Dysregulated myelopoiesis and enhanced glutaminase-dependent neutrophil survival in preclinical obesity

Abstract

Obesity is increasingly recognized as a state of chronic low-grade inflammation associated with altered immune cell function, yet the mechanisms driving these changes remain incompletely understood. This study investigated myeloid cell subpopulations and neutrophil behavior in adult participants exhibiting preclinical obesity (body mass index [BMI] 32–51 kg/m², n = 12) compared to normal-weight controls (BMI 21–24 kg/m², n = 9), correlating findings with metabolic and inflammatory markers. Peripheral blood samples were analyzed by flow cytometry to quantify myeloid cell populations and TLR4/IL-1R surface expression. Neutrophils were cultured under normoxic (18% O₂) or hypoxic (1% O₂) conditions, with or without glutaminase inhibition, to assess spontaneous neutrophil death. Participants with preclinical obesity exhibited increased monocyte numbers and eosinophils, whereas total neutrophil numbers were not significantly different between groups, together with a higher percentage of HLA-DR^-/low monocytes and activated immature (CD16⁻CD11b⁺CD10⁻) neutrophils. In participants under 60 yr of age, IL‑1R expression on monocytes was significantly increased in the obesity group. Significant metabolic differences were also noted, including higher A1c (5.9 ± 0.1% vs 5.3 ± 0.1%), hs-CRP (9.01 ± 3.33 vs 0.69 ± 0.17 mg/L), and alkaline phosphatase (102.33 ± 9.67 vs 65.75 ± 5.54 U/L) in the preclinical obesity cohort, alongside decreased mean cell hemoglobin. Ex vivo neutrophil culture revealed that hypoxia reduced spontaneous neutrophil death in both groups; however, this effect was significantly reduced by glutaminase inhibition specifically in neutrophils from participants with preclinical obesity, suggesting a heightened reliance on glutamine metabolism for survival under hypoxia. These findings demonstrate dysregulated myelopoiesis and altered neutrophil behavior in preclinical obesity, providing mechanistic insight into the early immune consequences of metabolic dysfunction.

Introduction

Obesity, characterized by excess adiposity, is a serious, common, and costly chronic health condition. By 2030, 50% of adult Americans are projected to be obese.¹ The preclinical state of obesity without abnormal organ or tissue function can transition into clinical obesity with altered function of tissues and organs.² This transition is driven by multifaceted causes, including low-grade chronic inflammation and abnormal metabolism, which can lead to life-altering complications.² Mechanistic studies indicate that obesity-associated adipose tissue stress promotes recruitment and activation of innate immune cells—particularly monocytes/macrophages and neutrophils—driving cytokine/adipokine signaling that impairs insulin action and propagates inflammation in adipose tissue and liver, thereby accelerating metabolic dysfunction.^3–5

Large cohort studies have identified that obesity and its associated conditions are associated with myeloid lineage bias, as shown by changes in white blood cell populations. The absolute numbers of blood leukocytes and neutrophils have been associated with obesity.^6–9 An increased number of blood monocytes has been observed in obese individuals, including a high incidence (82%) of type 2 diabetes.¹⁰ The neutrophil-to-lymphocyte ratio is correlated with the severity of co-existing metabolic disturbance in obese individuals,¹¹ and the monocyte-to-lymphocyte ratio is associated with mortality in a large study.¹² Additionally, the white blood cell counts, and its subcomponent cell count had a significant positive correlation with the severity of metabolic syndrome.^10,13 Moreover, neutrophil and monocyte subpopulations are linked with the development of clinical obesity with overt organ dysfunction, and their reduction is linked with improved clinical signs upon clinical intervention.¹⁴ While these blood parameters can be obtained from a routine CBC differential test and analyzed retrospectively from medical records, their utility as biomarkers for the state of obesity may be limited because of confounding factors behind CBC data. Single-cell analysis, including flow cytometry and single-cell RNA sequencing, allows us to identify subpopulations of blood cells and their cellular signatures at the single-cell level. Monocyte subpopulations such as classical monocytes, intermediate monocytes, and non-classical monocytes, based on their CD14 and CD16 expression profiles, have been associated with inflammation or metabolic state in obesity.^10,13 In addition to flow cytometry, low-density neutrophils (LDN) separated by density gradation, which are increased in autoimmune inflammatory diseases, are increased in morbidly obese patients.¹⁵

