Monocytes have been extensively employed in the study of trained immunity (14, 24). Whole blood containing EDTA was purchased from BioIVT (Gray, Tennessee). The blood samples were de–identified except for age, gender, and race. BioIVT provided blood that was collected from healthy individuals in 3 age groups, i.e. 20–30, 31–59, >60 years of age. There were 35 individuals in each age group and a total of 105 individuals in the study. The age range for the entire cohort was 20–71 years. A detailed description of the subjects is provided in Table 1. Monocytes were isolated from whole blood using immunomagnetic negative selection with the EasySep™ Direct Human Monocyte Isolation Kit (StemCell #19669). Monocyte purity was assessed by flow cytometry. This method yields CD14+ monocyte purity of up to 97.1%.

Peripheral blood was obtained from sepsis patients at our participating hospital, Johnson City Medical Center, following informed consent (IRB study # 098–98s–MSHA). All patients met the Sepsis–3 criteria (25). A detailed description of the included sepsis patients is provided in Table 2. Of the 5 patients included in the study, 2 were female and 3 were male. The patients ranged in age from 23–78 with the average age being 58.6. The average SOFA score of the cohort was 1.6, and the average Apache II score was 8.4. Peripheral blood was collected in BD Vacutainer^® CPT™ (BD #362760) tubes. Monocytes were isolated from total mononuclear cells using the EasySep™ Human Monocyte Isolation Kit (StemCell #19359) following manufacturer instructions.

Monocytes were adjusted to 5 x 10⁵ cells/mL in serum–free RPMI 1640 with 1% Antibiotic–Antimycotic (Gibco 15240–062) and treated with 10 μg/mL β–glucan, isolated from Candida albicans as previously described (26). The C. albicans β–glucan employed in this study was >95% pure. The β–glucan was depyrogenated to remove any residual endotoxin, and it was sterile. The lack of endotoxin in the β–glucan preparation was established with the HEK–Blue LPS Detection kit 2 (Invivogen). Sterility of the β–glucan was confirmed by microbiological testing. Isovolumetric PBS was added to monocyte cultures as a control. Monocytes from the same individual were divided between both treatments. Glucan was washed out after 24h (Figure 1A), and media was replaced with RPMI 1640 with 1% Antibiotic–Antimycotic and supplemented with 10% human serum. The media was replenished at day 3. On day 7, the media volume was reduced, and cells were treated with 10 ng/mL LPS (Escherichia coli O55:B5, Sigma L6529–1MG) or isovolumetric PBS to serve as baseline control for 24 hours. On day 8, the supernatants were harvested for cytokine analysis, and the cells were stained with antibodies for cell surface markers to be analyzed by flow cytometry.

Metabolic analysis of isolated monocytes was performed using the Agilent Seahorse XFp Analyzer (Santa Clara, CA). Monocytes were seeded at 1 x 10⁵/well and trained using the protocol defined above. The XFp Cell Energy Phenotype test template was used according to the manufacturer’s specifications. Oxygen consumption rate (OCR) was measured to approximate respiration, and extracellular acidification rate (ECAR) was measured to approximate glycolysis. Baseline OCR/ECAR was calculated by averaging the first three measurements prior to oligomycin/FCCP injection. Stressed OCR/ECAR was calculated by averaging the five measurements taken after oligomycin/FCCP injection. Final in well concentrations of oligomycin (Sigma 495455–10MG) and FCCP (Sigma C2920–10MG) were both 1µM.

TNF–α and IL–6 concentrations in conditioned media were analyzed by ELISA (TNA–α: BioLegend 430204, IL–6: BioLegend 430504). Cell culture supernatants were collected and stored at –80°C prior to analysis. ELISAs were performed following manufacturer instructions.

Trained and control cells were harvested, blocked using block buffer (PBS, 5% rabbit serum, 0.5% goat serum albumin, 5 mM EDTA, 0.1% NaN₃) and stained for flow cytometry using the following mouse anti–human antibodies: CD14 FITC, Clone TÜK4 (Bio–Rad MCA1568F), HLA–DR PerCP–Cy5.5, Clone G56–6 (BD Biosciences 560652), CD123 PE, Clone 9F5 (BD Biosciences 555644), CD11b/Mac–1 PE–Cy7, Clone ICRF44 (BD Biosciences 557743), CD40 APC, Clone 5C3 (BD Biosciences 555591), CD16 Pacific Blue™, Clone 3G8 (BD Biosciences 558122). All flow cytometry experiments were performed on a BD LSRFortessa flow cytometer. Cells were gated by forward scatter, side scatter, and CD14 positivity (Figure S1B). Cell gating was adjusted to account for glucan particles that remained from the initial treatment, given that some residual particles remain after multiple wash steps (Figure S1A). Mean fluorescence of isotype controls was subtracted from sample mean fluorescence to normalize between experimental replicates.

