Association of monocyte TNFα expression and serum inflammatory biomarkers with walking impairment in PAD

Abstract

Objective

Inflammation contributes to the development of peripheral artery disease (PAD) and may contribute to intermittent claudication by adversely affecting vascular and skeletal muscle function. We explored the association of inflammation to maximal walking time (MWT) in patients with claudication.

Methods

Circulating inflammatory biomarkers, including tumor necrosis factor α (TNFα), C-reactive protein (CRP), interleukin-6 (IL-6), and soluble intercellular adhesion molecule-1 (sICAM) were measured in 75 subjects with intermittent claudication, as well as 43 healthy subjects. Real-time PCR was used to quantify mRNA expression of TNFα, IL-6, IFN gamma, and CD-36 from peripheral blood monocytes. Treadmill testing was performed in PAD subjects to assess MWT.

Results

Compared with healthy subjects, PAD subjects had higher levels of circulating TNFα (p<0.0001), CRP (p=0.003), sICAM (p<0.0001), and IL-6 (p<0.0001). Expression of both IL-6 (p=0.024) and CD36 (p=0.018) was greater in PAD than healthy subjects. Among subjects with PAD, higher gene expression of TNFα was associated inversely with MWT (p=0.01). MWT was also associated inversely with greater levels of circulating TNFα (p=0.028), CRP (p=0.024), IL-6 (p=0.03), and sICAM (p=0.018).

Conclusions

Systemic inflammation, as indicated by TNFα inflammatory gene expression in peripheral blood monocytes and by circulating biomarker levels, is associated with impairment in walking time in patients with PAD and intermittent claudication.

Introduction

Inflammation contributes to the pathobiology of atherosclerosis in patients with peripheral artery disease (PAD) and also may contribute to the pathophysiology of intermittent claudication. Insufficient perfusion resulting from hemodynamically significant arterial stenosis defines PAD and is required for the development of intermittent claudication.¹ However, because stenosis correlates imperfectly with walking capacity in patients with PAD, non-hemodynamic factors affecting vascular and skeletal muscle function may be important determinants of claudication severity. This notion is supported by the finding that non-invasive therapies for claudication, such as exercise training and cilostazol, improve walking time without significantly affecting the degree of conduit vessel stenosis.²

Previous studies have demonstrated that systemic inflammation is associated with PAD. Several inflammatory biomarkers, including high sensitivity C-reactive protein (CRP), soluble adhesion molecules (such as sICAM), and tumor necrosis factor α (TNFα), are associated with both the incidence and severity of PAD.^3–6 Inflammation increases the risk of adverse graft outcomes in PAD patients who have undergone peripheral bypass surgery,⁷ and greater systemic inflammation is associated with increased cardiovascular morbidity and mortality in individuals with PAD.^{8, 9} The contribution of inflammation to the pathophysiology of intermittent claudication has been suggested by studies showing that elevated inflammatory biomarkers are inversely associated with physical performance,¹⁰ and that HMG CoA reductase inhibitors (‘statins’) improve walking time in PAD.^{11, 12}

Circulating peripheral blood monocytes (PBMCs) are known drivers of the inflammatory milieu that promotes development of systemic atherosclerosis.¹³ Whether elaboration of inflammatory cytokines from monocytes fuels the inflammatory cascade in PAD subjects and contributes to walking impairment has not been previously explored. We examined the association of expression of inflammatory genes from peripheral blood monocytes with impaired walking time in PAD patients with claudication.