Perturbations of metabolic pathways are linked to immune cell dysfunctions in obesity-associated conditions. The immunometabolic states in human obesity without overt organ dysfunction and their impact on blood myeloid cell populations and functions are not completely understood. Neutrophils are known to quickly undergo spontaneous cell death after isolation. Increased numbers of immature neutrophils have been found in the blood of obese individuals. Metabolic pathways involving glucose and glutamine have been implicated in human neutrophil death. Therefore, we studied myeloid subpopulations and neutrophil death in culture, and their association with blood parameters and clinical characteristics, including the degree of obesity and insulin resistance. A recent study analyzing changes in metabolites, proteins, and microbes revealed immunometabolic changes around the age of 45.¹⁶ We recruited participants who were aged 45–65 yr, had a body mass index (BMI) 18.5–24.9 or above 30 kg/m², had HbA1c <6.5% (without established type 2 diabetes), and did not have overt organ dysfunctions.

Materials and methods

Participants

Participants with obesity or normal weight were recruited at the Clinical Research Unit of Upstate Medical University and were included if they were aged 45–65 yr, had a BMI 18.5–24.9 or above 30 kg/m², and HbA1c <6.5%. Individuals were excluded if they had an acute infection, sleep disorders, weight <110 lb, renal or hepatic impairment, or recent use of diabetic medications, antibiotics, or anti-inflammatory drugs. In the low BMI group (n = 12), 2 were male and 10 were female, aged 45–64, all were white. In the high BMI group (n = 9), 3 were male and 6 female, aged 47–65, which included 7 white, one black/African American, and one Asian participant.

Blood samples from the participants included whole blood for Hemoglobin A1C, whole blood with EDTA for complete blood count (CBC), and plasma for a comprehensive metabolic panel and high-sensitivity C-reactive protein (hs-CRP) analysis. The waist-to-hip ratio was determined by trained personnel using a standardized method. Blood samples and clinical assessments were taken within 4 h of waking up (considering circadian rhythms) in a non-fasting state. A summary of participant characteristics is shown in Table 1 and details in Table S1. The study protocol was approved by the Institutional Review Board of Upstate Medical University (#2102492), and written informed consent was obtained from all participants prior to enrollment.

Table 1.

Comparisons between normal and obese participants were performed using appropriate statistical tests. Data are presented as mean ± SE. P values were determined using the Mann–Whitney U test.

	Normal	Obese	P-value
No. (Sex, F/M)	12 (10/2)	9 (6/3)	ND
Age, y	54.6 ± 2.0	55.8 ± 2.2	NS
Body weight, kg	66.5 ± 1.7	106.4 ± 6.9	<0.0001
BMI, kg/m2	23.5 ± 0.4	36.9 ± 2.0	<0.0001
Waist circumference, cm	80.1 ± 1.9	114.3 ± 5.3	<0.0001
Hip circumference, cm	95.9 ± 1.8	120.0 ± 4.8	<0.0001
Waist-to-hip ratio	0.8 ± 0.02	1.0 ± 0.02	0.0014
HbA1c, %	5.3 ± 0.1	5.9 ± 0.1	0.0015
CRP, mg/L	1.4 ± 0.4	9.0 ± 3.3	0.0006
Alkaline Phosphatase, U/L	62.7 ± 4.1	102.3 ± 9.7	0.0009
Mean cell hemoglobin, pg	30.9 ± 0.3	29.5 ± 0.5	0.019
Red cell distribution width, %	13.1 ± 0.1	13.8 ± 0.3	0.0421
Glucose, mg/dL	77.9 ± 3.0	104.8 ± 6.0	0.0001
Osmolality, cal, mOsm/kg	285 ± 1.0	288.4 ± 1.3	0.0498

Flow cytometry

We used immunophenotyping by flow cytometry to obtain differential blood cell counts. Whole blood cells were treated with 1X RBC lysis buffer (Biolegend, cat. no. 420301). Following lysis and thorough washing, blood monocytes were stained with CD45 (2D1), LIN—CD3, CD19, CD20, CD56 (UCHT1, HIB19, 2H7, 5.1H11), CD14 (63D3), HLA-DR (Tü36), CD16 (3G8), CD11c (3.9), IL-1R (Goat igG), CD284—TLR4 (HTA125), and IL-6R (502). Blood neutrophils were stained with CD16 (3G8), CD10 (HI10a), CD11b (CBRM1/5), CD14 (63D3), CD62L (DREG-56), CD66b (G10F5), CD15—SSEA-1 (W6D3), CD45 (2D1), and CD284—TLR4 (HTA125). Samples were analyzed by flow cytometry on the Cytek Aurora. Full panels can be found in Table S2.

