All animal studies included herein were approved by the Oregon National Primate Research Center (ONPRC) Institutional Animal Care and Use Committee (IACUC). Throughout the duration of the study, every effort was made to minimize animal discomfort per the regulations stipulated by the USDA and OLAW. For these studies, animals are trained to voluntarily present their leg for blood draws, allowing samples to be obtained within 3–5 minutes. The amount of blood taken from each animal is highly regulated, as each draw is reported on the electronic medical records. The monkeys are on a standard operating procedure for voluntary alcohol self-administration with water also available concurrently with alcohol (4% w/v, diluted in water, a concentration that is not avoided by rhesus monkeys). Moreover, they are provided food in 3 meals/day and kept in a positive weight gain throughout the protocol. They are provided 1–2 hrs a day of direct social contact and 24 hrs/day with visual and auditory access to all monkeys in the housing room. The daily drinking sessions occur within the housing cage and begin at the same time for all monkeys on protocol in the housing room.

Blood samples were collected from 12 rhesus macaques (4 females, 8 males) before and after 6mo of open access to alcohol through the Monkey Alcohol Tissue Research Resource (MATRR; www.matrr.com; cohorts 7a, 16, and 18) (49). In this model, rhesus macaques (4–6 years old) are trained to use operant drinking panels and to present their legs in their home cages for blood collection. Animals go through an induction phase of 4 mo of schedule-induced polydipsia protocol with 16hr of drinking sessions followed by an “open access” phase of 6 mo with 22h/day of voluntary drinking session (4% ethanol w/v solution or water choice) as described in detail (50) and the drinking category of ethanol consumption is determined as previously described (51). Animals are individually housed in quadrant cages with constant room temperature (20-22°C), humidity (65%) and an 11hr light cycle. Cohort demographics with drinking categorization are outlined in more detail in Table 1 .

PBMC were stained with: CD3-BV510 (Clone OKT3), CD20-BV510 (Biolegend: Clone 2H7), CD14-AF700 (Clone M5E2, Biolegend), CD16-PB (Clone 3G8, Biolegend), HLADR-APC-Cy7 (Clone L243, Biolegend), CX3CR1-PE (Clone 2A9-1, Biolegend), CD86-BV605 (Clone IT2.2, Biolegend), CD163-PerCP-Cy5.5 (Clone GHI/61, Biolegend), CD169-Pe-Dazzle 594 (Clone. 7-239, Biolegend), TLR4-APC (Clone HTA125, Biolegend), and True-Stain Monocyte Block (Biolegend) for 30 min at 4°C. Samples were acquired on the Attune NxT Cytometer (ThermoFisher Scientific) and were analyzed using FlowJo (BD).

PBMCs were stimulated overnight (~16 hours) with a bacterial cocktail of 1ug/mL Pam3CSK4 (TLR1/2 ligand, InvivoGen), 0.5ug/mL of FSL-1 (TLR2/6 ligand, InvivoGen), and 0.5ug/mL of LPS (TLR4 ligand, InvivoGen) in the presence of Brefeldin-A (Biolegend). The following day, the cells were surface stained with CD3-FITC (Clone Sp34, Biolegend), CD20-FITC (Clone 2H7, Biolegend), CD14-AF700 (Clone M5E2, Biolegend), CD16-PB (Clone 3G8, Biolegend), HLADR-APC-Cy7 (Clone L243, Biolegend), and True-Stain Monocyte Block (Biolegend). They were then fixed, permeabilized, and antibodies against TNFα-APC (Clone Mab11, Biolegend) and IL-6-PE (Clone MQ2-6A3, BD) were added. Data were acquired on the Attune NxT Cytometer (ThermoFisher Scientific) and analysis was completed using FlowJo (BD).

5x10⁵ PBMC were incubated with pHrodo Red E. Coli BioParticles (Thermo Fisher Scientific) for 2 hours. Cells were then stained with Ghost Dye Violet 510 (Tonbo) for 30 minutes to determine viability and then surfaced stained with CD3-FITC (Clone Sp34, BD), CD20-FITC (Clone 2H7, Biolegend), CD14-AF700 (Clone M5E2, Biolegend), CD16-PB (Clone 3G8, Biolegend), HLADR-APC-Cy7 (Clone L243, Biolegend), and True-Stain Monocyte Block (Biolegend), acquired on the Attune NxT Cytometer (ThermoFisher Scientific) and analyzed using FlowJo (BD).

