https://orcid.org/0000-0002-6362-7535
Figures
Abstract
Background
Intermittent claudication (IC) is a common manifestation of peripheral arterial disease. Some patients with IC experience a rise in Urinary N-acetyl-β-D-Glucosaminidase (NAG)/ Creatinine (Cr) ratio, a marker of renal injury, following exercise. In this study, we aim to investigate whether peripheral blood mononuclear cells (PBMC) from patients with IC who exhibit a rise in urinary NAG/ Cr ratio following exercise exhibit differential IL-10/ IL-12 ratio and gene expression compared to those who do not have a rise in NAG/ Cr ratio.
Methods
We conducted a single center observational cohort study of patients diagnosed with IC. Blood and urine samples were collected at rest and following a standardised treadmill exercise protocol. For comparative analysis patients were separated into those with any rise in NAG/Cr ratio (Group 1) and those with no rise in NAG/Cr ratio (Group 2) post exercise. Isolated PBMC from pre- and post-exercise blood samples were analysed using flow cytometry. PBMC were also cultured for 20 hours to perform further analysis of IL-10 and IL-12 cytokine levels. RNA-sequencing analysis was performed to identify differentially expressed genes between the groups.
Results
20 patients were recruited (Group 1, n = 8; Group 2, n = 12). We observed a significantly higher IL-10/IL-12 ratio in cell supernatant from participants in Group 1, as compared to Group 2, on exercise at 20 hours incubation; 47.24 (IQR 9.70–65.83) vs 6.13 (4.88–12.24), p = 0.04. 328 genes were significantly differentially expressed between Group 1 and 2. The modulated genes had signatures encompassing hypoxia, metabolic adaptation to starvation, inflammatory activation, renal protection, and oxidative stress.
Citation: Normahani P, Boyle JJ, Cave L, Brookes P, Woollard KJ, Jaffer U (2022) Peripheral blood mononuclear cell gene expression and cytokine profiling in patients with intermittent claudication who exhibit exercise induced acute renal injury. PLoS ONE 17(3): e0265393. https://doi.org/10.1371/journal.pone.0265393
Editor: Christian Schulz, Ludwig-Maximilians-Universitat Munchen, GERMANY
Received: April 5, 2021; Accepted: March 2, 2022; Published: March 17, 2022
Copyright: © 2022 Normahani et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data cannot be shared publicly because of sensitive nature and risk of identifying patients. Data are available from the Imperial College London Institutional Data Access / Ethics Committee (contact via Ms Becky Ward, Research Governance Manager, E: becky.ward@imperial.ac.uk) for researchers who meet the criteria for access to confidential data.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Peripheral arterial disease (PAD) is characterised by narrowing or occlusion of peripheral arteries and is estimated to effect over 230 million people worldwide [1]. Intermittent claudication (IC), which is defined as exertional pain in the lower extremity that is relieved by rest, is a common manifestation of PAD. Patients with IC have a three- to six- fold increased mortality from cardiovascular events compared with age matched controls [2]. This elevated risk may be related to the systemic endothelial dysfunction and inflammation associated with PAD which can accelerate existing cardiovascular disease [3–6]. This may be linked to reduced vessel wall shear stress and repeated ischemia-reperfusion injury [7] which has been observed to improve following revascularisation [8]. In patients with IC, exercise results in tissue ischaemia with local accumulation of anaerobic metabolites and oxygen free radicals. This episodic ischaemia results in these metabolites circulating and resulting in a transient systemic inflammatory response and end organ injury [9–11].
We have previously shown that a subgroup of patients with IC experience a rise in Urinary N-acetyl-β-D-Glucosaminidase (NAG)/ Creatinine (Cr) ratio, a sensitive and specific marker of renal injury, following exercise [11]. This group also demonstrated a lower level of baseline oxidative stress, higher post exercise increase in oxidative stress and macro-vascular endothelial function [11]. We proposed that this group may represent a phenotypically distinct subset of patients with IC who are more susceptible to acute kidney injury, and possibly other end organ injury, following exercise.
Human peripheral blood mononuclear cells (PBMC) are composed of lymphocytes (T cells, B cells, and NK cells), monocytes, and dendritic cells. They are an important component of innate immunity and play a key role in inflammation and the oxidative stress response. Differences in PBMC gene expression have been demonstrated between patients with and without PAD [12]. Furthermore, both IL-10 (Interleukin 10) and IL-12 (Interleukin 12) have important roles in the regulation of the innate immune system [13] and have been associated with the inflammatory state of atherosclerosis observed in PAD [14]. In this study, we aim to investigate whether PBMC, from patients diagnosed with IC, who exhibit a rise in urinary NAG/ Cr ratio following exercise exhibit differential IL-10/ IL-12 and gene expression compared to individuals who do not have a rise in NAG/ Cr ratio. This may help identify useful biomarkers for patients with IC who are susceptible to exercise induced end organ injury, and also potentially provide useful insights into underlying molecular mechanisms.