The absolute number of cell populations was calculated by multiplying the frequency of CD45⁺ cells by the total white blood cell (WBC) count from the CBC data divided by 100. The median fluorescence intensity (MFI) was calculated by subtracting the MFI of unstained cells from the MFI of stained cells (IL-1R2 or TLR4) to quantify the cytokine receptor expression on the cell surface.

Live/dead viability dyes were not included at the time of acquisition. Instead, non-viable and debris-like events were excluded using FSC-based gating, with lymphocytes used as an internal reference. FSC^low events as well as FSC^very-high doublets and aggregates were excluded. All samples were processed under identical RBC lysis conditions (5–10 min at room temperature), and post hoc sensitivity analyses showed impact of the excluded events on neutrophil frequencies and marker expression despite cohort-dependent differences in the FSC-based exclusion rate (see Figs. S3 and S4).

Neutrophil isolation and culture

Neutrophils were isolated using the EasySep™ Direct Human Neutrophil Isolation Kit from (StemCell Technologies, cat. no. 19666). Cells were cultured in Human Plasma-Like Medium (HPLM) (Gibco™, cat. no. A4899101) supplemented with 1% penicillin-streptomycin (U/mL) (P/S) (Gibco™, cat. no. 15-140-122) and 10% dialyzed fetal bovine serum (dFBS) (Neuromics; cat. no. FBSS005-D; Lot: N18G51-D). Cell suspensions were diluted to a concentration of 70,000 cells/33 µL and seeded into a 96-well U-bottom non-treated culture plate. An additional 33 µL of inhibitor or media was added to reach a final culture volume of 66 µL, with the outer well of each plate filled with 200 µL of UltraPure™ distilled water to minimize evaporation. Cells were treated with a range of metabolic inhibitors and activators prepared as 2X working solutions, including Bis-2-(5-phenylacetamido-1,3,4-thiadiazol-2-yl)ethyl sulfide (BPTES; 10 µM and 100 µM) (Sigma-Aldrich, cat. no. SML0601), CB-839 (10 µM and 100 µM) (Sigma-Aldrich, cat. no. 5.33717), 3-nitropropionic acid (3-NPA; 1 mM) (Sigma-Aldrich, cat. no. 164603), 6-aminonicotinamide (6-AN; 100 µM) (Sigma-Aldrich, cat. no. A9400), DL-Buthionine-(S, R)-sulfoximine (BSO; 150 µM) (Sigma-Aldrich, cat. no. B2640), and N-Acetyl-L-cysteine (NAC; 1 mM) (Sigma-Aldrich, cat. no. A9165). All compounds were prepared in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, cat. no. D2650) or ethanol (EtOH) at a final solvent concentration of 6:1000 (DMSO) or 2:1000 (EtOH). Plates were placed in either normoxia (18% O₂, 5% CO₂) or in hypoxia (1% O₂, 5% CO₂) for 24 h.

After incubation, apoptosis staining was performed using Apotracker Green (BioLegend, cat. no. 427403) and LIVE/DEAD Fixable Far Red viability dye (Invitrogen, cat. no. L10120). Apotracker stock (80 μΜ) was diluted 1:20 in PBS, then further diluted 1:1 to yield a 4 μΜ working solution; 10 μL of this solution was added to 190 μL of cell suspension. Cells were incubated for 15 min at room temperature, centrifuged, and washed with PBS. LIVE/DEAD staining was prepared by making a 1:2000 dilution, and 200 μL of this solution was added to each sample. Cells were incubated on ice for 30 min, centrifuged, and washed twice with FACS buffer before being resuspended in the same volume. Samples were then analyzed immediately using an Attune flow cytometer.

Data analysis

Statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, Inc., San Diego, California, USA). Data distribution was first assessed using normality testing. Depending on the distribution, comparisons between two groups were analyzed using either the Wilcoxon signed-rank test for paired samples or the Mann–Whitney U test for unpaired samples. Correlation values were determined using rank-based methods, and pathology data were categorized into quartiles to identify the highest and lowest groups. Associations of monocyte or neutrophil subpopulations, as well as eosinophils, with clinical and laboratory parameters were further adjusted for total leukocytes, total monocytes, and classical monocytes to account for potential confounding factors.

