key: cord-0252955-2ke2uzb9
authors: Seufert, AL; Traxler, SK; Hickman, JW; Peterson, RM; Lashley, SJ; Shulzhenko, N; Napier, RJ; Napier, BA
title: Dietary palmitic acid induces trained immunity that controls inflammation and infection
date: 2021-06-16
journal: bioRxiv
DOI: 10.1101/2021.06.15.448579
sha: 49f6c50436d79fc90e4f21f3d5fd307cd79e0a79
doc_id: 252955
cord_uid: 2ke2uzb9

Trained immunity is epigenetic reprogramming that occurs in innate immune cells in response to primary inflammatory stimuli and leads to enhanced inflammation upon secondary challenge with homologous or heterologous stimuli. We find exposure to high-fat diets confers a hyper-inflammatory response to systemic LPS and enhanced mortality, independent of microbiome. Ketogenic diet (KD) does not alter homeostatic inflammation, but enhances the response of immune cells to LPS challenge ex vivo. Lipidomics identified dietary palmitic acid (C16:0; PA) may be acting as a primary inflammatory stimulus in our model. Here we show PA induces a hyper-inflammatory response to LPS challenge in cultured macrophages and in vivo, correlating with increased endotoxemia mortality and enhanced resistance to C. albicans infection in RAG-/- mice. Our study identifies PA is an inducer of trained immunity that leads to a hyper-inflammatory response to secondary heterologous stimuli, and is deleterious during systemic inflammation, but enhances resistance to infection.

Diets enriched in saturated fatty acids increase endotoxemia severity and mortality

We have previously found that mice fed Western Diet (WD) showed increased disease severity and 88 mortality in a model of endotoxemia, independent of the WD-dependent microbiome or associated weight gain

(1). Considering the WD is enriched in dietary saturated fatty acids (SFAs), which have been shown to 90 enhance production of inflammatory cytokines from innate immune cells in vitro (14) (15) (16) (17) . Thus, we sought to 91 understand if enriched dietary SFAs were sufficient to drive enhanced endotoxemia severity and mortality in 92 vivo.

To examine the immune effects of chronic exposure to diets enriched in SFAs on endotoxemia, we fed 94 mice either a WD (enriched in SFAs and sucrose), a ketogenic diet (KD; enriched in SFAs and low-95 carbohydrate), or standard chow (SC; low in SFAs and sucrose), for 2 weeks (wk) prior to endotoxemia 96 induction (Table S1 ). We defined 2 wk of feeding as chronic exposure, because this is correlated with WD-or 97 KD-dependent microbiome changes, and confers metaflammation in WD mice (1), sustained altered blood 98 glucose levels (Fig S1A) , and elevated levels of the ketones Acetoacetate (AcAc) (Fig S1B) in the urine and β -99 hydroxybutyrate (BHB) in the blood of KD-fed mice ( Fig S1C) (18). We then induced endotoxemia by a single 00 intraperitoneal (i.p.) injection of LPS in order to induce pathophysiology that resembles symptoms of acute 01 septic shock in humans, including systemic arterial hypotension and increased circulating levels of TNF and IL-02 6 (19). We measured temperature loss, or hypothermia, as a measure of disease severity and survival to 03 determine outcome (1, 20, 21) . WD-and KD-fed mice showed significant and prolonged hypothermia, starting 04 at 10 hours (h) post-injection (p.i.), compared to the SC-fed mice that experienced mild and transient 05 hypothermia (Fig 1A) . In accordance with these findings, WD-and KD-fed mice displayed 100% mortality by 26 06 h p.i. compared to 100% survival of SC-fed mice ( Fig 1B) . LPS-induced hypoglycemia is a known driver of 07 endotoxemia mortality, and each of these diets has varying levels of sugars and carbohydrates (Table S1 ) (22, 08 23). However, mice in all diet groups displayed similar levels of hypoglycemia during disease (Fig S1D) ,

indicating that potential effects of diet on blood glucose were not a driver of enhanced endotoxemia severity.