Methods

Patient population

Patients with PAD, aged 40–85 years, were recruited as part of a randomized clinical trial evaluating the effect of statin and thiazolidinedione therapy on the contribution of inflammation and insulin resistance to the pathophysiology of intermittent claudication (NCT00225940). Data from the initial baseline visits were used for this analysis. A diagnosis of PAD was based on an ankle-brachial index (ABI) measurement ≤ 0.90 at rest in the symptomatic leg. ABI was calculated as the ratio of the higher of the ankle systolic pressures for a given leg and the higher of the right and left brachial artery systolic pressures. Subjects were eligible if they had stable claudication symptoms confirmed by the San Diego Claudication Questionnaire. Patients were excluded if they had any of the following: peripheral or coronary revascularization within six months; any unstable cardiovascular process (unstable angina, recent myocardial infarction, stroke, or congestive heart failure); or significant renal insufficiency (serum creatinine > 2.5 mg/dL). Individuals with diabetes were included but only if they were not using insulin or insulin-sensitizing medications. Older healthy subjects were recruited by advertisement and were eligible if they had no known medical problems, were taking no medications, and were not active smokers.

The study was approved by the institutional review boards at Brigham and Women’s Hospital and the Veterans Administration Boston Healthcare Service from which patients were recruited. All subjects provided written informed consent prior to participation.

Study protocol

Patients with PAD included in this analysis were among those recruited for participation in a randomized controlled trial of the effect of reducing inflammation and insulin resistance on walking time in patients with intermittent claudication. Because one of the primary study hypotheses was the effect of reducing inflammation with a statin on walking time, a four-week statin washout period was required for all patients who were previously on a statin. A subset of PAD subjects who met eligibility criteria went on to participate in the randomized portion of the trial, the data from which will be reported separately.

Baseline characteristics

Demographics and baseline clinical history were based on self-report. Race and ethnicity are recorded as non-Hispanic white or other. Individuals were assigned a diagnosis of hypertension if they reported a physician diagnosis of hypertension or reported use of anti-hypertensive medications. Hyperlipidemia was similarly assigned based on physician diagnosis or use of lipid-lowering medications prior to participation. A diagnosis of diabetes was based on self-report, use of diabetic medications, or if fasting plasma glucose measured ≥ 126 mg/dL. Personal history of coronary artery disease, myocardial infarction, stroke or transient ischemic attack, and family history cardiovascular disease were all based on self-report. Smoking status was recorded as current, former, or never smoking.

Laboratory measures

Analyses of total cholesterol, low density lipoprotein (LDL) cholesterol, high density lipoprotein (HDL) cholesterol, triglycerides, glucose, insulin, creatinine, and complete blood count were performed using standard laboratory techniques. Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as fasting insulin (mmol/L) x fasting glucose (μU/mL)/22.5. Measurement of inflammatory biomarkers was performed as follows: high sensitivity C-reactive protein (hsCRP) concentration was determined using an immunoturbidimetric assay on the Hitachi 917 analyzer (Roche Diagnostics, Indianapolis, IN). Interleukin-6 (IL-6), soluble intercellular adhesion molecule-1 (sICAM-1), and tumor necrosis factor α receptor II (TNF-RII) were measured by ELISA assay from R & D Systems (Minneapolis, MN).

Quantification of gene expression in peripheral blood mononuclear cells

To measure monocyte expression of inflammatory genes (including TNFα, interleukin-6 (IL-6), interferon-gamma (IFN gamma), and CD-36), peripheral blood mononuclear cells were isolated from fresh whole blood using density gradient ultracentrifugation with Percoll. Cells were plated on a 10cm dish for 2 hours, after which non-adherent cells were washed away, and the remaining monocytes were harvested for RNA using a Qiagen RNA isolation kit. For cDNA synthesis, 200 ng of RNA was reverse transcribed (Promega). Real time PCR (RT-PCR) was performed in a 384 well format with the LightCyler 480 Real-Time PCR system (Roche) using SYBR Green I (Bio-Rad). The mRNA levels of genes of interest were normalized to the 36B4 internal control housekeeping gene and relative fold change was calculated using the ΔΔCt method.

