Lipopolysaccharides (LPS) are endotoxins present in the outer membrane of Gram-negative bacteria. Elevated levels of LPS, called metabolic endotoxaemia, can exert detrimental health effects(Reference Hersoug, Møller and Loft1,Reference Cani, Amar and Iglesias2) . The pathological LPS effects are related to its capacity to activate metabolic cascades and trigger proinflammatory cytokine secretion(Reference Milan, Pundir and Pileggi3). Inflammation as a result of metabolic endotoxaemia has been a topic of intense debate among the scientific community, since it is considered a risk factor for obesity and other chronic diseases, such as insulin resistance, diabetes and CVD(Reference Hersoug, Møller and Loft1,Reference Michalski, Vors and Lecomte4,Reference Neves, Coelho and Couto5) . Moreover, high LPS concentrations were also associated with metabolic syndrome components(Reference Lassenius, Pietiläinen and Kaartinen6).
Increased intestinal permeability favours the occurrence of increased LPS translocation to the systemic circulation, due to factors such as high-fat diets consumption and dysbiosis with decreased intestinal bacterial diversity(Reference Cani, Bibiloni and Knauf7). The incorporation of LPS to chylomicrons may also contribute to plasma LPS concentration increase(Reference Ghoshal, Witta and Zhong8). Once in the circulation, endotoxin can bind to LBP (lipopolysaccharide binding protein). Due to its longer half-life compared with LPS, LBP is an important marker of plasma endotoxin concentration, as well as CD14 receptor (cluster of differentiation 14). The LPS–LBP complex associated with CD14 can generate an important metabolic impact, since they can activate inflammation via toll-like receptor 4 (TLR4), leading to the secretion of inflammatory cytokines(Reference Laugerette, Furet and Debard9). Interestingly, the proinflammatory effect of SFA is in part due to the ability to interact with TLR4 receptors(Reference Rocha, Caldas and Oliveira10).
It has been shown that the consumption of high-fat diets is associated with an increase in postprandial LPS concentration(Reference Harte, Varma and Tripathi11–Reference Vors, Pineau and Drai13). However, little is known about the role of different types of fatty acids on endotoxaemia modulation(Reference Hersoug, Møller and Loft1). Apparently, postprandial chylomicronaemia may increase the extra-hepatic exposure to LPS(Reference Ghoshal, Witta and Zhong8), and it may be affected by the quantity and quality of the fat consumed, evidencing a link between lipaemia and endotoxaemia(Reference Moreira, Teixeira and Alves14).
It has been suggested that a meal fatty acid profile, rather than its fat content, affects postprandial LPS plasma concentrations(Reference Milan, Pundir and Pileggi3). In addition, high LPS concentrations were observed after the consumption of SFA, whereas lower concentrations were observed after the ingestion of n-3 PUFA. Apparently, meal fatty acid profile may alter circulating endotoxin concentrations in a different way according to the type of fat consumed(Reference Lyte, Gabler and Hollis15).
Therefore, the purpose of this systematic review was to critically analyse studies that assessed the relationship between dietary fat quality and metabolic endotoxaemia in humans since the characterisation of the effect of specific nutrients on endotoxaemia can help in the identification of nutritional strategies capable to prevent or treat the damage caused by endotoxaemia and associated co-morbidities.
Methods
Protocol and registration
This systematic review was planned and conducted according to the PRISMA recommendations(Reference Moher, Liberati and Tetzlaff16) and registered in the International Prospective Register of Systematic Reviews – PROSPERO (CRD: 42018104349).
Literature search
The search of the original articles included in this review was performed on 10 February 2019, by two authors (T. L. N. C. and L. E. S.) independently in three electronic databases: MEDLINE (PubMed, www.pubmed.com), Cochrane (www.cochrane.org) and Scopus (www.scopus.com).
Keywords were chosen from the Medical Subject Headings, and the following search strategy was used: (endotoxemia OR endotoxins OR lipopolysaccharides OR lipopolysaccharide-binding protein OR bacterial endotoxin) AND (fatty acids OR saturated fatty acids OR monounsaturated fatty acids OR polyunsaturated fatty acids OR high fat diet OR omega-6 fatty acids OR omega-3 fatty acids OR dietary fat OR dietary fat unsaturated) AND humans AND (epidemiologic studies OR population-based OR survey OR representative OR cross-sectional OR case-control studies OR observational OR clinical trials OR double-blind method OR comparative study) NOT reviews. The search strategy was not restricted by date and included studies published in English, Spanish and Portuguese.