2. Materials and methods
We conducted a single center observational cohort study. The study was approved by the United Kingdom National Research Ethics Committee (REC 12/WA/0196). All recruited patients provided written informed consent.
2.1 Patient recruitment
Potential participants were identified from the prospectively managed arterial database of patients with IC undergoing routine clinical follow up at Hammersmith Hospital, London. All patients with evidence of haemodynamically significant lower limb PAD on duplex ultrasonography and documented current symptoms of intermittent claudication were invited to participate. Patients with signs or symptoms suggestive of chronic limb threatening ischaemia (CLTI) were not eligible for inclusion. Patients were excluded in the presence of concomitant disease severe enough to preclude treadmill exercise (such as musculoskeletal pain, heart disease or lung disease). Dialysis-dependent renal failure patients were also excluded. Patient demographics, comorbidities and medications as well as ankle-brachial pressure indices were recorded.
2.2 Treadmill exercise protocol
A standard testing protocol was followed: Ambulant patients were asked to empty their bladder before a 90-minute rest period (to avoid interference from the exercise of walking to the vascular laboratory). Following rest, urine and blood samples were collected and temporarily stored on ice. The participant was then asked to walk briskly on a treadmill (10 per cent incline; 4 miles/ hour) until they could not tolerate their typical claudication pain. Further blood and urine samples were taken and temporarily stored on ice.
Protease inhibitor cocktail (5% weight/volume; Sigma-Aldrich, Gillingham, UK) was added to urine samples which were stored at -70°C and analysed collectively. Urinary NAG and Cr levels were measured using ELISA. The NAG/Cr ratio was calculated to normalise for differences in urine output rate.
2.3 Peripheral blood mononuclear cell (PBMC) separation and immunofluorescence flow cytometry
A modified Boyum’s density gradient separation method was used to isolate PBMC. Briefly, 20 ml of blood collected in an EDTA tube was layered onto an equal volume of Lymphoprep (Axis-Shield UK, Cambridgeshire, UK) and centrifuged at 2800 rpm for 20 minutes. The ‘buffy coat’ was removed and diluted to 20 ml with Gibco RPMI 1640 medium (Life technologies ltd, Manchester, UK). This was further centrifuged at 2200 rpm, supernatant tipped off and cell pellet re-suspend and allowed to stand for 5 minutes in 5 ml of 0.85% Ammonium chloride to cause red cell lysis. The cell suspension was again diluted to 20 ml with RPMI and centrifuged at 1500 rpm for 5 minutes. Again, supernatant was tipped off and cells re-suspend in 5 ml of RPMI with 5% foetal calf serum, 1% glutamine, and 1% penicillin/ streptomycin (Sigma-Aldrich; Dorset, England). Cell concentration was determined using a haemocytometer and 5000 cell events were recorded for each sample.
For flow cytometric analysis 200μl of cell suspension was pipetted into 5 wells of V bottomed 96 well plates, which were centrifuged at 1500 rpm for 5 minutes. Supernatant was discarded and cell pellet re-suspended in 100μl of phosphate buffered saline (PBS) and again centrifuged at 1500 rpm for 5 minutes. The cell pellets were re-suspended in 40μl PBS and either 10μl of PBS or a two-colour direct immunofluorescence reagent in a 1:5 dilution (BD: Simultest™ CD3/CD16 + CD56; Simultest™ LeucoGATE™ (CD45/ CD14); Oncomark CD14/ CD64; Becton Dickson, Oxford, UK) was added and incubated at 4°C for 30 minutes. Samples were transferred to test tubes and 600μl PBS added. The percentage of fluorescent cells of each leukocyte population was quantified using a Beckman Coulter FC500 instrument (Beckman-Coulter, High Wycombe, UK).
2.4 Cytokine assay from PBMC
0.5ml of the PBMC suspension, containing 1 X 106 cells, was seeded in 92mm polystyrene cell culture dishes (Nunclon; Fisher Scientific, Loughborough, UK) and incubation was performed after Signorelli et al. [15] for 20 hours at 37°C in a moist chamber containing 5% CO2. 50μl of supernatant was collected after 1-, 6- and 20-hours incubation and stored at -70°C for ELISA. A custom quantitative multiplex ELISA protocol modified for low sample volume was used to analyse IL-12p70 (dynamic range 3000 to 5.35 pg/ml) and IL-10 (dynamic range 2680 to 4.78 pg/ml) from supernatant samples (Q-plex; Quansys Biosciences, Logan, Utah, USA).
