The Functionality of High-Density Lipoproteins Is Impaired in People with Diabetes Who Require Minor Amputations

Abstract

Objective:

High-density lipoprotein (HDL) functionality is emerging as a novel predictor of disease outcomes. The role of HDL functionality in patients with diabetes-related amputations is unexplored. We aimed to assess changes in HDL functionality in people with diabetes who had minor amputations and to determine the relationship between HDL functionality and wound closure (WC).

Approach:

Thirty patients with diabetes mellitus (DM) and 11 without (non-DM) undergoing minor amputations were recruited. Blood was collected at baseline (DM, n = 30; non-DM, n = 11), 1 month (DM, n = 22; non-DM, n = 5), and 6 months (DM, n = 14; non-DM, n = 3) postamputation and from 20 healthy gender- and age-matched control participants. HDL was isolated from plasma and functionality markers of macrophage cholesterol efflux, and anti-inflammatory and proangiogenic capacities in endothelial cells were assessed. The study adhered to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Results:

HDL-cholesterol (HDL-c) levels positively correlated with WC (r = 0.44, p < 0.05). Cholesterol efflux capacity of DM HDL was reduced 1-month postamputation (−44% vs. non-DM HDL, d = −1.5, p < 0.05). Following inflammatory stimulation, DM HDL-treated cells had elevated levels of C-X3-C motif chemokine ligand 1 (+92%, d = 0.9), C-C motif chemokine ligand 2 (+49%, d = 0.8), vascular cell adhesion molecule 1 (+67%, d = 0.9), and intracellular adhesion molecule 1 (+58%, d = 0.9) (vs. non-DM HDL, p < 0.05). The capacity of endothelial cells to form tubules reduced linearly with time following DM HDL treatment (p < 0.05), concomitant with reduced vascular endothelial growth factor A (−47% vs. non-DM HDL, d = −1.2, p < 0.05). Endothelial cells treated with HDL from participants with delayed WC (<50%, 1-month postamputation) exhibited increased v-rel avian reticuloendotheliosis viral oncogene homolog A (RELA) expression (+52%, d = 1.2, p < 0.05). The impairment of HDL functionality in DM was concomitant with reduced apolipoprotein AI (−34%, d = 1.2) and paraoxonase 1 (−32%, d = −1.3, vs. non-DM, p < 0.05) protein in HDL 1-month postamputation.

Innovation:

An HDL panel, including HDL-c and HDL functionality and composition, could be used in early screening to identify patients who may benefit from early therapeutic intervention, guiding wound care management.

Conclusion:

Impaired HDL functionality, mediated through changes in HDL composition, may contribute to delayed wound healing.

Graphical abstract

Keywords

HDL functionality angiogenesis inflammation diabetes diabetes-related foot ulcer wound healing

INTRODUCTION

People with diabetes experience an increased risk of cardiovascular disease (CVD) and vascular complications. Furthermore, the development of diabetes-related foot ulcers (DFUs) is a common major complication. More than 25% of patients with diabetes will experience a DFU in their lifetime, of which 20% will require amputation.¹ Despite the clinical burden of DFUs, there are currently no effective tools to predict wound healing. While markers exist for glycemic control (HbA1c), peripheral arterial function, and infection, there is currently no single biological marker that predicts whether a wound will heal in a timely manner or develop into a chronic nonhealing wound. Due to the weak quality of evidence for predicting DFU outcomes, it is strongly recommended by the International Working Group on the Diabetic Foot (IWGDF) that personalized systems are used for predicting individual patient wound healing outcomes.² Predicting the divergence in wound trajectories is difficult due to the contribution of a large range of factors to wound healing rate. This highlights an unmet clinical need to identify novel biomarkers that predict wound healing outcomes to guide optimal wound care and ultimately reduce the prevalence of amputations.

Christina A. Bursill, BSc (Hons 1), PhD

Wound healing is a complex process consisting of distinct yet overlapping phases of hemostasis, inflammation, proliferation, and vascular remodeling. Inflammation plays a key role in facilitating effective wound healing by mediating the immune response to infection. However, prolonged or exacerbated inflammation is a key cause of chronic nonhealing wounds in diabetes.³ Angiogenesis is important at all stages of wound healing⁴ but especially during the proliferative phase for effective revascularization of healing tissue. Angiogenic responses to wound hypoxia are essential for healing and are impaired in people with diabetes.⁵

High-density lipoprotein (HDL) exhibits several vascular protective effects (e.g., anti-inflammatory and proangiogenic properties) as well as the ability to efflux cholesterol (including from macrophages in atherosclerotic plaques).^6–9 The cholesterol efflux capacity (CEC) of HDL is also thought to aid tissue repair through removal of lipids from dead cells after acute injury.¹⁰ The anti-inflammatory and proangiogenic properties of HDL are mechanisms that underpin HDLs’ beneficial role in promoting wound healing in diabetes. Previous studies have shown that topical application of reconstituted HDL (rHDL) improves wound healing in diabetic mice. This is shown to be mediated through the uptake of HDL into key wound cells, including fibroblasts, macrophages, keratinocytes, and endothelial cells, where it regulates the expression of inflammatory genes, including Rela and Ccl2.⁶ Furthermore, rHDL has been shown to rescue diabetes-impaired wound healing by increasing wound angiogenesis through regulation of the pyruvate dehydrogenase kinase 4/pyruvate dehydrogenase complex axis response to wound hypoxia⁷ and key angiogenesis regulators including hypoxia inducible factor 1α (HIF-1α) and vascular endothelial growth factor A (VEGFA).⁹

It is well-established that circulating HDL-cholesterol (HDL-c) levels have an inverse relationship with CVD¹¹ and are found to be associated with the risk of amputation in patients with DFUs.¹² HDL functionality, however, is emerging as an improved predictor of disease outcomes over HDL-c concentration alone and is found to be impaired in patients with coronary artery disease.¹³ However, whether HDL functionality changes in patients with diabetes-related amputations remains unknown. The objectives of this study sought to assess HDL functionality in people with diabetes who had minor amputations and to determine the relationship between HDL functionality and wound closure (WC).

INNOVATION

We previously found HDL exerts anti-inflammatory and proangiogenic effects, mechanisms that underpin HDL’s beneficial role in promoting wound healing in diabetes. Circulating HDL-c levels inversely associate with amputation risk in patients with DFUs, with HDL functionality emerging as a predictor of disease outcomes. This study assessed HDL functionality in patients with diabetes-related amputations. We found that HDL functionality is impaired in people with diabetes who had minor amputations, driven by altered HDL composition, which is associated with delayed healing. These findings suggest that an HDL panel, including HDL-c, and HDL functionality and composition may support the identification of patients who would benefit from early therapeutic intervention, guiding wound care management approaches.

CLINICAL PROBLEM ADDRESSED

More than 25% of patients with diabetes develop DFUs. One in five patients with moderate-to-severe DFU (WIfI stage 3 or 4 based on classification of wound severity [location, size, and depth], presence of ischemia, and presence and severity of foot infection)¹⁴ will require lower-limb amputation, after which survival is exceedingly poor at <50% after 5 years. Currently, there are no effective tools to predict whether a patient will develop a chronic nonhealing DFU, highlighting an unmet clinical need. Circulating HDL cholesterol levels are inversely associated with CVD and amputation risk in patients with DFUs. For CVD, HDL functionality is emerging as an improved predictor of outcomes. The key functions of HDL include anti-inflammatory and proangiogenic properties and the ability to efflux cholesterol, all important processes in wound healing. The discovery of novel blood-based predictors of wound healing success will guide optimal wound care for improved patient outcomes.

MATERIAL AND METHODS

Electronic laboratory notebook was not used.

Study design and cohort characteristics

Blood samples were collected from patients with type 1 or type 2 diabetes mellitus (DM group) and without diabetes (non-DM group) admitted to the Royal Adelaide Hospital Vascular Surgery department, Adelaide, SA, for a minor (ankle preserved) amputation (i.e., ray amputation that includes excision of the toe and metatarsal head), approved by the Central Adelaide Local Health Network Human Research Ethics Committee, HREC/19/CALHN/349, and abide by the Declaration of Helsinki principles. Participants were identified from the elective or emergency vascular surgical list and were approached before surgery. For non-DM patients, indications for amputation included local infection or infected gangrene of the toes and chronic limb-threatening ischemia. No amputations were related to trauma.