Mice fed KD experience a shift towards nutritional ketosis, a metabolic state regulated by the liver when 11 blood glucose levels are low. During ketosis endogenous and exogenous FAs are used to synthesize the ketone bodies acetoacetate, β -hydroxybutyrate, and acetate, which are then distributed to other tissues for 13 energy (24). Thus, we wanted to understand if our phenotype was dependent on nutritional ketosis. Thus, mice 14 were fed for 2 wk (chronic exposure) with KD, SC supplemented with saccharine and 1,3-butanediol (SC + 15 BD), a compound that induces ketosis independent of diet (18), or SC-fed with the saccharine vehicle solution 16 as a control (SC + Veh). Next, we induced endotoxemia and found KD-fed mice showed significantly greater 17 temperature loss, and a significant survival defect compared to SC + BD-fed (SC + BD), and SC + Veh-fed 18 mice (SC + Veh) (Fig S1E, F) . Though not significant when compared to SC + Veh, the SC + BD mice did 19 confer an increase in hypothermia and decrease in survival suggesting that nutritional ketosis may play a minor 20 role in KD-dependent susceptibility to LPS lethality (Fig S1E, F ). Together these data suggest that diets 21 enriched in SFAs promote enhanced endotoxemia severity and this is independent of diet-dependent 22 hypoglycemic shock or nutritional ketosis.

Diets enriched in SFAs induce a hyper-inflammatory response to LPS and increased immunoparalysis 25 Endotoxemia mortality results exclusively from a systemic inflammatory response, characterized by an 26 acute increase in circulating inflammatory cytokine levels (ex: TNF, IL-6, and IL-1β) from splenocytes and 27 myeloid derived innate immune cells (monocytes and macrophages) (5, (25) (26) (27) . Pre-treatment of monocytes 28 and macrophages with dietary SFAs has been shown to enhance inflammatory pathways in response to 29 microbial ligands, including IL-1β and TNF expression and protein levels (15, 28, 29) . Considering this, we 30 hypothesized that exposure to enriched dietary SFAs within the WD and KD would enhance the inflammatory 31 response to systemic LPS during endotoxemia. Thus, we induced endotoxemia and measured the differences 32 in the systemic inflammatory response via expression of inflammatory cytokines in the blood (tnf, il-6, and il-1β) 33 every 5 h from 0 -20 h p.i. At 5 h p.i., mice fed all diets showed induction of tnf, il-6, and il-1β expression in the 34 blood (Fig 1C-E) . However, WD-and KD-fed mice experienced significantly higher expression of tnf, il-6, and 35 il-1β in the blood at 5 h p.i., compared with SC-fed mice (Fig 1C-E) , indicating that diets enriched in SFAs are 36 associated with a hyper-inflammatory response to LPS.

Sepsis patients often present with two immune phases: an initial amplification of inflammation, followed-38 by or concurrent-with an induction of immune suppression (immunoparalysis), that can be measured by a 39 systemic increase in the anti-inflammatory cytokine 31) . Interestingly, there was significantly increased il-10 expression in WD-and KD-fed mice at 20 h p.i., compared to SC-fed mice ( Fig 1F) . Further, in 41 septic patients, a high IL-10:TNF ratio equates with the clinical immunoparalytic phase and correlates with 42 poorer sepsis outcomes (32, 33). Here we saw, WD-and KD-fed mice had significantly higher il-10:tnf ratios at 43 10 and 15 h compared to SC-fed mice ( Fig 1G) . These data conclude that mice exposed to diets enriched in 44 SFAs show an initial hyper-inflammatory response to LPS, followed by an increased immunoparalytic 45 phenotype, which correlates with enhanced disease severity, similar to what is seen in the clinic.

Diets enriched in SFAs drive enhanced responses to systemic LPS independent of the microbiome

We have previously shown that WD-fed mice experience increased endotoxemia severity and mortality 49 independent of the microbiome (1). In order to confirm the increases in endotoxemia severity and mortality that 50 correlated with KD were also independent of KD-associated microbiome changes, we used a germ free (GF) 51 mouse model. Male and female C57BL/6 GF mice were fed SC, WD, and KD for 2 wk followed by injection 52 with 50 mg/kg of LPS, our previously established LD 50 for C57BL/6 GF mice injected with LPS (1). As we saw 53 in the conventional mice, at 10 h p.i. WD-and KD-fed GF mice showed significant loss of body temperature, 54 compared to SC-fed GF mice, indicating enhanced disease severity (Fig 2A) . Additionally, WD-and KD-fed GF 55 mice also displayed 100% mortality compared to only 50% mortality of SC-fed GF mice ( Fig 2B) . These data 56 indicate that, similar to WD-fed mice, KD-associated increase in endotoxemia severity and mortality is 57 independent of the microbiome.