2.5 Peripheral blood mononuclear cell (PBMC) RNA-sequencing
Isolated PBMC were lysed in RNA-lysis buffer, RNA purified and analysed by RNA-Sequencing (Fig 1). After RNA-sequencing was performed, read files were quality checked, processed and analysed using Partek® Flow® software (Version 7, Partek Inc., St. Louis, MO, USA). Read quality tests showed that both the average base quality score per position and the average quality score per read was above 30 (i.e., 99.9% base call accuracy), indicating that files were of good quality. Alignment was performed using the STAR aligner tool [16]. Post-alignment quality tests demonstrated that at least 75% of reads were aligned to the genome per sample and that samples had an average read depth of a covered region of at least 30. The aligned reads were quantified to the Partek® E/M annotation model producing number of counts per gene. Gene counts were normalised to read depth per sample, using the counts per million (CPM) method. The statistical comparison test, gene-specific analysis, was performed between the two test groups resulting in a fold change value, p-value and FDR-value per gene.
Blue boxes, processes as indicated. Blue circles, data types as indicated.
Significantly regulated genes were analysed using hierarchical cluster analysis and principal component analysis, Gene Ontology analysis by PANTHER GO-Slim (Version 14) [17–19], network analysis by mapping onto the STRING database (Version 11) [20] and graphing in Cytoscape (Version 3.8) [21], and transcription factor binding site analysis, using a summation of rank scores from OPOSSUM with JASPAR core matrices, OPOSSUM with JASPAR position weight matrices, X2K with each of ENCODE+ChEA, ENCODE, ChEA2015, JASPAR + TRANSFAC, ChEA2016, ARCHS4, ENRICHR, ChEA3 and PASTAA.
2.6 Statistical analysis
Data were analysed using Microsoft Excel (2010 edition) and SPSS (version 20). Difference between groups was tested using the Mann Whitney U test (two-tailed) for numerical values and Chi-squared for categorical data. p values of ≤0.05 were considered statistically significant.
3. Results
Twenty patients participated in this study. Patient demographics are presented in Table 1. Overall, across all participants median urinary NAG/Cr rose from a pre-exercise level of 8.94 (IQR, 6.73 to 14.27) to a post exercise level of 12.90 (IQR 9.07 to 17.72); p = 0.0003. Patients were separated into two groups for comparative analysis; those with any rise in NAG/Cr ratio post exercise (Group 1; n = 8) and those with no rise in NAG/Cr ratio (Group 2; n = 12).
3.1 PBMC composition
The pre-exercise total mononuclear cell counts were similar between participants in Group 1 (3.825 x 106; 2.575 x106 to 4.925 x106) and Group 2 (2.975 x106; 2.500 x106 to 4.950 x106) (cells/ ml, median; IQR). The post-exercise cell counts were also similar between participants in Group 1 (3.400 x106; 1.475 x106 to 9.338 x106) and Group 2 (4.800 x106; 3.575 x106 to 6.388 x106) (cells/ ml, median; IQR). Typical gates employed for flow cytometric analysis are demonstrated (Fig 2). Flow cytometric analysis demonstrated that the PBMC percentage composition was similar between participants in Group 1 and Group 2 (Fig 3).
A) CD3+/16-/56- cells (gate E), CD16+/56+/3- cells (gate G); B) CD45+/14+ cells (gate I); C) CD14+/64+ cells (gate L); D) CD19+/3- cells (gate J). We gated for the whole population from the Lymphoprep isolation; forward and side scatter plots is available in the S1 Fig.
3.2 PBMC supernatant cytokine profile
Supernatant from PBMC from pre-exercise blood samples did not demonstrate any statistically significant difference in IL-10 or IL-12 levels or IL-10/IL-12 ratio between participants in Group 1 and Group 2. However, post-exercise sample supernatant contained higher levels of IL-10 in Group 1 as compared to Group 2 at 20 hours following incubation (44.36, 15.11 to 96.27 versus 7.27, 5.24 to 38.74; pcg/ml/106 cells; median, IQR; p = 0.045; Fig 4A). IL-12 levels did not exhibit any difference between Group 1 and 2: 1.64, 1.17 to 3.47 versus 1.27, 1.09 to 2.43 (pcg/ml/106 cells; median, IQR; p = 0.724; Fig 4B).
A) IL-10 (pcg/ ml), * p = 0.045 B) IL-12 (pcg/ ml), C) IL-10/IL-12 ratio from cell medium supernatant of PBMC cultured from patients with (Group 1) and without (Group 2) rise in urinary NAG/Creatinine ratio following exercise, * p = 0.04. Statistical comparisons made using Mann Whitney U test.
Interestingly, there was a significantly higher IL-10/IL-12 ratio in cell supernatant from participants in Group 1, as compared to Group 2, on exercise at 20 hours incubation: 47.24, 9.70 to 65.83 versus 6.13, 4.88 to 12.24 (median, IQR; p = 0.04; Fig 4C). There was no difference in IL-10/IL-12 ratio from supernatant of PBMC taken from pre-exercise blood at one, six or 20 hours of incubation between participants in Group 1 and 2. Similarly, there were no differences seen in supernatant taken from post exercise samples at one- or six-hours following incubation.