Eligibility criteria for participants were as follows: recent minor amputation; for the DM group: diagnosis of diabetes (types 1 and 2) and HbA1c<108 mmol/mol (12%). Lipid-lowering drugs (e.g., statins and ezetimibe) were allowed. The exclusion criteria included history of pancreatitis or cancer with ongoing treatment or prognosis of <5 years, hyperbaric oxygen treatment (HBOT) within 6 months, and use of negative pressure wound therapy (NPWT) for the current wound or below-knee amputation. We excluded patients undergoing NPWT and HBOT as they are conditionally recommended by the IWGDF as adjunct therapies for wound healing² and may thus have affected wound healing rates. Patients with severe ischemia at the time of screening for the study (toe pressure < 40mm Hg and/or ankle-brachial index < 0.4) were excluded, as wound healing is unlikely in this cohort of patients. All study participants gave informed consent. At recruitment, participant demographics and clinical data were collected. Blood was collected at baseline (within 1 to 2 weeks of amputation surgery; DM, n = 30; non-DM, n = 11, Fig. 1A), 1 month (DM, n = 22; non-DM, n = 5), and 6 months (DM, n = 14; non-DM, n = 3) postamputation. Blood samples were also obtained from 19 healthy gender- and age-matched participants without diabetes and ulcers (healthy group). These timepoints were chosen as they are clinically meaningful with respect to wound healing. Most acute wounds are expected to heal within 4 weeks, whereas chronic wounds may take at least 6–12 months to heal. Wounds were photographed and analyzed using a 3D WoundVue camera (WoundVue, LBT Innovations),¹⁵ allowing topological wound area assessment, at baseline and 1-month postsurgery, or until completely healed. The frequency of follow-up visits was determined by clinical need.

Figure 1.

Flow diagram of participant recruitment; plasma HDL-c and HbA1c levels correlate with wound closure. Wound area was captured and analyzed using a 3D WoundVue camera. (A) Percent change in wound closure at 1 month postamputation (n = 3–20/group). Correlation between wound closure at 1 month and (B) HbA1c (%, mmol/mol) at baseline, and (C) plasma HDL-c (mM) at 1 month. Linear correlation was assessed by Pearson’s or Spearman’s correlation, with significance set as p < 0.05; r is the correlation coefficient. Participants were stratified by HDL-c level clinical cut-off (low: <1.0 mM; high: >1.0 mM). (D) Percent change in wound closure in participants with high and low HDL-c levels (n = 6–17/group). Data are expressed as mean ± SD. *p < 0.05 by unpaired independent t-test. HDL-c, high-density lipoprotein cholesterol; SD, standard deviation.

Measurement of clinical parameters

HbA1c, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and full blood count were analyzed by SA Pathology. Total cholesterol, HDL-c, and triacylglycerol concentrations were determined enzymatically (Roche Diagnostics, Indianapolis, Indiana, USA). Low-density lipoprotein-cholesterol (LDL-c) was calculated using the Friedewald formula.¹⁶

Preparation of HDL

HDL was isolated from plasma samples following polyethylene glycol precipitation of apolipoprotein B-containing lipoproteins.⁹

Enzyme-linked immunosorbent assay (ELISA)

Isolated HDL was assessed for the concentration of apolipoprotein AI (apoAI), serum amyloid A (SAA), and paraoxonase 1 (PON1) using commercially available Enzyme-linked immunosorbent assay kits (RayBiotech: ApoAI, ELH-ApoA1; SAA, ELH-SAA-1; ThermoFisher: PON1, EH376RB).

Cholesterol efflux assay in macrophages

Immortalized murine bone marrow-derived macrophages (iBMDMs) were kindly provided by Dr Ashley Mansell (Hudson Institute of Medical Research, Victoria, Australia). iBMDMs were seeded at 1 × 10⁵ cells/well and incubated for 24 h. iBMDMs were loaded with [³H] radiolabeled cholesterol (2 µCi/mL) and incubated for 24 h and then equilibrated with 0.2% BSA for 18 h. Cells were incubated with phosphate-buffered saline (PBS) (control) or participant HDL (25 µg/mL) for 4 h. Movement of radiolabeled cholesterol from cells to media containing HDL was quantified using the Tri-Carb 4810TR scintillation counter (PerkinElmer, Waltham, Massachusetts, USA) and expressed as a percentage of total cell [³H] cholesterol content.

HDL functionality in endothelial cells

Human coronary artery endothelial cells (HCAECs; Cell Applications, San Diego, California, USA) were seeded at 1.5 × 10⁵ cells/well, cultured for 24 h, and then treated for 18 h with PBS (control) or HDL from participant cohorts (0.6 mg/mL). To assess anti-inflammatory properties of HDL, HCAECs were stimulated with tumor necrosis factor α (TNFα, 0.6 ng/mL; Sigma Aldrich, St. Louis, Missouri, USA) for 4.5 h and then harvested for RNA analysis. To determine HDL angiogenic capacity, treated HCAECs were exposed to normal or high glucose for 48 h and then seeded at 1.5 × 10⁴ cells/well on polymerized growth factor-reduced Matrigel (Corning Inc., Corning, New York, USA) and incubated for 6 h. Tubules were imaged at 2.5× magnification under light microscopy. Number of tubules, branch points, and total tubule area was assessed using a bio-network computational model as described previously.¹⁷

Quantitative PCR

Quantitative PCR was performed to measure C-X3-C motif chemokine ligand 1 (CX3CL1), C-C motif chemokine ligand 2 (CCL2), CCL5, RELA, vascular cell adhesion molecule 1 (VCAM1), intracellular adhesion molecule 1 (ICAM1), and VEGFA (Supplementary Table S1). Relative changes in gene expression were calculated using the ^ΔΔC_t method and normalized to B2M.

Statistics

Data are expressed as mean ± standard deviation and analyzed on GraphPad Prism software (v10.0, San Diego, California, USA). Normal distribution of the data was determined using the Shapiro–Wilk test. Differences between groups were assessed using one-way analysis of variance (Tukey’s post hoc), Kruskal–Wallis test (Dunn’s post hoc), independent t-test (unpaired), or Mann–Whitney test, where appropriate. Linear correlation analysis was performed with Pearson’s correlation or Spearman’s correlation tests where appropriate. Statistical significance was set at p < 0.05.

RESULTS

Study participant characteristics

During the study, four individuals required below-knee amputations, two individuals died, four individuals required NPWT, and four individuals required wound suturing, which prevented wound area assessment, four individuals withdrew from the study, and six individuals were lost to follow-up, resulting in lower participant numbers in the later timepoints (Fig. 1A).

Table 1 summarizes the study participant characteristics. Participants with DM had higher levels of HbA1c and CRP compared with both the healthy and non-DM groups at baseline and 1 month postamputation. At baseline, the DM group had elevated levels of ESR and triacylglycerols and reduced total cholesterol and HDL-c levels compared with healthy controls. Interestingly, HDL-triacylglycerols were significantly higher in the non-DM group compared with healthy controls, with a nonsignificant trend observed in the DM group.

Table 1.