Additionally, we next wanted to confirm that the hyper-inflammatory response to systemic LPS was 59 independent of the WD-and KD-dependent microbiome, we measured systemic inflammation during 60 endotoxemia via the expression of tnf, il-6, and il-1β in the blood at 0-10 h p.i. Similar to what we saw in 61 conventional mice, WD-and KD-fed GF mice displayed significantly enhanced expression of tnf, il-6, and il-1β 62 at 5h, compared to SC-fed GF mice (Fig 2C-E) . Interestingly, il-10 expression and il-10:tnf were not 63 significantly different throughout all diets, suggesting the SFA-dependent enhanced immunoparalytic 64 phenotype is dependent on the diet-associated microbiomes in WD-and KD-fed mice (Fig 2F, G) . Together, 65 these data suggest that the early hyper-inflammatory response, but not the late immunoparalytic response, to 66 LPS associated with enriched dietary SFAs is independent of the diet-dependent microbiota.

It has been shown that circulating monocytes and splenocytes are necessary for induction of systemic 70 inflammatory cytokines during endotoxemia (26, 27) . Additionally, we see feeding a diet enriched only in SFAs 71 (KD) leads to enhanced expression of tnf and il-6 in the blood during endotoxemia (Fig 1C, D) . However, it 72 remains unclear if the KD induces in vivo reprogramming of monocytes and splenocytes leading to an 73 enhanced response to LPS. Thus, we next sought out to determine if the chronic exposure to KD alters the 74 response of monocytes and splenocytes to LPS ex vivo. First, we fed mice SC or KD for 2 wk (chronic 75 exposure), isolated bone marrow monocytes (BMMs) from the femurs and tibias of mice and determined 76 homeostatic inflammation of monocytes via expression of tnf and il-6. We found that prior to ex vivo LPS 77 stimulation, BMMs isolated from mice chronically exposed to SC-or KD showed no significant difference in tnf 78 expression, and il-6 expression was significantly decreased in BMMs from KD-fed mice ( Fig 3A) . However,

when BMMs were stimulated with LPS for 2h ex vivo, those from KD-fed mice showed significantly higher 80 expression of tnf and il-6 ( Fig 3A) . These data suggest that chronic exposure to KD does not enhance basal 81 inflammatory status, but reprograms BMMs to respond with enhanced inflammation to LPS.

Similarly, we isolated splenocytes from SC-and KD-fed mice and found no difference between 83 homeostatic inflammation of splenocytes between diets (0h), but an enhanced production of tnf in the 84 splenocytes of KD-fed mice challenged with LPS, compared to splenocytes from SC-fed mice ( Fig 3B) .

Together, these data suggest that BMMs and splenocytes from KD-fed mice are not inherently more Our data show that diets enriched in SFAs correlate with a hyper-inflammatory response to LPS in vivo 91 and in ex vivo monocytes and splenocytes. We next wanted to identify target dietary SFAs enriched in the 92 blood of mice that may be altering the host inflammatory response to LPS. It is known that the SFAs consumed 93 in the diet determine the SFA profiles in the blood (35) (36) (37) . Considering this, we used mass spectrometry 94 lipidomics to create diet-dependent profiles of circulating fatty acids in SC-and KD-fed mice (38). Mice were 95 fed SC or KD for 2 wk, then serum samples were collected via cardiac puncture and analyzed using qualitative tandem liquid chromatography quadrupole time of flight mass spectrometry (LC-QToF MS/MS). We used 97 principal component analysis (PCA) to visualize how samples within each data set clustered together 98 according to diet, and how those clusters varied relative to one another in abundance levels of free fatty acids 99 (FFA), triacylglycerols (TAG), and phosphatidylcholines (PC). For all three groups of FAs, individual mice 00 grouped with members of the same diet represented by a 95% confidence ellipse with no overlap between SC-01 and KD-fed groups (Fig 4A-C) . These data indicate that 2 wk of KD feeding is sufficient to significantly alter 02 circulating FFAs, TAGs, and PCs, and that SC-and KD-fed mice display unique lipid blood profiles. Similarly, 03 the relative abundance of sphingolipids (SG) in SC-and KD-fed mice displayed unique diet-dependent profiles 04 with no overlapping clusters ( Fig S2A) . Though the independent role of each FFA, TAG, PC, and SG species 05 has not been clinically defined, each are classes of lipids that when accumulated is associated with metabolic 06 diseases, which have been shown to enhance susceptibility to sepsis and exacerbate inflammatory disease 07 (39-42).