Eligibility criteria and data extraction
The eligibility criteria were applied independently for all included studies, and divergent opinions were settled by consensus. We selected original studies that met the following inclusion criteria: (1) dietary intervention in which a high-fat diet/meal was offered to the study participants; (2) authors described the amount and quality of the dietary fats offered in the diets/meals to the participants (saturated, monounsaturated and/or polyunsaturated fats); (3) LPS and/or LBP concentrations and responses to the dietary interventions were evaluated; (4) metabolic endotoxaemia markers were evaluated by limulus amoebocyte lysate or ELISA methods. We excluded animal and in vitro studies, LPS infusion assays, literature reviews, letters, comments, book chapters and abstracts.
For each study included, the following information was extracted: author’s name, year of publication, country where the study was performed, study purpose, sample size, participants’ sex, age and nutritional status, presence of chronic diseases, study duration, lipid profile of the diet offered to the experimental group, as well as results regarding LPS and/or LBP concentrations and its correlations with lipid-related and inflammation markers.
Quality, representativeness and risk of bias assessment
Quality assessment was conducted according to the Jadad score(Reference Jadad, Moore and Carroll17), using a five-point scale focused on evaluating the methodological quality and validity. Each study was selected by the random sequence generation and allocation concealment, blinding of participants and staff, and the participants withdrawal were evaluated. Studies with scores between 0 and 1 were considered of low quality, scores between 2 and 3 were considered of moderate quality and a score of 4 or higher were considered of high quality.
The representativeness was evaluated according to external validity. Thus, a study was considered representative if information such as eligibility criteria, sample size calculation, probability of error and power of the sample were presented.
The risk of bias was also assessed in each study included in this review, using the ‘Methods Guide for Effectiveness and Comparative Effectiveness Reviews’ criteria from the Agency for Healthcare Research and Quality(Reference Viswanathan, Ansari and Berkman18). Selection, performance, attrition, detection and reporting bias were evaluated. The studies were classified as: (1) low risk of bias, when more than 70 % of the questions were answered as ‘yes (low risk)’; (2) moderate risk of bias, when 40–69 % of the questions were answered as ‘yes (low risk)’ and (3) high risk of bias, when <40 % of the answers were ‘yes (low risk)’.
Data analysis
Due to the heterogeneity among the included studies, it was not possible to conduct a statistical meta-analysis. Thus, in accordance with the Cochrane handbook(Reference Higgins and Altman19), the authors opted to perform a systematic narrative review for the analysis of the compiled data. In order to present the results in a more comprehensive manner, the main characteristics and results of each study were described in tables and organised chronologically by year of publication.
Results
Study selection
We identified a total of 735 studies in the three searched databases, of which 112 were duplicates. From the 623 remaining studies, 590 were excluded after analysing the titles and abstracts. After reading the full text of the remaining thirty-three studies, eleven met all criteria of the systematic review. The main reasons for exclusion were: in vitro studies (n 298), animal studies (n 84), studies that administered LPS infusion (n 32), studies which did not evaluate LPS and/or LBP concentrations and responses to dietary interventions (n 53), lack of dietary intervention with offer of high-fat diet/meal to the subjects (n 83) and studies that did not report the amount and/or quality of the dietary fats in the diets/meals offered to the subjects (n 16). The reasons for studies exclusion are detailed in Fig. 1.
Flow chart of the study-selection process. LPS, lipopolysaccharide; LBP, lipopolysaccharide binding protein.
Description of included studies
All studies included in this review are randomised controlled trials, which evaluated the effect of SFA-, MUFA- and/or PUFA-rich diets on metabolic endotoxaemia markers. Eight out of the eleven included studies evaluated the postprandial LPS response exclusively, only two studies were long-term intervention studies (4 and 12 weeks) and one study evaluated the ratio of LBP:sCD14 during 8 weeks. It is noteworthy to emphasise that studies addressing this research subject are relatively recent and that the oldest article included in this review was published in 2010. The studies contained data on a total of 387 participants, 76 % male and 24 % female (Table 1). Three studies did not present information regarding the subject’s sex(Reference Deopurkar, Ghanim and Jay Friedman20–Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22).
Characteristics and main results of lipopolysaccharides (LPS) and lipopolysaccharide binding protein (LBP) concentrations in subjects who received diets containing different types of fats
CHO, carbohydrate; TLR4, Toll-like receptor; SOCS3, cytokine-3 signalling suppressor; HOMA, homeostatic model assessment; ΔTAG, change in TAG; ALA, α-linolenic acid.