Results

Demographic and clinical characteristics of the participants

The clinical characteristics of obese and normal-weight individuals are shown in Table 1. Obese individuals had higher BMI kg/m² and waist-to-hip ratio. Altered biochemical parameters include elevated Hemoglobin A1c (HbA1c), high-sensitivity C-reactive protein (CRP), Alkaline Phosphatase (ALP), Red Cell Distribution Width (RDW), and osmolality as well as decreased Mean Cell Hemoglobin (MCH). These biochemical parameter changes can reflect physiological alterations in preclinical obesity and have many effects on metabolism and inflammatory responses. All parameters are included in Table S3.

The total number of monocytes increases in obesity

Compared to the normal-weight group, obese individuals exhibited elevated absolute numbers of total monocytes (Fig. 1A). The number of classical and intermediate monocytes tended to be higher in the obese group than in the normal weight group, whereas non-classical monocytes were unchanged between the groups (Fig. 1B–D). Neutrophils were similar between groups, whereas eosinophil numbers were significantly increased in the obese group (Fig. 1E and F). We also analyzed neutrophil-to-lymphocyte and monocyte-to-lymphocyte ratios and found no significant differences between the normal weight and obesity groups (Fig. S1), likely influenced by the higher lymphocyte numbers in the obesity group in our cohort (Fig. S1).

Figure 1.

Circulating leukocyte subsets in lean and obese subjects. Peripheral blood cells from lean (n = 9) and obese (n = 12) subjects were analyzed by flow cytometry to quantify monocyte and granulocyte populations. (A) Total monocytes (HLA-DR⁺CD14⁺). (B) Classical monocytes (CD14⁺CD16⁻). (C) Intermediate monocytes (CD14⁺CD16⁺). (D) Non-classical monocytes (CD14^lowCD16⁺). (E) Neutrophils (CD66b⁺CD15⁺). (F) Eosinophils (CD16⁺CD45^low). (G) Neutrophil-to-lymphocyte ratio calculated from absolute neutrophil and lymphocyte counts. (H) Monocyte-to-lymphocyte ratio calculated from absolute monocyte and lymphocyte counts. Absolute counts are shown. Data are presented as mean ± SEM, with each dot representing one subject. Statistical significance was determined using the Mann–Whitney U test. NS, not significant.

HLA-DR ^−/low monocytes and an immature neutrophil population are increased in obesity

Next, we analyzed monocyte subpopulations using HLA-DR, TLR4, IL-1R2, and IL-6R, as shown in Fig. 2. Monocytic myeloid-derived suppressor cells (M-MDSCs) were mostly identified within the classical monocyte subset by CD14⁺/HLA-DR^−/low expression, whereas majority of intermediate and non-classical monocytes showed uniformly high HLA-DR levels (Fig. 2 and 3A). The absolute number and frequency of M-MDSCs was significantly higher in obese individuals than in normal-weight individuals (Fig. 3A and B). In neutrophil subpopulations, there was significant increase in number of CD16^-/CD11b⁺/CD10^- activated immature neutrophils in preclinical obesity (Fig. 3C and D). No significant changes were found in the MFIs for TLR4, IL-1R2, and IL-6R in monocytes and TLR4 in neutrophils (Fig. S2).

Figure 2.

Gating strategies for identification of monocyte, neutrophil, and lymphocyte subpopulations in peripheral blood. Representative flow cytometry plots showing the gating strategies used to identify leukocyte populations in freshly isolated peripheral blood. (A) Monocyte and lymphocyte gating strategy. Cells were gated on singlets and CD45⁺ leukocytes. From the CD45⁺ gate, two main branches were defined: one for monocytes and one for lymphocytes. Monocytes were identified as HLA-DR⁺CD14⁺ cells and further subdivided into classical (CD14⁺CD16⁻), intermediate (CD14⁺CD16⁺), and non-classical (CD14^lowCD16⁺) subsets. A separate branch was used to identify CD14⁺HLA-DR^low cells. Lymphocytes were gated from the LIN⁺SSC^low fraction and analyzed in 2 plots: CD45/CD16 to identify NK cells, and HLA-DR/CD16 to identify active versus inactive lymphocytes. (B) Neutrophil and granulocyte gating strategy. Cells were gated on singlets and CD45⁺ leukocytes, followed by identification of granulocytes and neutrophils. Granulocytes were further subdivided into eosinophils (CD16⁺CD45^low/−) and neutrophils (CD66b⁺CD15⁺). Neutrophils were then subdivided into CD10⁺ and CD10⁻ populations, each of which was further analyzed based on CD16 and CD11b expression, with CD10⁺ neutrophils additionally used to define a CD62L^low population. Plots are representative of all subjects analyzed.

