Monocytes Display Altered Inflammatory Profiles in Individuals with Diabetic Foot Ulcers Indicating Monocyte Dysfunction

ABSTRACT

Diabetic foot ulcers (DFUs) often exhibit impaired healing. Macrophage imbalance driven by persistent inflammation is one contributing factor, with a predominance of inflammatory over anti‐inflammatory macrophages in non‐healing ulcers. Monocytes, precursors of macrophages, are skewed towards an inflammatory phenotype in diabetes. Here, we investigated whether monocyte inflammatory skewing is further exacerbated in individuals with DFUs and explored how it relates to an individual's metabolic (HbA1c and lipids) profile. Monocyte inflammatory profile was assessed by examining phenotypic markers, gene expression, and cytokine production. We observed an elevation in total monocytes and the intermediate subset in the DFU group compared with healthy individuals. In addition, an impaired monocyte profile was seen in the DFU group, with a suppression of both inflammatory (CD86, TLR2, and TLR4) and anti‐inflammatory (CD163) markers, accompanied by the downregulation of distinct inflammatory and anti‐inflammatory genes. Despite a suppressed phenotype, DFU monocytes showed an increased basal TNF and IFN‐γ production, indicating low‐grade inflammation. Further, in exploratory analysis, monocyte CD86 and CD163 inversely correlated with HbA1c, while CD163 negatively correlated with cholesterol/HDL‐C and triglycerides, suggesting links between metabolic control and monocyte phenotype. Taken together, the impaired monocyte profile indicates underlying pathogenic inflammation and monocyte dysfunction in people with DFUs, suggesting the monocyte profile could serve as a potential biomarker for immunomodulation in DFU.

Monocytes display an altered inflammatory profile in individuals with DFU, indicating dysfunctional monocytes. The downregulation of both inflammatory and anti‐inflammatory markers, together with increased production of pro‐inflammatory cytokines, suggests the presence of low‐grade inflammation and dysregulated monocyte activation, which may impair healing responses in DFU.

Abbreviations

Apo

apolipoprotein

CD

cluster of differentiation

CVD

cardiovascular disease

DFU

diabetic foot ulcers

HDL‐C

high density lipoprotein cholesterol

IFN

interferon

IL

interleukin

LDL‐C

low‐density lipoprotein cholesterol

LOD

limit of detection

TC

total cholesterol

TLR

Toll‐like receptors

TNF

tumour necrosis factor

1. Introduction

The global incidence rate of diabetes has surged threefold in the past two decades, now affecting approximately 600 million people worldwide [1]. Consequently, there is a parallel rise in the incidence of associated complications. Despite significant progress in treatment strategies, diabetic foot ulcers (DFUs) markedly exhibit poor healing, leading to considerable morbidity and mortality [2]. The lifetime risk of developing a foot ulcer in individuals with diabetes is estimated as 19%–34%, though this varies with ethnicity and socio‐economic status [3]. While there are ongoing efforts to identify high‐risk individuals requiring ulcer prevention strategies, the increasing incidence of DFUs demands an understanding of underlying pathological mechanisms to develop improved diagnostics and therapeutics.

Although poor glycaemic control and prolonged diabetes are well‐established risk factors for developing complications, the role of underlying inflammation in diabetes is increasingly recognised [4]. A heightened pro‐inflammatory environment is evident in diabetes, with elevated circulating pro‐inflammatory cytokines and an altered immune cell activation [5]. In this respect, monocytes and macrophages are crucial in sustaining inflammation through the adoption of a pro‐inflammatory state, triggering the release of inflammatory cytokines. While there is a broad spectrum of phenotypes that macrophages can adopt, an imbalance between M1‐proinflammatory and M2‐anti‐inflammatory macrophages is clinically relevant, with a predominance of pro‐inflammatory adipose tissue macrophages in obesity and a higher proportion of CD86+ (M1) macrophages associated with atherosclerotic plaque instability [6, 7]. Even in DFUs, reduced Arg+ (M2) macrophages in ulcer biopsies are linked to decreased healing efficacy [8].

Wound macrophages mostly derive from blood monocytes, which can also exhibit inflammatory alterations. Elevations in circulating monocytes and shifts in subset proportions occur in disease states. Specifically, a rise in intermediate monocytes is associated with an increased risk of cardiovascular disease (CVD) events [9, 10, 11]. Furthermore, monocytes adopt an inflammatory phenotype as seen by altered expression of activation markers [12, 13, 14]. Monocyte inflammatory state may, in turn, influence the macrophage phenotype adopted amid a complex interplay of multiple factors [15].

Studies assessing monocyte inflammatory skewing in diabetes consistently showed unbalanced inflammatory states that were likely to negatively influence clinical outcomes [12, 13, 16, 17]. However, it is unclear whether the monocyte inflammatory state is further exacerbated in individuals with DFUs than in those with diabetes without ulcers. Here we investigated the monocyte inflammatory profile by examining monocyte phenotype, transcriptional profile, and cytokine production. We also sought to explore the factors that may be influencing monocyte phenotype, with a focus on metabolic profile.

2. Materials and Methods

2.1. Study Cohort

This study was approved by the Western Sydney Local Health District Human Research Ethics Committee. Written informed consent was obtained from all the participants before enrolling them in the study. The participants were recruited in three groups: healthy controls (HC), individuals with diabetes without active foot ulcers (DM), and individuals with active diabetic foot ulcers (DFU). Based on power calculations using the control cohort from a related study, a sample size of 18 per group would provide 80% power to detect large differences in monocyte markers between groups (p < 0.05), as well as to identify moderate to strong (r = 0.5–0.9) correlations when groups were combined for exploratory analysis.

