Mycobacterium vaccae NCTC 11659, a Soil-Derived Bacterium with Stress Resilience Properties, Modulates the Proinflammatory Effects of LPS in Macrophages

Inflammatory conditions, including allergic asthma and conditions in which chronic low-grade inflammation is a risk factor, such as stress-related psychiatric disorders, are prevalent and are a significant cause of disability worldwide. Novel approaches for the prevention and treatment of these disorders are needed. One approach is the use of immunoregulatory microorganisms, such as Mycobacterium vaccae NCTC 11659, which have anti-inflammatory, immunoregulatory, and stress-resilience properties. However, little is known about how M. vaccae NCTC 11659 affects specific immune cell targets, including monocytes, which can traffic to peripheral organs and the central nervous system and differentiate into monocyte-derived macrophages that, in turn, can drive inflammation and neuroinflammation. In this study, we investigated the effects of M. vaccae NCTC 11659 and subsequent lipopolysaccharide (LPS) challenge on gene expression in human monocyte-derived macrophages. THP-1 monocytes were differentiated into macrophages, exposed to M. vaccae NCTC 11659 (0, 10, 30, 100, 300 µg/mL), then, 24 h later, challenged with LPS (0, 0.5, 2.5, 250 ng/mL), and assessed for gene expression 24 h following challenge with LPS. Exposure to M. vaccae NCTC 11659 prior to challenge with higher concentrations of LPS (250 ng/mL) polarized human monocyte-derived macrophages with decreased IL12A, IL12B, and IL23A expression relative to IL10 and TGFB1 mRNA expression. These data identify human monocyte-derived macrophages as a direct target of M. vaccae NCTC 11659 and support the development of M. vaccae NCTC 11659 as a potential intervention to prevent stress-induced inflammation and neuroinflammation implicated in the etiology and pathophysiology of inflammatory conditions and stress-related psychiatric disorders.