Importantly, we identified a significant increase in multiple circulating FFAs within the KD-fed mice, 09 compared to the SC-fed mice, including a significant increase in free palmitic acid (PA; C16:0), a SFA that is 10 found naturally in animal fats, vegetable oils, and human breast milk (43), and is enriched in both WD and KD 11 (Table S1 ). Additionally, PA-containing TAGs and PCs were significantly elevated in KD-fed mice serum, 12 compared to SC-fed mice (Fig 4D-G) . These data indicate that KD feeding not only enhances levels of freely 13 circulating PA, but also enhances the frequency PA is incorporated into other lipid species in the blood.

14 Further, these data show that the KD induces significantly altered serum lipid profiles in mice within two weeks, 15 and KD-dependent circulating fatty acids may have the potential to reprogram innate immune pathways that 16 lead to a hyper-inflammatory response to LPS challenge.

Palmitic acid enhances macrophage response to LPS challenge

We have found free PA and PA-saturated lipids are significantly up-regulated in the blood of KD-fed 20 mice, and both WD and KD are enriched in PA with 12% and 23% of total kcal respectively, whereas the SC 21 contains only 3% PA (Table S1 ). Additionally, many groups have shown that PA induces expression and 22 release of inflammatory cytokines in macrophages and monocytes (28, 44) . Thus, we hypothesized PA may be 23 the SFA mediating KD-dependent hyper-inflammatory response to LPS. We next wanted to determine if pre-exposure to physiologically relevant concentrations of PA altered macrophage response during a secondary 25 challenge with LPS. Serum PA levels can differ between people, and depend not only on fasting and 26 postprandial states, but also metabolic health of the individual (45, 46) . Current literature indicates a wide 27 range of serum PA levels, between 0.7 and 3.6 mM, reflect a high-fat diet in humans (47-50). We aimed to use 28 a physiologically relevant concentration of PA for our in vitro studies, and decided on 1mM in order to stay 29 within a physiological range and avoid high levels of cytotoxicity in our macrophage model. Thus, we treated 30 primary bone marrow-derived macrophages (BMDMs) with and without 1mM of PA for 12h, removed the 31 media, subsequently treated with LPS (10 ng/mL) for an additional 24 h, and measured expression and release 32 of TNF, IL-6, and IL-1β. We found that BMDMs pre-treated with 1mM of PA for 12 h and then with LPS 33 expressed significantly higher levels of tnf and il-6, compared to naïve BMDMs treated with LPS ( Fig 5A, B) . il-34 1β expression was significantly lower in cells pre-treated with 1mM PA, suggesting a bifurcation in the 35 temporal transcriptional regulation of tnf/il-6 and il-1β by PA ( Fig 5C) . However, secretion of TNF, IL-6 and IL- 

Considering 1mM concentration of PA reflects the higher range of physiologically relevant serum levels, 42 we wanted to challenge BMDMs with a concentration of PA reflected in the lower range of physiologically 43 relevant serum levels. Thus, we treated BMDMs with and without 0.5mM of PA for 12 h or 24 h, removed the 44 media, subsequently treated with LPS (10 ng/mL) for an additional 24 h, and measured expression and 45 secretion of TNF, IL-6, and IL-1β. We found that 12 and 24 h pre-treatment with 0.5mM of PA induced 46 significantly higher expression of tnf and il-6, and il-1β after 24 h challenge with LPS, compared to naive 47 BMDMs treated with LPS ( Fig S4A-F) . Thus, lower physiological levels of PA enhance production and 48 secretion of inflammatory cytokines during secondary LPS challenge, further defining a novel role for PA in 49 regulating a hyper-inflammatory response to a subsequent challenge with a microbial ligand.