The HC were generally healthy individuals who were non‐smokers, without any diagnosed acute or chronic inflammatory conditions, including hypertension and prediabetes. DM participants were recruited from those attending the Westmead outpatient complex diabetes clinic, excluding individuals with known infections, malignancies, immunosuppression, and severe renal impairment. DFU cohort included individuals presenting to the Westmead high‐risk foot clinic with active diabetic foot ulcer(s), excluding those with ulcers rated PEDIS infection grade >2 and/or surgical wounds. Relevant medical history was obtained from participants or from their medical records.

2.2. Biochemical and Lipid Measurements

Peripheral venous blood samples (non‐fasting) were collected from all participants and processed within 1–2 h of collection for all analyses and storage. White blood cell counts (WBC) with differential were obtained using a haematology analyser (Beckman Coulter DXH 520). Lipid profile (total cholesterol (TC), low‐density lipoprotein cholesterol (LDL‐C), high‐density lipoprotein cholesterol (HDL‐C), non‐HDL‐C, triglycerides, Apolipoprotein (Apo)A and ApoB), HbA1c and C‐reactive protein were measured using Roche Cobas Pro by NSW Health Pathology at Westmead Hospital.

2.3. Measurement of Monocyte Subset Proportions and Surface Marker Expression

Monocyte subset proportions and surface marker expression were determined by whole blood flow cytometry [18]. An aliquot of blood from a K2EDTA blood tube (within 1–2 h of collection) was stained with appropriately titrated anti‐CD14‐V450 (BD Biosciences), anti‐CD16‐APC (Invitro technologies), and anti‐HLA‐DR‐PerCP (BioLegend) to identify monocyte subsets, along with PE conjugated antibodies against inflammatory (CD86, CD120b, Toll‐like receptor (TLR)2 and TLR4) or anti‐inflammatory (CD163 and CD93) markers each separately (Table S1). After 30 min of incubation at 4°C in the dark, the erythrocytes were lysed with Optilyse C (BioLegend) for 10 min. The samples were then washed with PBS and fixed with 1% formaldehyde. Single fluorochrome‐stained tubes were prepared for compensation. Data were acquired on an Attune NxT Acoustic flow cytometer (ThermoFisher Scientific) and analysed using FlowJo version 10.8.1 (BD Biosciences). A spectral overlap compensation matrix was created in FlowJo and applied to samples before analysis. Monocyte gating was performed as per the published strategy (Figure S1) [18]. The median fluorescence intensity (MFI) of specific surface markers was measured on each monocyte subset relative to fluorescence minus one (FMO) control (Figure S2).

2.4. Assessment of Monocyte Gene Expression

Monocytes (CD14++) were isolated from fresh blood using EasySep Direct Human Monocyte Isolation Kit (StemCell Technologies) as per manufacturer's instructions, with an additional magnetic enrichment step to increase purity. Approximately 1 × 10⁶ monocytes were harvested, and the cell pellet was lysed and homogenised in RNeasy lysis buffer (Buffer RLT) (QIAGEN), then stored at −80°C until RNA extraction. Total RNA was extracted from monocytes (purity >80% as assessed by flow cytometry) using the RNeasy mini kit (QIAGEN) according to the manufacturer's kit protocol, and gene expression profiling was performed by Nanostring nCounter Elements TagSets assay. RNA samples (N = 12 per group) that passed the quality control with A260/A280>1.8 and RNA integrity >8.0 were selected for the assay. A customised oligo probe sequence panel to target genes of interest was designed in consultation with the NanoString Bioinformatics team and procured as pooled Probe A and Probe B (Integrated DNA Technologies, Singapore). Three sets of 144‐plex NanoString nCounter Elements TagSets (NanoString Technologies, Inc.) were used for 36 RNA samples, with each set including 12 reaction aliquots plus controls. The assay was performed by staff at the Westmead Research Hub Genomics facility according to the manufacturer's instructions. Data analysis was performed using nSolver Analysis Software version 4.0 (NanoString Technologies, Inc.).

2.5. Assessment of Circulating Cytokine Profile and Monocyte Cytokine Production

Isolated CD14++ monocytes were cultured in RPMI‐1640 medium (Sigma‐Aldrich) supplemented with 5% standard human AB serum (Sigma‐Aldrich) for 24 h. Cell‐free culture supernatants were collected and stored at −80°C until further analysis. Cytokine levels were measured in culture supernatants and serum samples using LEGENDplex Multi‐analyte Flow assay kit—Human Essential Immune Response Panel (13‐plex) (BioLegend) by Westmead Research Hub Core facility as per the manufacturer's kit instructions. Data were acquired on the BD FACS Canto II (BD Bioscience) and analysed using LEGENDplex Data Analysis Software (BioLegend), according to the manufacturer's instructions. The cytokine concentrations were determined from the standard curve generated for each analyte and were normalised for 1 × 10⁶ cells. For values below the lower limit of detection (LOD) or above the upper limit of detection (ULOD), single imputations (with LOD/Sqrt2 or ULODx1.7 respectively) were used before normalisation [19].

2.6. Statistical Analysis

All statistical analyses were performed using GraphPad Prism 10 (Version 10.1.0) with no outlier exclusion. The normality of the datasets was assessed by the Anderson–Darling test and the Shapiro–Wilk test. Comparisons of continuous variables among three groups were performed by one‐way ANOVA with post hoc Tukey's test for normal or Kruskal‐Wallis test with Dunn's post hoc test for skewed data. Two‐group comparisons used an independent T‐test or Mann–Whitney U test, as appropriate. Categorical data were compared using a Chi‐squared test or Fisher's exact test, with Bonferroni corrections for multiple pairwise comparisons. Post hoc sensitivity testing using subgroups was performed for major findings. Correlations between two continuous variables were determined as exploratory analysis by two‐tailed Pearson's or Spearman's rank coefficient without any corrections for multiple comparisons. p < 0.05 was considered statistically significant for all the analyses.