Treadmill walking time

Exercise testing was performed in PAD subjects using a modified Gardner protocol. Subjects walked at a constant speed of 2 miles per hour (mph) beginning at 0% grade and increasing grade by 2% every two minutes. Maximal walking time (MWT) was recorded in seconds. To ensure reproducibility, subjects were required to perform up to three run-in treadmill tests with < 20% variability in MWT between sequential tests in order to be eligible for participation. An average of the final two run-in treadmill tests was used as the baseline exercise measurement.

Statistical methods

Statistical analyses were performed with SAS version 9.2 (SAS Institute, Inc., Cary, NC). Categorical variables were compared by the χ² test and continuous variables were compared using two-sided t-test or Wilcoxon rank sum test depending on the normality of the distribution. Categorical variables are expressed as percent and continuous variables as mean ± standard deviation (SD) or as median and inter-quartile range. Comparison of mean values across quartiles was achieved using linear regression. Multivariable models included age, gender, smoking status, and ABI values as covariates. P-values < 0.05 were considered statistically significant.

Results

Baseline characteristics

Baseline characteristics of PAD and healthy subjects are shown in table I. There were 75 subjects with PAD and stable intermittent claudication and 43 healthy subjects who were eligible for inclusion in this analysis. Subjects with PAD were older (68.5 ± 7.9 vs. 58.8 ± 10.0 yrs, p<0.0001), had higher resting blood pressure, glucose and insulin levels, and more abnormal cholesterol profiles than healthy subjects (table I). By the nature of the eligibility criteria, no healthy subjects had any underlying medical conditions. Of the subjects with PAD, diabetes was present in 20.0%, coronary artery disease in 34.7%, and 42.7% had undergone prior lower extremity revascularization.

Table I.

Baseline Characteristics

Baseline characteristics	Healthy (n=43)	PAD (n=75)	p-value
Age, years	58.7 ± 10.0	68.5 ± 7.9	<0.0001
Male gender	50%	77.3%	0.0025
Race/ethnicity (% non-Hispanic white)	88.1%	88.0%	0.38
Smoking (current or former)	42.9%	97.3%	<0.0001
Diabetes	0%	20%	0.0019
Coronary artery disease	0%	34.7%	<0.0001
Stroke	0%	20%	0.0019
Prior lower extremity intervention	0%	42.7%	<0.0001
Systolic blood pressure, mm Hg	128 ± 13	152 ± 21	<0.0001
Diastolic blood pressure, mm Hg	80 ± 8	80 ± 9	0.78
Waist-hip ratio	0.88 ± 0.14	0.99 ± 0.10	<0.0001
Ankle-brachial index	1.13 ± 0.09	0.71 ± 0.16	<0.0001
Maximal walking time (seconds)	n/a	438 ± 262	n/a
Creatinine, mg/dL	0.87 ± 0.15	1.08 ± 0.3	<0.0001
Glucose, mg/dL	84 ± 10	107 ± 42	0.0008
Insulin, mg/dL	5.8 ± 4.2	9.4 ± 5.8	0.0008
HOMA-IR	1.28 ± 1.1	2.57 ± 2.2	0.0007
LDL cholesterol, mg/dL	108 ± 25	131 ± 37	0.0007
HDL cholesterol, mg/dL	55 ± 16	46 ± 17	0.005
Triglycerides, mg/dL	107 ± 71	176 ± 125	0.001
Total cholesterol, mg/dL	184 ± 30	213 ± 40	0.0002

Inflammatory biomarkers

PAD subjects had significantly higher levels of biomarkers of inflammation compared to healthy subjects (table II). The median level of TNFα was 2499 pg/mL [2055, 3405] vs. 1831 pg/mL [1582, 2184] (p<0.0001); hsCRP was 2.49 mg/L [IQR 1.01, 4.81] vs. 1.41 mg/L [0.7, 2.42] (p=0.003); sICAM was 257.8 ng/mL [216, 306] vs. 195.4 ng/mL [171.5, 222.3] (p<0.0001), and IL-6 was 2.87 pg/mL [1.6, 3.7] vs. 1.11 pg/mL [0.86, 1.7] (p<0.0001) in PAD versus healthy subjects, respectively.