* Data referring to the percentage of energy from lipids.
† Data referring to the percentage of energy from lipids in the intervention diet. Except for: Deopurkar et al. (Reference Deopurkar, Ghanim and Jay Friedman20); Clemente-Postigo et al. (Reference Clemente-Postigo, Queipo-Ortuño and Murri21); Perez-Herrera et al. (Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22); Moreira et al. (Reference Rocha, Caldas and Oliveira10), which refer to the centesimal composition of the fatty acids of the intervention diet.
‡ Age and mean BMI presented for each group: (a) conventional peanut; (b) peanuts with high oleic content and (c) biscuit.
§ Age and mean BMI presented for each group: (a) fat-rich diet SFA; (b) diet rich in MUFA; (c) low fat and rich in CHO complexes; (d) low-fat and complex CHO-rich diet, supplemented with n-3.
|| Age and mean BMI presented for each group: (a) not obese and (b) obese.
Nine studies included overweight and obese adults. The other two included obese children/adolescents(Reference Alayón, Rivadeneira and Herrera27) and elderly participants(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25). Furthermore, one trial included subjects with the metabolic syndrome(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24) and one trial included subjects with type 2 diabetes, hypertension or hyperlipidaemia(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25).
In most of the dietary interventions, the researchers offered the subjects high-fat diets containing the three types of fatty acids of interest for this review(Reference Laugerette, Alligier and Bastard12,Reference Moreira, Teixeira and Alves14,Reference Clemente-Postigo, Queipo-Ortuño and Murri21–Reference Alayón, Rivadeneira and Herrera27) . Meanwhile, two studies tested diets containing only SFA and PUFA(Reference Lyte, Gabler and Hollis15,Reference Deopurkar, Ghanim and Jay Friedman20) .
In the acute studies (1–6 h postprandial response), the fat content of the meals ranged from 35 to 69 % of the total energy(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22–Reference Alayón, Rivadeneira and Herrera27) , therefore above the 20–35 % range recommendation(28). Moreover, among the long-term studies, the duration varied from 4 to 12 weeks, with the diet fat content of the prescribed diets ranging between 30 and 38 % of the total energy intake(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24,Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25) . Although in two studies, the percentage of dietary fat offered to the subjects was not disclosed, the researchers provided the centesimal composition of the test meals(Reference Deopurkar, Ghanim and Jay Friedman20,Reference Clemente-Postigo, Queipo-Ortuño and Murri21) .
Quality, representativeness and risk of bias assessment
According to the Jadad score(Reference Jadad, Moore and Carroll17), only one study included in this review presented high methodological quality(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25), five presented moderate quality(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22–Reference López-Moreno, García-Carpintero and Jimenez-Lucena24) and five presented low quality(Reference Laugerette, Alligier and Bastard12,Reference Deopurkar, Ghanim and Jay Friedman20,Reference Clemente-Postigo, Queipo-Ortuño and Murri21,Reference Morandi, Fornari and Opri26,Reference Alayón, Rivadeneira and Herrera27) .
Only two studies were classified as representative of the target population. Regarding the risk of bias assessment, six studies were considered as having a low risk of bias and three were classified as having a moderate risk of bias. The major limitations were related to the selection bias(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference Deopurkar, Ghanim and Jay Friedman20–Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25,Reference Alayón, Rivadeneira and Herrera27) , performance bias(Reference Clemente-Postigo, Queipo-Ortuño and Murri21,Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22) and detection bias(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference Deopurkar, Ghanim and Jay Friedman20-Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25,Reference Alayón, Rivadeneira and Herrera27) (Table 2).
Type of study, geographic distribution, quality assessment and risk of bias of the selected studies
Results of individual studies
The mean plasma LPS activity ranged from 0·23 to 1·40 endotoxin units (EU)/ml at the baseline and from 0·28 to 1·70 EU/ml after the ingestion of SFA-rich test meals, demonstrating an increase in LPS postprandial concentrations in response to SFA intake when compared with the fasting state(Reference Laugerette, Alligier and Bastard12,Reference Deopurkar, Ghanim and Jay Friedman20,Reference Schmid, Petry and Walther23,Reference Alayón, Rivadeneira and Herrera27) . In the study reported by Laugerette et al. (Reference Laugerette, Alligier and Bastard12), the AUC values for LPS concentrations, during 8 weeks, were higher after the overfeeding period compared with the response to the same SFA-rich meal provided before the overfeeding period. Also, the intake of SFA-rich meals led to an increase in LPS concentrations compared with MUFA- and PUFA-rich meals(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference López-Moreno, García-Carpintero and Jimenez-Lucena24) .