Monocytes were isolated using the EasySep™ APC Positive Selection Kit II (Stemcell) following staining of PBMCs with CD14-APC (Clone M5E2, Biolegend). CD14+ cells were pooled based on sampling timepoint and then incubated with or without LPS at 37°C with 5% CO2 for 4 hours. Cells were then loaded into a Chromium Controller (10x Genomics) at a concentration of 1,600 cells/uL; with the final target recovery of 30,000 cells. Library preparation was completed using the v3.1 Chromium Single Cell 3′ Kit (10x Genomics) following the manufacturer’s instructions. Libraries were sequenced at Novogene on the NovaSeq X Plus with a target of 20,000 reads per cell.

Samples were aligned to the rhesus macaque genome (Mmul_10) using the Cell Ranger Software (version 7.2) and downstream analysis was carried out in R 4.1.1 using Seurat (version 5.1). Aligned reads underwent initial quality control analysis by removing droplets with ambient RNA (cells with <400 detected genes), doublets (cells with >4000 detected genes) and dying cells (>5% total mitochondrial gene expression). All samples were integrated using Seurat and then normalized using NormalizeData.

The FindNeighbors and FindClusters (resolution = 0.3) function in Seurat were used to cluster the cells using the first 10 principal components and visualized via uniform manifold approximation and projection (UMAP). Clusters were identified using canonical markers from the FindAllMarkers functions ( Supplementary Table 1 ) and contaminating B and T cells were removed for downstream analysis.

Differentially expressed genes were determined via DESeq2 under default settings in Seurat and enriched using Metascape. Only statistically significant genes (average log(fold-change) cutoff >0.58 or <-0.58; adjusted p-value ≤ 0.05) were included in downstream analysis ( Supplementary Table 2 ).

Gene scoring analysis was conducted by Seurat’s AddModuleScore function, leveraging pathways and gene signatures lists from KEGG (www.genome.jp/kegg/pathway.html) or Ensembl BioMart. Specific gene expression was determined using the AggregateExpression function. For both analysis, cells from their respective sampling timepoint/stimulation condition were randomly assigned to three groups, and the average module or expression of the group was plotted.

Slingshot was used to reconstruct monocyte trajectories across pseudotime (52) after the UMAP underwent dimensional reduction in Seurat. Cluster Mono0 was set as the root state for all downstream analysis as it is the most abundant cluster in the baseline/non-stimulated sample. DEG between groups along their respective lineages were determined using the condiments workflow adopted from (53).

Monocytes were isolated from freshly thawed PBMC using EasySep™ APC Positive Selection Kit II (Stemcell). 1.6x10⁵-2x10⁵ monocytes were used for chromatin immunoprecipitation sequencing via the CUT&Tag approach. Sample preparation for H3k4me3 (Cat # 39016, Active Motif) was constructed using Active Motif’s CUT&Tag-IT™ Assay Kit (Cat # 53160, Active Motif) following manufacturer’s instructions. Libraries were sequenced by Novogene on the NovaSeq X Plus at a depth of 10 million reads/sample.

Paired-end reads were quality-checked using FASTQC, trimmed using Trimmomatic, and aligned to the Macaca mulatta genome (Mmul10) using Bowtie2. Peak calling was carried out using MACS2. The peaks were merged using bedtools across biological replicates within timepoints (baseline and 6mo) to enhance statistical power for differential peak analysis. Due to the lack of available annotation databases for rhesus macaques, a liftover of the Rhesus macaque genome (mmul10) to the human genome (hg38) was completed using LiftOver. EpiCompare was used for the comparisons between genome-wide histone modification profiles. All peaks overlapped with hg38 blacklisted regions were removed before downstream analysis with the ENCODE identifier ENCFF356LFX. The peak overlaps heatmap and TSS plots were generated using the ENCODE identifier ENCFF133YDZ as the reference. Differential Peak analysis was completed using the HOMER’s getDifferentialPeaks function with a cutoff of Poisson-based p-value < 0.0001. All peaks were annotated using ChIPseeker (54) in R 4.1.1. Differentially abundant peaks at baseline and 6mo ( Supplementary Table 3 ) were enriched using Metascape and Gene Set Enrichment Analysis (GSEA).

Monocytes were isolated from freshly thawed PBMC using positive selection magnetic activated cell sorting (Miltenyi Biotec). Histones were extracted from isolated monocytes using Abcam’s Histone Extraction kit (Abcam) per manufacturer’s instructions and final quantification of protein concentration was assessed using the Qubit Protein Assay Kit. The level of total H3 histone and respective histone modifications were measured with Abcam’s Histone H3 Modification Multiplex Assay Kit (colorimetric) (Abcam). Only select modifications were tested (H3K4me2, H3K4me3, H3k9me1, H3k9me2, H3K9me3, H3K27me1, H3K27me2, H3K27me3), utilizing 500ng of protein from each sample per reaction. Data were acquired using a SpectraMax iD3 multi-mode plate reader and final calculations using OD values were completed as outlined per the manufacturer.

