Fructose Reprogramming of Glutamine-Dependent Oxidative Metabolism

Fructose and Inflammation

Abstract

Fructose consumption has increased in developed countries and is linked to obesity, type 2 diabetes, and non-alcoholic fatty liver disease.
The study investigates how fructose impacts immune cells like monocytes and macrophages.
Fructose reprograms metabolic pathways, increasing glutaminolysis and oxidative metabolism, which supports inflammatory cytokine production in LPS-treated human monocytes and mouse macrophages.
Fructose increases mTORC1 activity, driving the translation of pro-inflammatory cytokines in response to LPS.
Monocytes treated with fructose rely on oxidative metabolism and show reduced metabolic flexibility.
Mice exposed to fructose have elevated IL-1$\beta$ levels after LPS challenge, demonstrating a pro-inflammatory role for dietary fructose.
Fructose's pro-inflammatory effect occurs at the expense of metabolic flexibility.

Introduction

The innate immune system requires metabolic rewiring towards glucose metabolism.
Monocytes are exposed to different carbon sources, including fructose.
Fructose is metabolized via:
- Ketohexokinase, producing fructose-1-phosphate (liver, kidneys, intestines).
- Hexokinase (HK), converting it to fructose-6-phosphate (at a lower rate than glucose).
Fructose intake has increased due to sucrose and high fructose corn syrup consumption, worsening conditions like obesity and type 2 diabetes.
Chronic fructose consumption drives hepatic fructolysis and enhances lipogenic gene expression.
Physiological fructose levels range from $0.04$ to $0.2$ mM, but can exceed $1$ mM in hematological malignancies and reach up to $5$ mM in bone marrow microenvironments.
Alterations in the glucose to fructose ratio enable acute myeloid leukemia blasts to enhance fructose uptake.
Localized tissues like the liver and kidneys have elevated fructose metabolism.
The impact of elevated fructose on the immune system is underexplored.
Chronic fructose exposure in rats leads to a more inflammatory phenotype in bone marrow mononuclear cells.
LPS-stimulated dendritic cells produce more pro-inflammatory cytokines when cultured in fructose.
The study characterizes how human monocytes and mouse macrophages respond to fructose metabolically and functionally.
Activated mononuclear phagocytes show plasticity in engaging fructose metabolism, but become metabolically inflexible.
Fructose exposure favors glutaminolysis and oxidative metabolism, supporting an inflammatory phenotype.
A short-term high fructose diet promotes inflammation in vivo.

Results

Fructose Promotes an Oxidative Phenotype

The study investigated the metabolic response to fructose in activated human monocytes compared to glucose, galactose, or sugar withdrawal. Galactose promotes OXPHOS in T cells by reducing glycolysis.
Using a Seahorse Bioanalyzer, monosaccharides were injected, followed by LPS stimulation after 1 hour. ECAR and OCR were measured to assess glycolysis and OXPHOS.
Glucose-incubated monocytes: Robust increase in basal and LPS-induced ECAR (Fig. 1A).
Fructose or galactose-treated monocytes: Low baseline glycolysis levels, slightly greater than no sugar controls (Fig. 1A).
Modest increase in ECAR post-LPS exposure in fructose, galactose, or no sugar treatments (Fig. 1A).
Glucose-treated cells reduced OCR upon activation, consistent with a switch from OXPHOS to glycolysis (Fig. 1B).
Fructose, galactose, or no sugar: Initial burst of increased oxygen consumption, maintaining higher OCR (Fig. 1B), demonstrating metabolic flexibility towards OXPHOS.
After 2-DG introduction, ECAR was reduced in all conditions (Fig. 1C).
OCR increased in glucose-treated monocytes (compensatory response), while fructose, galactose, or no sugar showed decreased OCR (Fig. 1D).
LPS-stimulated monocytes treated with fructose or galactose direct pyruvate towards the mitochondrial TCA cycle for OXPHOS, while glucose directs pyruvate to lactate production.
Extracellular lactate levels are reduced in fructose compared to glucose-treated monocytes (Supplementary Fig. 1A).
LPS-stimulated human monocytes treated with fructose maintained an elevated oxidative phenotype with low ECAR, like galactose or no sugar treatment.
Glucose availability maintained elevated glycolysis levels at the expense of reduced oxygen consumption (Fig. 1E).
LPS-stimulated monocytes cultured with galactose or no sugar had reduced viability compared to glucose (Fig. 1F).
No difference in viability between glucose and fructose treatment (Fig. 1F).
Monocytes were incubated with uniformly labelled $^{13}C6$ -fructose or $^{13}C6$ -glucose and performed stable isotope tracer analysis (SITA).
Activated monocytes transported fructose into the cell (via GLUT5 expression - Supplementary Fig. 1B) and incorporated comparable carbon levels into intracellular lactate.
Increased incorporation into TCA cycle intermediates and amino acids compared to glucose (Supplementary Fig. 1C–E).
Cells can metabolize fructose carbon and use it in the TCA cycle.
Fructose treatment promotes a low glycolytic rate without compromising cell viability.
Demonstrates metabolic flexibility and the ability to utilize an alternative carbon source.