Table II.

Circulating inflammatory biomarkers

	Healthy (n=43)	PAD (n=75)	p-value
TNFα, pg/mL	1831 [1582, 2184]	2499 (2055, 3405]	<0.0001
hsCRP, mg/L	1.41 [0.7, 2.42]	2.49 [1.01, 4.81]	0.003
IL-6, pg/mL	1.11 [0.86, 1.7]	2.87 [1.6, 3.7]	<0.0001
sICAM, ng/mL	195.4 [171.5, 222.3]	257.8 [216, 306]	<0.0001

Tumor necrosis factor α (TNFα), high sensitivity C reactive protein (hsCRP), interleukin-6 (IL-6), and soluble intercellular adhesion molecule (sICAM). Data are shown as median and interquartile range. Statistical comparisons achieved by Wilcoxon rank sum test.

Gene expression in peripheral blood monocytes

Expression of inflammatory genes was measured in peripheral blood monocytes obtained from a subset of PAD subjects (n=49) and healthy volunteers (n=9). There were no qualitative differences in the baseline characteristics of these patients compared to the groups as a whole. There was significantly higher expression of IL-6 (relative expression 1.85 [0.86, 3.81] vs. 0.74 [0.38, 1.53], p=0.024) and CD36 (1.99 [1.25, 3.04] vs. 0.79 [0.43, 1.37], p=0.018) in PAD compared to healthy subjects. There was no significant difference in expression of TNFα or IFN gamma in patients with PAD compared with healthy subjects (table III). There was no significant correlation between monocyte gene expression and circulating biomarker levels (for TNFα, r=0.22, p=0.13 and for IL-6, r=0.15, p=0.33).

Table III.

Gene expression in peripheral blood monocytes

	Healthy (n=9)	PAD (n=49)	p-value
TNFα	0.72 [0.49, 1.07]	1.1 [0.66, 1.56]	0.27
IL-6	0.74 [0.38, 1.53]	1.85 [0.86, 3.81]	0.024
INF-gamma	0.67 [0.31, 1.23]	0.39 [0.07, 1.30]	0.19
CD-36	0.79 [0.43, 1.37]	1.99 [1.25, 3.04]	0.018

Tumor necrosis factor α (TNFα), interleukin-6 (IL-6), and interferon-gamma (INF-gamma). Data shown as median and interquartile range. Statistical comparisons achieved by Wilcoxon rank sum test.

Association of gene expression and inflammatory markers with walking time

We explored whether expression of inflammatory genes in peripheral blood monocytes and circulating biomarker levels was associated with walking time in subjects with PAD. Average maximal walking time among PAD subjects was 438 ± 262 seconds. As shown in figure 1, monocyte TNFα gene expression was associated inversely with MWT. Relative TNFα expression levels were highest in those patients in the lowest quartile of walking time. Increasingly lower expression was noted in PAD subjects with greater walking time (1.69 ± 1.29 in quartile 1, 1.30 ± 0.90 in quartile 2, 1.13 ± 0.58 in quartile 3, and 0.88 ± 0.38 in quartile 4, p=0.016). This relationship persisted even after adjustment for age, gender, and smoking status (p=0.04), though no longer significant after adjustment for ABI (p=0.10). There was no significant relationship between maximal walking time and expression of other inflammatory genes in monocytes.

Figure 1.

The association between walking time and expression of tumor necrosis factor α (TNFα) in patients with PAD. Subjects were divided by quartile of maximal walking time and comparison of mean values across quartiles was achieved using linear regression. Data are shown as mean ± standard deviation.