Pentadecanoic acid (PDA, C15:0) is a SFA containing 1 less carbon than PA and is likewise found in 51 milkfat, which is the primary fat source in the KD used in our in vivo studies. To understand if this hyper-inflammatory response is specific to PA, we performed identical assays using a physiologically relevant 53 concentration of PDA (50μM) followed by a secondary challenge with LPS (50). In accordance with our PA-54 treated BMDMs, 12 and 24 h PDA treatments resulted in significantly increased expression of tnf, il-6, and il-1β 55 upon secondary stimulation with LPS ( Fig S4G-L) . This indicates that like PA, PDA alters macrophage 56 responses and leads to significantly enhanced expression of inflammatory cytokines when challenged with 57 LPS, suggesting this SFA-dependent regulation of macrophage response to LPS is not specific to PA.

Together, these data show dietary SFAs common to both WD and KD have the capacity to alter the induction 59 of inflammatory cytokines within macrophages upon secondary stimulation.

Palmitic acid is sufficient to increase endotoxemia severity and mortality.

Considering the drastic effect of PA on macrophage response to secondary LPS challenge, we next 63 wanted to understand if PA is sufficient to induce a hyper-inflammatory response to LPS in vivo. We answered 64 this question by first mimicking post-prandial systemic PA levels (1mM) by a single i.p. injection of ethyl 65 palmitate and then challenging with LPS (i.p.) (51). Thus, mice were fed SC for 2 wk and then injected with a 66 vehicle solution or ethyl palmitate (51). We rested the mice 12 h and then induced endotoxemia. PA-treated 67 mice experienced increased endotoxemia severity as indicated by their significant decline in temperature 68 compared to Veh mice ( Fig 6A) . Similar to WD-and KD-fed mice, PA-treated mice also exhibited 100% 69 mortality, compared to 20% mortality seen in Veh mice ( Fig 6B) . Importantly, mice injected with PA for shorter 70 time periods (0, 3, and 6 h) and then challenged with LPS did not exhibit increased disease severity or poor 71 survival outcome ( Fig S5A, B ), concluding that a 12 h pre-treatment with PA is required for an increase in 72 disease severity. Next, we measured systemic inflammatory status during endotoxemia via the expression of 73 tnf, il-6, il-1β, and il-10 in the blood between 0 and 20 h p.i. We found, similar to WD-and KD-fed mice, the 12 74 h PA-pre-treated mice showed significantly enhanced expression of tnf and il-6 5 h post-LPS challenge, 75 compared to Veh control (Fig 6C, D) . Expression of il-1β was moderately up-regulated in 2 of 3 12 h PA-pre-76 treated mice, compared to Veh-treated mice ( Fig 6E) . Thus, a 12 h pre-treatment with PA is sufficient to drive 77 enhanced disease severity in mice challenged with LPS and that this PA-specific effect is dependent on length 78 of exposure.

Enriched dietary PA induces trained immunity and resistance to fungal infection 81 Our data show that PA enhances endotoxemia severity in vivo, and enhances inflammatory responses 82 of macrophages to a secondary and heterologous stimulus (LPS) in vitro. This form of regulation resembles 83 trained immunity, which is described as "non-antigen specific innate immune cell memory"; however, it remains 84 unclear if PA is inducing innate immune cell memory by priming or trained immunity. Priming occurs when the 85 first stimulus enhances transcription of inflammatory genes and does not return to basal levels before the 86 secondary stimulation (52). In contrast, trained immunity occurs when the first stimulus changes transcription 87 of inflammatory genes, the immune status returns to basal levels, and challenge with a heterologous stimulus 88 enhances transcription of inflammatory cytokines at much higher levels than those observed during the primary 89 challenge (52). Thus, we evaluated the basal level expression of tnf, il-6, and il-1β in mice treated with 1mM of 90 PA or Veh i.p. for 12h, before stimulation with LPS. Interestingly, we did not see significant differences in tnf, il-91 6, or il-1β expression at 12 h p.i. with PA ( Fig 6F) , which suggests that circulating immune cells of these mice 92 were not in a primed state at these time points prior to LPS injection. Thus, we conclude that PA induces 93 trained immunity, and not priming, however the time point of initial inflammation induced by PA remains 94 unknown and most likely will be different for each inflammatory cytokine. Importantly, as a control we looked at 95 LPS-induced hypoglycemia in PA-treated mice, and 12 h pre-treatment with PA did not alter LPS-induced 96 hypoglycemia (Fig 6G) , indicating that low blood glucose was not a driver of endotoxemia severity in 12 h PA 97 mice.