There was considerable variability in the acute LPS concentration responses to MUFA-rich diets. In two studies, the researchers verified an increase on LPS concentrations after the consumption of a MUFA-rich meal(Reference Clemente-Postigo, Queipo-Ortuño and Murri21,Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25) . On the other hand, two clinical trials demonstrated a decrease on LPS concentrations after the ingestion of similar MUFA content meals(Reference Moreira, Teixeira and Alves14,Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22) . Regarding the effects of long-term MUFA-rich diet, López-Moreno et al. (Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25) showed a decrease on LPS concentrations after a 4-week intervention. However, the test diet in this study contained n-3 PUFA supplementation. It also presented lower fat (30 %) and higher carbohydrate (55 %) contents than the other diets tested by the researchers(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25). Thus, drawing conclusions about the effects of MUFA on LPS concentrations is not possible yet.
Regarding the consumption of PUFA-rich diets, the mean plasma LPS activity among the studies varied from 0·43 to 1·10 EU/ml at the baseline and from 0·26 to 0·70 EU/ml after the test meal ingestion. The assessment of the LPS response to PUFA-rich diets in two clinical trials demonstrated lower concentrations of that maker when compared with SFA(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15) . According to the results of the two studies mentioned above, both n-6 PUFA (conventional peanuts(Reference Moreira, Teixeira and Alves14)) and n-3 PUFA (fish oil(Reference Lyte, Gabler and Hollis15)) can reduce LPS concentrations in the bloodstream. Conversely, the ingestion of sunflower oil (PUFA) was associated with higher LPS concentrations(Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22). However, in that study, the oil was heated before being ingested. Therefore, changes in its chemical properties cannot be discarded, which may have compromised the results and influenced the LPS concentrations(Reference Perez-Herrera, Delgado-Lista and Torres-Sanchez22) (Table 3).
Effect of fat profile on endotoxaemia and main variables correlated with lipopolysaccharides (LPS) or lipopolysaccharide binding protein (LBP)
↑, Higher concentration; ↓, lower concentration; ADIPOQ, adiponectin, C1Q and collagen domain containing; ADIR2, nuclear protein complex; APS, antigen human prostate specific; CANX, calnexin; CARL, calreticulin; CAV1, caveolin protein coding type 1; EU, endotoxin units; iAUC, incremental AUC; IκBa, factor κB inhibitor; MCP-1, monocyte chemoattractant protein-1; ME2, NAD-dependent malic enzyme; MIF1, migration inhibitory factor macrophage; NRF2, nuclear factor erythrocytes derived from type 2; PDIA3, protein disulfideisomerase family A, member 3; PIK3C, protein coding phosphoinositide 3 kinase; PLIN, perilipin; sCD14, soluble cluster of differentiation 14; sVCAM, soluble vascular cell adhesion molecule-1; UCP2, uncoupling protein mithocondrial type 2; XBP1, X box binding protein.
a,b,c Unlike letters in the same column indicate significant differences between groups in the same study. Unlike letters on the same line indicate statistical difference between baseline and postprandial values.
* Regarding lipid in the highest amount in the intervention diet.
† Data are means and standard deviations, except for Moreira et al. (Reference Moreira, Teixeira and Alves14); the data are medians and interquartile ranges.
Regarding the LBP concentrations, López-Moreno et al. (Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25) demonstrated a postprandial (4 h) decrease in this marker after the consumption of a MUFA-rich meal. Similarly, another clinical trial reported lower LBP concentrations 1 h after ingestion of a fatty acid-balanced meal (31·5 % energy SFA, 35·0 % energy MUFA and 33·5 % energy PUFA) in comparison with the baseline values(Reference Morandi, Fornari and Opri26). However, the AUC values for LBP concentrations, during a 5 h postprandial period, remained unchanged(Reference Morandi, Fornari and Opri26). On the other hand, Laugerette et al. (Reference Laugerette, Alligier and Bastard12) demonstrated increased LBP:sCD14 ratio after the test meal, suggesting that variations related to LBP and sCD14 may be linked to pro-inflammatory LPS activity and low-grade inflammation.