3. Results

3.1. Participant Characteristics

Participant characteristics (Table 1) showed the cohorts differed in age, with DM and DFU being older than HC. The female‐to‐male ratio was significantly different between the groups, with more males in the DFU group. Most of the DM and DFU participants were overweight or obese with a higher BMI compared with HC. History of smoking was prominent in the diabetes groups, with at least half being current or ex‐smokers, compared with less than one‐quarter in HC being ex‐smokers. Regarding CVD risk factors, the DM and DFU cohorts were well matched, showing no significant differences in hypertension or history of CVD events, while diabetes duration was higher in DFU. Both DM and DFU cohorts were predominantly receiving insulin and/or metformin treatment, and most of them were on lipid‐modifying medications, with no difference between the groups. However, the use of anti‐thrombotic medication was higher in the DFU group. The metabolic parameters (HbA1c and lipid factors) showed inadequate glycaemic control and altered lipid profiles in DM and DFU participants. DFU participants had TC, LDL‐C, and non‐HDL‐C levels within the normal range and significantly lower than those of HC. However, high triglycerides, low HDL‐C and low ApoA1 were noted in both DM and DFU. In DFU, ulcers were either uninfected (58% PEDIS grade 0–1) or had surface infection (42% PEDIS grade 2). University of Texas depth scores were mostly superficial (grade 0, 63%), with some to tendon or joint capsule (grade 2, 32%), and only one to joint or bone (grade 3, 5%). Mean ulcer size was 110.0 mm² and mean duration was 29.3 weeks.

TABLE 1.

Participant demographics and clinical characteristics.

Characteristics	Healthy (HC), N = 18	Diabetes (DM), N = 20	Diabetic Foot Ulcer (DFU), N = 19	Global p‐value
Age (years)	44.6±12.0 (25–64)	59.5±12.2*** (37–78)	69.8±11.7****†† (36–88)	<0.0001
Sex (female/male)	12/6	9/11	2/17*	0.002
BMI	22.96 (21.53–25.39)	28.58** (26.73–34.22)	31.79**** (26.81–35.32)	<0.0001
Smoking (Never/Ex/Current)	15 (83%)/3 (17%)/0 (0%)	10 (50%)/7 (35%)/3 (15%)	8 (42%)/11 (58%)/0 (0%)	0.014
Diabetes Duration (Years)	—	13.6±9.7 (1.1–35)	25.5±14.3** (2–54)	0.005
Hypertension	—	12 (60%)	16 (89%)	0.093
CVD events	—	4 (20%)	7 (39%)	0.243
Medications
Insulin therapy	—	8 (40%)	12 (67%)	0.148
Metformin	—	15 (75%)	14 (78%)	>0.999
Lipid Lowering agents	—	19 (95%)	18 (94.7%)	>0.999
Anti‐hypertensive agents	—	11 (55%)	14 (77%)	0.224
Anti‐thrombotic	—	7 (35%)	15 (83%)**	0.006
HbA1c and Lipids
HbA1c (%)	4.9±0.4 (4.2–5.4)	7.6±1.4**** (5.4–10.0)	8.6±1.8**** (6.2–12.2)	<0.0001
HbA1c (mmol/mol)	31±3.9 (22–36)	59±15.8**** (36–86)	71±20.1**** (44–110)	<0.0001
Total cholesterol (mmol/L)	4.6±0.8 (3.1–6.4)	4.5±1.6 (2.6–7.6)	3.5±0.6***† (2.1–4.4)	0.005
LDL‐C (mmol/L)	2.5 (2.1–3.0)	1.8 (1.4–3.1)	1.6**** (0.9–1.8)	0.0002
HDL‐C (mmol/L)	1.4 (1.3–1.8)	0.9*** (0.7–1.2)	1.0** (0.8–1.3)	0.0005
Cholesterol:HDL‐C	2.9 (2.6–3.6)	3.9** (3.4–5.9)	3.3† (2.5–4.0)	0.002
Non‐HDL‐C (mmol/L)	2.9 (2.6–3.5)	3.1 (2.2–4.7)	2.3*† (2.0–2.5)	0.004
Triglycerides (mmol/L)	0.9 (0.7–1.3)	2.7**** (1.7–4.7)	1.4* (1.2–3.0)	<0.0001
ApoA1 (g/L)	1.60 (1.52–1.86)	1.32** (1.20–1.47)	1.32* (1.14–1.56)	0.004
ApoB (g/L)	0.76 (0.66–0.87)	0.87 (0.66–1.20)	0.76 (0.63–0.89)	0.250
ApoA1:ApoB	2.00 (1.89–2.49)	1.65* (1.16–2.05)	1.83 (1.46–2.38)	0.043

Note: Continuous data are presented as either mean ± SD (range) for normal distribution or median (IQR) for non‐normal distribution. Nominal data are presented as number (%). Statistical significance represented as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. HC; †p < 0.05, ††p < 0.01, †††p < 0.001 vs. DM.

Abbreviations: LDL‐C, low‐density lipoprotein cholesterol; HDL‐C, high‐density lipoprotein cholesterol; Apo, Apolipoprotein.

3.2. WBC Counts and Monocyte Proportions Are Altered in Individuals with Diabetes

To evaluate systemic inflammation, we examined WBC counts and monocyte subset proportions (Table 2). Both the DM and DFU groups showed elevated WBC counts compared with HC, an indication of heightened inflammation. Monocytes, neutrophils, and basophils were significantly increased in DFU compared with HC, whereas no significant differences were observed in DM. The inflammatory biomarkers neutrophil‐to‐lymphocyte ratio (NLR) and monocyte‐to‐lymphocyte ratio (MLR) were also higher in DFU than in both HC and DM. Proportions of monocyte subsets differed significantly across the cohorts. Intermediate monocytes were elevated in both DM and DFU, while non‐classical monocytes were reduced in DFU compared with both HC and DM. No significant changes were observed in classical monocytes.