Study participant characteristics

Characteristic	Healthy	n	Non-DM	n	DM	n
Age (years)	67 ± 11	19	77 ± 12	11	68 ± 9	30
Male sex (%)	17 (89.5)	19	8 (72.7)	11	26 (86.7)	30
Smoker (%)	N/A		4 (36.4)	11	3 (10.7)	28
Peripheral Arterial Disease (PAD) (%)	N/A		9 (81.8)	11	13 (46.4)	28
Baseline postoperative wound size (mm²)	N/A		266.8 (133.5–943.7)	6	186.3 (100.0–534.8)	22
HbA1c
Baseline (mmol/mol)	37 ± 3.4	18	35.8 ± 4.9	5	69.3 ± 20.2^****,†††	24
Baseline (%)	5.5 ± 0.3	18	5.4 ± 0.4	5	8.5 ± 1.8^***,†††	24
1 month (mmol/mol)	N/A	—	38.0 (30.0–38.0)	3	61 (52.5–65.0)††	16
1 month (%)	N/A	—	5.6 (4.9–5.6)	3	7.7 (7.0–8.1)††	16
ESR (mm/h)
Baseline	5.00 (4.00–6.75)	18	13.00 (1.00–25.00)	2	60.50 (25.75–115.08)^***	8
1 month	N/A	—	61.33 ± 55.52	3	37.77 ± 24.71	13
CRP (mg/L)
Baseline	0.45 (0.40–0.73)	17	8.40 (6.30–41.10)	9	19.10 (5.33–89.43)^****,††	22
1 month	N/A	—	4.60 (2.35–70.90)	5	2.80 (1.10–12.65)	20
Total cholesterol (mM)
Baseline	4.23 (3.34–4.86)	18	3.39 (2.81–4.26)	11	3.16 (2.40–3.70)^*	30
1 month	N/A	—	4.06 ± 1.56	5	4.03 ± 1.49	21
HDL-cholesterol (mM)
Baseline	1.99 ± 0.39	18	1.29 ± 0.36^****	11	1.09 ± 0.37^****	30
1 month	N/A	—	1.46 ± 0.42	5	1.29 ± 0.44	21
LDL-cholesterol (mM)
Baseline	2.03 ± 1.06	18	1.82 ± 0.96	11	1.83 ± 0.99	30
1 month	N/A	—	2.18 ± 1.53	5	2.31 ± 1.14	21
Triacylglycerol (mM)
Baseline	0.96 ± 0.49	18	1.48 ± 0.73	11	1.67 ± 0.72^**	30
1 month	N/A	—	2.14 ± 0.94	5	2.17 ± 0.96	21
HDL-triacylglycerol (mM)
Baseline	0.06 (0.57-0.12)	18	0.11 (0.10–0.49)^*	11	0.09 (0.06–0.44)	30
1 month	N/A	—	0.14 (0.08–0.60)	5	0.13 (0.07–0.43)	21

Data are expressed as either mean ± SD or median (interquartile range).

p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001 versus healthy; ††p < 0.01, †††p < 0.001, ††††p < 0.0001 versus non-DM by one-way ANOVA (Tukey’s post hoc) or Kruskal–Wallis test (Dunn’s post hoc).

ANOVA, analysis of variance; CRP, C-reactive protein; DM, diabetes mellitus; ESR, erythrocyte sedimentation rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; N/A, not applicable; SD, standard deviation.

Baseline total white cell and neutrophil counts were higher in both the DM and non-DM groups compared with healthy controls (Table 2). While no difference was observed in baseline lymphocyte numbers, 1-month lymphocyte numbers were higher in DM compared with non-DM participants. Baseline monocytes and eosinophils were higher in non-DM compared with healthy. Baseline basophil counts were lower in DM compared with both healthy and non-DM groups. Baseline and 1-month platelet numbers were higher in the non-DM group. Baseline hemoglobin, packed cell volume, and red cell distribution were reduced in both DM and non-DM compared with healthy, whereas red cell distribution was elevated in non-DM compared with DM. Clinically, this suggests that patients with DM experience systemic inflammation and elevated lipid profiles.

Table 2.

Study participant blood count

Characteristic	Healthy	n	Non-DM	n	DM	n
White cell count (×10⁹/L)
Baseline	6.6 (5.8–7.2)	17	9.0 (8.0–12.2)^**	10	8.2 (7.0–11.4)^**	30
1 month	N/A	—	8.6 ± 1.5	5	8.7 ± 1.7	20
Neutrophils (×10⁹/L)
Baseline	3.8 (3.0–4.3)	17	6.7 (5.2–8.4)^**	10	5.5 (4.1–8.0)^**	30
1 month	N/A	—	6.3 ± 1.6	5	5.3 ± 1.4	20
Lymphocytes (×10⁹/L)
Baseline	1.9 (1.3–2.2)	17	1.3 (1.1–1.7)	10	1.6 (1.3–2.1)	30
1 month	N/A	—	1.2 ± 0.5	5	2.2 ± 0.6††	18
Monocytes (×10⁹/L)
Baseline	0.5 (0.5–0.7)	17	0.8 (0.6–1.0)^*	10	0.6 (0.4–0.8)	30
1 month	N/A	—	0.6 (0.3–0.8)	5	0.5 (0.4–0.8)	20
Eosinophils (×10⁹/L)
Baseline	0.1 (0.07–0.3)	17	0.3 (0.2–0.4)^*	10	0.1 (0.07–0.2)††	30
1 month	N/A	—	0.4 ± 0.2	5	0.2 ± 0.2	20
Basophils (×10⁹/L)
Baseline	0.06 ± 0.02	17	0.06 ± 0.02	10	0.04 ± 0.02^**,†	30
1 month	N/A	—	0.05 (0.06–0.08)	5	0.04 (0.03–0.08)	19
Platelet count (×10⁹/L)
Baseline	232.5 ± 54.1	17	495.8 ± 250.9^****	10	264.1 ± 99.8††††	30
1 month	N/A	—	551.2 ± 325.5	5	271.9 ± 107.9††	20
Red blood cells (×10⁹/L)
Baseline	4.7 ± 0.2	17	4.0 ± 1.2	10	4.2 ± 0.7	30
1 month	N/A	—	4.1 ± 1.7	5	4.3 ± 0.6	20
Hemoglobin (g/L)
Baseline	143.6 ± 6.2		113.9 ± 28.0^***	10	118.7 ± 18.7^***	30
1 month	N/A		113.6 ± 37.1	5	122.7 ± 17.3	20
Packed cell volume (L/L)
Baseline	0.7 ± 1.0	17	0.4 ± 0.08^**	10	0.4 ± 0.05^**	30
1 month	N/A	—	0.3 ± 0.10	5	0.4 ± 0.05	20
Mean cell volume (fL)
Baseline	84.5 ± 21.9	17	90.9 ± 8.2	10	87.9 ± 8.1	30
1 month	N/A	—	92.9 ± 9.0	5	88.6 ± 6.0	18
Mean cell hemoglobin (g/L)
Baseline	31.0 (30.0–31.5)	17	29.5 (26.0–31.0)	10	29.0 (27.5–31.0)^*	30
1 month	N/A	—	29.2 ± 3.1	5	28.5 ± 3.5	20
Red cell distribution (%)
Baseline	12.8 ± 0.9	17	16.5 ± 2.9^****	10	14.2 ± 1.7^**,†	30
1 month	N/A	—	15.6 ± 1.1	5	14.5 ± 1.6	20
Mean platelet volume (fL)
Baseline	10.3 ± 0.8	16	10.0 ± 1.3	10	10.3 ± 1.0	30
1 month	N/A	—	9.2 ± 0.2	5	10.5 ± 1.0^†	20

Data are expressed as either mean ± SD or median (interquartile range).

p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001 versus healthy; †p < 0.05, ††p < 0.01, ††††p < 0.0001 versus non-DM by one-way ANOVA (Tukey’s post hoc) or Kruskal–Wallis test (Dunn’s post hoc). DM, diabetes mellitus; N/A, not applicable; SD, standard deviation.

Plasma HDL-c and HbA1c levels correlate with WC

Wound areas were measured at baseline and 1 month postamputation and expressed as the percentage change from the starting wound area. The rate of WC was slower in participants with DM (42.5 ± 37.7%) compared with non-DM (77.3 ± 23.1%), although this did not reach statistical significance (p = 0.091, Fig. 1B). Consistent with previous studies,¹⁸ baseline HbA1c (%) levels inversely correlated with WC (Fig. 1C). Plasma HDL-c levels measured at 1 month postamputation positively correlated with WC (Fig. 1D). Furthermore, when participants were stratified according to 1.0 mM, the clinical cut-off for normal HDL level classification, WC rates were 69% higher in participants with HDL-c levels > 1.0 mM compared with those with HDL-c < 1.0 mM (p < 0.05, Fig. 1E). No correlations were found between WC and 1-month HbA1c levels or with platelet or lymphocyte counts (Supplementary Fig. S1). Furthermore, 1-month ESR and CRP did not correlate with WC or any HDL functionality markers (Supplementary Figures S2, S3, and S4). Our findings suggest that patients with reduced HDL-c may experience slower WC.