Introduction
Inflammatory conditions, such as allergic asthma, and conditions in which inflammation is considered a risk factor, e.g., stress-related psychiatric disorders including anxiety disorders, mood disorders, and trauma-and stressor-related disorders, such as posttraumatic stress disorder (PTSD) [1,2], are prevalent and are a significant cause of disability worldwide [3,4]. One strategy to reduce the risk of these conditions would be to use antiinflammatory or immunoregulatory approaches to mitigate inflammatory responses [5,6]. As posited by the "Hygiene" hypothesis or "Old Friends" hypothesis, industrialized nations have higher rates of chronic low-grade inflammation, inflammatory diseases, and psychiatric disorders due to urbanization resulting in decreased exposures to immunoregulatory microorganisms (i.e., microorganisms that promote a balanced expression of regulatory and effector T cells) [7,8]. Indeed, growing up on a farm provides protection against allergic asthma [8], while healthy young persons raised in a city, without daily exposure to pets for the first 15 years of life, respond to a psychosocial stressor with exaggerated increases in numbers of circulating peripheral blood mononuclear cells (PBMCs) and increases in circulating concentrations of interleukin 6 (IL-6), a proinflammatory cytokine, relative to those who were raised on farms in close proximity to farm animals [9].
One mechanism that has been implicated in exaggerated inflammation in the context of both allergic asthma and stress-related psychiatric disorders is the recruitment of inflammatory monocytes to the airways or central nervous system [1,[10][11][12]. Monocytederived macrophages in house dust mite-allergic mice or persons with asthma have an inflammatory phenotype, characterized by increased IL-6 and IL-12 production [13]. Likewise, studies in rodents have shown that IL-6 is increased in individuals that respond to psychosocial stress with a susceptible phenotype following subsequent chronic stress exposure. Indeed, individual variability in ex vivo stimulated IL-6 from leukocytes prior to stress exposure predicts stress susceptibility versus resilience [14,15]. Furthermore, the transfer of hematopoietic progenitor cells from stress-susceptible mice to IL-6-deficient mice was sufficient to transfer a stress-susceptible phenotype, highlighting the importance of bone marrow-derived leukocytes to psychosocial stress-induced anxiety-like defensive behavioral responses and depressive-like responses [15]. In particular, monocytes were elevated in individuals that developed a susceptible phenotype after stress exposure, and monocyte numbers were negatively correlated with behavioral responses [15,16]. In clinical populations, circulating monocyte numbers, relative abundance, or inflammatory gene expression patterns in monocytes have been associated with risk of anxiety disorders [17], mood disorders [18][19][20][21], and suicide risk [18,19,22], and have been identified among topranking features predicting future PTSD symptoms [23]. Furthermore, monocytes from persons with depression show an altered response to lipopolysaccharide (LPS) [24].
Studies in persons with a diagnosis of depression have shown elevated circulating concentrations of other proinflammatory cytokines, including IL-12, a heterodimer between IL-12p35 and IL-12p40, encoded by IL12A and IL12B, respectively. For example, individuals with depression have elevated plasma IL-12 concentrations relative to healthy controls [25][26][27][28], which are decreased after treatment [25][26][27]. Likewise, the ratio of IL-12 to anti-inflammatory cytokines, such as transforming growth factor β1 (TGF-β1), is decreased following antidepressant treatment [26]. Indeed, meta-analyses of inflammatory markers in depression reveal increases in a diverse range of proinflammatory cytokines, including IL-12 [29,30]. Finally, treatment with tumor necrosis factor (TNF) and IL-12/IL-23 blockers can improve depressive symptoms in persons with inflammatory disease [31].
One strategy, therefore, is to develop interventions that restore adequate anti-inflammatory and immunoregulatory responses and, thus, reduce the risk of immune-mediated conditions and stress-related psychiatric disorders, in which inflammation is a risk factor. One of these immunoregulatory interventions is Mycobacterium vaccae (M. vaccae) NCTC 11659, which has anti-inflammatory, immunoregulatory, and stress-resilience properties (for review, see [5]. However, little is known about how M. vaccae NCTC 11659 affects specific immune cell targets, including monocyte-derived macrophages that can drive inflammatory and neuroinflammatory responses and, thus, increase risk of immune-mediated conditions and stress-related psychiatric disorders [1,16,32]. Monocytes that differentiate into macrophages with a proinflammatory bias [1], or exosomes derived from these cells [33,34], can infiltrate the central nervous system (CNS) and induce an inflammatory neuroimmune environment, which is a risk factor for the development of stress-related psychiatric disorders, including anxiety disorders, mood disorders, and trauma-and stressor-related disorders, such as PTSD [1,32]. However, macrophages with an anti-inflammatory bias that infiltrate the CNS are vital in CNS repair [35,36] and might be key in suppressing inappropriate inflammation in the CNS. In line with recommendations to avoid the use of the terms "resting" and "activated" innate immune cells, and dichotomous "M1-like" and "M2-like" categorization of these cells [37][38][39][40][41], here we simply refer to proinflammatory or anti-inflammatory bias, or polarization of macrophages, and describe the specific changes in gene expression.
This study used a whole-cell, heat-killed preparation of M. vaccae NCTC 11659 . M. vaccae NCTC 11659 is a saprophytic environmental bacterium with antiinflammatory and immunoregulatory properties. The original strain, designated R877, was isolated by Dr. John L. Stanford and Roger C. Paul (University College London and Middlesex Hospital Medical School, London), from mud in the Kyoga Nile valley near Lake Kyoga in Uganda [42]. Multiple single-colony isolations were conducted by John Stanford and John M. Grange to clone a stable rough variant of M. vaccae R877, designated R877R. The rough variant was deposited as M. vaccae NCTC 11659 and a heat-killed preparation of M. vaccae NCTC 11659 was initially evaluated in clinical studies of leprosy and tuberculosis (for review, see [5]. Subsequent studies in mice demonstrated that immunization with M. vaccae NCTC 11659 provided long-term protection from allergic airway inflammation in a manner dependent on induction of regulatory T cells (Treg) and production of anti-inflammatory cytokines, including IL-10 and TGF-β1 ( [43]; for review, see [5,44]. We have since described in detail the effects of immunization with a heat-killed preparation of M. vaccae NCTC 11659 to prevent: (1) stress-induced increases in spontaneous colitis; (2) stress-induced exaggeration of secretion of proinflammatory cytokines, including IL-6, from mesenteric lymph node cells treated with anti-CD3 antibody ex vivo; and (3) stressinduced increases in anxiety-like defensive behavioral responses [45][46][47][48][49][50][51][52][53][54][55], including effects relevant to three of the four major diagnostic criteria for PTSD (see Supplementary Table S4 in [45]; see Table 1 in [56].
We set out to determine if M. vaccae NCTC 11659 could shift human monocyte-derived macrophages to a less inflammatory phenotype that would respond to subsequent immune challenges with LPS with a bias toward anti-inflammatory responses, characterized by increased ratios of genes encoding anti-inflammatory cytokines, i.e., IL-10 and TGF-β1, to proinflammatory cytokines, i.e., IL-12 and IL-23 [57,58]. For a list of abbreviations for all gene symbols, see Supplemental S1.

Cell Viability
Cell viability was assessed using flow cytometry gating of live singlets based on forward and side scatter (FSC/SSC), which revealed no effect of either M. vaccae NCTC 11659 or LPS on cell viability (see Supplemental Results S2.1, Figure S1, and Supplemental Materials and Methods S3 in the Supplementary Material). Sample sizes are listed in Table S1. This was supported by an assessment of ACTB mRNA expression using real-time reverse transcription quantitative-polymerase chain reaction (real-time RT-PCR), which revealed no consistent effects of either M. vaccae NCTC 11659 or LPS on cell viability as assessed by the expression of this reference gene (see Section 2.3 below).

Effects of Heat-Killed M. vaccae NCTC 11659 (300 µg/mL) and LPS (250 ng/mL) on Gene Expression Using NanoString nCounter ® Inflammation v2 Panel
To determine if there was any effect of treatment condition on reference genes, we analyzed counts of the six reference genes standardized using negative and positive control probes (i.e., CTLC, GAPDH, GUSB, HPRT1, PGK1, and TUBB). Two-way ANOVA revealed main effects for M. vaccae NCTC 11659, LPS, or interactions between these factors for each of the six reference genes (see Supplemental Results S2.2 and Figure S2 in the Supplementary Material). Thus, all normalization was done using Removing Unwanted Variation-III (RUV-III), specifically designed to address this issue in the NanoString nCounter gene expression assay [59].

