis, dsDNA break repair, and immune/inflammatory responses (34, 35, 37). SIRT1 exerts these protective effects by deacetylating transcription factors such as p53, FOXO3, NF-κB, DNA repair enzymes, and histone proteins (35). Two distinct classes of molecules, namely, RNA binding proteins and noncoding RNAs, have been shown to posttranscriptionally regulate SIRT1 expression (34, 45). Additionally, the viral protein Tat was shown to activate the expression of miR-34a and miR-217, which then directly targeted and downregulated SIRT1 expression (43, 44). This suggests that viral replication in the intestine could contribute significantly to SIRT1 downregulation through enhancing miR-34a and miR-217 expression. More recently, miR-142-5p was also demonstrated to target and downregulate SIRT1 expression in macrophages/microglia from rhesus macaques with SIV-induced encephalitis (46). Along similar lines, SIRT1 protein expression decreased in colon of mice with DSS-induced colitis but was restored following treatment with resveratrol, a SIRT1–stimulating polyphenol (37). Consistent with findings in DSS-induced colitis mice, SIRT1 mRNA and protein expression decreased significantly in both colonic epithelium and LPL compartments of chronically SIV-infected macaques. However, we had to test five anti-human SIRT1 Abs before we found one that cross-reacted with the rhesus macaque for IF studies. This is a challenge facing the nonhuman primate research community, as almost all Abs are made primarily for use in the human or mouse. The results of the luciferase reporter assay together with the significantly elevated miR-34a expression in both colonic epithelium and LPLs provide a potential mechanism wherein miR-34a can directly bind to two different sites on the rhesus macaque SIRT1 3′ UTR and inhibit its expression in both compartments.

Enhanced miR-34a expression has been previously associated with cellular senescence in renal (23), endothelial (24), and colonic crypt epithelial cells (28), and thus we were intrigued to determine whether similar senescence-associated changes occurred in the colonic epithelium during chronic SIV infection. We focused on the colonic epithelium, as this compartment is highly proliferative and plays a central role in nutrient absorption and barrier function. Accordingly, senescence-induced changes can reduce the proliferation capacity and interfere with restoration of the epithelial barrier. Although the accumulation of phospho-γH2A.X foci in response to persistent DDR is considered a major trigger for cellular senescence (33), the strong expression of Ki67, a cell proliferation marker by crypt epithelial cells, suggested that these cells did not progress to irreversible cell cycle arrest (senescence) and retained the capacity to proliferate. This may be because SIV-infected rhesus macaques are necropsied when they lose >20% of body weight, which supports the argument that the duration of exposure to oxidative stress secondary to inflammation is not long enough to induce cellular senescence, unlike in IBD patients, who live with the disease for decades (36). Overall, our findings suggest that inflammation-induced DDR in the colonic epithelium as evidenced by enhanced γH2A.X expression results in elevated prosenescent miR-34a and significantly reduced antisenescent SIRT1 mRNA and protein expression. Nevertheless, low-level viral replication–induced chronic intestinal inflammation/immune activation reported to occur in patients receiving long-term antiretroviral therapy (47) can generate a persistently maintained DDR state similar to IBD patients (36), eventually causing cellular senescence. Future studies are needed to determine whether senescence-associated changes occur in long term antiretroviral therapy–treated individuals, as they can cause accelerated aging of the intestine, leading to epithelial barrier dysfunction, microbial translocation, and AIDS progression.

In the LPL compartment, miR-34a expression increased and SIRT1 expression decreased significantly. Because the LPL compartment is enriched for immune cells and primarily involved in mediating immune/inflammatory responses, we focused on the RelA/p65 subunit of NF-κB, a transcription factor central to immune and inflammatory responses (48). Most notably, posttranslational modifications such as phosphorylation, acetylation, and SUMOylation are necessary for modulating NF-κB–mediated gene expression (48). Among the different modifications, acetylation of the p65 subunit catalyzed by p300/CBP and p300/CBP-associated factor can occur on seven distinct lysine residues (lysine 122, 123, 218, 221, 310, 314, and 315) that can either activate or inhibit NF-κB signaling depending on which lysine residue is acetylated (48). Among these, acetylation of lysine 310 is of significant interest because it is crucial for full activation of NF-κB transcriptional potential (48). Equally important is the finding that SIRT1 (48) and HDAC3 (48) are two major enzymes that can deacetylate lysine 310 on p65, thereby inhibiting its transcriptional activity. Additionally, Tat-mediated miR-34a upregulation inhibited SIRT1 expression in TZM-bl cells, resulting in enhanced acetylated p65 (lysine 310) expression (43). Furthermore, Tat has also been shown to directly bind to the deacetylase domain of SIRT1 and suppress its ability to deacetylate lysine 310 on the p65 subunit of NF-κB, resulting in enhanced acetylated p65 activity (49). Therefore, we probed the downstream effects of miR-34a–mediated SIRT1 downregulation in the LPL compartment by determining the expression of acetylated p65 (lysine 310). Using IF studies, acetylated p65 (lysine 310) expression was found to be markedly increased solely in the LPL compartment of chronic SIV-infected macaques. Interestingly, acetylated p65⁺ cells were negative for CD68, CD163, and CD3. Subsequently, using anti–rhesus macaque IgG and IgA Abs, we successfully localized acetylated p65 (lysine 310) to plasma cells. Furthermore, markedly increased numbers of IgA⁺/acetyl p65⁺ and IgG⁺/acetyl p65⁺ cells were detected in the colons of chronic SIV-infected macaques. These findings were not surprising, as DDR signaling evidenced by strong phospho-γH2A.X in other lamina propria cells such as lymphocytes and macrophages also activates the JAK-STAT3 pathway for efficient dsDNA repair (50). We previously demonstrated constitutive activation of the phosphorylated form of STAT3 (tyrosine 705) in CD3⁺ lymphocytes and CD68⁺ macrophages (10). These results, together with our previously published findings (10), suggest that whereas plasma cells activate p65 through acetylation of lysine 310, lymphocytes and macrophages can constitutively activate STAT3 by phosphorylation of tyrosine 705 in response to oxidative stress–induced DDR. Moreover, using in situ hybridization we successfully localized miR-34a expression to the colonic epithelium and plasma cells in the lamina propria. These interesting findings are consistent with a previous study that identified miR-34a to be significantly upregulated following B cell activation in vitro with 1 μM CpG (51). More recently, activated B cells isolated from multiple sclerosis patients were found to express significantly high miR-132 and low SIRT1 protein levels (52). Similar to miR-34a, miR-132 has two predicted binding sites on the 3′ UTR of SIRT1 mRNA. Furthermore, SIRT1 downregulation resulted in increased production of lymphotoxin and TNF-α. Interestingly, in vitro treatment of B cells from multiple sclerosis patients with resveratrol, a SIRT1 activator, increased SIRT1 protein and normalized lymphotoxin and TNF-α production (52). Although our results do not demonstrate causal relationships, they may partially explain the intestinal B cell hyperactivity reported in HIV-infected individuals characterized by the presence of increased numbers of IgA- and IgG-producing plasma cells (53). Overall, our findings demonstrate an interesting association between miR-34a upregulation and elevated acetylated p65 expression, and they provide a potential mechanism that may cause hyperactivation and possible dysfunction of the intestinal B cell system.