Figure 3.

HLA-DR^−/low monocytes and CD10- neutrophil subpopulations in lean and obese subjects. Peripheral blood leukocytes from lean (n = 9) and obese (n = 12) participants were analyzed by flow cytometry. (A) Representative plots showing HLA-DR^−/low monocytes gated from CD14⁺ total monocytes. (B) Absolute number and percentage of HLA-DR^−/low monocytes. (C) Representative plots showing CD10^- neutrophil subpopulations gated from CD66b⁺CD15⁺ total neutrophils. (D) Absolute number and percentage of CD16^-CD11b⁺CD10⁻ neutrophils. Data are shown as mean ± scatter plots, with each dot representing one participant. P values were determined using the Mann–Whitney U test. NS, not significant.

Obesity in the <60 years showed increased IL1R2⁺ monocytes

Age-related changes in immune receptor expression differed between the groups. Normal-weight individuals exhibited increased IL-1R2 expression in monocytes with age. In contrast, obesity was associated with decreased IL-1R2 levels in monocytes (Fig. 4A). There was a trend of increasing TLR4 in neutrophils as normal-weight individuals aged (Fig. 4B). To mitigate the confounding effects of advanced age, the analyses were restricted to participants under 60 yr of age. Within this cohort, IL-1R2 expression in monocytes was elevated in individuals with preclinical obesity (Fig. 4C). We also found a significant increase in the number of eosinophils in obese participants by excluding participants over 60 yr of age (Fig. 4D).

Figure 4.

TLR4 and IL-1R expression in peripheral blood leukocytes. Peripheral blood leukocytes from lean (n = 9) and obese (n = 12) participants were analyzed by flow cytometry. TLR4 and IL-1R levels were measured as median fluorescent intensity (MFI) of antibody-stained cells minus unstained controls. (A) Correlation between IL-1R expression in total monocytes and participant age. (B) Correlation between TLR4 expression in neutrophils and participant age. Participants aged ≥60 yr were excluded from panels (C) and (D), which show (C) IL-1R expression in total monocytes and (D) absolute number of eosinophils. Data are shown as mean ± scatter plots, with each dot representing one participant. P values were determined using simple linear regression (A and B) or Mann–Whitney U test (C and D). NS, not significant.

Myeloid cellular parameters are associated with metabolic, inflammatory and erythropoietic parameters

We next assessed the correlation between obesity-increased subpopulations and metabolic/inflammatory parameters available in our cohort (Table 2). In our cohort, the numbers of monocytes, classical monocytes, intermediate monocytes, HLA-DR^-/low monocytes, neutrophils, and eosinophils correlated with BMI, WHR, and HbA1c. The number of monocytes, classical monocytes, and intermediate monocytes was correlated with WHR and HbA1c levels. Neutrophil and eosinophils were correlated with BMI and CRP levels in obesity. MCH, an erythropoiesis parameter, correlated with monocytes, classical monocytes, intermediate monocytes, and HLR-DR^−/low monocytes in our obese cohort. ALP was correlated with intermediate monocytes and eosinophils and weakly with HLR-DR^−/low monocytes and neutrophils in all participants but not in our obese cohort.

Table 2.

Correlations between myeloid subsets and parameters of obesity, HbA1c, C-reactive protein (CRP), alkaline phosphatase (ALP), and mean cell hemoglobin (MCH) in the entire cohort (top) and obese participants. R² and P values were determined using a simple linear regression. NS, not significant.

All Participants
	Monocytes	Classical monocytes	Intermediate monocytes	HLA-DR-/low monocytes	Neutrophils	Eosinophils
BMI	0.2040, 0.0001	0.1340, 0.0021	0.1488, 0.0012	0.1840, 0.0004	0.2929, 0.0009	0.5631, <0.0001
WHR	0.1880, 0.0002	0.08285, 0.0173	0.2472, <0.0001	0.1605, 0.0009	0.1932, 0.0093	0.3107, 0.0006
HbA1c	0.5287, <0.0001	0.4318, <0.0001	0.2653, <0.0001	0.2506, <0.0001	0.2332, 0.0038	0.2312, 0.0040
CRP	0.07984, 0.0196	0.06169, 0.0411	0.03634, NS	0.03239, NS	0.3464, 0.0003	0.4973, <0.0001
ALP	0.05144, NS	0.007039, NS	0.1188, 0.0040	0.07193, 0.0308	0.1494, 0.0239	0.3500, 0.0002
MCH	1.070e−5, NS	0.008172, NS	0.02163, NS	0.01216, NS	0.09797, NS	0.1155, 0.0493