Glycolysis and OXPHOS are Coupled in Fructose-Treated Human Monocytes

To confirm that the observed increase in OCR in fructose-treated monocytes was due to OXPHOS (as opposed to an increase in other oxygen-consuming processes), the ATP synthase inhibitor, oligomycin, was utilized.
Monocytes were treated with either glucose or fructose and allowed to rest before LPS exposure.
Oligomycin was later injected, and the bioenergetic changes were monitored over time.
While the OCR of both glucose and fructose LPS-simulated monocytes decreased to the same level upon oligomycin treatment (Fig. 2A), the drastic reduction of OCR in fructose-cultured monocytes reflects a greater reliance on OXPHOS.
Cells reliant on OXPHOS may demonstrate metabolic flexibility upon oligomycin treatment by increasing glycolysis.
However, surprisingly, ECAR in fructose-cultured monocytes decreased in response to oligomycin, suggesting a lack of metabolic adaptation in these cells (Fig. 2B, C).
Secondly, to establish whether the elevated ECAR levels post-LPS treatment reflected glycolytic activity as opposed to other acidifying processes, a lactate dehydrogenase inhibitor (GSK2837808A; LDHi) was used.
The increased ECAR in response to LPS stimulation was reduced in glucose-treated monocytes upon LDHi treatment (Fig. 2D).
By contrast, LDHi modestly impacted ECAR in fructose-treated cells, arguing that fructose-mediated glycolysis is coupled to OXPHOS. The low level of ECAR under this condition is most likely due to acidification of the media by an alternative source to lactate (Fig. 2D).
This was not due to changes in LDH phosphorylation in fructose- versus glucose-treated cells (Supplementary Fig. 1F).
The corresponding OCR levels increased in the glucose-treated monocytes upon LDHi treatment as pyruvate is directed towards OXPHOS, again demonstrating the bioenergetic flexibility of human monocytes.
By contrast, OCR in fructose-treated monocytes remained unchanged for the duration of the assay (Fig. 2E, F).
These data further demonstrate that glycolysis and OXPHOS are tightly coupled in fructose-treated cells, revealing impaired metabolic flexibility in comparison to glucose-treated cells (Fig. 2C, F).

Fructose Treatment Enhances LPS-Induced Inflammation

Given the distinct metabolic characteristics of human monocytes exposed to fructose (Fig. 1A), the impact on monocyte function was investigated.
Fructose-treated monocytes produced elevated levels of a panel of secreted cytokines, namely interleukin-1$\beta$ (IL-1$\beta$), IL-6, IL-8, IL-10 and tumor necrosis factor (TNF) (Fig. 3A–E), with IL-1$\beta$, IL-8, IL-10 and TNF reaching statistical significance in comparison to glucose treatment.
Despite elevated levels of cytokine secretion, there were no differences in various surface markers associated with monocyte activation (HLA-DR, CD80, CD86, CD62L, CCR5 and CCR2) between the glucose- or fructose-treated monocytes (Supplementary Fig. 2A).
To determine whether the increased production of cytokines was a consequence of increased transcription, RNA-sequencing (RNA-seq) analysis of LPS-stimulated monocytes treated with glucose or fructose was performed.
Only five genes that were significantly changed between the two conditions were observed, with no alteration to transcript levels for the cytokines of interest (Supplementary Fig. 2B).
Consistent with this, the expression level of genes encoding cytokines was also comparable in monocytes treated with either glucose or fructose (Fig. 3F), suggesting that the increased cytokine production was not through transcriptional regulation.
Fructose has recently been reported to activate mTORC1 via dihydroxyacetone phosphate sensing20.
Consistent with this, phosphorylation of the downstream mTORC1 target, S6 ribosomal protein, was elevated significantly in LPS-stimulated monocytes treated with fructose (Fig. 3G).
No differences in the induction of AMPK (AMP-activated protein kinase) signaling between the two groups were observed (Supplementary Fig. 2C).
LPS stimulation in the presence of fructose enhances inflammatory cytokine production, in part, through increased mTORC1-mediated translation.