Higher LPS concentrations showed a consistent correlation with increased plasma TAG among studies(Reference Moreira, Teixeira and Alves14,Reference Clemente-Postigo, Queipo-Ortuño and Murri21,Reference Schmid, Petry and Walther23) (Table 3). Furthermore, postprandial LPS concentrations presented positive associations with adipose tissue inflammation markers, such as, PPARg and IL-6, while presenting inverse correlations with adiponectin C1Q and collagen domain containing (ADIPOQ), perilipin (PLIN), calnexin (CANX), nuclear factor erythrocytes derived from type 2 (NRF2), X box binding protein (XBP1), uncoupling protein mitochondrial type 2 (UCP2) and NAD-dependent malic enzyme (ME2)(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24). Moreover, LBP concentrations demonstrated a direct association with caveolin protein coding type 1 (CAV1), nuclear protein complex (ADIR2), protein coding phosphoinositide 3 kinase (PIK3CA) and antigen human prostate specific (APS); in spite of being negatively associated with XBP1, calreticulin (CARL), CANX and protein disulfideisomerase family A, member 3 (PDIA3)(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24). With regard to peripheral mononuclear cells’ inflammation markers, the authors verified positive correlations between LPS and IKBA and MIF1 and negative correlations between LBP and NFκB. In addition, LPS-P-selectin and SVCAM were positively correlated with plasma inflammation markers(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24).
According to some authors(Reference Vors, Pineau and Drai13), postprandial IL-6 is correlated with fasting LBP in lean and obese volunteers, and others(Reference Laugerette, Furet and Debard9) observed IL-6 response after overfeeding with butter, cheese and almonds due to the ratio between LBP and sCD14 in plasma. Finally, López-Moreno et al. (Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25) verified positive correlations between fasting LPS concentrations and the IκB gene expression, in addition to correlations between LBP and monocyte chemoattractant protein-1 (MCP-1). Moreover, the authors demonstrated correlations between LPS and MCP-1 expression in peripheral mononuclear cells in the postprandial period(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25).
Discussion
The results from the studies included in this systematic review indicate that the dietary fat profile seems to modulate the LPS concentrations in the bloodstream. While the consumption of SFA-rich meals was associated with an increase postprandial in LPS concentrations(Reference Laugerette, Alligier and Bastard12,Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15,Reference Deopurkar, Ghanim and Jay Friedman20,Reference Schmid, Petry and Walther23–Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25,Reference Alayón, Rivadeneira and Herrera27) , PUFA-rich meals were associated with lower LPS concentrations(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15) , both in eutrophic and overweight individuals (Fig. 2). In a 4-week high-fat diet study involving animals, increased plasma LPS concentration was verified two to three times during the meal(Reference Cani, Amar and Iglesias2). However, the mechanisms involved between dietary fat profile and metabolic endotoxaemia in humans are not fully understood. Thus, here we aim to discuss the possible implications of dietary fatty acids on metabolic endotoxaemia (Fig. 3).
Postprandial effect of the intake of different fatty acid types (SFA and PUFA) on lipopolysaccharide (LPS) concentrations in eutrophic and overweight subjects. SFA response was obtained by mean values assessed at baseline, 4 h and 5 h postprandial. Delta values (Δ = final – baseline value) were obtained by the mean results obtained in all studies. (a) LPS concentrations in eutrophic subjects. (b) LPS concentrations in overweight subjects. (c) LPS concentrations delta (Δ = LPS final – LPS baseline) in eutrophic subjects after SFA and PUFA intake. (d) LPS concentrations delta (Δ = LPS final – LPS baseline) in overweight subjects after SFA and PUFA intake. (a and b) 
, MUFA; 
, SFA; 
, PUFA. (c and d) 
, SFA; 
, PUFA. EU, endotoxaemic units.
Possible mechanisms explaining changes in lipopolysaccharide (LPS) concentrations. ↑, Higher concentration; ↓, lower concentration; ROS, reactive oxygen species; CD14, cluster of differentiation 14; TLR4, Toll-like receptor; MD2, myeloid differential protein-2.
High-fat meal intake stimulates liver production of bile acids, which aid in the digestion and absorption of fats through micelles formation(Reference De Aguiar Vallim, Tarling and Edwards29). Due to its lipid A fraction, LPS is incorporated into the micelles in the intestinal lumen and, thus, the endotoxin is carried out to the enterocytes and incorporated into the chylomicrons(Reference Michalski, Vors and Lecomte4,Reference Ghoshal, Witta and Zhong8) . After a single exposure to a high-fat load, obese subjects, who had subtle impairment in barrier function, had a greater increase in the permeability of the small intestine compared with non-obese patients. That result suggests that fat themselves are capable of altering paracellular permeability, by directly damaging the tight junctions(Reference Genser, Aguanno and Soula30).