Circulating levels of TNFα were also significantly associated with walking time subjects with PAD (figure 2), p=0.028. MWT was lowest in those subjects with the highest levels of circulating TNFα (531 ± 323 seconds in quartile 1, 522 ± 288 seconds in quartile 2, 385 ± 270 seconds in quartile 3, and 309 ± 158 seconds in quartile 4). A similar relationship was noted between MWT and circulating hsCRP (p=0.028), IL-6 (p=0.03), and sICAM (p=0.018) (figure 2). The associations between all circulating inflammatory biomarkers and MWT remained significant even after adjustment for age, gender, smoking status, diabetes, and ankle-brachial index (for TNFα, p=0.002; for hsCRP, p=0.037, for IL-6, p=0.01, and for sICAM, p=0.0028). Using linear regression, we found that there was a significant relationship between MWT and TNFα (p=0.012), IL-6 (p=0.004), CRP (p=0.04), and s-ICAM (p=0.005). These data remained statistically significant even after multivariable adjustment for age, gender, smoking status, and ABI for TNF alpha (p=0.003); IL-6 (p=0.004); and s-ICAM (p=0.0002), but not for hsCRP (p=0.058).

Figure 2.

The association between maximal walking time (MWT) and circulating inflammatory biomarkers, including tumor necrosis factor α (TNFα), high sensitivity C reactive protein (hsCRP), interleukin-6 (IL-6), and soluble intercellular adhesion molecule (sICAM). Comparisons of mean walking time across biomarker quartiles was achieved by linear regression. Data for MWT are shown as mean ± standard deviation.

Quartiles values for inflammatory biomarkers are as follows: for TNFα, Q1: 2025.7 pg/mL, Q2: 2025.7–2528.6 pg/mL, Q3: 2528.6–3592.4 pg/mL, and Q4: ≥ 3592.4 pg/mL; for hsCRP, Q1: < 1.01 mg/L, Q2: 1.01–2.47 mg/L, Q3: 2.47–4.81 mg/L, and Q4: ≥ 4.81 mg/L; for sICAM, Q1: < 214.7 ng/mL, Q2: 214.7–255.3 ng/mL, Q3: 255.3–295.4 ng/mL, and Q4: ≥ 295.4 ng/mL; and for IL-6, Q1: < 1.61 pg/mL, Q2: 1.61–2.87 pg/mL, Q3: 2.87–3.49 pg/mL), and Q4: ≥ 3.49 pg/mL.

Discussion

Systemic inflammation promotes the development and progression of peripheral artery disease. Peripheral blood monocytes are key players in the pathogenesis of atherosclerotic plaque formation,¹³ and monocyte-derived pro-inflammatory cytokines may exacerbate the inflammatory cascade that fosters atherosclerosis in patients with PAD. Whether monocyte-derived inflammatory cytokines contribute to impaired functional capacity in patients with PAD has not been previously explored. Here we demonstrate that monocyte mRNA expression of the pro-inflammatory cytokine TNFα is associated with impaired maximal walking time in patients with PAD and intermittent claudication. We also demonstrate that circulating levels of pro-inflammatory cytokines, including TNFα, hsCRP, IL-6, and sICAM are significantly elevated in patients with PAD, and also correlate strongly with impaired maximal walking time, even after adjustment for demographic parameters, atherosclerotic risk factors including smoking and diabetes, and degree of hemodynamic stenosis, as indicated by the ABI.

Intermittent claudication, and importantly, improvement in symptoms in response to therapies such as exercise training, do not necessarily relate to changes in the severity of arterial stenosis, raising questions about other contributors to symptomatic PAD and other mechanisms by which known therapies afford a clinical benefit. Inflammation, including release of cytokines from peripheral blood monocytes, may contribute to claudication itself. One can theorize that monocytes promote an inflammatory cascade leading to impaired vascular function in the lower extremities, worsening insulin resistance, and advancing skeletal muscle dysfunction - all factors which may contribute to claudication.^14–16 In patients with PAD, routine daily physical activity, in and of itself, may also exacerbate the inflammatory state, as skeletal muscle injury from repetitive limb ischemia with walking can increase circulating levels of cytokines and adhesion molecules, increasing oxidant stress and endothelial dysfunction.^17–19 Acute physical activity in patients with PAD also increases local proportion of activated neutrophils and myeloperoxidase content, additional markers of oxidant stress and inflammation, in the affected limb specifically.^20–22 Taken together, over-expression of pro-inflammatory genes in peripheral blood monocytes as demonstrated in our study may be both a cause and a result of impaired walking in patients with PAD.