All statistical comparisons in this manuscript were completed in R 4.3.1 unless otherwise stated in the methods. Differences between timepoints (baseline vs. 6mo of CAC) or response to challenge (non-stimulated vs. stimulated) were determined using a linear mixed-model approach. All model assumptions, including normality, were assessed using a combination of visual plots, including residual plots and histograms. Differences in cell cluster frequencies in the scRNA-seq data set were determined using a one-way ANOVA with unmatched samples. The Šídák test was used to correct for multiple comparisons. Values for module scores were tested using an unpaired t-test with Welch’s correction. All bar plots depict the mean ± SEM. A p-value of <0.05 was considered significant.

To understand how short-term durations of heavy alcohol use affect the profile of circulating monocytes, PBMC obtained before and after 6mo of daily, heavy ethanol consumption were phenotyped using key canonical surface markers ( Figures 1A, B ). After 6mo of drinking, there was a modest decrease in the frequency of circulating monocytes after alcohol use compared to baseline ( Figure 1C ) that was accompanied by a significant increase in the proportion of nonclassical monocytes. ( Figure 1D ). While the frequency of CD163+ monocytes, representing a hypo-inflammatory state, was not significantly changed after 6mo of daily alcohol use, the expression level of CD163 was significantly enhanced by alcohol use, indicated by increased mean fluorescent intensity (MFI) ( Figures 1E, F ). Additionally, no differences were noted in the frequency of cells expressing CD169, CX3CR1, CD86, or TLR4 or MFI of expression (data not shown).

Next, we challenged the cells with a cocktail of bacterial ligands (Pam3CSK4, FSL-1, and LPS) to emulate the complexity of bacterial pathogens and measured production of IL-6 and TNFα ( Figure 2A ). Following stimulation, a lower percentage of monocytes produced TNFα and IL-6 after 6mo of heavy drinking compared to baseline ( Figure 2B ). Moreover, levels of IL-6, but not TNFα, produced by monocytes were also reduced after 6mo of drinking compared to baseline ( Figures 2C, D ). This alteration in immune mediator production stemmed from a significant reduction TNFα and IL6 production by CD14^high monocytes, as there were no changes observed in the CD14^low subset ( Figures 2E, F , Supplementary Figure 1A ). Additionally, the overall ability of monocytes to phagocytose E. coli conjugated bioparticles was decreased after short-term alcohol consumption ( Figures 2G, H ). Both CD14^high and CD14^low subsets contributed to this reduced functional output ( Supplementary Figures 1A, B ). Taken together, these data suggest that 6mo of heavy drinking results in increased generation of hypo-inflammatory monocytes, reduced production of inflammatory mediators, and attenuated phagocytic potential.

Next, we sought to understand the molecular mechanisms driving phenotypic and functional changes after 6mo of heavy drinking by performing single-cell RNA sequencing (scRNAseq) on isolated monocytes before and after 6mo of ethanol consumption in the presence and absence of LPS stimulation, a well-studied and effective activator commonly used to assess monocyte function following alcohol exposure. Uniform manifold approximation and projection (UMAP) revealed six distinct clusters with varying frequencies across timepoint and stimulation conditions ( Figures 3A, B ). All clusters expressed HLADR, with minimal or no FCGR3 (CD16) expression, suggesting they are predominantly classical monocyte clusters ( Figure 3C ). Interestingly, we were unable to detect a cluster with high expression of FCGR3 (CD16) in this data set. Cluster Mono 1 was characterized by high expression of interferon-stimulated genes (ISG; ISG15, ISG20, IFI16), whereas Mono 2 expressed high levels of chemokines (CCL2, CCL3, CCL4L1, NFKB2). Mono 3 marker genes included predominantly alarmins (S100A4, S100A6, S100A9, S100A10), while marker genes of Mono 4 play critical roles in cell-cycle (JARID2, PPARG, PRKCA, SMARCC1, VAV3). Marker genes of Mono 5 are involved in the resolution of inflammation (ERBB4, ROBO1, NRG1, PTRRT). Finally, cells in Mono 6 expressed high levels of genes that are involved in key signaling pathways/cell activation (PIK3CB, EREG, HIPK2, FOSL1) ( Figure 3C ). Functional enrichments of marker genes specific to each cluster were used to gain insight into the functions of the clusters. In line with the high expression of ISG, genes defining Mono 1 enriched to gene ontology (GO) terms associated with anti-viral immune responses ( Figure 3D ). Marker genes of Mono 2, Mono 3, and Mono 6 mapped to pathways involved in anti-bacterial response and inflammatory signaling ( Figure 3D ). Marker genes of Mono 4 mapped to cell cycling while those of Mono 5 enriched to hypo-inflammatory feedback mechanisms ( Figure 3D ).