The composition of HDL from people with diabetes contains less apoAI and PON1

HDL composition, including changes to protein content, is altered in disease states, such as diabetes.^19,20 We measured levels of apoAI, PON1, and SAA in patient HDL. The level of apoAI in HDL isolated from people with diabetes overall was not significantly different between groups (Fig. 2A). Pairwise timepoint comparisons showed that apoAI in baseline HDL from people with DM was −34% lower than those non-DM but did not reach statistical significance (DM: 29.6 ± 9.8 ng/mL vs. non-DM: 45.0 ± 26.7 ng/mL, p = 0.0592; Fig. 2B). However, apoAI levels in 1-month HDL from DM were significantly reduced by 32% compared with HDL from the non-DM group (DM: 35.9 ± 15.5 ng/mL vs. non-DM: 53.0 ± 10.6 ng/mL, p < 0.05, Fig. 2C). HDL PON1 levels in DM reduced linearly with time postamputation (p < 0.05, Fig. 2D). Pairwise timepoint comparisons showed no change in HDL PON1 at baseline (Fig. 2E) but a significant 33% reduction in PON1 in 1-month HDL from DM compared with the non-DM group (DM: 7.8 ± 2.4 ng/mL vs. non-DM: 11.6 ± 4.5 ng/mL, p < 0.05, Fig. 2F). SAA was not significantly different across groups (Fig. 2G).

Figure 2.

The composition of HDL from people with diabetes who require minor amputation contains less apoAI and PON1, which is associated with delayed healing postamputation. Participant HDL was assessed for protein levels of apoAI, PON1, and SAA by ELISA. (A–C) ApoAI in HDL, (D–F) PON1 in HDL, and (G) SAA in HDL. Wound area was captured and analyzed using a 3D WoundVue camera. Participants were stratified by percent change in wound closure (WC) defined as nondelayed (WC > 50%) or delayed (WC < 50%) closure at 1 month postamputation. (H) ApoAI in HDL from people with delayed versus nondelayed healing, (I) PON1 in HDL from people with delayed versus nondelayed healing, (J) SAA in HDL from people with delayed versus nondelayed healing. Results are expressed as mean ± SD (healthy, n = 9, non-DM baseline, n = 10 to 11, non-DM 1 month, n = 5, DM baseline, n = 19–21, DM 1 month, n = 19, DM 6 months, n = 10 to 11). ^*p < 0.05, by either one-way ANOVA (Tukey’s post hoc), Kruskal–Wallis test (Dunn’s post hoc), Mann–Whitney test, or unpaired independent t-test. †p < .05, linear trend as assessed by one-way ANOVA. ANOVA, analysis of variance; iBMDMs, immortalized bone marrow derived macrophages; DM, diabetes mellitus; ELISA, Enzyme-linked immunosorbent assay; HDL, high-density lipoprotein;

Participants were stratified by nondelayed WC (>50%) versus delayed WC (<50%) at 1 month postamputation. This threshold was chosen as it aligns with previous studies.²¹ Baseline HDL from people with nondelayed WC (>50%) had 41% higher levels of apoAI (45.3 ± 21.8 ng/mL) compared with those with delayed WC (<50%: 26.7 ± 12.8 ng/mL, p < 0.01, Fig. 2H). Similarly, baseline HDL from people with nondelayed WC (>50%) had 30% higher levels of PON1 protein (119.4 ± 52.3 ng/mL) compared with those with delayed WC (<50%: 84.1 ± 42.6 ng/mL, p < 0.05, Fig. 2I). There was no difference in the amount of SAA in HDL from either of the two groups (Fig. 2J). Clinically, our findings show patients with low HDL apoAI and PON1 may be at increased risk for delayed WC.

HDL from people with diabetes who had a minor amputation has impaired CEC but recovers with time postamputation

HDL functionality, measured by cholesterol efflux, is inversely associated with incidence of coronary heart disease and is emerging as a better predictor of disease outcomes than HDL-c.²² We found that the CEC of healthy HDL (184.8 ± 117.4%, p < 0.05) and baseline non-DM HDL (187.3 ± 89.6%, p < 0.05) was higher than the PBS control (100 ± 27.2%, Fig. 3A). CEC of baseline and 1-month DM HDL was not different from PBS, suggesting an inability to augment cholesterol efflux. However, 6-month DM HDL CEC (181.8 ± 75.2%) was significantly higher (+82%) than PBS (p < 0.05). Furthermore, CEC of DM HDL increased linearly with time following amputation (p < 0.05). Pairwise timepoint comparisons showed no difference between DM and non-DM HDL at baseline (Fig. 3B). However, CEC of 1-month HDL was significantly lower (−41%) in DM HDL (131.0 ± 38.6%) compared with non-DM HDL (222.9 ± 133.7%, p < 0.01, Fig. 3C). No correlations were found between WC and CEC at baseline and 1 month postamputation (Supplementary Fig. S5).

Figure 3.

HDL from people with diabetes who had a minor amputation has impaired cholesterol efflux capacity but recovers with time postamputation. iBMDMs were loaded with [³H] radiolabeled cholesterol (2 µCi/mL, 24 h) and treated with PBS or participant HDL (25 µg/mL, 4 h). (A) Cholesterol efflux capacity was quantified using a scintillation counter. Pairwise comparisons of cholesterol efflux capacity of (B) baseline and (C) 1-month HDL. Each experiment was performed independently in triplicate. Results are expressed as mean ± SD (PBS, n = 11, healthy, n = 16, non-DM baseline, n = 10, non-DM 1 month, n = 5, DM baseline, n = 28, DM 1 month, n = 20, DM 6 months, n = 13). ^*p < 0.05, ^**p < 0.01 by either one-way ANOVA (Tukey’s post hoc), Kruskal–Wallis test (Dunn’s post hoc), or unpaired independent t-test. †p < 0.05, linear trend as assessed by one-way ANOVA. ANOVA, analysis of variance; iBMDMs, immortalized bone marrow derived macrophages; DM, diabetes mellitus; HDL, high-density lipoprotein; PBS, phosphate-buffered saline; SD, standard deviation.

HDL from people with diabetes who had a minor amputation exhibits a higher inflammatory response, which is exacerbated with time postamputation

HDL has anti-inflammatory properties and regulates the expression of inflammatory mediators, which play important roles in wound healing, including nuclear factor-ĸB p65 (coded for by RELA), CX3CL1, CCL2, CCL5, VCAM1, and ICAM1.^23,24 We next investigated the anti-inflammatory properties of participant HDL by assessing its ability to suppress TNFα-stimulated inflammatory gene expression in endothelial cells (Fig. 4). CX3CL1 expression was strikingly elevated in TNFα-stimulated cells exposed to healthy (1,002.0 ± 888.6%, p < 0.05), baseline DM HDL (2873.0 ± 3461.0%, p < 0.001), 1-month DM HDL (5805.0 ± 5989%, p < 0.001), and 6-month DM HDL (6271 ± 6816%, p < 0.001) compared with no TNFα cells (100 ± 12.3%, Fig. 4A). Interestingly, we observed a positive correlation between CX3CL1 expression and time postamputation (p < 0.05). Pairwise timepoint comparison of 1-month HDL showed that CX3CL1 expression was 93% higher in DM HDL cells compared with non-DM HDL cells (p < 0.05, Fig. 4B).

Figure 4.

HDL from people with diabetes who had a minor amputation exhibits a higher inflammatory response, which is exacerbated with time postamputation. HCAECs were treated with PBS or participant HDL (0.6 mg/mL, 18 h) and then stimulated with TNFα (0.6 ng/mL, 4.5 h) and harvested for RNA. mRNA levels of (A and B) CX3CL1, (C and D) CCL2, (E and F) VCAM1, and (G and H) ICAM1 were normalized to B2M. Each experiment was performed independently in triplicate. Results are expressed as mean ± SD (PBS, n = 12–14, PBS + TNFα, n = 11–14, healthy, n = 14–19, non-DM baseline, n = 8 to 9, non-DM 1 month, n = 4–3, DM baseline, n = 25–27, DM 1 month, n = 20–21, DM 6 months, n = 14). ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001 by either one-way ANOVA (Tukey’s post hoc), Kruskal–Wallis test (Dunn’s post hoc), unpaired independent t-test, or Mann–Whitney test. †p < 0.05, linear trend as assessed by one-way ANOVA. ANOVA, analysis of variance; DM, diabetes mellitus; ELISA, Enzyme-linked immunosorbent assay; HCAEC, human coronary artery endothelial cell; HDL, high-density lipoprotein; PBS, phosphate-buffered saline; SD, standard deviation.