Dietary fatty acid profile seems to influence the chylomicron–LPS complex transport into the bloodstream. Therefore, considering that chylomicrons are formed after food intake, variations on postprandial LPS concentrations may be linked to lipaemia(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25).
Some authors observed that the ingestion of long-chain fatty acids, such as n-3 PUFA, resulted in lower postprandial TAG concentrations and, consequently, lower lipaemia(Reference López-Moreno, Garcia-Carpintero and Gomez-Delgado25,Reference Bravo, Napolitano and Botham31,Reference Oscarsson and Hurt-Camejo32) . These results can be attributed to the DHA and EPA ability to increase chylomicron clearance and reduce VLDL serum concentration(Reference Oscarsson and Hurt-Camejo32). Similarly, the ingestion of fish oil, rich in n-3 PUFA resulted in lower postprandial lipaemia, when compared with a SFA-rich oil mix(Reference Jackson, Zampelas and Knapper33) and rapeseed oil, rich in MUFA(Reference López-Miranda, Williams and Larion34). On the other hand, the ingestion of SFA lead to a higher lipaemia in comparison with both MUFA and PUFA intake(Reference Bravo, Napolitano and Botham31,Reference Botham and Wheeler-Jones35) . It is noteworthy that SCFA and medium-chain fatty acid are not incorporated into chylomicrons and therefore have a limited effect on postprandial lipaemia(Reference Mu and Høy36). However, more research must be carried out to clarify the effects of different fatty acids with different chemical structures on endotoxaemia.
According to a body of literature(Reference Moreira, Teixeira and Alves14,Reference Deopurkar, Ghanim and Jay Friedman20) , other dietary factors may influence the lipidaemic response and LPS concentrations. Deopurkar et al. (Reference Deopurkar, Ghanim and Jay Friedman20) demonstrated that the intake of orange juice, associated with a high-fat and high-carbohydrate diet, did not change the LPS concentrations in the bloodstream, neither oxidative stress and inflammation markers concentrations. These results can be attributed to the orange juice flavonoid content, since it can have antioxidant and anti-inflammatory effects(Reference Deopurkar, Ghanim and Jay Friedman20), as well as to the resveratrol and fibres, which may promote a hypolipidaemic effect, through bile acid reduction and sterol reabsorption(Reference Moreira, Teixeira and Alves14). In this regard, it is important to consider that fatty acids are inserted in a food matrix and the composition of a food and/or a meal as a whole can influence fat digestion and absorption, and consequently their effects on metabolic endotoxaemia(Reference Astrup, Bertram and Bonjour37). In addition, the fat structure can also impact postprandial absorption and endotoxaemia. Vors et al. (Reference Vors, Drai and Pineau38) demonstrated that the consumption of emulsified fat by obese men increased the transport of LPS by kilomicrons and led to a more efficient clearance when compared with fat spread. That effect if probably due to the formation of larger chylomicrons from emulsification.
The dietary fatty acid profile also seems to influence the apo composition, as well as the number and size of TAG-rich lipoproteins, such as the chylomicrons(Reference Botham and Wheeler-Jones35). According to Sakr et al. (Reference Sakr, Attia and Haourigui39), PUFA can form larger chylomicrons particles than SFA and these particles are hydrolysed more rapidly by lipoprotein lipase. In addition, the remaining PUFA and MUFA chylomicrons are absorbed faster by the liver than the SFA-enriched particles, in which the clearance is impaired due to the Apo C-III accumulation(Reference Botham, Avella and Cantafora40,Reference Bravo, Ortu and Cantafora41) . Therefore, larger PUFA-rich chylomicrons are purified more efficiently than small ones, rich in SFA, and may contribute to decrease LPS concentrations in the circulation and favour metabolic endotoxaemia control(Reference Xiang, Cianflone and Kalant42).
