Single cell transcriptomics reveals opioid usage evokes widespread suppression of antiviral gene program

Abstract

Chronic opioid usage not only causes addiction behavior through the central nervous system, but also modulates the peripheral immune system. However, how opioid impacts the immune system is still barely characterized systematically. In order to understand the immune modulatory effect of opioids in an unbiased way, here we perform single-cell RNA sequencing (scRNA-seq) of peripheral blood mononuclear cells from opioid-dependent individuals and controls to show that chronic opioid usage evokes widespread suppression of antiviral gene program in naive monocytes, as well as in multiple immune cell types upon stimulation with the pathogen component lipopolysaccharide. Furthermore, scRNA-seq reveals the same phenomenon after a short in vitro morphine treatment. These findings indicate that both acute and chronic opioid exposure may be harmful to our immune system by suppressing the antiviral gene program. Our results suggest that further characterization of the immune modulatory effects of opioid is critical to ensure the safety of clinical opioids.

Fig. 1. scRNA-seq revealed a widespread suppression of antiviral genes in opioid-dependent individuals.

a Experimental workflow schematic. Peripheral blood from opioid-dependent individuals and control individuals were collected, PBMCs were isolated, and microdroplet-based scRNA-seq was performed using Chromium Controller (10X Genomics). b t-SNE plot of naive (51,041) and LPS (100 ng/mL)-treated (21,873) PBMCs were clustered (cells were filtered based on >300 and <2000 genes per cell, <10,000 UMIs per cell; see Methods) and identified into immune populations (top) and visualized by control individuals and opioid-dependent individuals in each state (bottom): naive state control samples 1–7 (C1–C7), naive state opioid-dependent samples 1–7 (O1–O7), LPS-treated control samples 1–3 (C1–C3 (LPS)), LPS-treated opioid-dependent samples 1–3 (O1–O3 (LPS)). c Volcano plot showing fold change of gene expression (log2 scale) for downregulated (Control) and upregulated (Opioid) genes for opioid-dependent cells compared to non-dependent controls for naive state populations: monocytes and LPS-treated populations: NK cells, CD4+ T cells, and activated T cells (x-axis) with a significance of 0.05 (y-axis, −log10 scale). Significant genes shown with black dots, significant antiviral genes shown with green dots, and insignificant genes shown with gray dots. d Pathway enrichment analysis of significant differential genes across all naive and LPS-treated cell types evaluated by −log10(p value) as indicated by blue-purple scale (x-axis: cell type/state, y-axis: pathways). White represents an analysis which did not provide enrichment results for the specific pathway. Source data listing genes and expression values for c and d are provided in Source Data file. Similar findings were observed in repeat experiments using different patient samples.

Fig. 2. LPS-stimulated antiviral gene program were consistently suppressed in opioid-dependent individuals.

a Evaluation of the three innate immune response gene programs stimulated by LPS: antiviral, peaked inflammatory, and sustained inflammatory. b Left: Heatmap of scaled expression of core antiviral and inflammatory response genes observed in control sample cells (C1–C3) and opioid-dependent sample cells (O1–O3) in LPS-treated populations: CD4+ T cells, activated T cells, NK cells, and B cells. Color scale for heatmap indicates scaled gene expression. Yellow indicates positive scaled gene expression, purple indicates negative scaled gene expression, and while black represents zero scaled gene expression. Labeled key antiviral and inflammatory genes expression across LPS sample cells in the LPS-treated populations (Supplementary Figs. S6 and S7). Right: Average expression of all genes in a geneset (log expression) for each cell, grouped by control samples (C1–C3) and opioid-dependent samples (O1–O3) for LPS-treated populations: CD4+ T cells (C1–C3: 5211 cells and O1–O3: 6378 cells), activated T cells (C1–C3: 2456 cells and O1–O3: 1578 cells), NK cells (C1–C3: 268 cells and O1–O3: 351 cells), and B cells (C1–C3: 1527 cells and O1–O3: 747 cells). Inset box plots show the median, lower and upper hinges that correspond to the first quartile (25th percentile) and third quartile (75th percentile), and the upper and lower whiskers extend from the smallest and largest hinges at most 1.5 times the interquartile range. For CD4+ T cells, two-tailed T-test with comparison tests between control and opioid-dependent groups for each geneset: core antiviral (p < 2.22e−16), peaked inflammation (p < 2.22e–16), sustained inflammation (p < 2.22e–16). For activated T cells, two-tailed T-test with comparison tests between control and opioid-dependent groups for each geneset: core antiviral (p < 2.22e–16), peaked inflammation (p = 2.7e–08), sustained inflammation (p < 2.22e−16). For NK cells, two-tailed T-test with comparison tests between control and opioid-dependent groups for each geneset: core antiviral (p < 2.22e−16), peaked inflammation (p = 0.44), sustained inflammation (p = 0.00054). For B cells, two-tailed T-test with comparison tests between control and opioid-dependent groups for each geneset: core antiviral (p < 2.2e−16), peaked inflammation (p = 3.7e−6), sustained inflammation (p = 0.0033). ^nsp > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Similar findings were observed in repeat experiments using different patient samples.

Fig. 3. Short exposure to morphine resulted in suppression of antiviral genes upon LPS treatment.

a, b Evaluation of ISG15 mRNA expression after morphine treatment. PBMCs from a healthy, non-opioid-exposed individual were pretreated with morphine (0, 10, 100 μM) for 24 h (a) or 3 h (b) followed by LPS (100 ng/mL) stimulation for 3 h. Interferon pathway gene ISG15 expression was evaluated by RT-qPCR. Values displayed as fold increase (log10) to gene expression in LPS-treated cells over unstimulated cells, plus or minus one standard deviation. Error bars here represent technical variability; experiments were repeated at least three times with similar results. c Cell hashing scRNA-seq of healthy PBMCs pretreated with morphine for 24 h followed by LPS (100 ng/mL) treatment for 3 h. Left: Heatmaps of scaled expression of core antiviral response genes observed in LPS-treated populations: CD4+ T cells, CD8+ T cells, and NK cells. Color scale for heatmap indicates scaled gene expression. Yellow indicates positive scaled gene expression, purple indicates negative scaled gene expression, and while black represents zero scaled gene expression Right: Average expression of all genes in a geneset (log expression) for each cell, grouped by mock-treated and morphine-treated cells of LPS-treated populations: CD4+ T cells (LPS (534 cells), Morphine+LPS (605 cells)), CD8+ T cells (LPS (152 cells), Morphine+LPS (158 cells)), and NK cells (LPS (37 cells), Morphine+LPS (9 cells)). Inset box plots show the median, lower and upper hinges that correspond to the first quartile (25th percentile) and third quartile (75th percentile), and the upper and lower whiskers extend at most 1.5 times the interquartile range. All comparisons use two-tailed T-tests. For CD4+ T cells: comparison between control and opioid-dependent groups for each geneset: core antiviral (p < 2.22e−16), peaked inflammation (p = 0.91), sustained inflammation (p = 0.16). For CD8+ T cells: comparison between control and opioid-dependent groups for each geneset: core antiviral (p = 6.3e−07), peaked inflammation (p = 0.91), sustained inflammation (p = 0.85). For NK cells: comparison between control and opioid-dependent groups for each geneset: core antiviral (p = 0.0053), peaked inflammation (p = 0.23), sustained inflammation (p = 0.00039). ^nsp > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data for a and b detailing expression values are provided in Source Data file.