CCL2 expression (Fig. 4C) was elevated in TNFα-stimulated cells exposed to PBS (1123.0 ± 939.4%, p < .001), healthy (800.9 ± 302.0%, p < 0.01), and all DM HDL (baseline: 1025.0 ± 692.1%; 1 month: 1232.0 ± 722.9%; 6 months: 1595.0 ± 892.5%, all p < 0.0001) compared with no TNFα cells (100.0 ± 9.7%). There was a stepwise increase in CCL2 expression with time postamputation in DM HDL-treated cells (p < 0.05). Pairwise comparison of baseline HDL showed a significant 44% increase in CCL2 in DM HDL cells compared with non-DM HDL cells (p < 0.05, Fig. 4D).

The adhesion molecules VCAM1 and ICAM1 followed similar trends in expression to CCL2. VCAM1 (Fig. 4E) and ICAM1 (Fig. 4G) were elevated following TNFα stimulation in cells treated with PBS (VCAM1: 1254.0 ± 1204.0%; ICAM1: 1258.0 ± 1179.0%, both p < 0.01), healthy HDL (VCAM1: 946.0 ± 460.4%; ICAM1: 1062.0 ± 549.2%, both p < 0.001), and DM HDL at baseline (VCAM1: 1371.0 ± 1131.0%; ICAM1: 1221.0 ± 907.0%, both p < 0.0001), 1 month (VCAM1: 1647.0 ± 902.7%; ICAM1: 1718.0 ± 1201.0%, both p < 0.0001), and 6 months (VCAM1: 1722.0 ± 1266.0%; ICAM1: 1760.0 ± 1174.0%, both p < 0.0001) postamputation compared with No TNFα cells (VCAM1: 100.0 ± 3.6%; ICAM1: 100.0 ± 6.9%). Pairwise comparison of baseline HDL showed a significant increase in VCAM1 (+67%) and ICAM1 (+58%) in DM HDL cells compared with non-DM HDL cells (p < 0.05, Fig. 4F and H).

Expression of RELA was elevated in cells treated with healthy HDL compared with PBS in TNFα-stimulated cells (p < 0.05), but there was no change across the other groups (Supplemental Fig. 6A). Expression of CCL5 remained unchanged across the groups (Supplemental Fig. 6B). No significant correlations were found between WC and inflammatory gene expression at baseline or 1 month postamputation (Supplementary Figs. S7 and S8).

HDL from people with diabetes requiring minor amputation exhibits impaired angiogenic effects on endothelial cells

We next investigated the angiogenic properties of participant HDL. The ability of DM HDL to induce endothelial tubule formation in cells exposed to 5 mM glucose diminished with time postamputation (Fig. 5A). Cells treated with DM HDL had a trend for a reduction in tubule number (p = 0.077, Fig. 5B) and significantly lower tubule area over time from baseline to 6 months postamputation (p < 0.05, Fig. 5C), but there was no change in branch points (Fig. 5D). There was 47% less VEGFA expression in endothelial cells treated with baseline DM HDL (68.2 ± 38.0%) compared with baseline non-DM HDL (129.3 ± 80.8%, p < 0.05, Fig. 5E). HDL had similar effects on tubule formation in cells exposed to high glucose (25 mM; Fig. 6). In high glucose, cells treated with DM HDL exhibited a significant trend over time for a reduction in tubule number (p < 0.05), tubule area (p < 0.05), and branch points (p < 0.05) postamputation (Fig. 6A–D). However, there was no difference in VEGFA expression between groups in high glucose conditions (Fig. 6E). No correlations were found between WC and tubule formation or VEGFA expression in 5 and 25 mM glucose (Supplementary Figs. S9 and S10).

Figure 5.

HDL from people with diabetes requiring minor amputation exhibits impaired angiogenic effects on endothelial cells in normal glucose conditions. HCAECs were treated with PBS or participant HDL (0.6 mg/mL, 18 h) and then exposed to normal glucose conditions (5 mM, 48 h). Cells underwent the Matrigel tubulogenesis assay or were harvested for total RNA. (A) Representative images of tubules were photographed at 5× magnification by light microscopy. (B) Tubule number, (C) tubule area, and (D) branch points were analyzed using a bio-network computational model.¹⁷ (E) VEGFA mRNA levels, normalized to B2M. Each experiment was performed independently in quadruplicate. Scale bars are 500 µm. Results are expressed as mean ± SD (PBS, n = 12 to 13, healthy, n = 16–18, non-DM baseline, n = 6 to 7, DM baseline, n = 24–26, DM 1 month, n = 18–21, DM 6 months, n = 13 to 14). ^*p < 0.05 by either one-way ANOVA (Tukey’s post hoc) or Kruskal–Wallis test (Dunn’s post hoc). †p < 0.05, linear trend as assessed by one-way ANOVA. ANOVA, analysis of variance; DM, diabetes mellitus; HCAEC, human coronary artery endothelial cell; HDL, high-density lipoprotein; PBS, phosphate-buffered saline; SD, standard deviation; VEGFA, vascular endothelial growth factor A.

Figure 6.

HDL from people with diabetes requiring minor amputation exhibits impaired angiogenic effects on endothelial cells in high-glucose conditions. HCAECs were treated with PBS or participant HDL (0.6 mg/mL, 18 h) and then exposed to high glucose conditions (25 mM, 48 h). Cells underwent the Matrigel tubulogenesis assay or were harvested for total RNA. (A) Representative images of tubules were photographed at 5× magnification by light microscopy. (B) Tubule numbers, (C) tubule area, and (D) branch points were analyzed using a bio-network computational model.¹⁷ (E) VEGFA mRNA levels, normalized to B2M. Each experiment was performed in quadruplicate. Scale bars are 500 µm. Results are expressed as mean ± SD (PBS, n = 11 to 12, healthy, n = 16 to 17, non-DM baseline, n = 6, DM baseline, n = 23–27, DM 1 month, n = 15–20, DM 6 months, n = 11–13). †p < 0.05, linear trend was assessed by one-way ANOVA. ANOVA, analysis of variance; DM, diabetes mellitus; HCAEC, human coronary artery endothelial cell; HDL, high-density lipoprotein; PBS, phosphate-buffered saline; SD, standard deviation.

HDL from people who experience delayed wound healing postamputation has impaired angiogenic and anti-inflammatory effects in endothelial cells

Participants were stratified by nondelayed WC (>50%) versus delayed WC (<50%) at 1 month postamputation. Despite a 28% increase in VEGFA expression in endothelial cells treated with baseline HDL from people with nondelayed WC (>50%, 103.4 ± 52.4%) compared with those with delayed WC (<50%: 74.2 ± 24.1%, Fig. 7A), this was not statistically significant. There was also no difference in VEGFA with 1-month HDL-treated cells (Fig. 7B). However, 1-month HDL from people with nondelayed WC induced 52% lower endothelial expression of the inflammatory gene RELA (110.2 ± 59.0%) compared with those with delayed WC (230.7 ± 141.5%, p < 0.05, Fig. 7C). No differences were observed in CCL2, CX3CL1, VCAM1, and ICAM1 expression between the two groups (Fig. 7D–G).

Figure 7.

HDL from people that experience delayed wound healing postamputation has impaired anti-inflammatory effects in endothelial cells. Wound area was captured and analyzed using a 3D WoundVue camera. Participants were stratified by percent change in wound closure (WC) defined as nondelayed (WC > 50%) or delayed (WC < 50%) closure at 1 month postamputation. HCAECs were treated with PBS or participant HDL (0.6 mg/mL, 18 h) and then exposed to normal glucose conditions (5 mM, 48 h) and were harvested for total RNA. VEGFA mRNA levels of (A) baseline and (B) 1-month HDL. HCAECs were treated with PBS or participant HDL (0.6 mg/mL, 18 h) and then stimulated with TNFα (0.6 ng/mL, 4.5 h) and harvested for RNA. mRNA levels of (C) RELA, (D) CCL2, (E) CX3CL1, (F) VCAM1, and (G) ICAM1 of 1-month HDL. Gene expression was normalized to B2M. Each experiment was performed in triplicate. Results are expressed as mean ± SD (n = 8–13/group). ^*p < 0.05 by unpaired independent t-test or Mann–Whitney test. HCAEC, human coronary artery endothelial cell; HDL, high-density lipoprotein; PBS, phosphate-buffered saline; VEGFA, vascular endothelial growth factor A.

Taken together with the changes to HDL CEC and the anti-inflammatory and proangiogenic properties of HDL, this suggests that patients with dysfunctional HDL may be at increased risk for delayed WC.