Trained and control monocytes were adjusted to a density of 5 x 10⁶ cells/mL in 12 x75 mm tubes in RPMI 1640. Cells were treated with 10 μg/mL β–Glucan or vehicle for 1 or 2 hours at 37°C. Cells were fixed using BD Cytofix (BD Biosciences 554655), permeabilized with BD Phosflow Perm Buffer III (BD Biosciences 558050) and stained for 1 hr. with the following antibodies: Alexa Fluor^® 488 Mouse anti–Akt1, Clone 55/PKBa/Akt (BD Biosciences 560048, Alexa Fluor^® 647 Mouse anti–Akt (pS473), Clone M89–61 (BD Biosciences 560343), PE Mouse Anti–mTOR (pS2448), Clone O21–404 (BD Biosciences 563489), PerCP/Cyanine5.5 anti–ZAP70 Phospho (Tyr319)/Syk Phospho (Tyr352) Antibody, Clone 1503310 (Biolegend 683710). Cells were gated for CD14 positivity on a BD LSRFortessa flow cytometer (Figure S2). Mean fluorescence from fluorescence–minus–one controls was subtracted from the sample mean fluorescence to normalize between experimental replicates.

To assess H3K4 methylation, human monocytes were isolated and adjusted to a density of 1x10⁶ cells/mL in RPMI 1640, and 2 x10⁵ monocytes were plated per well in a 96–well plate. Wells were treated with 10 μg/mL β–glucan or PBS. Glucan was washed out after 24h on day 1, and media was replaced with RPMI 1640 supplemented with 10% human serum. The media was replenished at day 3. On day 7, cells were assayed for H3K4 methylation using the EpiQuik In Situ Histone H3K4 Methylation Assay Kit (EpiGentek P–3015–096). To assess H3K27 acetylation, 1 x10⁶ human monocytes were plated in a 12–well plate at 0.5 x 10⁵ cell/mL in RPMI 1640 as previously described. Wells were treated with 10 μg/mL β–glucan or PBS. Glucan was washed out after 24 hrs. on day 1, and media was replaced with RPMI 1640 supplemented with 10% human serum. The media was replenished at day 3. On day 7, histones were isolated using the EpiQuik Total Histone Extraction Kit (EpiGentek OP–0006–100) and frozen at –80°C. H3K27 acetylation was assayed using the EpiQuik Global Acetyl Histone H3K27 Quantification Kit (EpiGentek P–4059–96).

All statistical analysis was performed in GraphPad Prism 9 or R. All data are presented as individual values or mean ± SEM represented by error bars unless otherwise specified. The effect of age and immune training was evaluated using a mixed–effect model in which all treatment conditions of a blood sample were paired. Šídák’s multiple comparisons test was used for all post hoc tests. When the assumption of constant variance was not met, a natural log transformation of the data was performed. Simple linear regression was used to determine the effect of age on metabolic rates and cytokine production. Effect of sex on metabolic rates and cytokine production were assessed using a PERMANOVA because the assumption of constant variance was not met. The effect of trained immunity in sepsis was analyzed using a paired T–test. When the assumption of constant variance was not met, a Wilcoxon test was performed. An α = 0.05 was used to determine the statistical significance of all measurements.

This study was approved by the East Tennessee State University Institutional Review board (IRB study # 098–98s–MSHA). Written informed consent was obtained prior to participation of patients in this study.

Cellular metabolism plays a significant role in immune cell function (27–29). Decreases in respiration and glycolysis have been linked to declining immune systems in aging individuals (27, 30, 31). We found that immune training with β–glucan increased baseline and stressed OCR in each age group (Figure 1B). Immune training also significantly increased baseline and stressed ECAR in all three age groups (Figure 1C). There was no significant difference in the metabolic status of control and immune trained monocytes across the age groups. We also examined the effect of immune training on monocyte metabolism as a function of gender (Figures 2A, B) or age (Figures 2C, D). We observed no significant correlation as a function of age or gender (Figures 2A–F).

Increased cytokine production in response to immune training with β–glucan is a hallmark of the immune trained phenotype (18, 32). To assess monocyte function, we measured IL–6 and TNF–α secretion following immune training. We found that immune training with β–glucan increased TNF–α expression in all three age groups (Figure 3A). There was no significant effect of gender or age on TNF–α expression (Figures 3C, E). There was no effect of immune training on IL–6 expression across all age groups. (Figure 3B), indicating a differential effect of immune training on cytokine expression. However, IL–6 expression positively correlated with age in untrained monocytes (Figure 3F). While there was a trend toward increased IL–6 production as a function of age in trained monocytes (p = 0.0518), it did not reach statistical significance (Figure 3F). There was no significant effect of gender on IL–6 expression (Figure 3D).