Another mechanism that seems to be involved in LPS modulation in response to dietary fat profile refers to bile acid metabolism. Evidence suggests a dynamic and robust interaction between diet, bile acids and intestinal microbiota. Intestinal bacteria participate in the conversion of primary bile acids into secondary ones. In turn, secondary bile acids can modify the bacterial community, by promoting the growth of bacteria capable to metabolise bile acids or exerting bactericide effects(Reference Ridlon, Kang and Hylemon43,Reference Wahlström, Sayin and Marschall44) . In this context, an experimental study in mice demonstrated that the consumption of an isoenergetic diet associated with high-SFA intake modified the composition of bile acids through taurine conjugation. The researchers were unable to show similar results when the consumption of the same diet was associated with PUFA intake. Higher concentrations of taurocholate, besides favouring Bilophila wadsworthia growth, a Gram-negative and sulphidogenic bacteria, which compromised intestinal barrier function, were observed with the consumption of a high SFA diet(Reference Devkota and Chang45). On the other hand, fish oil, rich in n-3 PUFA, supplementation inhibited the growth of B. wadsworthia, suggesting that the type of fatty acid ingested may play a role in the modulation of bile acid composition(Reference Devkota, Wang and Musch46).
Increased luminal bile concentrations can lead to higher intestinal permeability by suppressing tight junction proteins expression, and thus favouring paracellular transport of LPS(Reference Moreira, Texeira and Ferreira47). According to Willemsen et al. (Reference Willemsen, Koetsier and Balvers48), PUFA intake increases the intestinal epithelial barrier integrity, while the intake of saturated palmitic acid (C16 : 0) showed detrimental results. These results were associated with increased occludin expression, a tight junction protein, responsible for paracellular intestinal permeability regulation. Moreover, PUFA may influence lipid transportation through the intestinal barrier cells phospholipid membranes, by altering its structure, and therefore preventing the passage of endotoxins into the bloodstream(Reference Willemsen, Koetsier and Balvers48,Reference Bellenger, Bellenge and Escoula49) .
Although the studies included in this review showed a reduction in LPS concentrations after the ingestion of n-3 and n-6 PUFA(Reference Moreira, Teixeira and Alves14,Reference Lyte, Gabler and Hollis15) in humans, that same effect was not observed in animal studies(Reference Wahlström, Sayin and Marschall44). Kaliannan et al. (Reference Kaliannan, Wang and Li50) demonstrated that the consumption of n-3 PUFA increases the endogenous activity of intestinal alkaline phosphatase and decreases LPS production and intestinal permeability improvement, which resulted in decreased metabolic endotoxaemia(Reference Kaliannan, Wang and Li50). On the other hand, rats fed with n-6 PUFA-rich diets presented high concentrations of LPS and LBP(Reference Kaliannan, Wang and Li50). Similarly, the authors observed a reduction in metabolic endotoxaemia and inflammation markers concentration after fish oil supplementation for 2 months (n-3 PUFA)(Reference Kaliannan, Wang and Li50). Regarding the different SFA types, we observed that the studies included in this review predominantly used palmitic and lauric acids, found in palm oil, coconut oil, milk and derivatives and that the consumption of both increased LPS concentrations. However, different impacts on endotoxaemia can be observed due to the addition of other dietary components (as for example, emulsifiers) to oils rich in SFA(Reference Lecomte, Couedelo and Meugnier51). In a study in which rats were fed diets rich in palm oil, the addition of soya lecithin had no impact on endotoxaemia. On the other hand, the addition of milk phospholipids led to a reduction in endotoxaemia(Reference Lecomte, Couedelo and Meugnier51).
Dysbiosis is another factor that can contribute to increased passage of LPS to the circulatory system, since changes in the intestinal microbiota can change the diversity and abundance of bacteria, increasing Gram-negative bacteria and generating large amounts of endotoxin after bacterial lysis, in addition to affecting the integrity of the barrier(Reference Cândido, Valente and Grześkowiak52). In an experimental study with rats, the intake of a SFA-rich diet, reduced the intestinal barrier resistance and increased the abundance of H2S (sulphide acid) producing bacteria, such as Bilophila and Desulfovibrio (Reference Lam, Ha and Hoffmann53). On the other hand, Patterson and colleagues(Reference Patterson, Doherty and Murphy54) observed an increase in the number of bifidobacteria the after ingestion of n-3 PUFA. Increase in bifidobacteria may have a protective effect against endotoxaemia induced by a high-fat diet, since in rats fed a prebiotic-enriched high-fat diet (oligofructose), an increase in the amount of bifidobacteria was observed, which was negatively correlated with endotoxaemia(Reference Cani, Neyrinck and Fava55). The bifidogenic effect may be related to lower endotoxin concentrations in the bloodstream, since these bacteria cannot degrade glycoproteins mucus and preserve the barrier function(Reference Griffiths, Duffy and Schanbacher56).