Effects of LPS
Previous studies have shown that LPS (Escherichia coli 0111:B4) is a potent inducer of inflammation in human monocyte-derived macrophages [60]. After 24 h of exposure to 250 ng/mL LPS, 107 of the 249 inflammation-related genes in the CodeSet (i.e., 43%) were differentially expressed (p < 0.05 and fold change either less than -1.5 or greater than 1.5) within the BBS vehicle groups between LPS conditions (i.e., BBS/250 ng/mL LPS vs. BBS/RPMI 1640 vehicle). Of these 107 differentially expressed genes, 57 were upregulated and 50 were downregulated ( Figure S3). A detailed table of all differentially expressed genes can be found in Table S2 in the Supplementary Material. Overall, the response to LPS was consistent with previous studies showing that the majority of genes responsive to LPS in THP-1-derived macrophages contain known or putative NF-κB sites, including CCL3, CCL4, CXCL1, CXCL2, CXCL8 (IL8), CXCL10, CXCR4, IL1RN [61]. As LPS signals through the toll-like receptor 4 (TLR4)-MD-2 complex [62], the differentially expressed genes found in the Toll-Like Receptor Signaling KEGG pathway, including TLR4, can be found in Figure S4 in the Supplementary Material.

Effects of M. vaccae NCTC 11659 in LPS-Challenged Cells
Previous studies have shown that immunization with M. vaccae NCTC 11659 attenuates stress-induced exaggeration of inflammatory gene expression in vivo [45,46]. Following forty-eight hours of exposure to 300 µg/mL M. vaccae NCTC 11659 or vehicle and 24 h of exposure to 250 ng/mL LPS, 115 of the 249 genes in the CodeSet (i.e., 46%) were differentially expressed (p < 0.05 and fold change either less than −1.5 or greater than 1.5) within the LPS groups between M. vaccae NCTC 11659 conditions (i.e., 300 µg/mL M. vaccae NCTC 11659/250 ng/mL LPS vs. BBS/250 ng/mL LPS). Out of these 115 differentially expressed genes, 52 were upregulated and 63 were downregulated ( Figure 1B). A detailed table of all differentially expressed genes can be found in Table S4 in the Supplementary Material. Consistent with a downregulation of TNF and CREB1, MAP2K4, MAP3K5, RIPK1, TRADD, and TRAF2 (involved in transcriptional responses to TNF signaling) in M. vaccae NCTC 11659 exposed, LPS-challenged macrophages, a number of genes downstream of TNF signaling were also downregulated, including CCL2, CXCL10, CSF1, PTGS2, and TNFAIP3. In contrast, genes involved in IL-10/STAT3 signaling were upregulated, including IL10 and IL10RB, while IL-10 responsive genes, e.g., IL1A, IL12B, IRF7, STAT1 and TNF [67,68], were downregulated. This is consistent with previous studies showing that, at the transcriptional level, IL-10 attenuates expression of a subset of genes activated by TLR signaling, accounting for 20-25% of mRNAs induced after LPS stimulation including Il1a, Il12b, Il18, and Tnf [69]. The effects of M. vaccae NCTC 11659 in LPS-challenged cells on differentially expressed genes found in the Toll-Like Receptor Signaling KEGG pathway can be found in Figure S6 in the Supplementary Material.

Effects of LPS in M. vaccae NCTC 11659-Exposed Cells
When the effects of LPS were evaluated among M. vaccae NCTC 11659-exposed cells, i.e., M. vaccae NCTC 11659/LPS challenge versus M. vaccae NCTC 11659/RPMI 1640 exposed cells, 148 of the 249 genes in the CodeSet (i.e., 59%) were differentially expressed (p < 0.05 and fold change either less than -1.5 or greater than 1.5) between the LPS groups within M. vaccae NCTC 11659 conditions (i.e., 300 µg/mL M. vaccae NCTC 11,659/250 ng/mL LPS vs. 300 µg/mL M. vaccae NCTC 11659/RPMI 1640). Of these 148 differentially expressed genes, 65 were upregulated and 83 were downregulated ( Figure S7). A detailed table of all differentially expressed genes can be found in Table S5 in the Supplemental Information. Among M. vaccae NCTC 11659-exposed cells, the LPS challenge resulted in the downregulation of a number of genes involved in proinflammatory responses, particularly NF-κB responsive genes including BCL2L1, BIRC2, IL-12B, NFKB2, NFKBIA, TNF, TNFAIP3, and TRAF2. The effects of LPS in M. vaccae NCTC 11659-exposed cells on the differentially expressed genes found in the Toll-Like Receptor Signaling KEGG pathway can be found in Figure