Plasma cells that increase substantially in chronically inflamed tissues (54) can activate the NF-κB pathway in response to proinflammatory factors present in their immediate environment, leading to the transcription of NF-κB–responsive genes (55). In this context, the presence of NF-κB–binding/response elements on the miR-34a promoter may facilitate NF-κB proteins to bind and drive miR-34a expression (56). miR-34a may then directly interact and bind to the 3′ UTR of SIRT1 and downregulate its expression. Because SIRT1 is required to deacetylate lysine 310 residue on p65, its reduced expression can maintain p65 in its transcriptionally active form and in this way potentially generate a feed-forward mechanism involving NF-κB, miR-34a, and SIRT1, leading to constitutive activation and transcription of NF-κB–responsive genes. Given that the p65 subunit regulates transcription (57) and assembly of Ig H and L chains (58, 59) and is indispensable for IgA and IgG production (60), its constitutive activation may partly explain the IgA overproduction reported in the intestine of chronic HIV-infected individuals (53) and, possibly, prolong the survival of plasma cells (long-lived plasma cells) in niches such as the inflamed colonic lamina propria (53). From the pathogenesis standpoint, IgG-producing long-lived plasma cells in the colonic lamina propria of IBD patients were found to produce matrix metalloproteinase-3 (54), an NF-κB–responsive matrix metalloproteinase shown to cause tissue destruction (61). More recently, IgG plasma cells in the colonic mucosa of IBD patients were shown to induce TNF-α and IL-1β production by CD14⁺ macrophages via IgG–IC–FcγR signaling, thereby aggravating colonic inflammation (62). Furthermore, because IgG has strong complement-fixing properties, its overproduction may exacerbate intestinal inflammation and structural damage and contribute to disease progression (53). Furthermore, plasma cells have also been demonstrated to produce vascular endothelial growth factor, IL-6, and TNF-α in various chronic inflammatory conditions (54). Collectively, these findings suggest that intestinal plasma cells may have broader functions in modulating inflammatory responses that can potentially contribute to epithelial barrier disruption, and the miR-34a–SIRT1-mediated constitutive NF-κB activation may considerably assist in these functions.

In summary, our results demonstrate marked miRNA dysregulation in the intestine during chronic SIV infection. Whereas miR-34a was significantly upregulated in both colonic epithelium and LPL compartments, SIRT1 mRNA and protein expression significantly decreased in both compartments. Elevated miR-34a and reduced SIRT1 expression were associated with constitutive expression of acetylated p65 in colonic IgA⁺ and IgG⁺ plasma cells. The findings suggest a possible role for the miR-34a–SIRT1–p65 axis in causing hyperactivity and dysfunction of the mucosal B cell system. The recent demonstration that miR-34a expression is elevated following in vitro B cell activation using CpG molecules (51) and its ability to regulate B cell development/differentiation by modulating Foxp1 expression (63) strengthens the possibility. Apart from SIRT1, miR-34a has also been shown to regulate the expression of several cell cycle regulators (NMYC, CCND1, CDK4, CDK6, MET) (22), and future studies involving cell sorting protocols and high-throughput proteomics are needed to identify and validate additional targets (∼640 predicted) to understand the full functional significance of its upregulation in the colon during HIV/SIV infection. Additionally, in vivo miR-34a modulation studies are needed to determine whether a causal relationship exists between miR-34a upegulation and constitutive acetylated p65 expression. Nevertheless, importantly note that direct in vivo miRNA inhibition using antagomirs is currently not feasible owing to lack of an efficient in vivo delivery system that can selectively target specific segments of the intestine. However, the recent demonstration that a 12 mo supplementation of resveratrol, an SIRT1-activating anti-inflammatory polyphenol present in grape extracts, modulated miR-34a expression in PBMCs of type 2 diabetes and hypertensive patients (64) is very encouraging, as it highlights an alternative approach preferably in combination with antiretroviral drugs to modulate miR-34a expression and chronic persistent inflammation.

We thank Ronald S. Veazey, Maurice Duplantis, Yun Te Lin, Faith R. Schiro, Cecily C. Midkiff, Christopher Monjure, and Coty Tatum for technical assistance.

The authors have no financial conflicts of interest.