Fructose Treatment Drives a Sustained Oxidative Phenotype

LPS-stimulated monocytes in the presence of fructose clearly enhance oxygen consumption in the short term (Fig. 1C).
Next, it was determined whether heightened oxygen consumption was sustained long term in activated monocytes.
Here, using the Seahorse Bioanalyzer, a mitochondrial stress test with a final injection of the ionophore, monensin21, was performed following a 24-h incubation period with either glucose or fructose.
Elevated levels of oxygen consumption were indeed sustained long term in fructose-treated LPS-stimulated monocytes (Fig. 4A).
This was characterized by increased levels of oxidative parameters such as basal respiration, ATP-linked respiration, and a higher percentage of coupling efficiency in comparison to glucose (Fig. 4B–D and Supplementary Fig. 3A–D).
Corresponding ECAR levels were unsurprisingly higher in glucose-treated monocytes in comparison to fructose-treated cells (Fig. 4E, F and Supplementary Fig. 3E).
Fructose-treated monocytes had a significantly higher ratio of basal OCR/ECAR ratio in comparison to glucose, indicating their commitment to oxidative metabolism (Fig. 4G).
Next, the contributions of glycolysis-derived and OXPHOS-derived ATP production was assessed.
These analyses show glucose-treated monocytes as glycolytic cells functioning at their maximal glycolytic ATP production rate at baseline (Fig. 4H).
At maximal bioenergetic capacity, they demonstrate flexibility towards ATP production from OXPHOS but still derive the majority of their ATP from glycolysis and therefore remain classed as glycolytic cells (Fig. 4H and Supplementary Fig. 3F, G).
By contrast, fructose-treated monocytes are oxidative (deriving the majority of their ATP from OXPHOS) at baseline, and this phenotype is further exacerbated when cells are at maximal bioenergetic capacity (Fig. 4H and Supplementary Fig. 3F, G).
They do, however, have some metabolic scope to increase ATP production from glycolysis (Fig. 4H).
Differential mitochondrial properties might explain the heightened oxidative phenotype of fructose-treated monocytes.
However, no difference in mitochondrial content (MitoTracker Green), membrane potential (tetramethylrhodamine ethyl ester) or mitochondrial-derived reactive oxygen species (ROS) was observed (Fig. 4I).
Levels of the individual respiratory complexes were assessed by immunoblot.
Again, there was no difference in the abundance of complexes I–IV between fructose- or glucose-treated monocytes (Fig. 4J).
A high fructose diet has been shown to increase de novo lipogenesis in the liver10.
This correlates with increased mitochondrial ATP production, which may support this energy-demanding process.
A potential explanation for the elevated ATP-linked respiration observed is that fructose-treated monocytes are supporting a higher level of lipogenesis.
No differences in phosphorylation of enzymes that catalyse the citrate-derived fatty acid synthesis steps; ATP citrate lyase (ACLY) or acetyl-CoA carboxylase was observed (Supplementary Fig. 3H).
However, fructose-cultured, LPS-stimulated monocytes have increased levels of the lipid mediator, prostaglandin E2, and greater sensitivity to the ACLY inhibitor, BMS303141, with regards to cytokine production (Supplementary Fig. 3I, J).
Fructose treatment elevates ATP-linked oxygen consumption, independent of mitochondrial ROS and content.

Fructose Increases TCA Cycling and Anaplerosis

SITA coupled with gas chromatography-mass spectrometry (GC-MS) was used to further characterize the metabolic activity of the mitochondria in glucose- and fructose-treated cells.
Human monocytes were activated with LPS for 24 h and incubated with either $^{13}C6$ -glucose or $^{13}C6$ -fructose (Fig. 5A).
Mass isotopologue distribution (MID) analysis of the TCA cycle intermediates— succinate, fumarate and malate—and amino acids—glutamate and aspartate—highlighted that there was increased cycling in the fructose-cultured monocytes compared to glucose.
This was indicated by a reduced proportion of the unlabelled form of the metabolite (m + 0) and an increased proportion of the labelled form (predominantly represented by m + 2) (Fig. 5B, C and Supplementary Fig. 4A).
The TCA cycle relies on other metabolites to replenish it, a process termed anaplerosis. Glutamine-derived carbon enters the TCA cycle and contributes to TCA-mediated amino acid biosynthesis, such as aspartate.
To investigate whether fructose treatment increases glutamine anaplerosis, SITA with $^{13}C_5$ -glutamine was performed.
Monocytes were incubated with $^{13}C_5$ -glutamine in the presence of either glucose or fructose and activated with LPS (Fig. 5C).
The ass isotopologue distribution analysis of the TCA cycle metabolites—succinate, fumarate and malate—demonstrated an increase in the percentage of $^{13}C$ into these intermediates in the presence of fructose (represented as m + 4) (Fig. 5D and Supplementary Fig. 4B).
Fructose-treated monocytes are able to incorporate elevated amounts of glutamine-derived carbon to the TCA cycle intermediates and amino acids in comparison to glucose- treated monocytes (Supplementary Fig. 4C–G).
Fructose treatment increases the proportion of both sugar and glutamine carbon into the TCA cycle to support the observed increased rates of OXPHOS.