TABLE 2.

White blood cell and monocyte subset proportions.

Characteristics	Healthy (HC), N = 18	Diabetes (DM), N = 20	Diabetic foot ulcer (DFU), N = 19	Global p‐value
WBC (106/mL)	5.33 (4.65–6.77)	7.10* (6.17–8.02)	8.52*** (6.72–11.05)	0.0002
Lymphocytes (106/mL)	1.92 (1.33–2.15)	2.31 (1.81–2.94)	1.76 (1.39–2.22)	0.041
Monocytes (106/mL)	0.39 (0.34–0.56)	0.50 (0.41–0.55)	0.68*** (0.52–0.82)	0.001
Neutrophils (106/mL)	3.15 (2.24–4.02)	4.08 (3.38–5.06)	5.48*** (4.21–6.99)	0.0003
Eosinophils (106/mL)	0.14 (0.07–0.25)	0.17 (0.08–0.27)	0.22 (0.13–0.31)	0.130
Basophils (106/mL)	0.02 (0.01–0.02)	0.02 (0.02–0.03)	0.04****† (0.03–0.05)	<0.0001
Neutrophil to lymphocyte ratio (NLR)	2.01±0.94 (0.65–4.47)	2.00±0.87 (0.73–4.28)	3.40±1.58**†† (0.97–7.30)	0.0006
Monocyte to lymphocyte ratio (MLR)	0.26 (0.17–0.30)	0.23 (0.18–0.28)	0.44*†† (0.23–0.51)	0.007
Classical monocytes (%)	81.3 (77.9–87.9)	81.6 (77.3–82.8)	83.7 (78.6–84.7)	0.223
Intermediate monocytes (%)	7.40 (6.09–9.90)	10.15* (7.66–13.85)	9.96* (8.65–12.40)	0.016
Non‐classical monocytes (%)	9.41±4.07 (3.43–18.80)	9.33±3.62 (4.21–18.20)	6.78±2.11*† (2.37–10.20)	0.032

Note: Data presented as either mean ± SD (range) for normal distribution or median (IQR) for non‐normal distribution. Statistical significance represented as *p < 0.05, **p < 0.01, ***p < 0.001 vs. HC; †p < 0.05, ††p < 0.01, †††p < 0.001 vs. DM.

3.3. Monocyte Phenotype Is Altered in DFU with a Decreased Expression of Inflammatory and Anti‐Inflammatory Markers

To assess inflammatory skewing of monocytes, phenotypic marker expression was determined on each subset. An impaired monocyte phenotype was observed in DFU, with a diminished expression of both inflammatory and anti‐inflammatory markers on classical monocytes. More specifically, the inflammatory markers CD86, TLR2, and TLR4 were significantly reduced compared with HC, with CD86 and TLR4 further lower than DM (Figure 1A). The anti‐inflammatory marker CD163 was also significantly reduced in both DM and DFU compared with HC, but no difference was detected between DM and DFU (Figure 1B).

FIGURE 1.

Monocyte surface marker expression. Marker expression determined by whole blood flow cytometry and compared between HC (N = 18), DM (N = 20) and DFU (N = 19). (A) Inflammatory markers: CD64, CD86, CD120b, TLR2, and TLR4 (B) Anti‐inflammatory markers: CD163, CD93 (C) CD86/CD163. Data presented as scatter plots with bars indicating mean with SD or median with IQR. Statistical significance was determined by one‐way ANOVA followed by the post hoc Tukey's test or Kruskal–Wallis test followed by Dunn's test, as appropriate. *p < 0.05; **p < 0.01; ***p < 0.001. All p‐values are unadjusted (no FDR applied).

While phenotypic changes were more evident in the classical subset, a few changes were observed in other subsets as well, with an elevation of CD86 on non‐classicals in DM compared with HC. In addition, CD64 was significantly elevated on non‐classical monocytes of DFU compared with HC and DM. Assessing the CD86/CD163 ratio to determine inflammatory/anti‐inflammatory balance, no significant differences were detected, though it trended higher in the DM group compared with HC (Figure 1C).

3.4. Monocyte Gene Expression Is Suppressed in DM and DFU

To obtain a broader understanding of monocytic changes, we assessed inflammatory and anti‐inflammatory gene expression in CD14++ monocytes. Many inflammatory and anti‐inflammatory genes were downregulated in DM and DFU relative to HC (Figure 2). This included a reduction in pro‐inflammatory chemokine receptor genes (CCR4, CXCR4) and inflammatory gene regulators (TLR4, cREL, and HIF1A) in both DM and DFU. Transcription factor IRF5 and pro‐inflammatory cytokine IL‐12A were also downregulated in DM and trended lower in DFU. Several anti‐inflammatory genes linked to tissue remodelling and reparative functions (MRC1, TGFB1, TGFBR1, VEGFA, and VEGFB) were also reduced in both DM and DFU. Anti‐inflammatory metabolic/nuclear receptor PPARA was downregulated in DM and DFU, while PPARD was reduced in DM and trended low in DFU.

FIGURE 2.