Canonical inducers of non-antigen specific innate immune cell memory (e.g., BCG or β -glucan) induce 99 long-lived enhanced innate immune responses to secondary inflammatory stimuli (9, 53). Considering WD 00 followed by a reversion to SC has been shown to reprogram monocyte precursors in atherosclerotic mice long-01 term (2), we hypothesized that exposure to PA feeding would reprogram the inflammatory response in vivo and 02 that this program would persist even after mice were "rested". In order to determine if PA alone can induce 03 long-lived trained immunity, we injected SC mice with a vehicle solution or 1mM of PA i.p. once a day for 9 04 days (to mimic 1 high-fat meal per day) and then rested the mice for 1 wk (Veh or PA SC). When challenged 05 with systemic LPS, PA SC showed an increase in endotoxemia severity and mortality compared to Veh SC 06 mice (Fig 6H-I) , indicating that PA alone can induce long-lived immune memory.

Lastly, the most commonly studied models for inducing trained immunity are immunization with BCG or 08 with β -glucan. These models of trained immunity have been shown to protect mice from systemic Candida 09 albicans infection via lymphocyte-independent epigenetic alterations that lead to decreased kidney fungal 10 burden (54). Thus, we next wanted to test if PA treatment induces lymphocyte-independent clearance of C. 11 albicans infection. For these experiments, we treated in female Rag knockout (Rag -/-) mice with or without PA 12 for 12 h and subsequently infected i.v. with 2x10 6 C. albicans. In accordance with canonical trained immunity 13 models, mice treated with PA for 12 h showed a significant decrease in kidney fungal burden compared to Veh 14 mice, 24 h post-infection ( Fig 6J) . These data find that PA is sufficient to induce lymphocyte-independent 15 trained immunity in vivo and enhances host resistance to systemic C. albicans infection. 

In this study we showed WD-and KD-fed mice experience increased endotoxemia disease severity that 27 correlates with an acute hyper-inflammatory response to LPS treatment, and poor survival outcome compared 28 to SC-fed mice (Fig 1) . Changes in blood glucose levels were not significant between WD-, KD-and SC-fed 29 mice during disease, and artificially inducing ketosis in mice did not recapitulate the same results as KD 30 feeding, indicating the KD-associated disease phenotypes are independent of hypoglycemic shock and ketosis 31 ( Fig S1) . Furthermore, we repeated our experiments in GF mice fed the same diets, and showed that 32 increased disease severity, enhanced inflammation, and poor survival still occurred in GF WD-and KD-fed 33 mice, indicating that SFA-dependent disease phenotypes are also independent of diet-associated microbiota 34 (Fig 2) .

Immunoparalysis during sepsis is associated with high mortality rates in humans, and is indicated by 36 enhanced tnf:il-10 ratios in the blood (32, 33). Following the initial hyperinflammatory response to LPS-induced 37 endotoxemia, WD-and KD-fed mice show significantly higher tnf:il-10 ratios in their blood compared to SC-fed 38 mice, indicating enhanced immunoparalysis. Interestingly, we found WD-and KD-fed GF mice did not show 39 significant alterations in blood tnf:il-10 ratios compared to SC-fed GF mice, and we conclude that SFA-40 dependent enhanced immunoparalysis is dependent on the microbiome (Fig 1G, 2G) . It is important to explore 41 other dietary constituents of SFA-enriched diets that may be altering the microbiota to drive immunoparalysis in 42 vivo.

Other studies have highlighted the role of enriched dietary SFAs in driving disease severity. WD-fed 44 atherosclerotic mice show long-term alterations to monocyte precursors in the bone marrow, and enhanced 45 innate immune responses to LPS treatment (2). Additionally, short-term (3 day) KD-fed mice also exhibit 46 enhanced endotoxemia-induced death (5). Our study adds to these compelling discoveries, and is unique in 47 that we have found a specific dietary SFA associated with WD and KD that may be underlying immune 48 regulation prior to and during disease. We used a lipidomics approach to determine the fatty acid profiles of 49 blood in KD-fed mice, and showed significantly altered serum lipids compared to SC-fed mice. Interestingly, we 50 found that many of the augmented lipids (TAGs, PCs, and SGs) were saturated with palmitic acid (C16:0; PA) 51 (Fig 4, S2) . PA is one of the most abundant SFAs in human serum, and has been found to be elevated in the 52 blood of those with metabolic syndrome, diabetes and obesity (50, (59) (60) (61) (62) . This finding, along with our ex vivo 