Obese participants
	Monocytes	Classical monocytes	Intermediate monocytes	HLA-DR − /low monocytes	Neutrophils	Eosinophils
BMI	0.004523, NS	3.179e-5, NS	0.008795, NS	0.00175, NS	0.3343, 0.0120	0.5371, 0.0005
WHR	0.3575, 0.0001	0.2570, 0.0016	0.2702, 0.0012	0.01053, NS	0.03572, NS	0.5335, 0.0006
HbA1c	0.4657, <0.0001	0.4338, <0.0001	0.1356, 0.0271	0.08614, NS	0.3357, 0.0117	0.07884, NS
CRP	0.003063, NS	0.003164, NS	01.426e-5, NS	0.01641, NS	0.3710, 0.0073	0.3861, 0.0063
ALP	0.03879, NS	0.1700, 0.0125	0.03277, NS	0.01385, NS	0.8208, NS	0.2184, NS
MCH	0.3189, 0.0003	0.1539, 0.0180	0.3607, 0.0001	0.3029, 0.0009	0.02201, NS	0.01068, NS

Aging significantly altered cellular/metabolic parameters in study participants. Individuals with preclinical obesity exhibited a progressive increase in MCH levels with age, whereas normal-weight individuals demonstrated a corresponding decrease in MCH over time (Fig. 5A). ALP levels showed an inverse trend: obesity was associated with declining ALP concentrations with increasing age, whereas normal-weight individuals displayed elevated ALP levels as they aged (Fig. 5B). High ALP levels were further associated with an increased number of eosinophils, and high HbA1c levels corresponded with an elevated monocyte count (Fig. 5C and D).

Figure 5.

Associations between blood parameters and leukocyte counts or receptor expression in peripheral blood. Peripheral blood from lean (n = 9) and obese (n = 12) participants was analyzed. TLR4 and IL-1R expression were measured as median fluorescent intensity (MFI) of antibody-stained cells minus unstained controls using flow cytometry. (A) Correlation of mean cell hemoglobin (MCH) with participant age. (B) Correlation of alkaline phosphatase (ALP) levels with participant age. (C) Eosinophil counts in participants with ALP levels in the lowest versus highest quartiles. (D) Monocyte counts in participants with HbA1c levels in the lowest versus highest quartiles. (E) Eosinophil counts in participants aged 45–59 with glucose levels in the lowest versus highest quartiles. (F) Eosinophil counts in participants aged 45–59 with MCH levels in the lowest versus highest quartiles. Data are shown as mean ± scatter plots, with each dot representing one participant. P values were determined using simple linear regression (A and B) or Mann–Whitney U test (C–F). NS, not significant.

Restricting analyses to participants under 60 yr of age revealed an increased eosinophil count in individuals with high glucose levels, as well as a correlation between high glucose and low MCH levels (Fig. 5E and F). Although our limited number of cohorts may be influenced by an outlier,¹⁷ considering the effect of aging on immunometabolic parameters¹⁶ can highlight obesity-associated cellular/metabolic parameters in the blood in a certain age group.

Neutrophil death in culture

Measurements of spontaneous neutrophil death in culture upon exposure to hypoxia revealed a high batch effect and inter-individual variation within our study population. Consistent with a previous report,¹⁸ we observed hypoxia-prolonged survival of isolated neutrophils in the normal weight and obese groups (Fig. 6A and B). To mitigate the batch effect, we analyzed data based on normal versus obese pairs in the same batch. The extent of the hypoxia effect (% increase/decrease from normoxia control) was similar between the normal and obese groups (Fig. 6C). To gain insight into the metabolic pathways regulating neutrophil death or survival in culture, we tested known metabolic inhibitors. The glycolysis inhibitor 2-DG strongly accelerated neutrophil death (data not shown), confirming the reliance of neutrophils on glucose utilization for prolonging survival. Interestingly, neutrophils in obesity showed less hypoxia-prolonged survival when glutaminase was mildly inhibited by either BPTES or CB-839 (Fig. 6D–G). Briefly, 6-AN (a pentose phosphate pathway inhibitor), 3-NPA (a succinate dehydrogenase inhibitor), BSO (a γ-glutamylcysteine synthetase inhibitor), and NAC (a ROS inhibitor) had no effect on hypoxia-induced survival (Fig. 6H–K). Therefore, neutrophils in obesity may rely on glutaminase-mediated metabolism for survival more than those in normal-weight individuals under hypoxia.