Monocyte gene expression of inflammatory and anti‐inflammatory markers. Determined by Nanostring Elements TagSet assay and compared between HC (N = 12), DM (N = 12), and DFU (N = 12). (A) Heatmap of z‐scores illustrating the expression patterns of genes across the study groups. Scale shows z‐score values −2 (low, green) to 4 (high, red). (B) Differently expressed genes based on corrected p‐value for multiple testing (Benjamini and Hochberg FDR). Data presented as scatter plots with bars indicating mean with SD or median with IQR. Pairwise group differences were tested by one‐way ANOVA followed by post hoc Tukey's test or Kruskal–Wallis test followed by Dunn's test, as appropriate. *p < 0.05; **p < 0.01, ****p < 0.0001.

3.5. Monocyte Basal Cytokine Production Is Higher in DFU

We further investigated whether low‐grade inflammation was present in DFU by assessing cytokine production in unstimulated CD14++ monocytes. While a small proportion of CD16+/CD16++ monocytes was present within the enriched CD14++ samples, the proportions did not differ between groups and were therefore unlikely to influence cytokine production. Levels of several cytokines were either below the lower LOD or above the ULOD across the groups, with <25% of values outside LOD for most cytokines, ≤50% for MCP‐1, IL‐17A, and IL‐12(p70) and 98% for TGF‐β. The average level of cytokine production (except IP‐10) was higher in DFU monocytes than in DM and HC (Table 3). This included both inflammatory and anti‐inflammatory cytokines, but significant differences were only observed for TNF and IFN‐γ compared with HC. The cytokine profile of DM monocytes showed no significant differences from that of HC. Serum cytokine profile showed significantly elevated levels of IL‐10 (anti‐inflammatory) and IL‐6 (pro‐inflammatory) in DFU compared with HC. However, no significant differences were observed for serum TNF and IFN‐γ, in contrast to monocytes.

TABLE 3.

Monocyte and serum cytokine profile.

Monocyte cytokines (pg/mL)	Healthy (HC), N = 17	Diabetes (DM), N = 18	Diabetic foot ulcer (DFU), N = 17	Global p‐value
IL‐2	3.96 (2.37–8.03)	4.94 (2.36–8.57)	8.13 (3.19–11.08)	0.062
IP‐10	109.7 (14.81–486.0)	23.43 (11.54–126.6)	26.43 (16.04–124.4)	0.447
IL‐1β	17.61 (9.60–57.98)	28.78 (7.18–1420)	84.47 (16.15–867.7)	0.181
TNF	4.38 (3.13–11.40)	9.83 (4.18–80.98)	107.8* (3.65–1367)	0.045
MCP‐1‡	4806 (362.8–44,061)	13781 (770.5–43426)	43343 (2507–55,995)	0.121
IL‐17a	2.53 (2.04–4.15)	2.69 (2.03–4.83)	4.70 (2.46–5.88)	0.140
IL‐6‡	327.4 (86.98–1786)	324.5 (122.5‐52839)	2556 (82.10‐52789)	0.306
IFN‐γ	6.09 (3.67–10.86)	6.57 (4.80–12.81)	9.96* (6.39–14.46)	0.036
IL‐12†	5.24 (4.24–9.74)	5.34 (4.23–12.09)	9.27 (5.12–16.96)	0.134
IL‐8‡	11897 (1507–32253)	13859 (2873–20,430)	16180 (5956–89,093)	0.255
IL‐4	3.16 (1.82–6.53)	3.12 (1.86–8.20)	6.03 (2.61–10.32)	0.119
IL‐10	3.66 (2.70–15.72)	4.08 (2.70–33.99)	15.38 (2.78–124.15)	0.501
TGF‐β†	17.95 (17.56–21.19)	17.89 (17.24–22.24)	22.01 (18.85–22.29)	0.246
Serum cytokines
IL‐2	1.23 (1.23–2.62)	1.23 (1.23–3.10)	3.16 (1.23–6.81)	0.201
IP‐10	111.5 (74.08–132.9)	116.3 (77.91–170.3)	139.5 (111.0–229.6)	0.052
IL‐1β†	7.63 (7.63–10.14)	7.63 (7.63–16.65)	7.63 (7.63–23.61)	0.426
TNF†	2.32 (1.52–4.39)	1.52 (1.52–3.95)	1.52 (1.52–6.14)	0.819
MCP‐1	160.7 (110.2–218.0)	174.3 (103.6–228.8)	209.1 (120.3–283.4)	0.292
IL‐17A	4.76 (2.35–6.35)	2.85 (1.89–4.18)	3.16 (1.89–5.70)	0.222
IL‐6	2.64 (2.64–4.66)	2.64 (2.64–4.59)	7.11* (3.26–17.29)	0.028
IFN‐γ†	8.03 (8.03–8.03)	8.03 (8.03–8.03)	8.03 (8.03–15.19)	0.148
IL‐12†	2.78 (2.78–4.21)	2.78 (2.78–4.82)	4.31 (2.78–5.77)	0.190
IL‐8†	3.35 (3.35–4.09)	3.35 (3.35–3.35)	3.35 (3.35–4.08)	0.227
IL‐4	2.81 (1.08–5.87)	1.64 (1.08–6.80)	5.58 (1.86–15.34)	0.155
IL‐10†	2.87 (2.87–2.87)	2.87 (2.87–5.98)	4.12* (2.87–7.74)	0.036
TGF‐β	103.6 (38.93–196.0)	89.22 (52.82–140.4)	38.93 (38.93–197.7)	0.432

Note: Data presented as median (IQR). Statistical significance represented as *p < 0.05 vs. HC. † indicates cytokines with >50% of values below LOD; ‡ indicates cytokines with values above ULOD; percentages indicate the proportion of values outside detection range: Monocyte: IL‐2 (8%), IP‐10, IL‐1β (13%), TNF (31%), MCP‐1, IL‐17A (50%), IL‐6 (21%), IFN‐γ (25%), IL‐12 (56%), IL‐8 (17%), IL‐4 (15%) and TGF‐β (98%). Serum: IL‐2, IL‐6 (50%), IL‐1β (69%), TNF (58%), IL‐17A (35%), IFN‐γ (85%), IL‐12 (62%), IL‐8 (83%), IL‐4 (31%), and TGF‐β (38%).