Fructose-Treated Human Monocytes are Vulnerable to Metabolic Challenge

Fructose treatment promotes a reduced glycolytic rate and enhanced OXPHOS.
Glucose and fructose are both metabolised by the enzyme HK, producing metabolites glucose 6-phosphate or fructose-6-phosphate, respectively.
The expression levels of the two predominant isoforms of HK: HKI and HKII were investigated.
No difference in expression levels of HKI was observed; however, HKII levels were increased in fructose-treated, LPS-stimulated monocytes, in comparison to glucose (Fig. 6A, B).
To further delineate the role of HK in monocyte function, the HK inhibitor, 2-DG was utilized.
Activated monocytes in glucose or fructose were treated for 24 h with 2-DG before measuring cytokine production.
Following treatment with 2-DG, production of IL-1$\beta$, IL-6 and TNF was largely unaffected in glucose-treated cells, with the exception of IL-10. By contrast, fructose-treated cells exhibited a dose-dependent decrease in all cytokines (Fig. 6C).
Fructose-treated cells reduced OXPHOS after treatment with 2-DG.
These results suggest the oxidative metabolic phenotype induced by fructose supports monocyte function.
Monocytes were treated with 2-DG to determine the level of cell viability.
Fructose-treated monocytes were acutely sensitive to 2-DG-mediated cell death in comparison to glucose-treated monocytes that were largely unaffected (Fig. 6D).
Next, it was assessed whether the cells were similarly sensitive to inhibition of oxidative metabolism.
The viability of LPS-stimulated monocytes treated with glucose was minimally affected by inhibition of complex I (with rotenone), III (with antimycin A) and V (with oligomycin).
However, monocytes treated with fructose were completely unable to tolerate incubation with these drugs, reflected by a striking reduction in viability (Fig. 6E).
Fructose treatment renders LPS-stimulated monocytes metabolically inflexible and dependent on oxidative metabolism.
Sensitivity to both 2-DG and mitochondrial inhibitors suggest that glycolysis and oxidative metabolism become inextricably coupled upon exposure to fructose.