Interestingly, physiologically relevant levels of a similarly sized SFA, pentadecanoic acid (PDA), also 60 enhanced inflammatory cytokines during LPS challenge in vitro (Fig S4) . Plasma levels of PDA are considered 61 a biomarker for dietary intake of milkfat (63), and similar to PA, PDA is common to both WD and KD. Our 62 findings indicate that SFA-dependent regulation of macrophage response to LPS is not specific to PA, and that WD and KD contain multiple SFAs that are capable of reprogramming macrophages, and enhancing the 64 response of inflammatory cytokines upon secondary stimulation with a microbial ligand. Therefore, WD and KD 65 may contain cocktails of SFAs that induce trained immunity, and increase susceptibility to inflammatory 66 diseases. Determining plasma concentrations of these immune modulating SFAs may be clinically useful as 67 biomarkers for inflammatory disease vulnerability, and nutritional therapeutic interventions may be beneficial in 68 preventing or treating disease.

It is possible that 1mM PA may be regulating levels of the Krebs Cycle metabolite, succinate, which has 70 been shown to stabilize the transcription factor, HIF-1 , and regulate IL-1β expression in macrophages (64). If

PA disrupts the Krebs Cycle, this could lead to the accumulation of metabolites such as succinate, itaconate, 72 or alpha-ketoglutarate, which are known to impact epigenetic markers associated with inflammatory regulation 73 of macrophages (65). While PA has been shown to modulate macrophage metabolism (28), the impact of 74 these alterations on the epigenome is unknown. The interplay between macrophage metabolism and 75 epigenetics will be important to consider in future trained immunity studies where PA serves as the primary 76 stimulus.

We found here that PA is capable of inducing non-antigen specific innate immune memory trained 78 immunity, which is known to enhance monocyte and macrophage responses to homologous and/or 79 heterologous stimuli (52). This would explain why diets high in PA correlate with hyper-inflammatory responses 80 to LPS in vivo, and why PA-treated BMDMs show enhanced expression of inflammatory cytokines upon LPS 81 treatment. Importantly, our work is the first to show that PA treatment followed by a 1wk resting period, both 82 induce long-term immune reprogramming that leads to significantly impaired survival in mice with endotoxemia 83 (Fig 6) . Our findings also align well with recent studies that suggest diets enriched in SFAs induce trained 84 immunity in an atherosclerotic mouse model, and non-microbial stimuli induce trained immunity in human 85 monocytes (2, 12, 13).

Our findings here also align with the growing body of evidence indicating that trained immunity is a 87 double-edged sword, where the phenomenon can be beneficial for resistance to infection, but detrimental in 88 the context of inflammatory diseases (10). We conclude that PA-induced trained immunity may exacerbate the 89 acute phase of sepsis and contribute to tissue damage brought on by enhanced inflammation. However, we 90 are aware that trained immunity is a key feature of BCG vaccination, which has been shown to enhance innate immune responses to subsequent infections, and may be responsible for increased resistance to severe 92 66) . Strikingly, we show here that RAG -/mice injected with PA 12 h prior to C. albicans infection 93 show significantly enhanced clearance of kidney fungal burden compared to Veh-injected RAG -/mice. This is 94 the first study to show that PA is capable of inducing non-antigen specific innate immune memory and 95 enhancing inflammation that can be beneficial or detrimental depending on the disease state, and completely 96 independent from mature lymphocytes. 

For all panels, *, p < 0.05; **, p < 0.01; ***, p < 0.001. n=3 per group. Veh + rest PA + rest

Western diet regulates immune status and the response to LPS-driven sepsis 02 independent of diet-associated microbiome

Western Diet Triggers NLRP3-Dependent Innate Immune Reprogramming

Site-Specific Reprogramming of Macrophage Responsiveness to Bacterial 06 Lipopolysaccharide in Obesity

Dyslipidemic Diet-Induced Monocyte "Priming" and Dysfunction in Non-Human 08

Primates Is Triggered by Elevated Plasma Cholesterol and Accompanied by Altered Histone 09

Opposing Effects of Fasting Metabolism on Tissue Tolerance in Bacterial and Viral 11 Inflammation

Induction of proinflammatory cytokines by 13 long-chain saturated fatty acids in human macrophages

Complement pathway amplifies caspase-11-dependent cell death and endotoxin-43 induced sepsis severity