Summary of NanoString nCounter Inflammation Panel
As data from the NanoString nCounter Inflammation Panel suggested that M. vaccae NCTC 11659 upregulated IL10 expression and downregulated IL12B expression in LPSchallenged macrophages, and the anti-inflammatory effects of M. vaccae NCTC 11659 in vivo have been shown to depend on IL-10 and TGF-β1 signaling, we followed up in detail to assess the effects of M. vaccae NCTC 11659 and LPS on IL10, TGFB1, IL12A, IL12B, and IL23A using real-time RT-PCR. Here we describe results from experiments designed to measure the expression of genes encoding major inflammatory cytokines in macrophages by extracting total RNA from THP-1 monocyte-derived macrophages and performing real-time PCR as described above. Analysis of the housekeeping gene, ACTB, for differential expression across treatment groups found no meaningful differences in expression ( Figure S12 in the Supplemental Information). Overall, similar effects of M. vaccae NCTC 11659 and LPS were observed on IL12A, IL12B, IL23A, IL10, and TGFB1 mRNA expression using real-time RT-PCR as in Experiment 1 (Figures S13 and S14).

Effects of Heat-Killed M. vaccae NCTC 11659 and LPS on IL10:IL12A mRNA Expression Ratio
Expression levels of IL12A, IL12B, IL23A, and IL10 are important in their own right to characterize the immune environment, but the ratio between anti-inflammatory and inflammatory cytokines in the environment is an important feature of macrophage polarization [70]. For example, high expression of IL-12 and IL-23, but low expression of IL-10, is characteristic of macrophages with proinflammatory polarization [57,58]. In contrast, high expression of IL-10, but low expression of IL-12, is characteristic of macrophages with an anti-inflammatory polarization, particularly macrophages implicated in anti-inflammatory and immunoregulatory effects [57,58]. This suggests that the ratios of IL10 to IL12A, IL12B, and IL23A may be meaningful measures of macrophage polarization. The ratio of IL10 to IL12A expression was calculated in THP-1 monocyte-derived macrophages after incubating them with various concentrations of M. vaccae NCTC 11659 and LPS as described below. In line with the low expression of IL12A mRNA in the vehicle (0 ng/mL LPS) condition, 70% of IL10:IL12A mRNA expression ratio data were real-time PCR non-detects (Table S9 in Table S10 in the  Supplementary Material. greater than IL12A mRNA expression (Figure 2A). Exposure of THP-1 monocyte-derived macrophages to higher concentrations of M. vaccae NCTC 11659 in the presence of 0.5 ng/mL LPS decreased the IL10:IL12A mRNA expression ratio (Figure 2A), whereas exposure of THP-1 monocyte-derived macrophages to the highest concentration of M. vaccae NCTC 11659 (300 µ g/mL) in the presence of 250 ng/mL LPS increased the IL10:IL12A mRNA expression ratio (Figure 2A). Exposure of THP-1 monocyte-derived macrophages to LPS in the absence of M. vaccae NCTC 11659 did not affect the IL10:IL12A mRNA expression ratio (Figure 2A). Post hoc p values for planned pairwise comparisons can be found in Table S10 in the Supplementary Material.  As mentioned above, expression levels of IL12A, IL12B, IL23A, and TGFB1 are important in their own right to characterize the immune environment, but the ratio between anti-inflammatory and inflammatory cytokines in the environment is an important feature of macrophage polarization. For example, high expression of IL-12, but low expression of TGF-β1, is characteristic of macrophages with a proinflammatory bias [57,58]. In contrast, high expression of TGF-β1, but low expression of IL-12, is characteristic of macrophages with an anti-inflammatory bias, particularly macrophages implicated in anti-inflammatory and immunosuppressive effects [57,58]. This suggests that the ratio of TGFB1 to IL12A, IL12B, and IL23A may be a meaningful measure of macrophage polarization. The ratio of TGFB1 to IL12A expression was calculated in THP-1 monocyte-derived macrophages after incubating them with various concentrations of M. vaccae NCTC 11659 and LPS as described above. The analysis revealed similar effects of M. vaccae NCTC 11659 and LPS on these ratios as the IL10:IL12A, IL10:IL12B, and IL10:IL23A mRNA expression ratios ( Figure S15A-C). Post hoc p values for planned pairwise comparisons can be found in Table S10 in the Supplementary Material.

Discussion
Here we report evidence that M. vaccae NCTC 11659, a soil-derived bacterium with anti-inflammatory, immunoregulatory, and stress resilience properties, modulates the proinflammatory effects of LPS with M. vaccae NCTC 11659 inducing "adjuvant-like" effects in the absence of LPS or at low concentrations of LPS (0, 0.5 ng/mL) and mitigating the effects of LPS at higher concentrations of LPS (250 ng/mL). These data suggest that M. vaccae NCTC 11659 might function to induce an anti-inflammatory or immunoregulatory phenotype under hyper-inflammatory conditions, which are observed in response to allergic airway inflammation, in response to extreme trauma or stress, or in individuals with a diagnosis of trauma-and stressor-related disorders, such as PTSD [71]. M. vaccae NCTC 11659 alone had adjuvant-like effects, based on increases in proinflammatory cytokine mRNA expression and decreases in anti-inflammatory to proinflammatory cytokine mRNA ratios in the absence of subsequent LPS challenge, or challenge with a low concentration of LPS (0.5 ng/mL). LPS alone was a reliable stimulator of inflammation based on concentration-dependent increases in proinflammatory cytokine mRNA expression and decreased anti-inflammatory to proinflammatory cytokine mRNA expression. Interestingly, pretreatment with M. vaccae NCTC 11659 prior to LPS challenge resulted in attenuation of LPS-induced exaggeration of inflammation based on decreased proinflammatory cytokine mRNA levels and increased anti-inflammatory to proinflammatory cytokine mRNA ratios. Specifically, at the highest concentration of LPS (250 ng/mL), pretreatment with M. vaccae NCTC 11659 shifted THP-1 monocyte-derived macrophages from a proinflammatory phenotype (IL10 low and IL12B high) toward an anti-inflammatory phenotype (IL10 high and IL12B low).