The importance of reducing inflammation as a potential therapeutic target for patients with claudication is supported by several clinical trials of statin therapy to improve impaired walking time in PAD. Statins, known to reduce systemic inflammation, have been shown to improve walking time in individuals with claudication,^{11, 12, 23} although there are conflicting data from trials using lower dose statins or other lipid lowering therapies.^{24, 25} The contribution of inflammation to impaired physical functioning has also been suggested from cross-sectional studies in patients with PAD, both with and without claudication, in whom biomarkers of inflammation, such as CRP, D-dimer and fibrinogen are associated with poorer performance on the six-minute walk test and other measures of strength and mobility.¹⁰ Our study differs from this prior work in that it is the first to explore gene expression in monocytes, a cell type recognized to be a source of circulating inflammatory biomarkers, and uses exercise treadmill testing as a gold standard measure of functional capacity in PAD patients.

There are several limitations to this work that merit consideration. We did not find a significant correlation between expression of pro-inflammatory genes in monocytes and circulating levels of inflammatory biomarkers. The source of elevated circulating inflammatory biomarkers in PAD is no doubt multifactorial. Thus, the lack of a significant association between monocyte expression and circulating levels of these biomarkers may reflect the fact that peripheral blood monocytes are only one of several sources of production of circulating inflammatory biomarkers, with tissue-based monocytes also contributing to the inflammatory milieu. Failure to detect a significant association also may reflect the small sample size. In addition, in this cross-sectional analysis, we cannot determine whether monocyte TNFα gene expression is a direct cause of impaired walking time, whether limitations in physical activity in patients with PAD may be the primary driver of increased inflammation and insulin resistance, or whether inflammation may be exacerbated by acute bouts of physical activity in patients with PAD.^26–29 Nevertheless, this data does support the prospect that modulating inflammation, and specifically inflammation in monocytes, might favorably impact PAD and claudication symptoms. Finally, we recognize that there are differences in clinical characteristics beyond the presence of PAD alone that may account for some of the differences in gene expression in PAD compared to healthy subjects. It is not possible to separate the effects of age and other co-morbid diseases which, by definition, were more prevalent in the PAD population. Future studies with additional comparator groups, such as individuals with other manifestations of atherosclerosis but not involving the lower extremities, would be required to discern whether the findings are attributable to PAD per se or to other related factors.

In conclusion, over the last several decades, there has been increasing understanding of the contribution of inflammation to the pathophysiology of peripheral artery disease.³⁰ Inflammation is known to be associated with the initial development of PAD, contributes to adverse morbidity and mortality in individuals with PAD, and portends worse outcomes in PAD subjects undergoing peripheral bypass surgery, including restenosis and graft failure.^{4, 5, 7–9, 31} Inflammatory mediators, including TNFα, also promote skeletal muscle wasting resulting in decreased muscle mass and strength.^{32, 33} Here we demonstrate a clear association of walking impairment with circulating inflammatory biomarkers and with monocyte expression of the inflammatory gene, TNFα. Whether monocyte over-activation can be ameliorated with our existing therapies for claudication, such as exercise training, merits further evaluation, as suggested by our data here.

It has become increasingly clear is that the pathophysiology of intermittent claudication is multifactorial. Hemodynamic stenosis resulting from atherosclerotic plaque is necessary for development of claudication, but future studies are required to better understand the relative contribution of non-hemodynamic factors, such as inflammation, to intermittent claudication. Improved understanding of these factors will hopefully guide the development of novel targeted therapies to improve functional impairment in patients with PAD.