Figure 6.

Hypoxia alters survival of peripheral blood neutrophils and the effect of metabolic modulators. Freshly isolated neutrophils from lean (n = 9) and obese (n = 12) participants were cultured in human plasma-like media under normoxia (18% O₂, 5% CO₂) or hypoxia (1% O₂, 5% CO₂) for 24 h. Neutrophil viability was assessed using ApoTracker Green and LIVE/DEAD staining and analyzed by flow cytometry. (A) Effect of hypoxia on neutrophils from normal-weight participants. (B) Effect of hypoxia on neutrophils from obese participants. Each plot shows the average of triplicate cultures, and each connected line represents one participant. (C–K) Hypoxia-extended survival (Δ = live cells in hypoxia–live cells in normoxia) measured in each cohort following treatment with metabolic inhibitors or modulators: (D–E) CB-839, (F–G) BPTES (bis-2-(5-phenylacetamido-1,3,4-thiadiazol-2-yl) ethyl sulfide; glutaminase inhibitors), (H) 6-AN (6-aminonicotinamide; pentose phosphate pathway inhibitor), (I) 3-NPA (3-nitropropionic acid; succinate dehydrogenase inhibitor), (J) BSO (buthionine sulfoximine; γ-glutamylcysteine synthetase inhibitor), and (K) NAC (N-acetylcysteine; ROS inhibitor). Data are shown as mean ± scatter plots, with each dot representing one participant. P-values were determined using a paired Wilcoxon test. NS, not significant.

Discussion

While systemic inflammatory and metabolic changes are increasingly recognized as drivers of obesity-associated pathologies, how immunometabolic changes contribute to transitions from pre-clinical state without organ dysfunction to its clinical state with organ dysfunctions in humans remains to be determined. In a small cohort of obese adults without overt organ dysfunctions, our analysis using a flow cytometry panel identified changes in subpopulations of monocytes and neutrophils in obese individuals—such as classical monocytes and HLA-DR^−/low monocytes, and IL-1R2⁺ monocytes and TLR4⁺ immature neutrophils in younger (45–59 yr old) obese individuals along with erythropoietic parameters, RDW and MHC, and eosinophils.

A previous study showed increased classical monocytes and HLA-DR^−/low MDSC-like monocytes in high BMI individuals, with HLA-DR^−/low monocytes further increased in obese patients with type 2 diabetes.¹⁵ Similarly, in obesity with organ dysfunction, an increase in HLA-DR^−/low monocytes was associated with BMI, whereas CD14/CD16 subsets such as classical monocytes did not change,¹⁹ suggesting that HLA-DR^−/low monocytes may be further increased in developing organ dysfunctions such as type 2 diabetes, dyslipidemia, and/or liver and kidney dysfunction. In addition, we found CD16^-/CD11b⁺/CD10^- neutrophils, which are phenotypically immature and activated, and often found in low-density neutrophils, are increased in obesity. These increased immature myeloid populations indicate increased myelopoiesis in obesity. To account for the potential confounding effects of aging on immune and metabolic parameters, we conducted a sub-analysis excluding participants over 60 yr of age. This age-restricted cohort revealed more pronounced differences in myeloid cell subpopulations and inflammatory markers between obese and normal-weight individuals, including elevated eosinophil counts, increased TLR4 expression in immature neutrophils, and higher IL-1R2 expression in monocytes among those with preclinical obesity. Emerging evidence has shown that inflammatory signals, such as those through Toll-like receptors, IL-6, and IL-1, are modified myelopoiesis programs during acute viral infection and sepsis.^20,21 Obesity-induced and age-related changes in myelopoiesis programs require further study to have a better understanding of the downstream impact. Because blood collection occurred in a nonfasting state without restriction of recent food intake, postprandial metabolic responses may have influenced neutrophil phenotypes, including TLR4 expression.²² We therefore acknowledge the timing of food intake as a potential confounder and have noted this limitation in the interpretation of our findings.