M. vaccae NCTC 11659 Had "Adjuvant-like" Effects on THP-1 Monocyte-Derived Macrophages
Acute effects of M. vaccae NCTC 11659 on human monocyte-derived macrophages, in the absence of subsequent LPS stimulation, were largely proinflammatory when assessed 48 h later, shifting macrophages toward a proinflammatory macrophage phenotype. This is evidenced by concentration-dependent increases in IL12B and IL23A mRNA expression, hallmark cytokines of proinflammatory macrophages that together encode IL-23, a heterodimeric cytokine of IL-12p40 (encoded by IL12B) and IL-23p19 (encoded by IL23A) [58], and concentration-dependent decreases in the IL10:IL12B, IL10:IL23A, TGFB1:IL12B, and TGFB1:IL23A mRNA expression ratios [58] in human THP-1 monocytederived macrophages. The administration of M. vaccae NCTC 11659 alone, therefore, seems to act as its own adjuvant (i.e., ensuring immune activation before subsequent immune challenge). These effects are consistent with previous studies of M. vaccae NCTC 11659 in vivo. For example, in a murine model of allergic airway inflammation, intra-tracheal administration of M. vaccae NCTC 11659 increases pulmonary Il1b, Il6, and Tnf mRNA expression acutely, 12 h to 3 days following injection. In a mouse model of chronic psychosocial stress, administration of M. vaccae NCTC 11659, in single-housed control animals (i.e., not exposed to subsequent chronic psychosocial stress), M. vaccae NCTC 11659 increases IL-6 secretion, and the IL-6:IL-10 ratio, in freshly isolated mesenteric lymph node cells stimulated with anti-CD3 antibody in vitro, assessed over one month following the final immunization with M. vaccae NCTC 11659, suggesting that adjuvant-like effects of M. vaccae NCTC 11659 can be long-lasting [46]. Similarly, in a model of inescapable stress in rats, administration of M. vaccae NCTC 11659 to home-cage control animals increases hippocampal Il6 mRNA expression, assessed one week after the final immunization [45]. Together, the in vitro data reported here are consistent with previous studies demonstrating that, in the absence of subsequent psychosocial stress, inescapable stress, or immunological challenge, M. vaccae NCTC 11659 has persistent adjuvant-like effects.

Lipopolysaccharide (LPS) Was a Reliable Stimulator of Inflammatory Cytokine mRNA Expression in THP-1 Monocyte-Derived Macrophages
As expected, exposure of human monocyte-derived macrophages to LPS, in a concentrationdependent manner, shifted macrophages toward a proinflammatory macrophage phenotype. Similar to the administration of M. vaccae NCTC 11659 alone, LPS stimulation in the absence of M. vaccae NCTC 11659 increased IL12B and IL23A mRNA expression in a concentrationdependent manner (again, characteristic of proinflammatory macrophages) and decreased the IL10:IL12B, IL10:IL23A, TGFB1:IL12B, and TGFB1:IL23A mRNA expression ratios in a concentration-dependent manner in human THP-1 monocyte-derived macrophages. LPS stimulation at the higher concentrations studied (i.e., 2.5 ng/mL and 250 ng/mL) was a reliable stimulator of inflammatory cytokine mRNA expression and may be most relevant to how psychosocial stressors or immune stimulation with LPS are reliable stimulators of inflammatory cytokine mRNA expression in human monocyte-derived macrophages, as previously demonstrated [1].

Treatment with M. vaccae NCTC 11659 Prior to Immune Stimulation with LPS Resulted in Attenuation of an LPS-Induced Proinflammatory Phenotype in THP-1 Monocyte-Derived Macrophages
Exposure to M. vaccae NCTC 11659 attenuated subsequent proinflammatory responses to immune stimulation with LPS. Unlike how M. vaccae NCTC 11659 stimulated a proinflammatory response in the absence of LPS challenge, M. vaccae NCTC 11659, in a concentrationdependent manner, attenuated proinflammatory responses to subsequent LPS challenge. This is evidenced by the attenuation of LPS-induced inflammatory responses, such as attenuating LPS-induced increases in IL12B and IL23A mRNA expression. In addition, exposure to M. vaccae NCTC 11659 attenuated LPS-induced decreases in the IL10:IL12B, IL10:IL23A, TGFB1:IL12B, and TGFB1:IL23A mRNA expression ratios in human THP-1-derived macrophages. These reported effects are potentially relevant for previously reported effects of M. vaccae NCTC 11659 to attenuate allergic airway inflammation in a murine model of allergic airway inflammation, or to attenuate stress-induced inappropriate, highly inflammatory responses. These effects are consistent with what has been described in vivo [5,[43][44][45][46]72]) and may represent a form of innate immune tolerance.