At high concentrations in the bloodstream, LPS can activate inflammatory pathways signalling cascade, therefore favouring the development of chronic diseases(Reference Grunfeld and Feingold57). With the aid of proteins, such as LBP (LPS binding protein), CD14 and differentiating myeloid protein 2(Reference Lu, Yeh and Ohashi58), LPS binds to the TLR4(Reference Akira59) and induces cytokine and pro-inflammatory factors secretion, like NFκBp65 and IL-6(Reference Laugerette, Vors and Géloën60).
Likewise the LPS, SFA, such as lauric acid (C12 : 0), can promote the expression of pro-inflammatory factors via TLR4. These fatty acids can bind to CD14 and differentiating myeloid protein 2 and activate the TLR4 through the formation of CD14–TLR4–MD2, an inflammatory signalling complex(Reference Lee, Ye and Gao61). On the other hand, n-3 PUFA appears to exert an anti-inflammatory effect, mediated by the G protein-coupled receptor 120, due to its ability to inhibit the TLR4-induced signalling pathway(Reference Rocha, Caldas and Oliveira10,Reference Lee, Ye and Gao61) . Therefore, SFA seem to modulate the TLR4-induced inflammatory response, and this effect can be accentuated in the presence of LPS(Reference Rocha, Caldas and Oliveira10).
Furthermore, López-Moreno et al. (Reference López-Moreno, García-Carpintero and Jimenez-Lucena24) also observed a greater inflammatory response after the ingestion of SFA, and the researchers suggested that such outcome could be related to the postprandial increase in LPS concentrations. These findings are supported by positive correlations between LPS and postprandial gene expression of IkBa and MIF1 in peripheral mononuclear cells, both involved in inflammatory response regulation. In addition, a positive relationship was observed between LPS and the adhesion molecules P-selectin and VCAM, which could favour atherosclerosis development(Reference López-Moreno, García-Carpintero and Jimenez-Lucena24).
Limitations
Relatively few studies have evaluated the effect of the dietary fat profile on humans’ metabolic endotoxaemia. Although the acute LPS response studies have previously been explored, only Morandi et al. (Reference Morandi, Fornari and Opri26) evaluated the behaviour of LBP concentrations after dietary fat intake and used the AUC to analyse their data. Since the baseline values of LPS and LBP have not yet been established, we believe that the AUC is a more accurate method to identify and measure changes in these markers concentrations in response to dietary interventions.
It should be noted that the studies included in this review did not consider the different types of SFA and PUFA of the test meals/diets. Due to the heterogeneity of the studies in terms of the test diet, our comments were limited to the type of fatty acid present in greater quantity in the test meals. Another point to be considered is that the number of studies available at the moment is still very small for us to establish a strong conclusion.
Conclusions
Experimental studies and human clinical trials, evaluating the impact of specific nutrients on LPS concentrations, indicated that the dietary fatty acid profile might play a role in metabolic endotoxaemia modulation. According to the studies included in the present review, the intake of SFA-rich meals increased plasma LPS concentrations, while PUFA-rich meals reduce LPS concentrations. It is worth highlighting that such effects were observed only in studies evaluating the LPS postprandial response. On the other hand, these same results were not verified in long-term intervention studies included in our review. Therefore, SFA intake can be considered as a dietary risk factor for the development of metabolic endotoxaemia. In contrast, the intake of PUFA appears to exert a protective effect.
Finally, in order to decrease LPS concentrations in the bloodstream and consequently prevent chronic diseases associated with metabolic endotoxaemia development, like diabetes, obesity and arteriosclerosis, we believe that changes in the dietary fatty acid profile, such as lowering the intake of SFA-rich meals and increasing PUFA-rich meals ingestion may be a simple but effective strategy and it should therefore be recommended. However, the results need to be confirmed, given the small number of studies involving human subjects. Therefore, future studies on dietary fat quality and endotoxaemia are needed.
Acknowledgements
The authors thank the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG, Brazil), Coordenação de Aperfeiçoamento de Pessoal de Ensino Superior (CAPES, Brazil) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil).
The authors’ contributions are as follows: T. L. N. C., L. E. S. and J. F. T contributed to the study design, literature search, data analyses and writing the manuscript. R. C. G. A, A. C. M. C. and R. A. G. R. contributed to the interpretation of data, manuscript writing and editing of the article. All authors read and approved the final version of the manuscript.
The authors declare no conflicts of interest or financial from this study.