Non-stimulated monocytes at baseline (BL) were primarily composed of Mono1 and Mono 3 clusters ( Figures 3B, E ). However, 6mo of heavy drinking resulted in a significant decrease in Mono 1 and 3 while the frequency of Mono 4 (proliferating) and 5 (regulatory phenotype) increased ( Figures 3B, E ). To uncover the impact of 6mo of heavy drinking on monocyte transcriptome, we collapsed all clusters within their respective stimulation condition and sampling timepoint to ensure robust differential gene expression analysis ( Supplementary Table 2 ). First, to understand how 6mo of daily heavy drinking rewires the transcriptional landscape of monocytes, we compared the unstimulated transcriptional landscape of monocytes at baseline and after 6mo of drinking (BL_NS vs. 6mo_NS). We observed 241 differentially expressed genes (DEG) ( Figure 4A ), suggesting 6mo of ethanol consumption is sufficient to induce heterogeneity within the monocyte transcriptome. There were 89 genes with higher expression in BL samples, enriching to gene ontology (GO) terms associated with pathogen response pathways (ex. TREM1, ISG20, CD14, GBP3) ( Figures 4A, B, D ). In contrast, after ethanol consumption, we see an increase in the expression of 152 genes involved with inflammation regulation/resolution (ex. JARID2, KDM2A) and homeostatic metabolic pathways such as oxidative phosphorylation (ex. ND2, ND3, ND4) ( Figures 4A, C, D ).

Following LPS stimulation of BL samples, the frequency of the ISG (Mono 1), neutrophil-like (Mono 3), and proliferating (Mono 4) clusters decreased with a concomitant increase in inflammatory monocytes (Mono 2) ( Figures 3B, E ). Similarly, LPS stimulation of samples after 6mo of alcohol consumption, led to the increase of inflammatory monocytes (Mono 2) as well as Mono 6 which expressed high levels of signaling/activation genes. When comparing the frequency of clusters at BL and 6-months following LPS (BL_LPS vs 6mo_LPS), we observed a higher frequency of regulatory (Mono 4), resolution (Mono 5), and signaling/activation (Mono 6) clusters, while that of the inflammatory (Mono 2) subset was lower in the ethanol group ( Figures 3B, E ).

We then assessed the post-LPS transcriptional differences before and after 6mo of drinking (BL_LPS vs. 6mo _LPS) ( Figure 4E ). The 115 genes upregulated in the BL samples enriched to key innate immune cell processes, including cytokine production (ex. IL1B, TYROBP) and migration (ex. ICAM1, CCL4L1) ( Figures 4E, F, H ). These changes in gene expression align with our functional data showing increased functional response to LPS at BL ( Figure 2 ). The 293 genes upregulated in the 6mo samples played a role in epigenetic regulation (ex. KDM6A, ARID2), resolution of inflammation (ex. TGFB1, ASH1L, ILRUN) ( Figures 4E, G, H ) as well as immune signaling (ex. LRRFIP2, MDM2, COX2) ( Figures 4E, G, H ) in agreement with the dampened IL-6 production. In line with the differential gene expression, scores of gene modules associated with inflammation, viral/bacterial, and phagosome responses were decreased at 6mo, regardless of stimulation condition ( Figure 4I ).

Next, we compared the transcriptional response to LPS by evaluating changes in gene expression after stimulation respective to their non-stimulated control (BL_NS vs. BL_LPS) (6mo_NS vs. 6mo_LPS) and then comparing genes that were differently expressed ( Supplementary Figure 2 ). We identified 236 DEG in response to LPS at BL and 170 DEG following LPS stimulation after 6mo of heavy drinking, with the majority of the DEG shared between the two sampling time points ( Supplementary Figure 2A ). Shared DEG enriched to gene ontology (GO) terms associated with immune responses (ex. IL1B, CXCL8, CCL3, TGFB1) and signaling (ex. IRAK2, C5AR1, PIK3AP1) ( Supplementary Figure 2B ). DEG unique to BL enriched to pathways involved in IL-6 production (ex. SELENOK, RABGEF1, IL6R) and regulation of gene transcription (ex. EP300) ( Supplementary Figures 2C, D ). On the other hand, DEG exclusive to 6mo time point are involved in negative regulation of inflammatory signaling (ex. NFBK1, LYN) ( Supplementary Figure 2E ). These data strongly suggest that 6mo of ethanol consumption induces a hypoinflammatory state in circulating monocytes.