Dietary Fructose Increases Inflammation in a Mouse Model

The metabolic and mechanistic implications of fructose exposure in human monocytes cultured ex vivo were explored.
To further investigate the impact of physiological fructose exposure on inflammation in vivo, a mouse model was employed.
First, it was determined whether mouse LPS-challenged bone marrow-derived macrophages (BMDMs) phenocopied human monocytes when exposed to fructose in vitro.
To better recapitulate the in vivo microenvironment, mouse macrophages were incubated with either glucose alone or a 1:1 ratio of glucose to fructose (maintaining the equivalent concentration of total monosaccharide).
Consistent with human monocytes, LPS-stimulated mouse macrophages exposed to fructose produced elevated levels of cytokines TNF, IL-1$\beta$, IL-6 and IL-12 at the protein level, but not the messenger RNA level in comparison to glucose alone (Fig. 7A and Supplementary Fig. 5A, B).
This observation is not due to reduced glucose availability in the double monosaccharide condition (Supplementary Fig. 5C).
SITA analysis was performed using mouse macrophages cultured in universally labelled $^{13}C6$ -glucose alone or $^{13}C6$ -glucose with $^{13}C_1$ -fructose (Fig. 7B) in order to confirm fructose uptake in the presence of glucose.
Fructose enters the cells in the presence of glucose as depicted by the presence of the m + 1 isotopologue (Fig. 7C).
Similar to human monocytes, BMDMs exposed to fructose significantly increased glutamine uptake and phosphorylation of the mTORC1 target, S6 ribosomal protein (Fig. 7D, E).
Fructose treatment also led to an increase in Akt phosphorylation (Supplementary Fig. 5D).
Next, the role of glutamine metabolism in fructose-treated mouse macrophages and their response to LPS was investigated.
The glutaminase inhibitor CB-839 did not alter levels of TNF; however, it significantly reduced IL-1$\beta$ and IL-12 in BMDMs cultured in the presence of both monosaccharides, whereas cytokine production was unchanged in cells cultured with glucose alone (Fig. 7F).
CB-839 also reduced phosphorylation of S6 in fructose-exposed cells (Supplementary Fig. 5E), suggesting that increased glutamine metabolism supports mTORC1 activity in the presence of fructose.
Further exploring the role of mTORC1 in fructose-mediated inflammation, BMDMs were treated with the mTORC1 inhibitor, rapamycin.
Rapamycin treatment significantly reduced cytokine production in both glucose-alone and glucose–fructose-treated macrophages (Fig. 7G).
Mouse macrophages require mTORC1 activity for cytokine production regardless of sugar exposure, yet those exposed to both monosaccharides (in contrast to those exposed to glucose alone) rely on glutaminolysis to support the increased cytokine production promoted by fructose exposure.
To determine if fructose supplementation can influence LPS-induced inflammation in vivo, a mouse LPS model of systemic inflammation was used.
Mice were provided 10% glucose or 10% fructose/10% glucose solutions for 2 weeks (to preclude development of any metabolic disorders) prior to LPS challenge (Fig. 7H).
Serum IL-1$\beta$ levels were significantly increased in mice exposed to fructose, and an increasing trend of IL-6 and TNF was observed (Fig. 7I).
This increase in serum IL-1$\beta$ was not due to fructose-enhancing baseline IL-1$\beta$ secretion, as seen at weeks 1 and 4 of sugar-water treatment in unchallenged mice (Supplementary Fig. 6A, B).
The presence of fructose does not result in a global increase in serum inflammatory markers, as, while levels of CXCL5 were elevated in the presence of fructose, this did not reach statistical significance, and there was no change in CXCL1 and CCL11 (Supplementary Fig. 6C–E).
Short-term fructose supplementation enhances LPS-induced systemic inflammation, suggesting physiological repercussions of high fructose exposure in mammals.
Fructose-supported inflammation does not occur in T cells.
Heightened inflammation observations were not reflected in the mouse T cell compartment.
No differences in $CD4^+$ or $CD8^+$ T cell proliferation were observed (Supplementary Fig. 7A).
Fructose supplementation did not influence polarization of either Th1 or inducible regulatory T cells (Supplementary Fig. 7B, C).
Fructose has a specific effect on LPS-stimulated mononuclear phagocytes.

Discussion

Activated human monocytes and mouse macrophages respond metabolically and functionally to fructose exposure.
Mononuclear phagocytes from both species are metabolically plastic in engaging in the metabolism of an alternative carbon source and reprogram cellular pathways to favor oxidative metabolism.
The cells are left metabolically inflexible and vulnerable to further metabolic challenge.
Fructose exposure promotes elevated cytokine production in both human and mouse mononuclear phagocytes and that a high fructose diet promotes an inflammatory phenotype in vivo.
Fructose contributes to numerous metabolic disorders such as obesity, cancer and non-alcoholic fatty liver disease; however, to date, our understanding of its impact on the immune system is lacking.
LPS-stimulated human monocytes and mouse macrophages exposed to fructose have an enhanced inflammatory phenotype supported by oxidative metabolism and glutaminolysis.
Fructose-exposed cells have increased mTORC1 activity, and while this is required to support cytokine production regardless of sugar exposure, those cells exposed to fructose rely specifically on glutaminolysis to support their inflammatory phenotype (Fig. 7J).
We demonstrate that fructose-dependent metabolic reprogramming is maintained long term, but this is independent of changes to mitochondrial dynamics.
Our data show that fructose carbon can contribute to both glycolysis and the TCA cycle, and whether this is mediated by HK or ketohexokinase in both human and mouse mononuclear phagocytes requires further exploration.
In comparison to glucose, fructose-derived pyruvate is not converted to lactate.
Fructose exposure has also been linked extensively to lipid biosynthesis, for example, fructose as a substrate is 30% more efficient at synthesising fatty acids than glucose, a phenomenon that has been implicated in the pathophysiology of non-alcoholic fatty liver disease.
Consistent with this, cytokine production was more sensitive to ACLY inhibition and levels of prostaglandin E2 were elevated in our fructose-treated monocytes.
The increase in LPS-induced inflammation from dietary fructose was not due to an enhanced global inflammatory effect, with certain chemokines measured having no observable differences, in addition to no effect on the mouse T cell compartment.
Our work using metabolic inhibitors shows that fructose treatment leaves cells metabolically inflexible and acutely vulnerable to further metabolic challenge.
Our results have highlighted the metabolic plasticity of human monocytes in response to fructose exposure and have elucidated the metabolic mechanisms supporting fructose-induced inflammation.