DISCUSSION

Currently, there are no definitive biomarkers that identify patients at high risk of developing nonhealing DFUs, which hinders optimal wound care and increases the likelihood of amputation. The identification of biomarkers that predict wound healing outcomes in people with DFUs would be immensely beneficial. This is the first study to assess changes in the functionality of HDL isolated from patients who have had minor amputations. We report the following important findings: (1) plasma HDL-c positively correlated with the rate of WC postamputation and (2) individuals with >1.0 mM HDL-c had improved WC. (3) The composition of HDL from people with diabetes that had an amputation contained less apoAI and PON1 protein and this associated with delayed healing postamputation. (4) HDL from people with diabetes that had an amputation had impaired CEC and impaired anti-inflammatory and proangiogenic properties, suggesting impaired HDL functionality. However, 6 months postamputation, the CEC of DM HDL was partially restored. (5) In contrast, DM HDL induced greater inflammatory responses, and its angiogenic capacity was further impaired over time postamputation. (6) HDL from people with delayed healing elicited elevated proinflammatory properties, with higher RELA levels in endothelial cells. Taken together, these findings suggest impairment of the CEC, anti-inflammatory, and proangiogenic properties of HDL, mediated through changes to HDL composition, may contribute to delayed healing postamputation. A combination of HDL-related markers including HDL-c and HDL functionality and composition, has potential as a more sensitive guide for wound healing outcomes than HDL-c levels alone.

Despite a ∼50% reduction in the rate of WC at 1 month postamputation from baseline in people with diabetes compared with those without diabetes as may have been anticipated, it was not statistically significant, likely due to variations in wound healing rates. HbA1c was, however, inversely correlated with the rate of WC. This is consistent with previous studies that show elevated HbA1c and variability in HbA1c measurements are associated with reduced time to healing,^18,25 confirming that people with poorly controlled diabetes experience diminished healing capacity.

Dyslipidemia is a major risk factor for CVD and the development of peripheral artery disease,²⁶ which increases the risk of amputation in individuals with diabetes by fourfold.²⁷ Dyslipidemia consists of elevated total cholesterol, triacylglycerol, and LDL-c and reduced HDL-c.²⁸ In this study, HDL-c was lower in both the DM and non-DM groups with amputations. Low HDL-c is positively associated with increased risk of amputation and wound-related deaths¹² and reduced wound healing.²⁹ Consistent with this, we found that plasma HDL-c levels positively correlated with WC rate. Furthermore, participants with HDL-c levels below the clinical cut-off of 1.0 mM had delayed WC. Plasma triacylglycerol was elevated in the DM group, and the triacylglycerol content of HDL was higher in both the DM and non-DM groups. Previous studies have shown that hypertriglyceridemia contributes to alterations in HDL functionality, affecting the anti-inflammatory properties of HDL.³⁰ This suggests the higher triacylglycerol levels in DM participants may be a contributor to HDL dysfunction. Interestingly, total cholesterol was lower in DM participants, which is likely due to the lipid-lowering effects of their medications (Supplementary Table S2). Statins have been shown to mediate HDL functionality, increasing cholesterol efflux, but with no effect on the anti-inflammatory properties of HDL.³¹

HDL composition can be altered in disease states, such as diabetes.²⁰ We found that HDL from people with diabetes had lower amounts of apoAI and PON1 protein. This is consistent with previous studies showing depletion of apoA1 and PON1 in HDL from people with diabetes.^20,32 Furthermore, PON1 activity has been shown to be dependent on levels of apoAI as well as other apolipoproteins.¹⁹ Depletion of apoAI and PON1 from HDL alters HDL functionality, including the CEC and anti-inflammatory properties.^20,32,33 In addition, our data showed that lower amounts of apoAI and PON1 in HDL were associated with delayed healing postamputation. We propose that these changes to HDL composition are critical to its function and may contribute to impaired healing postamputation.

The role of circulating HDL levels is already well-established to be associated with the risk of CVD¹¹ and diabetes.³⁴ Numerous studies have found a strong association between HDL CEC and the prevalence of atherosclerosis and CVD incidence, independent of HDL-c levels.^35,36 HDL CEC from patients with type 2 DM (T2DM) is predominantly found to be lower than from healthy subjects.^36,37 However, the association of HDL CEC and T2DM is not unequivocal, as other studies have reported either no changes^38,39 or increased CEC in patients with T2DM.⁴⁰ We found that CEC was reduced in 1-month DM HDL compared with non-DM HDL. We postulate that HDL CEC impairment may contribute to delayed healing in people with diabetes through failure to mediate lipid removal from dead cells following injury.¹⁰ Interestingly, CEC improved over time, suggesting that this impairment may be transient and can be regenerated. While the molecular determinants remain largely unknown, acute inflammation has been shown to remodel HDL and impair CEC while chronic inflammation has more subtle effects.⁴¹ In this study, baseline circulating ESR and CRP levels were elevated in the DM group, suggesting that the persistent inflammatory milieu within diabetes may alter HDL composition and subsequently its function. Consistent with this, HDL from the DM group contained less apoA1, which can directly alter the CEC of HDL.²⁰ ESR and CRP levels then gradually declined over time. This gradual resolution of inflammation within the DM group may have restored HDL CEC function by 6 months postamputation.

The inflammatory response is critical for wound healing. However, in diabetes, exacerbation of this response and a failure to “switch off” prolongs inflammation and is a major contributor to chronic nonhealing wounds. HDL has well-established anti-inflammatory properties.³³ In this study, DM HDL had impaired anti-inflammatory properties, failing to attenuate TNFα-stimulated endothelial inflammatory response, and this was exacerbated over time. This is likely to contribute to prolonged wound inflammation and impaired healing in people with diabetes. Furthermore, elevated RELA, a key transcription factor controlling the expression of several inflammatory mediators such as CCL2, was associated with delayed healing postamputation. These changes to the proinflammatory effects of HDL are consistent with previous studies that show HDL loses its anti-inflammatory properties in diabetes.¹³ Moreover, HDL from people with diabetes increases endothelial VCAM1 expression, which is associated with elevated CRP.⁴² In poor metabolic environments with heightened inflammation, HDL may become proinflammatory. Consistent with this, HDL from the DM group had lower amounts of PON1, an antioxidant protein. In addition, our DM group had increased circulating ESR and CRP levels, concomitant with elevated white cell counts, indicative of elevated inflammation, which, in turn, contribute to poor wound healing in diabetes.

The angiogenic capacity of HDL was impaired in people with diabetes. Cells treated with DM HDL had reduced capacity to form tubules with time postamputation, concomitant with a reduction in the key angiogenesis mediator VEGFA. This is consistent with our previous work that showed HDL from a cohort of Australian Aboriginal people with diabetes-related vascular complications had reduced capacity to promote endothelial tubule formation.⁴³ This was associated with impaired induction of the transcription factor HIF1A, which regulates VEGFA expression.⁴³ Furthermore, HDL from people with diabetes was unable to adequately induce other key mediators of angiogenesis, including endothelial nitric oxide,⁴⁴ and reduced endothelial cell angiogenic functions of migration and proliferation.⁴⁵

Interestingly, we saw divergent recovery patterns postamputation, with HDL CEC recovering by 6 months postamputation, but conversely, the anti-inflammatory and proangiogenic effects worsened with time. The CEC properties of HDL are primarily driven through the apoAI protein.²⁰ ApoAI was reduced at 1 month postamputation in DM compared with non-DM, likely contributing to the impairment to CEC at baseline and 1 month. However, HDL apoAI levels remained in a steady state within the DM group and did not worsen with time, which suggests HDL apoAI may become partially restored with time, facilitating CEC recovery, whereas HDL PON1 reduced linearly with time in the DM group. The antioxidant properties of PON1 contribute to the anti-inflammatory and proangiogenic effects of HDL,^32,33 and these changes to HDL PON1 were concomitant with the changes to the anti-inflammatory and proangiogenic effects postamputation, which worsened with time, suggesting this is likely a key contributor to the longitudinal alterations in HDL functionality.

There were limitations in this study that should be noted. We were unable to retain a high number of participants in the study at the later timepoints postamputation, in large part due to the intensive morbidity and mortality of this cohort. We also excluded patients from the study who were treated with NPWT and HBOT, as they are conditionally recommended by the IWGDF as an adjunct therapy for wound healing.² We thus felt that including patients who were undergoing these therapies would complicate the assessment of wound healing, accepting that NPWT is relatively commonly used after major foot surgery. Regardless, our findings provide significant insight into the functionality of HDL and its potential contributions to delayed wound healing in people with diabetes postamputation.