We then performed a trajectory analysis to gain insight into which genes drive the differentiation of monocytes in response to alcohol consumption and LPS stimulation. The six monocyte clusters were ordered by pseudotime, originating in Mono 1. This cluster was used to root the analysis as Mono 1 had the highest cluster frequency in the baseline and non-stimulated sample. We identified two trajectories leading to Mono 2 or Mono 5 as their terminal points ( Figure 5A , Supplementary Figures 3A, B ). Lineage 1, ending in Mono 2, is driven by LPS stimulation, evidenced by the high contribution of LPS-stimulated cells at the end of the pseudotime ( Figure 5B ). DEG between non-stimulated and LPS-challenged groups across pseudotime indicates that this trajectory is driven by an increase in the expression of genes involved in antimicrobial responses including chemokines and components of the inflammasome (CCL3, CCL4L1, IL6, NLRP3) ( Figure 5C ). Lineage 2 culminates in Mono 5, a cluster with a high density of cells after 6mo of heavy alcohol use ( Figures 3B, E , 5D ). Differential gene expression for lineage 2 indicates a significant reduction in the expression of inflammatory genes (S100A8, S100A9, SOD2) across pseudotime ( Figure 5E ). These data show that 6mo of heavy drinking skew the transcriptome of monocytes towards a hypo-inflammatory state.

To further understand the molecular underpinnings of the transcriptional changes after 6mo of heavy drinking, we investigated the expression of post-translational histone modifications via a multiplex ELISA kit. As lysine modifications were predominantly reported to regulate monocyte function (55–57), we focused on lysine residues previously shown to be modulated with alcohol use (58, 59). The abundance of activation histone mark H3k4me3 and repressive histone marks H3k9me2 were decreased at 6mo, along with a modest decrease in the deposition of repressive mark H3k27me3 ( Figure 6A ).

To identify the genes impacted by the changes in H3k4me3 deposition, we performed the Cleavage Under Targets & Tagmentation (CUT&Tag) assay. As CUT&Tag is a newer technique used to evaluate DNA-protein interactions, we first benchmarked our data using EpiCompare against a published H3k4me3 profile of healthy monocytes from ENCODE ( Supplementary Figure 3 ). Overall, our peak recovery and feature distribution are comparable to the ENCODE reference and supporting literature (60). Although there is limited overlap between the peaks identified in the monocyte reference and the 6mo timepoint, a similar trend is also observed when comparing the baseline and 6mo timepoints. This suggests the discrepancy in peak calling is likely a result of ethanol consumption. When evaluating the H3k4me3 signature at the genomic level, we observed more differentially abundant genes at baseline compared to after 6mo of alcohol consumption ( Figure 6B ). Moreover, the deposition of H3k4me3 in the promoter region was decreased after alcohol use while the deposition in intergenic regions, exons, and introns increased ( Figure 6C ). We identified 5,870 peaks that were more abundant in the promoter region at baseline and 482 peaks that were abundant in the same region after 6mo of heavy drinking ( Supplementary Table 3 ). Peaks that were more abundant at baseline regulated genes important to innate immune responses and epigenetic remodeling such as IRAK4, TRAF6, MAPK1, and HDAC1 ( Figures 6D, E ). IRAK4 and TRAF6 are key components of the signaling cascade initiated by LPS binding to TLR4. Therefore, we compared their expression between baseline and after 6 months of ethanol access using the scRNA-seq data. A decrease in IRAK4 and TRAF6 transcript counts were detected after 6mo of ethanol consumption ( Figure 6F ). Moreover, TLR4 signaling module score was also reduced after 6 months of ethanol consumption indicating decreased expression of TLR4 signaling cascade components ( Figure 6G , Supplementary Table 4 ). In contrast, peaks that were more abundant after 6mo of heavy drinking overlapped genes important for key metabolic pathways such as ADP and NADH processes, which have been shown to orchestrate innate immune cells’ ability to regulate and resolve inflammation (61, 62) ( Figure 6H ).