Treatment with M. vaccae NCTC 11659 Prior to Immune Stimulation with LPS Resulted in the Promotion of an Anti-Inflammatory Phenotype and Inhibition of LPS-Induced Immune Activation in THP-1 Monocyte-Derived Macrophages
We report evidence supporting the hypothesis that M. vaccae NCTC 1169 induces features of macrophages involved in anti-inflammatory, immunoregulatory, and immunosuppressive effects [58,73]. M. vaccae NCTC 11659 did not directly upregulate IL10 or TGFB1 mRNA expression in cells subsequently challenged with LPS; however, it increased the mRNA expression ratios of IL10:IL12B, IL10:IL23A, TGFB1:IL12B, and TGFB1:IL23A upon subsequent immune stimulation with LPS. These findings have implications on how macrophages stimulated with M. vaccae NCTC 11659 may locally modulate T cell responses in the tissue, promoting immunoregulation directly at the site of inflammation [73]. A graphical illustration of our reported effects of M. vaccae NCTC 11659 and LPS on THP-1 monocyte-derived macrophages can be found in Figure 3.
inflammation [73]. A graphical illustration of our reported effects of M. vaccae NCTC 11659 and LPS on THP-1 monocyte-derived macrophages can be found in Figure 3.

Limitations
One limitation of our reported findings was the low expression of IL12A mRNA in THP-1 monocyte-derived macrophages, resulting in real-time RT-PCR non-detects (Cq > 40) that were excluded from the analysis. This resulted in low IL12A sample sizes, particularly in the RPMI 1640 condition, which affected the analysis of IL12A mRNA as well as the IL10:IL12A and TGFB1:IL12A mRNA ratios. The low IL12A expression in the RPMI 1640 condition is likely due to the lack of IFN-γ stimulation, an inducer of IL12A transcription especially important in macrophages [70,74,75]. Of course, as is common with in vitro studies, we were also limited by only specific cell types being present in the culture (only monocytes and macrophages), so cell-to-cell interactions could not directly be studied, such as T cell differentiation. Furthermore, human PMA-differentiated THP-1-derived macrophages may not accurately represent human monocyte-derived macrophages, as it has been shown that these cell types can differ in terms of LPS and IFN-γ-induced cytokine, chemokine, and growth factor secretion [76], although differences may be dependent on the specific protocol used [77]. For example, studies of mycobacterial infection have found that PMA-differentiated THP-1-derived macrophages and human monocyte-derived macrophages respond similarly in terms of bacterial uptake, viability and host response to drug-susceptible and drug-resistant mycobacterial infections [78]. Further, the

Limitations
One limitation of our reported findings was the low expression of IL12A mRNA in THP-1 monocyte-derived macrophages, resulting in real-time RT-PCR non-detects (Cq > 40) that were excluded from the analysis. This resulted in low IL12A sample sizes, particularly in the RPMI 1640 condition, which affected the analysis of IL12A mRNA as well as the IL10:IL12A and TGFB1:IL12A mRNA ratios. The low IL12A expression in the RPMI 1640 condition is likely due to the lack of IFN-γ stimulation, an inducer of IL12A transcription especially important in macrophages [70,74,75]. Of course, as is common with in vitro studies, we were also limited by only specific cell types being present in the culture (only monocytes and macrophages), so cell-to-cell interactions could not directly be studied, such as T cell differentiation. Furthermore, human PMA-differentiated THP-1-derived macrophages may not accurately represent human monocyte-derived macrophages, as it has been shown that these cell types can differ in terms of LPS and IFN-γ-induced cytokine, chemokine, and growth factor secretion [76], although differences may be dependent on the specific protocol used [77]. For example, studies of mycobacterial infection have found that PMA-differentiated THP-1-derived macrophages and human monocyte-derived macrophages respond similarly in terms of bacterial uptake, viability and host response to drug-susceptible and drug-resistant mycobacterial infections [78]. Further, the THP-1 cell line used was isolated from a single person (i.e., a 1-year-old male that had acute monocytic leukemia); thus, the extent to which our reported findings can be generalized to monocyte-derived macrophages in females, or other populations, cannot be inferred from these data. In addition, our reported findings were limited by measuring gene expression at a single timepoint. Therefore, it cannot be inferred that our conclusions apply to other timepoints, especially considering the temporally dynamic functions of the genes measured [79]. Finally, it will be important to confirm that transcriptional changes observed here are associated with changes in protein synthesis and release, again, taking into account the temporal dynamics of the response.