In conclusion, this study is the first to longitudinally track HDL functionality in people with diabetes postminor amputation. We found that the CEC, anti-inflammatory, and proangiogenic capacities of HDL are impaired in people with diabetes that require minor amputations. Furthermore, we observed divergent changes in DM HDL functionality postamputation, with CEC restored by 6 months, yet anti-inflammatory and proangiogenic properties declined over time. In addition, we found that changes to the composition of the HDL particle (reduced apoAI and PON1) were concomitant with the impairment to HDL functionality and associated with delayed healing postamputation. Our findings suggest that HDL dysfunction may contribute to impaired healing postamputation, and the combination of HDL functionality and composition with plasma HDL-c levels could provide a more sensitive guide for wound healing outcomes.

Clinically, HDL-related assays are applicable and would be particularly useful when dealing with challenging chronic wounds that do not heal within a timely manner. As some chronic wounds can take months to years to heal, lab assays assessing a range of HDL factors, including HDL-c and HDL functionality and composition, would give a more comprehensive representation of an individual’s HDL levels and functionality and could provide an alternate and more reliable predictive value for wound healing. This would allow clinicians to decide early in the treatment regimen, at the time of surgery, or up to 1 month postsurgery, if more intensive care is required.

KEY FINDINGS

•

Plasma high-density lipoprotein cholesterol (HDL-c) levels positively correlate with the rate of wound closure 1 month postamputation, making HDL-c a positive prognostic marker for wound healing.

The cholesterol efflux, anti-inflammatory, angiogenic capacity, and composition of HDL from people with diabetes are impaired postamputation.

•

Low HDL apoAI and PON1 may be early warning indicators for impaired HDL functionality and delayed wound healing.

•

HDL testing, including HDL-c and HDL functionality and composition, could become part of a postamputation risk panel to predict delayed healing and guide optimal wound care for improved outcomes.

Footnotes

ACKNOWLEDGMENTS

The authors would like to thank all the study participants.

AUTHORS’ CONTRIBUTION

Z.L.: Investigation, formal analysis, and writing—review and editing. E.L.S.: Investigation, formal analysis, visualization, and writing—original draft. J.T.M.T.: Methodology, investigation, formal analysis, validation, visualization, and writing—review and editing. V.A.N., L.S., and Z.S.: Investigation and writing—review and editing. M.D.: Software and writing—review and editing. G.P.: Resources and writing—review and editing. J.V., P.J.P., J.D., and N.M.: Validation and writing—review and editing. R.A.F.: Methodology, resources, validation, and writing—review and editing. C.A.B.: Conceptualization, formal analysis, funding acquisition, methodology, supervision, validation, visualization, and writing—review and editing.

AUTHOR DISCLOSURE AND GHOSTWRITING STATEMENT

No competing financial interests exist. The content of this article was expressly written by the authors listed. No ghostwriters were used to write this article.

FUNDING INFORMATION

This work was supported by the Hospital Research Foundation (THRF) BioMed City Translational Grant (to R.A.F. and C.A.B.; #THRF 2018/040-83100-01), Diabetes Australia General Grant (#Y24G-BURC to C.A.B. and J.T.M.T.), and Heart Foundation Lin Huddleston Fellowship (to C.A.B.).

ABOUT THE AUTHORS

Zahra Lotfollahi, PhD, is an academic podiatrist researching diabetic foot ulcer healing at the Queen Elizabeth Hospital and Royal Adelaide Hospital (RAH). Emma Solly, PhD, is a postdoctoral scientist at SAHMRI with expertise in wound healing and diabetes. Joanne Tan, PhD, is a postdoctoral scientist specialising in HDL biology and wound healing. Victoria Nankivell, PhD, is a postdoctoral scientist focussed on vascular inflammation. Mohsen Dorraki, PhD, is a computer scientist with expertise in medical machine learning. Guilherme Pena, MD, PhD, a vascular surgery registrar experienced in diabetic foot disease management. Johan Verjans, MD, PhD, is a clinician-scientist specialising in medical machine learning. Zygmunt Szpak, PhD, is a computer scientist specialising in medical machine learning. Joseph Dawson, MBBS, ChM, MD, MRCS, FRCS, FRACS, is a vascular surgeon experienced in diabetic foot disease management. Neil McMillan, PhD, is a medical scientist at Adelaide University, with expertise in diabetic foot disease management. Peter Psaltis, MBBS, PhD, FRACP, FCSANZ, is an interventional cardiologist with expertise in inflammation. Robert Fitridge, MBBS, MS, FRACS, is a consultant vascular surgeon specialising in diabetic foot disease management. Christina Bursill, PhD, directs the Vascular Research Centre at SAHMRI, Adelaide University and specialises in inflammation, HDL biology and wound healing.

Supplemental Material

Abbreviations and Acronyms

References

1. Boulton

AJM

, Vileikyte

, Ragnarson-Tennvall

, et al. The global burden of diabetic foot disease. Lancet 2005;366(9498):1719–1724; doi: 10.1016/S0140-6736(05)67698-2

2. Chen

, Vilorio

, Dhatariya

, et al. Guidelines on interventions to enhance healing of foot ulcers in people with diabetes (IWGDF 2023 update). Diabetes Metab Res Rev 2024;40(3):e3644; doi: 10.1002/dmrr.3644

3. Zhao

, Liang

, Clarke

, et al. Inflammation in chronic wounds. Int J Mol Sci 2016;17(12):2085; doi: 10.3390/ijms17122085

4. DiPietro

. Angiogenesis and wound repair: When enough is enough. J Leukoc Biol 2016;100(5):979–984; doi: 10.1189/jlb.4MR0316-102R

5. Brem

, Tomic-Canic

. Cellular and molecular basis of wound healing in diabetes. J Clin Invest 2007;117(5):1219–1222; doi: 10.1172/jci32169

6. Lotfollahi

, Tan

JTM

, Nankivell

, et al. Topical reconstituted high-density lipoproteins elicit anti-inflammatory effects in diabetic wounds. Adv Wound Care (New Rochelle) 2026;15(2):91–106; doi: 10.1089/wound.2023.0162

7. Primer

, Tan

JTM

, Sandeman

, et al. Reconstituted high-density lipoproteins rescue diabetes-impaired endothelial cell metabolic reprograming and angiogenic responses to hypoxia. Arterioscler Thromb Vasc Biol 2025;45(5):683–701; doi: 10.1161/ATVBAHA.124.320110

8. Qiu

, Zhao

, Zhou

, et al. High-density lipoprotein cholesterol efflux capacity is inversely associated with cardiovascular risk: A systematic review and meta-analysis. Lipids Health Dis 2017;16(1):212; doi: 10.1186/s12944-017-0604-5

9. Tan

, Prosser

, Dunn

, et al. High-Density lipoproteins rescue diabetes-impaired angiogenesis via scavenger receptor class B type I. Diabetes 2016;65(10):3091–3103; doi: 10.2337/db15-1668

10.

10. Luna-Luna

, Niesor

, Pérez-Méndez

. HDL as bidirectional lipid vectors: Time for new paradigms. Biomedicines 2022;10(5):1180; doi: 10.3390/biomedicines10051180

11.

11. Gordon

, Castelli

, Hjortland

, et al. High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am J Med 1977;62(5):707–714; doi: 10.1016/0002-9343(77)90874-9

12.

12. Ikura

, Hanai

, Shinjyo

, et al. HDL cholesterol as a predictor for the incidence of lower extremity amputation and wound-related death in patients with diabetic foot ulcers. Atherosclerosis 2015;239(2):465–469; doi: 10.1016/j.atherosclerosis.2015.02.006

13.

13. Morgantini

, Natali

, Boldrini

, et al. Anti-inflammatory and antioxidant properties of HDLs are impaired in type 2 diabetes. Diabetes 2011;60(10):2617–2623; doi: 10.2337/db11-0378

14.

14. Mills

Sr , Conte

, Armstrong

, et al.; Society for Vascular Surgery Lower Extremity Guidelines Committee. The society for vascular surgery lower extremity threatened limb classification system: Risk stratification based on wound, ischemia, and foot infection (WIfI). J Vasc Surg 2014;59(1):220–234.e1–2; doi: 10.1016/j.jvs.2013.08.003

15.