Increased H3K4 methylation and H3K27 acetylation are associated with immune training and improved response to infection (18). We found that immune training significantly altered monocyte H3K4 methylation (Figure 4A). However, only H3K4 methylation in the 20–30 age group was significantly increased (Figure 4A). We also observed that immune training had a significant effect on H3K27 acetylation (Figure 4B). Interestingly, H3K27 acetylation was significantly increased in immune–trained cells in the 60+ age group, but not in the younger groups (Figure 4B). Thus, we observed a differential effect of β–glucan induced histone modification in that H3K4 methylation was increased in monocytes derived from healthy young individuals, while H3K27 acetylation was significantly increased in monocytes from aging individuals.

Cheng and colleagues have shown that β–glucan induced immune training in monocytes is mediated via an Akt/mTOR/HIF1–α dependent mechanism (14). Dectin–1 is one of the primary receptors for β–glucan (33, 34) and may be required for β–glucan mediated trained immunity (14). Syk tyrosine kinase is an important downstream mediator of Dectin–1 dependent signaling and is critical in Dectin–1 mediated immune responses (35). We examined the effect of immune training, as a function of age, on Syk, Akt and mTor dependent signaling. We found that phosphorylated Syk was increased in monocytes treated for 1 hour with β–glucan across all age groups (Figure 5A). However, only trained monocytes from young individuals expressed significantly higher levels of phosphorylated Syk at 2 hrs. stimulation (Figure 5A). We found that AKT phosphorylation was upregulated in monocytes treated for 1 hour with β–glucan across all age groups (Figure 5B). Monocytes from the 20–30 and 31–59 age groups that were trained with β–glucan for 2 hrs. showed significantly higher phosphorylated AKT (Figure 5B). However, this effect was not observed in the 60+ age group. Phosphorylated mTOR was significantly increased in monocytes treated with β–glucan for one or two hours across all age groups (Figure 5C). Total AKT expression was significantly increased in β–glucan trained monocytes at 1 and 2 hrs., although there was no significant difference at 1 hr. in the 31–59 group (Figure 5D). The flow cytometry gating strategy for the intracellular signaling studies is shown in Figures S1, S2.

Previous studies have demonstrated that age plays a significant role in the phenotype of CD14+ monocytes (36–38). In addition, immune training with β–glucan has been reported to alter the expression of cell surface markers on monocytes (32). We measured the expression of cell surface markers in trained and control monocytes in the presence and absence of LPS secondary stimulation. We found that immune–trained cells in all age groups showed lower levels of CD16+ cells, indicating a higher relative abundance of intermediate or non–classical monocytes (CD14+, CD16+) (39). Interestingly, this effect was not observed after LPS stimulation (Figure 6A). CD11b/Mac–1 expression was only significantly altered in trained monocytes from the 20 – 30 age group and only following LPS stimulation (Figure 6B). CD123 expression was not altered in immune trained monocytes (Figure 6C). In contrast, immune training significantly decreased monocyte cell surface HLA–DR (Figure 6D). CD40 was significantly decreased by immune training in monocytes from the 60+ age group (Figure 6E). In the presence of LPS secondary stimulation, CD40 was decreased in all three age groups (Figure 6E). When taken together, these data are suggestive of differentiation toward an M2 macrophage phenotype.

The data above clearly show that monocytes from healthy aging and younger individuals can be trained in response to β–glucan. However, there is no data available on the ability of monocytes from septic patients to exhibit the immune trained phenotype. Therefore, we isolated monocytes from patients meeting the broad criteria of Sepsis–3. Monocytes isolated from septic patients were trained with β–glucan employing the protocol in Figure 1A and allowed to rest for 7 days. After 7 days, monocytes were assayed for OCR to approximate respiration and ECAR to approximate glycolysis using the Agilent XFp Cell Energy Phenotype Test protocol. Trained septic monocytes had significantly higher baseline and stressed OCR than control septic monocytes (Figure 7A). Baseline ECAR was increased in trained septic monocytes (Figure 7B). While stressed ECAR showed an increase, there was no statistically significant effect of immune training on stressed ECAR (Figure 7B). Additionally, immune trained monocytes from sepsis patients upregulated TNF–α (Figure 7C) and IL–6 (Figure 7D) upon LPS stimulation. These data clearly show that monocytes from septic patients are capable of exhibiting the immune trained phenotype and their functionality is increased, in a fashion similar to healthy aging subjects.