Clinical Implications
Previous studies suggest that inappropriate inflammation is a clinical feature of allergic asthma and a risk factor for the development of anxiety disorders, mood disorders, and trauma-and stressor-related disorders, such as PTSD (for review, see [56]. For example, elevated biomarkers of inflammation have been identified as risk factors for the future development of PTSD following exposure to trauma [23,80], and persons with a diagnosis of PTSD have decreased Treg [81] and enhanced spontaneous production of proinflammatory cytokines from PBMCs [82]. Meanwhile, persons with a diagnosis of PTSD have increased serum concentrations of LPS, lipopolysaccharide-binding protein (LBP; a biomarker of "leaky gut"), and HMGB1, a damage-associated molecular pattern (DAMP) [71], leading the authors to suggest that "Going forward, additional investigations to evaluate the microbiota and approaches to modify the intestinal microenvironment may be a useful adjunct approach to complement existing treatments for PTSD." The present study suggests that M. vaccae NCTC 11659 is a candidate for mitigating the physiological and psychological impacts of allergic asthma and stress-related psychiatric disorders, which can be highly comorbid [83,84].

Future Directions
Future directions include assessing the impact of M. vaccae NCTC 11659 on LPS activation of human primary monocytes isolated from donor PBMCs, compared to human PBMC monocyte-derived macrophages. This would enable confirmation of THP-1 work in human primary cells, with multiple donor testing possible. Future directions also include the assessment of additional genes using real-time RT-PCR, such as ARG1, a hallmark of macrophages with an anti-inflammatory phenotype, co-culturing THP-1 monocyte-derived macrophages with naïve T cells to directly assess the effects of M. vaccae NCTC 11659 on T cell differentiation and function, and assessment of the effects of M. vaccae NCTC 11659 on immune activation by immune challenges other than LPS. Furthermore, future studies should determine if M. vaccae NCTC 11659 effects are generalizable to other rapidly growing mycobacteria, or molecular constituents of M. vaccae NCTC 11659 that have been shown to have anti-inflammatory effects in macrophages, such as 10(Z)-hexadecenoic acid [85]. Likewise, it will be important to determine to what extent these results can be replicated in vivo. In mice, future studies should examine bone marrow-derived monocyte education following administration of M. vaccae NCTC 11659 in vivo, which could be informative in relation to understanding monocyte/innate immune cell differentiation, priming, and trained immunity, versus tolerance in the bone marrow compartment [86].

Conclusions
In 2013, in an article titled, "Harnessing monocyte-derived macrophages to control central nervous system pathologies: no longer 'if' but 'how'" Shechter and Schwartz wrote, "Taken together, these recent advances reveal a dramatic therapeutic opportunity for controlled harnessing of macrophages for repair of the damaged CNS following acute insults, in neurodegenerative conditions, and in psychiatric disorders" [35]. We feel that the current study highlights how microbe-based interventions might, in the near term, prove to be novel and effective approaches to the realization of these dramatic therapeutic opportunities.

Experiment 1, NanoString Platform
The overall experimental design for Experiment 1 is illustrated in Figure S16A in the Supplementary Material. Briefly, in Experiment 1, human THP-1 monocyte-derived macrophages were exposed to heat-killed preparations of M. vaccae NCTC 11659 (30 µg/mL or 300 µg/mL) or sterile borate-buffered saline (BBS) vehicle. Twenty-four hours later, the THP-1 monocyte-derived macrophages were challenged with lipopolysaccharide (LPS; Escherichia coli 0111:B4; 250 ng/mL) or RPMI 1640 vehicle. Finally, 24 h later, total RNA was extracted and assessed using the NanoString Platform. A timeline of Experiment 1 can be found in Figure S16A in the Supplementary Material.

NanoString Gene Expression Analysis
The NanoString CodeSet used was the nCounter Human Inflammation v2 Panel (Item No. 115000072; XT-CSO-HIN2-12, NanoString Technologies, Seattle, WA, USA). The NanoString nCounter ® Inflammation Panel is a multiplex gene expression analysis platform that includes 255 genes that represent a broad range of pathways related to inflammation including apoptosis, epidermal growth factor (EGF) signaling, interleukin signaling, Ras signaling, T cell receptor signaling, and toll-like receptor signaling. The CodeSet included 249 inflammation-related genes and 6 housekeeping genes. A number of previous studies have used the nCounter Human Inflammation v2 Panel (e.g., [87][88][89][90][91][92][93]). Total RNA was extracted using QIAGEN RNeasy spin columns according to the manufacturer-issued instructions (Cat. No. 74104, QIAGEN, Germantown, MD, USA). Once dissolved in nuclease-free water, RNA was quantified using a NanoDrop One machine (Cat. No. ND-ONE-W, ThermoFisher Scientific, Madison, WI, USA), diluted to 10 ng/mL, and frozen at −80 • C before submitting samples to the Veterans Health Administration, Rocky Mountain Regional Veterans Affairs Medical Center (RMRVAMC) Core Equipment facility for further processing according to the manufacturer-issued instructions.