Abbreviations: IL, interleukin; IP, interferon‐gamma‐inducible protein; TNF, tumour necrosis factor; IFN, interferon; MCP, monocyte chemoattractant protein; TGF, transforming growth factor.

3.6. Monocyte Phenotype Correlates with Altered Lipid Profile and Poor Glycaemic Control

As diabetes is characterised by a triad of poor glycaemic control, altered lipid profile, and inflammation, we explored whether monocyte phenotype associates with any of these parameters. Considering classical monocytes are the major subset, all correlations were performed on this subset. Differently expressed phenotypic markers (between the cohorts) were analysed relative to HbA1c and lipid factors (Table S2). HbA1c levels showed moderate to strong inverse correlations with monocyte CD86, TLR2, TLR4, and CD163 markers (Figure 3A). Exploring the relationship between monocyte phenotype and lipids revealed several associations; CD163 expression on monocytes correlated with HDL‐C and inversely with triglycerides and the cholesterol:HDL‐C ratio. In addition, LDL‐C showed correlations with TLR2 and TLR4 (Figure 3B).

FIGURE 3.

Correlation of monocyte phenotype with metabolic profile. (A) HbA1c (N = 56) with CD86, TLR2, TLR4, and CD163; (B) LDL‐C (N = 51) with TLR2 and TLR4. HDL‐C (N = 56), cholesterol:HDL‐C (N = 56) and triglycerides (N = 56) with CD163. Data presented as scatter plot: HC in orange circles, DM in red squares, and DFU in maroon triangles. Samples excluded from analysis, wherever lipid data were not available. Statistical significance was determined by two‐tailed Pearson's correlation or Spearman's rank correlation*; r/r* represents the correlation coefficient; p < 0.05 considered statistically significant.

3.7. Some Monocyte Changes in DFU May be Related to Advanced Age and BMI

Given the demographic differences between the groups, such as age, BMI, sex, and smoking status, we explored whether the observed monocyte changes were influenced by these variables (Table S3). Interestingly, advanced age was significantly associated with increased total monocytes, intermediate proportions and a reduction in TLR4 expression. Similarly, BMI was found to correlate positively with total monocytes and negatively with CD163 expression. Monocyte CD86, TLR4, and CD163 were found to be lower in males, whereas low CD86 and TLR4 were related to smoking (ex/current). Sensitivity analysis determined robustness of results (Figure S3). CD86 and TLR4 remained significantly lower in DFU in the male‐only subgroup, non‐obese (BMI <30) subgroup, and in the middle age tertile, indicating a more robust result. TLR2 only remained lower in DFU for the male‐only subgroup, and CD163 only remained lower in the middle age tertile, indicating less robust results.

4. Discussion

In this study, we examined monocyte phenotypic and functional alterations in diabetes, including individuals with DFUs. Although circulating monocyte counts and intermediate monocyte proportions were elevated in both DM and DFU, indicative of increased inflammation, monocytes did not appear to exhibit a more inflammatory phenotype. Indeed, several inflammatory markers, along with anti‐inflammatory CD163, were reduced in DFU, and both DM and DFU showed decreased expression of distinct inflammatory and anti‐inflammatory genes compared with HC. Despite this reduced gene expression, monocytes from DFU showed increased basal production of both inflammatory and anti‐inflammatory cytokines, with significantly higher levels of TNF and IFN‐γ in addition to elevated circulating IL‐10 and IL‐6, indicating low‐grade inflammation.

The demographic differences, such as older age, longer diabetes duration and poor glycaemic control in DFU participants, were not surprising, as these factors increase the risk of DFUs [20]. Likewise, the predominance of males in the DFU group is in line with DFUs being more prevalent in men [21]. The altered lipid profile (low HDL‐C, ApoA1 and high triglycerides) is typically seen in individuals with diabetes, whereas the reduced TC and LDL‐C in the DFU cohort likely reflects effective lipid‐lowering by medications [22].

The elevated WBC counts in DM and DFU align with several reports, showing associations with the incidence of diabetes and being a risk factor for chronic non‐healing DFUs [23, 24]. Further elevated monocytes and neutrophils in DFU, along with high NLR and MLR, reflect the systemic inflammatory response in individuals with DFUs. Notably, these parameters are important prognostic biomarkers for DFU healing [25, 26]. The expansion of intermediate monocytes in DM and DFU further supports enhanced inflammation, as intermediates are functionally classified as inflammatory [27, 28]. While other studies may not have reported subset alterations in diabetes [13, 29], our finding likely arises from using a standardised and validated gating approach to delineate the subsets [18]. However, intermediate monocytes being raised in both DM and DFU indicate diabetes as the cause rather than the presence of foot ulcers. While non‐classical monocytes are typically elevated in inflammatory conditions [28], reduced levels observed in our DFU cohort are consistent with other studies reporting fewer non‐classical monocytes in DFU with PAD than in DFU without PAD, in unhealed versus healed DFU, and in diabetes with complications compared with diabetes without complications [29, 30, 31]. This suggests a link between reduced non‐classical monocytes and more severe diabetes cases despite their inflammatory nature.