Despite our original intention to capture altered hematopoiesis by minimizing confounding metabolic changes, our obese participants had significantly elevated HbA1c, CRP, and ALP levels at preclinical levels. Myeloid cellular parameters increased by obesity were correlated with HbA1c, and CRP along with BMI and WHR, suggesting that insulin resistance and low-grade inflammation are closely linked with dysregulated hematopoiesis in preclinical obesity. Unfortunately, although we excluded uncontrolled dyslipidemia, detailed lipid profiles were not available in this study. A previous study showed that hematopoietic parameters combined with high-density lipoprotein (HDL) levels are associated with the severity of metabolic syndrome.²³ Overall, our data support an immune-metabolic interaction in preclinical obesity.

Neutrophils under inflammatory conditions show prolonged survival by reducing spontaneous apoptosis in culture.^24–26 Severe obesity (BMI > 40 kg/m²) increases LDNs,²⁷ which are increased in several inflammatory diseases. LDNs in inflammatory conditions are resistant to apoptosis,²⁸ whereas LDNs in healthy individuals seem to have unaltered survival capacity in culture.²⁹ We sought to determine whether potential increase in LDNs in obesity changes neutrophil survival in hypoxia. Although we did not find changes in neutrophil survival in hypoxia by obesity, neutrophils from obese individuals relying on glutaminase for prolonged survival in hypoxia have an implication in metabolic reprogramming in neutrophils by obesity. A previous study also found that glutamine is increased as neutrophils are cultured in a non-stimulated condition.³⁰ As neutrophils utilize mitochondria for survival in hypoxia,¹⁸ hypoxia-induced survival mechanisms will be of interest in developing cellular biomarkers in neutrophils in obesity.

The early identification of dysregulated myeloid populations in obese adults offers an opportunity to stratify individuals based on their metabolic risk before the onset of overt organ dysfunction. Beyond monocyte or neutrophil counts found in large cohorts of obesity with heterogeneous comorbidities, immunological phenotyping using surface marker expression by flow cytometry identified subpopulations correlated with metabolic parameters in obese individuals in a small cohort of this study, suggesting the utility of myeloid subpopulations as a potential biomarker to characterize younger aged (45–59 yr old) obesity without organ dysfunction. While the relatively small cohort of this study represents a limitation and constrains broad population-level generalizability, the study was designed as a feasibility and hypothesis-generating human investigation using controlled experimental conditions. Within this context, the observed close associations between myeloid subpopulations and metabolic, inflammatory, and erythropoietic parameters support the presence of immune–metabolic interactions during the transition from preclinical to clinical obesity. Despite modest cohort size, the consistency of the observed effects across matched obese and normal weight samples underscores the biological relevance of these findings. Similar cohort selection and study design may therefore be well suited for hypothesis-driven human immunometabolism research that requires intensive functional assays of short-lived primary cells.

Further investigation using single-cell transcriptomic and advanced sequencing approaches will enable the identification of gene signatures within relevant myeloid subpopulations. In parallel, the development of cell-based functional assays, such as the neutrophil culture system employed in this study, combined with metabolic measurements, will provide additional mechanistic insight into the pathogenesis of organ dysfunction in human obesity. This integrated approach also establishes a framework for targeted interventions aimed at preventing progression to clinical obesity and associated comorbidities. Limitations of the present study include the modest sample size, restriction to a defined age range and matched sex distribution, and the potential influence of batch effects in flow cytometric analyses; these will be addressed in future studies with expanded cohorts and dedicated validation efforts.

Author contributions

M.B. performed the experiments, analyzed the data, and wrote the manuscript. A.L. performed the experiments and analyzed the data. L.A. collected participant data and records. J.D. assisted experiments. R.W. recruited study participants, directed the Clinical Research Unit team and oversaw clinical data collections, and N.U. designed the experiments, wrote the manuscript, and supervised the project. All approved the final manuscript.

Madison Babcock (Data curation [Equal], Formal analysis [Equal], Investigation [Equal], Writing—original draft [Equal]), Aric Lechner (Data curation [Equal], Formal analysis [Equal], Investigation [Equal]), Julia Drolet (Methodology [Equal]), Lynn Agostini (Data curation [Equal], Investigation [Equal], Methodology [Equal]), Ruth S Weinstock (Project administration [Equal], Resources [Equal], Supervision [Equal]), and Norifumi Urao (Conceptualization [Equal], Formal analysis [Equal], Funding acquisition [Equal], Investigation [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Writing—original draft [Equal], Writing—review & editing [Equal])

Supplementary material

Supplementary material is available at ImmunoHorizons online.

Funding

This study was supported by awards from the National Institute of Health (R01DK111489 and R01GM144624 to N.U.).

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The data supporting the findings of this study are included in the article and Supplemental Material. Additional data will be made available by the corresponding author upon request.