Real-Time RT-PCR
After centrifuging 24-well cell culture plates at 500× g for 5 min at 4 • C, supernatants were collected and total RNA was harvested from differentiated cells using a standard method of phenol-chloroform extraction [94]. Briefly, after collecting 500 µL of supernatant from each well, 0.5 mL of TRIzol ® (Cat. No. 15596-026, Invitrogen, Waltham, MA, USA) was added to each well and plates were frozen at −80 • C for 3 days until they were processed further. After thawing the plates and extracting the aqueous layer using chloroform separation, but before the RNA was precipitated with isopropanol, 10 µg of glycogen (Cat. No. AM9510, Invitrogen) was added to the aqueous layer to increase the total RNA yield from precipitation. Once dissolved in nuclease-free water, the RNA was quantified, and purity was assessed (A260/A280 ratio) using a NanoDrop One machine (Cat. No. ND-ONE-W, ThermoFisher Scientific). cDNA synthesis was conducted with SuperScript II TM (Cat. No. 18064-014, Invitrogen) following manufacturer instructions using approximately 500 ng of the total RNA. Two µL of cDNA was used as template material. Real-time RT-PCR was conducted in duplicate using the CFX96 Touch Real-Time PCR Detection System (Cat. No. 1845097, Bio-Rad, Hercules, CA, USA) and SYBR Green master mix (Cat. No. 204145, QIAGEN, Hilden, Germany). All genes that were analyzed were normalized using the gene ACTB, encoding beta-actin. Real-time RT-PCR data were represented as a fold increase over the lowest amount of mRNA expressed for each gene using the delta-delta Ct method. Real-time RT-PCR was performed with an independent set of cells separate from those used for flow cytometry. Sample sizes for real-time RT-PCR are listed in Table  S9 in Supplementary Material. Low sample sizes in the 0 ng/mL LPS condition were typically a result of PCR non-detects in the presence of low expression of target genes under BBS/RPMI 1640 vehicle conditions. Primers cDNA sequences were obtained from Genbank at the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov (accessed on 8 August 2019)). Primer sequences (Table S11 in Supplementary Material) were designed using the PrimerQuest TM Tool from the Integrated DNA Technologies (IDT) website (https://www.idtdna.com/ pages (accessed on 8 August 2019)). Sequence specificity was tested using the Basic Local Alignment Search Tool at NCBI [95]. Primers were obtained from IDT. Primer specificity was verified by melt curve analyses. All primers were designed to span exon/exon boundaries and thus exclude amplification of genomic DNA.

Statistical Analysis
The data were analyzed as described below. 4.6.1. NanoString nCounter ® Inflammation Panel An analysis of reference genes in the NanoString platform revealed that treatment conditions altered the expression of all six reference genes in the nCounter Human Inflammation v2 Panel, resulting in inadequate normalization. As the "accuracy and reliability of gene expression results are dependent upon the proper normalization of the data against internal reference genes" [59,96], it was not possible to use the standard normalization methods within nSolver and ROSALIND ® . Instead, all normalization was done using Removing Unwanted Variation-III (RUV-III), specifically designed to address this issue in the NanoString nCounter gene expression assay [59]. RUV-III normalized data were imported into ROSALIND ® . Normalized data were analyzed by ROSALIND ® (https://rosalind.bio/ (accessed on 15 November 2022)) with a HyperScale architecture developed by ROSALIND ® , Inc. (San Diego, CA, USA). The limma R library [97] was used to calculate fold changes and p-values in ROSALIND ® . The statistical software program R was used to make volcano plots using the ggplot2 and ggrepel packages.

Real-Time RT-PCR
Real-time RT-PCR data were represented as either the relative gene expression (i.e., relative to the lowest expression of the gene of interest) or the ratio of one gene's expression to another gene's expression. Relative gene expression was calculated using the 2 −∆∆Ct method and the ratio of one gene to another gene was calculated by finding the Cq difference between the two genes and raising this value to the power of 2. For a detailed methodology of how mRNA expression ratios were calculated, see the supplementary files. Real-time RT-PCR data were analyzed using a two-way ANOVA followed by Dunnett's test, if appropriate, at a two-tailed alpha level of 0.05 using a single pooled error term for Dunnett's test. Specifically, in the presence of a main effect of M. vaccae NCTC 11659, planned pairwise comparisons between the BBS control condition and each concentration of M. vaccae NCTC 11659 (10 µg/mL, 30 µg/mL, 100 µg/mL, or 300 µg/mL) were conducted within the RPMI 1640 or LPS conditions (0.5 ng/mL, 2.5 ng/mL, or 250 ng/mL). In the presence of a main effect of LPS, planned pairwise comparisons between the RPMI 1640 control condition and each concentration of LPS (0.5 ng/mL, 2.5 ng/mL, or 250 ng/mL) were conducted within the BBS or M. vaccae NCTC 11659 conditions (10 µg/mL, 30 µg/mL, 100 µg/mL, or 300 µg/mL). In the presence of a significant interaction between M. vaccae NCTC 11659 and LPS, both types of planned pairwise comparisons were made.

Software
Generating real time RT-PCR relative expression values and real time RT-PCR statistical analysis was conducted using Python (version no. 3.8.1). All Python code is available on GitHub at https://github.com/evho3333/Statistical-Analysis.git (accessed on 8 July 2022). Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. We are grateful for support from an anonymous donor provided through Benefunder, which supported this project. Publication of this article was funded in part by the University of Colorado Boulder Libraries Open Access Fund.