Despite the alterations in intermediate and non‐classical proportions, the monocyte activation profile differed largely in the classical subset. Reduced CD163 expression in DM and DFU monocytes is in line with other studies in type 2 diabetes [12, 13, 17] and may indicate inflammatory skewing of monocytes. Indeed, increased soluble CD163 (sCD163) is observed in diabetes, which likely correlates with low surface CD163, as monocytes shed CD163 under inflammatory conditions [32]. Moreover, low CD163 reflects reduced monocyte differentiation into macrophages, particularly those adopting an anti‐inflammatory (M2‐like) phenotype. Notably, reduced CD163 expression in both DM and DFU aligns with studies reporting lower CD163 in a mixed cohort with diabetic complications, while no differences were detected between those with and without DFUs [16, 33]. This suggests that diabetes itself drives this inflammatory skewing even before ulcers develop, subsequently leading to impaired healing.

Considering low CD163 is indicative of monocyte inflammation, the initial expectation was higher inflammatory activation markers in DM and DFU. In contrast, no changes were detected in DM, whereas DFU monocytes showed reduced CD86, TLR2 and TLR4 expression. Monocyte CD86 expression was found to vary across studies, with it being elevated, reduced or unchanged in diabetes [17, 34, 35]. However, an elevation in TLR2/4 was reported in type 2 diabetes and diabetic neuropathy compared with healthy individuals [36, 37]. Reduced expression of these markers in our cohort could be due to metformin and/or lipid‐lowering therapies, which may dampen inflammation by lowering TLR2/4 levels [38, 39]. Indeed, our exploratory analysis showed that TLR2/4 correlated with LDL‐C, which may explain their decrease in DFU, where LDL‐C was significantly lower than in controls, unlike in DM. Taken together, the altered monocyte phenotype indicates a suppressed activation state, potentially impairing the subsequent macrophage phenotype and its functions, consistent with an insufficient macrophage response seen in early diabetic wounds [40]. Importantly, low monocyte CD163 may lead to fewer M2 macrophages, which likely contributes to poor healing in DFU, as M2 macrophages are necessary for collagen production and tissue remodelling. While the markers examined here can be classed as M1/M2 markers, macrophages exist as a continuum of functional states and exhibit significant heterogeneity and plasticity in various microenvironmental conditions. As such, the simplistic M1/M2 dichotomy lacks the potential to capture the full spectrum of macrophage functionality and plasticity in vivo [41, 42]. Indeed, monocytes and macrophages display coexistence of M1 and M2‐related markers with overlapping functions and can transform from one phenotype to another in the course of disease progression or in response to a new trigger. Moreover, functional polarisation of monocyte‐derived macrophages primarily occurs in the tissue and can adopt different opposing phenotypes based on microenvironmental cues [15, 43, 44]. Therefore, circulating monocyte phenotype may not predict tissue behaviour. Regardless, the specific markers expressed on monocytes do link to their cellular profile and functionality, irrespective of the macrophage phenotype they ultimately adopt.

The altered gene expression of monocytes broadly correlates with the impaired monocyte phenotype, at least in DFU. Specifically, the downregulation of anti‐inflammatory genes such as MRC1, TGFB1, TGFBR1, VEGFA, and VEGFB in DM/DFU indicates a reduced ability of monocytes to perform anti‐inflammatory and pro‐resolving functions, including tissue repair and remodelling [45, 46]. While this points to an inflammatory monocyte profile, the concomitant downregulation of inflammatory genes suggests a shift away from the pro‐inflammatory state, indicating dysregulated or exhausted monocytes, as seen in chronic infections [47]. In the DM group, inflammatory gene expression did not align with the monocyte phenotype. Post‐translational regulations or gene expression analyses that did not include the entire cohort may explain this disparity. Nevertheless, the suppressed gene expression in DFU monocytes implies defective activation, characterised by transcriptional impairments in cellular activation and immune response.

Despite a suppressed monocyte phenotype in DFU, the higher cytokine production by unstimulated DFU monocytes indicates low‐grade inflammation, likely resulting from prolonged metabolic stress, persisting even after withdrawing the challenge. While the increased production of TNF and IFN‐γ by DFU monocytes appears conflicting with the reduced inflammatory surface marker expression, it complements the low CD163 expression. TNF and IFN‐γ are known to downregulate monocyte CD163 expression and promote activation of CD163⁻ M1 macrophages [15, 48]. However, despite low CD163 in DM, unchanged IFN‐γ suggests a more complex relationship attributable to multiple factors. Nevertheless, the low‐grade inflammatory state of DFU monocytes, combined with an impaired monocyte phenotype, may indicate dysregulated monocyte activation and its pathological consequences in wound healing [49, 50]. While we saw elevated serum IL‐6 and IL‐10 in the DFU group, no changes were detected in serum TNF and IFN‐γ, which may be because serum cytokines are not solely produced by monocytes but by a range of other circulating cells. Nevertheless, elevated serum IL‐6 indicates low‐grade inflammation in DFU, consistent with other studies. However, increased IL‐10 appears at odds, as it is an anti‐inflammatory cytokine. Considering IL‐10 has an immunosuppressive effect, its elevation may contribute to the observed impaired monocyte activation in DFU. Similar findings have been reported in uncontrolled or severe diabetes and are likely an altered immune response to metabolic stress, indicating hyperinflammation [51].

Collectively, our data imply a mixed or altered dysfunctional monocyte profile in individuals with DFU, distinct from traditional dichotomous M1/M2 signatures. While displaying features of chronic low‐grade inflammation by constitutive cytokine production, a suppressed activation phenotype perhaps reflects a state of exhaustion. It is likely that upon prolonged or repeated exposure to challenges (like metabolic stress from hyperglycaemia, ageing, and acute infections), monocytes adopt a paradigm of inflammatory suppression reflecting a dysfunctional exhausted state as seen in sepsis, while maintaining or altering cytokine responses [52, 53]. This is well evidenced by reduced differentiation (low monocyte CD163), low‐grade inflammation (increased basal cytokine production), and immune suppression (noted by decreased activation markers CD86 and TLR2/4) in DFU. Alternatively, this pattern may also indicate a partial endotoxin tolerance phenotype, as reported in chronic low‐grade endotoxemia in diabetes [54, 55]. However, in our cohort, we observed an enhanced cytokine response, contrary to the blunted response in tolerance models. Hence, we propose that DFU monocytes show inflammatory reprogramming with partial tolerance features that would eventually result in dysfunctional monocytes.