Effects of aging on mortality, hypothermia, and 45 cytokine induction in mice with endotoxemia or sepsis

Lipopolysaccharide inhibition of glucose production through the Toll-like receptor-47 4, myeloid differentiation factor 88, and nuclear factor kappa b pathway

Depression of hepatic gluconeogenesis and the hypoglycemia of endotoxin 50 shock

Ketone bodies: a review of physiology, pathophysiology and application of monitoring to 52 diabetes

Current Murine Models of Sepsis

Participation of Monocyte Subpopulations in 56 Progression of Experimental Endotoxemia (EE) and Systemic Inflammation

Evidence that TLR4 Is Not a Receptor for Saturated Fatty Acids but Mediates

Lipid-Induced Inflammation by Reprogramming Macrophage Metabolism

Palmitate and Lipopolysaccharide Trigger Synergistic Ceramide Production in 64

Broad defects in the energy metabolism of leukocytes underlie immunoparalysis in 66 sepsis

Sepsis: Inflammation Is a Necessary Evil

Pro-versus anti-inflammatory cytokine profile in 70 patients with severe sepsis: a marker for prognosis and future therapeutic options

Anti-inflammatory 73 cytokine profile and mortality in febrile patients

Epigenetic programming of monocyte-to-macrophage differentiation and trained innate 75 immunity

Lipid and phospholipid fatty acid composition of 77 plasma, red blood cells, and platelets and how they are affected by dietary lipids: a study of normal 78 subjects from Italy, Finland, and the USA

Dietary-induced changes in fatty acid composition of human 80 plasma, platelet, and erythrocyte lipids follow a similar time course

Fatty acid composition of the diet: impact on serum lipids and atherosclerosis. Clin 82 Investig

Comprehensive analysis of phospholipids in the brain, heart, kidney, and liver: brain 84 phospholipids are least enriched with polyunsaturated fatty acids

Sphingolipids and phospholipids in insulin resistance and related 87 metabolic disorders

Changes in Plasma Free Fatty Acids Associated with Type-2 Diabetes

The Role of Ceramides in Insulin Resistance

The Role of Obesity in Sepsis Outcome among Critically Ill Patients: A 93

Retrospective Cohort Analysis

Biological and Nutritional Properties of Palm Oil and Palmitic Acid: Effects on Health

Unsaturated fatty acids prevent activation of NLRP3 inflammasome in human 25 monocytes/macrophages

Palmitate conditions macrophages for enhanced responses toward 27 inflammatory stimuli via JNK activation

Origins and evolution of the Western diet: health implications for the 21st century

Unbound free fatty acid levels in human serum

Serum Fatty Acids, Desaturase Activities and Abdominal Obesity -A Population

Based Study of 60-Year Old Men and Women

Fatty acid composition of serum lipids predicts the development of 35 the metabolic syndrome in men

Higher production of IL-8 in visceral vs. subcutaneous adipose tissue. Implication of 37 nonadipose cells in adipose tissue

A review of odd-chain fatty acid metabolism and the role of 39 pentadecanoic Acid (c15:0) and heptadecanoic Acid (c17:0) in health and disease

Succinate is an inflammatory signal that induces IL-1beta through HIF-1alpha. 42

Krebs Cycle Reborn in Macrophage Immunometabolism

Barillas-Mury, BCG vaccine protection from severe coronavirus 46 disease 2019 (COVID-19)

Extract and visualize the results of multivariate data analysis

ggplot2: Elegant graphics for data analysis

FactoMineR:an R package for multivariate analysis

PC(20:4_20:4) PC(p-40:6) / PC(o-40:7)A PC

2_18:2) PC(16:0_20:4) PC(p-40:3) / PC(o-40:4) PC(16:0_18:2) PC(18:0_22:5)B PC(18:0_22:6) PC(18:1_20:5) PC

:1) / PC(o-32:2) PC(18:1_18:1) PC(18:1_18:2) PC(16:0_20:3) PC(18:2_18:3)B PC 30:1 PC

PC(16:0_16:1) PC(18:0_22:5)A PC(16:0_18:1) PC(20:0_18:2) PC(18:2_20:4) PC(16:0_22:6) PC(14:0_20:4) PC

:1(d7)PC PC

2_18:3)A

PC(18:1_20:4) PC(16:0_22:5) PC(p-38:4) / PC(o-38:5)B PC(16:1_22:6) PC