15. Pena

, Kuang

, Szpak

, et al. Evaluation of a novel three-dimensional wound measurement device for assessment of diabetic foot ulcers. Adv Wound Care (New Rochelle) 2020;9(11):623–631; doi: 10.1089/wound.2019.0965

16.

16. Knopfholz

, Disserol

, Pierin

, et al. Validation of the friedewald formula in patients with metabolic syndrome. Cholesterol 2014;2014:261878; doi: 10.1155/2014/261878

17.

17. Fouladzadeh

, Dorraki

, Min

KKM

, et al. The development of tumour vascular networks. Commun Biol 2021;4(1):1111; doi: 10.1038/s42003-021-02632-x

18.

18. Christman

, Selvin

, Margolis

, et al. Hemoglobin A1c predicts healing rate in diabetic wounds. J Invest Dermatol 2011;131(10):2121–2127; doi: 10.1038/jid.2011.176

19.

19. Carnuta

, Stancu

, Toma

, et al. Dysfunctional high-density lipoproteins have distinct composition, diminished anti-inflammatory potential and discriminate acute coronary syndrome from stable coronary artery disease patients. Sci Rep 2017;7(1):7295; doi: 10.1038/s41598-017-07821-5

20.

20. Bonilha

, Zimetti

, Zanotti

, et al. Dysfunctional high-density lipoproteins in type 2 diabetes mellitus: Molecular mechanisms and therapeutic implications. J Clin Med 2021;10(11):2233; doi: 10.3390/jcm10112233

21.

21. Coerper

, Beckert

, Küper

, et al. Fifty percent area reduction after 4 weeks of treatment is a reliable indicator for healing–analysis of a single-center cohort of 704 diabetic patients. J Diabetes Complications 2009;23(1):49–53; doi: 10.1016/j.jdiacomp.2008.02.001

22.

22. Saleheen

, Scott

, Javad

, et al. Association of HDL cholesterol efflux capacity with incident coronary heart disease events: A prospective case-control study. Lancet Diabetes Endocrinol 2015;3(7):507–513; doi: 10.1016/s2213-8587(15)00126-6

23.

23. Bursill

, Castro

, Beattie

, et al. High-density lipoproteins suppress chemokines and chemokine receptors in vitro and in vivo. Arterioscler Thromb Vasc Biol 2010;30(9):1773–1778; doi: 10.1161/atvbaha.110.211342

24.

24. Cockerill

, Rye

, Gamble

, et al. High-density lipoproteins inhibit cytokine-induced expression of endothelial cell adhesion molecules. Arterioscler Thromb Vasc Biol 1995;15(11):1987–1994; doi: 10.1161/01.atv.15.11.1987

25.

25. Dhatariya

, Li Ping Wah-Pun Sin

, Cheng

JOS

, et al. The impact of glycaemic variability on wound healing in the diabetic foot – A retrospective study of new ulcers presenting to a specialist multidisciplinary foot clinic. Diabetes Res Clin Pract 2018;135:23–29; doi: 10.1016/j.diabres.2017.10.022

26.

26. Criqui

, Aboyans

. Epidemiology of peripheral artery disease. Circ Res 2015;116(9):1509–1526; doi: 10.1161/circresaha.116.303849

27.

27. Barnes

, Eid

, Creager

, et al. Epidemiology and risk of amputation in patients with diabetes mellitus and peripheral artery disease. Arterioscler Thromb Vasc Biol 2020;40(8):1808–1817; doi: 10.1161/atvbaha.120.314595

28.

28. Wu

, Parhofer

. Diabetic dyslipidemia. Metabolism 2014;63(12):1469–1479; doi: 10.1016/j.metabol.2014.08.010

29.

29. Chen

, Ma

, Chen

, et al. Association of high-density lipoprotein cholesterol and wound healing in patients with diabetic foot ulcers. Chin Med J (Engl) 2022;135(1):110–112; doi: 10.1097/cm9.0000000000001544

30.

30. Patel

, Puranik

, Nakhla

, et al. Acute hypertriglyceridaemia in humans increases the triglyceride content and decreases the anti-inflammatory capacity of high density lipoproteins. Atherosclerosis 2009;204(2):424–428; doi: 10.1016/j.atherosclerosis.2008.07.047

31.

31. Lee

, Choi

, Cheon

, et al. Effect of two lipid-lowering strategies on high-density lipoprotein function and some HDL-related proteins: A randomized clinical trial. Lipids Health Dis 2017;16(1):49; doi: 10.1186/s12944-017-0433-6

32.

32. Brites

, Martin

, Guillas

, et al. Antioxidative activity of high-density lipoprotein (HDL): Mechanistic insights into potential clinical benefit. BBA Clin 2017;8:66–77; doi: 10.1016/j.bbacli.2017.07.002

33.

33. Ansell

, Navab

, Watson

, et al. Anti-inflammatory properties of HDL. Rev Endocr Metab Disord 2004;5(4):351–358; doi: 10.1023/B:REMD.0000045107.71895.b2

34.

34. Femlak

, Gluba-Brzózka

, Ciałkowska-Rysz

, et al. The role and function of HDL in patients with diabetes mellitus and the related cardiovascular risk. Lipids Health Dis 2017;16(1):207; doi: 10.1186/s12944-017-0594-3

35.

35. Ebtehaj

, Gruppen

, Bakker

SJL

, et al. HDL (High-Density Lipoprotein) cholesterol efflux capacity is associated with incident cardiovascular disease in the general population. Arterioscler Thromb Vasc Biol 2019;39(9):1874–1883; doi: 10.1161/ATVBAHA.119.312645

36.

36. Soria-Florido

, Castaner

, Lassale

, et al. Dysfunctional high-density lipoproteins are associated with a greater incidence of acute coronary syndrome in a population at high cardiovascular risk: A nested case-control study. Circulation 2020;141(6):444–453; doi: 10.1161/CIRCULATIONAHA.119.041658

37.

37. Attia

, Nakbi

, Smaoui

, et al. Increased phospholipid transfer protein activity associated with the impaired cellular cholesterol efflux in type 2 diabetic subjects with coronary artery disease. Tohoku J Exp Med 2007;213(2):129–137; doi: 10.1620/tjem.213.129

38.

38. Denimal

, Benanaya

, Monier

, et al. Normal HDL cholesterol efflux and anti-inflammatory capacities in type 2 diabetes despite lipidomic abnormalities. J Clin Endocrinol Metab 2022;107(9):e3816–e3823; doi: 10.1210/clinem/dgac339

39.

39. Dullaart

, Annema

, de Boer

, et al. Pancreatic beta-cell function relates positively to HDL functionality in well-controlled type 2 diabetes mellitus. Atherosclerosis 2012;222(2):567–573; doi: 10.1016/j.atherosclerosis.2012.03.037

40.

40. Yassine

, Belopolskaya

, Schall

, et al. Enhanced cholesterol efflux to HDL through the ABCA1 transporter in hypertriglyceridemia of type 2 diabetes. Metabolism 2014;63(5):727–734; doi: 10.1016/j.metabol.2014.03.001

41.

41. Ronsein

, Vaisar

. Inflammation, remodeling, and other factors affecting HDL cholesterol efflux. Curr Opin Lipidol 2017;28(1):52–59; doi: 10.1097/MOL.0000000000000382

42.

42. Ebtehaj

, Gruppen

, Parvizi

, et al. The anti-inflammatory function of HDL is impaired in type 2 diabetes: Role of hyperglycemia, paraoxonase-1 and low grade inflammation. Cardiovasc Diabetol 2017;16(1):132; doi: 10.1186/s12933-017-0613-8

43.

43. Morrison

, Solly

, Shemesh

, et al. Elevated HDL-bound miR-181c-5p level is associated with diabetic vascular complications in Australian Aboriginal people. Diabetologia 2021;64(6):1402–1411; doi: 10.1007/s00125-021-05414-6

44.

44. Sorrentino

, Besler

, Rohrer

, et al. Endothelial-vasoprotective effects of high-density lipoprotein are impaired in patients with type 2 diabetes mellitus but are improved after extended-release niacin therapy. Circulation 2010;121(1):110–122; doi: 10.1161/circulationaha.108.836346

45.

45. Pan

, Ma

, Ren

, et al. Diabetic HDL is dysfunctional in stimulating endothelial cell migration and proliferation due to down regulation of SR-BI expression. PLoS One 2012;7(11):e48530; doi: 10.1371/journal.pone.0048530

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

2.42 MB

0.00 MB