Correlation of metabolic profile with monocyte phenotype indicates that poor glycaemic control and altered lipids may influence monocyte skewing. While the inverse correlation between HbA1c and monocyte CD163 was expected, the negative associations with inflammatory markers were somewhat counterintuitive [17]. We propose that poor glycaemic control may exert two‐pronged effects, potentially driving impaired monocyte activation. The positive correlation of HDL‐C with CD163 aligns with the anti‐inflammatory functions of HDL‐C [14, 27, 56]. Importantly, low HDL‐C and high cholesterol:HDL‐C are typical in individuals with diabetes, and their association with reduced CD163 highlights the clinical significance of increased CVD risk and delayed ulcer healing [57, 58]. The inverse association of triglycerides with CD163 indicates the pro‐inflammatory effects of triglycerides, which aligns with the association of sCD163 with high triglycerides and inflammatory biomarkers in people with diabetes [59]. Interestingly, TLR2/4 were the only markers that correlated with LDL‐C, aligning with the upregulation of monocyte TLR2/4 in response to LDL‐C [60]. Although we observed monocyte skewing relative to HbA1c and altered lipids, a subsequent confirmatory analysis with formalised hypotheses is needed to confirm these associations. Furthermore, it would be worthwhile to examine whether any causal relationship exists between metabolic control and monocyte phenotype.

Considering demographic differences between the study groups, it was valuable to explore whether these factors could contribute to the observed monocyte alterations. The significant association of age with monocyte count and intermediates suggests that advanced age of DM and DFU participants could contribute to these changes. The literature instead reports that monocyte counts are unchanged or decrease with age [61], indicating that the observed association is likely driven by the diabetes groups being older, rather than age itself. Elevated intermediate proportions, conversely, have been reported with advanced age [62]; thus, the elevated intermediate monocytes in DM and DFU are likely at least partially age‐related. BMI association with monocytes and CD163 expression is not surprising, as obesity strongly correlates with diabetes [12, 13], though even within the non‐obese subgroup, CD163 remained lower in the DFU group. The lower expression of CD86, TLR4 and CD163 in DM/DFU groups and their association with males makes it difficult to discern whether the reduction is due to predominance of males or presence of diabetes. We previously reported no sex effect on these markers in healthy individuals [27]. Moreover, in the male‐only subgroup analysis, all markers except CD163 remained lower in the DFU group, indicating sex imbalances between groups are not the cause for these differences.

4.1. Data Limitations

As with any research, this study has certain limitations. Notably, it may have been underpowered for some markers; although the sample size was predetermined by power calculation, the standard deviation for some of the markers exceeded our estimates. Moreover, the ideal detectable effect sizes between groups remain difficult to establish, meaning some interesting differences may have been undetectable. Additionally, the small sample size prevented correlations within each group, and associations observed in the combined cohort may have been influenced by demographic factors, potentially masking group‐specific correlations. More importantly, the correlation analyses performed were largely exploratory, which can result in ‘false positives’. Similarly, multiple‐testing correction (e.g., FDR) was not used for cell surface markers; therefore, results may include an increased risk of false positives. There was variability within sample groups, such as the DFU group including a mix of early and chronic ulcers, that may have influenced our findings. The modest sample size limits stratification by clinical parameters, which would otherwise offer a more detailed monocyte profile. While DM and DFU cohorts showed no major differences in diabetes medications and CVD‐related factors, human health is complex, and the potential influence of unassessed comorbidities or treatments cannot be excluded. Lastly, as cytokine analysis saw many values out of LOD, single imputations were applied, which may introduce bias or reduce the accuracy of the results. However, these findings provide direction for a larger study to confirm these differences and associations.

5. Conclusions

This study revealed suppression of both inflammatory and anti‐inflammatory markers in monocytes of individuals with DFU despite an increase in systemic inflammation, indicating monocyte dysfunction. Chronic low‐grade inflammation from metabolic stress could be driving this suppression, leading to altered monocyte activation, resulting in impaired immune responses/functions. The suppressed immune response could potentially affect the normal healing process in people with diabetes, making it refractory. However, further studies are needed to examine the broader functional aspects and the causal factors for this immune suppression and approaches to target these for intervention. Our findings highlight the predictive and prognostic scope of monocyte activation state and the necessity of expanding therapeutic intervention to address anti‐inflammatory strategies to improve outcomes.

Author Contributions

R.M.: study conception and design, acquisition of study data, analysis and interpretation, researched the data and wrote the manuscript. H.M.: study conception and design, reviewed/edited the manuscript. R.B.: study design, reviewed/edited the manuscript. M.V.: acquisition of study data, reviewed/edited the manuscript. S.L.: acquisition of study data, reviewed/edited the manuscript. J.G.: acquisition of study data, reviewed/edited the manuscript. S.L.: acquisition of study data, reviewed/edited the manuscript. H.W.: study conception and design, analysis, reviewed/edited the manuscript.

Funding

Funding was provided by NSWHP‐ICMPR‐Westmead Private Practice Fund and Westmead Hospital‐WSLHD, NSW Health.

Ethics Statement

This study was approved by the Western Sydney Local Health District Human Research Ethics Committee (HREC/2020‐ETH01856) and conducted in accordance with the ICH‐GCP guidelines. Written and informed consent was obtained from all the study participants.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: eji70258‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request.