3.3 PBMC RNA-sequencing
328 genes were significantly differentially expressed between Group 1 and 2 (Fig 5A). Significantly differentially expressed genes (Fig 6) were assessed by Gene Ontology (GO) analysis. Significant GO categories for Molecular Function are shown (Fig 7A). For space and clarity, they are reduced to the distinct and meaningful categories in the list. Signal transduction was a notable molecular function. Specific signal transduction activities included kinases, GPCR (G-Protein Coupled Receptor) activity, ion channels, cytokine activity, ubiquitinylation, transcription factor activity. Significant GO categories for Cellular Processes are shown (Fig 7B). These included a set of metabolic categories: Nitrogen compound metabolism; Heterocycle metabolic processes; Nucleic acid metabolic processes; RNA metabolic processes; Nitrogen compound biosynthesis; Ribonucleoprotein biosynthesis; Catabolic processes; ncRNA metabolic processes. Additional notable GO categories were cellular responses to stress, and to DNA damage.
A) hierarchical cluster analysis. Upregulated genes (red) and downregulated genes (green) as indicated. There is a clear separation between the groups with (POS) and without (NEG) urinary NAG/Creatinine rise following exercise. The group with a rise urinary NAG/ Creatinine has a cluster of upregulated genes to the right-hand side of the diagram. B) principal component analysis. There is a clear separation between the groups with (blue, POS) and without (red, NEG) urinary NAG/ Creatinine rise following exercise. The groupings are shown with the superimposed polygons.
A) Y-axis, important GO categories of Molecular Function; X-axis, probability expressed as negative log of the false discovery rate (FDR). B) Y-axis, important GO categories of Cellular Function; X-axis, expressed as negative log of the false discovery rate (FDR).
3.3.1 Network analysis.
Network analysis was assessed to help to identify the most connected genes in a process, which are more likely to be important regulators. As only a single time point was available, the differentially expressed genes were mapped onto the database of protein-protein interactions. The most highly connected genes (Fig 8A, 8B) included RPS2 (Ribosomal Protein S2), IL6 (Interleukin 6), PTGS2 (Prostaglandin Synthase 2), CCL5 (C-C Chemokine Ligand 5), RPP25 (Ribonuclease P and MRP subunit P25), SOD2 (Superoxide Dismutase 2), CCR7 (C-C Motif Chemokine Receptor 7), VEGF (Vascular Endothelial Growth Factor), and a family of ribosomal proteins. The two most highly connected nodes were IL6 and RPS2. These data were consistent with the GO data pointing to transmembrane signal transduction, cytokine signalling, kinase activity and ribonucleoprotein and ribosomal biogenesis. The number of connexions for each node (degree of connectedness) was then output from Cytoscape and collated into ‘bins’ to smooth out random variations in connections per node. These bins were on a logarithmic scale because node degree and number of nodes typically forms a straight line on that graph (the power law relationship). Then connectedness was graphed against the number of nodes (Fig 8C). The graph was highly non-linear, with 2 unexpected peaks, one in the mid-range and one comprising highly connected nodes.
A) overview of gene network. Blue, nodes in the network. Grey lines, known connections between nodes. Some nodes are very highly connected. B) alternate overview of gene network, redrawn to maximally display connections and the most highly connected nodes. Blue, nodes in the network. Grey lines, known connections between nodes. Some nodes are very highly connected. C) graph of node connections. X-axis, number of connections per node; Y-axis, number of nodes.
The STRING database also output a summary of the network (Fig 9). This indicated over-representation for regulation of immune system and transcription factor binding in Biological Process (Fig 9). Molecular Function showed enrichment for Transcription Factors (Fig 9).
The summary data are shown at the top and summarise the graphs in Fig 7.
3.3.2 Transcription factor binding site analysis.
We then assessed by transcription factor binding site (TFBS) analysis. Different TFBS analysis programs may yield different results. Therefore, multiple analysis programs were used. Output for PASTAA is shown (Fig 10A). This showed over-representation for binding sites for ATF, CREB (which is related to ATF), CREL (NFkB), SP1, EGR3, SRF, KROX, STAT, AHRARNT. Output TFBS scores were ranked for each and combined (Fig 10B). Top scoring TFs across 11 programs and databases indicated that the most highly scored TFBS representing subfamilies of transcription factors were CREB1 (which is related to ATF1 and CREM), NFkB (CREL), IRF, ZNF, P53, EGR, AP1, FOX, TCF, NR4A and MTF1.
A) list of predicted Transcription Factor binding sites (PASTAA) ranked by P-value (generated using the Max Planck Institute for Molecular Genetics Database), B) X-axis, score calculated as an aggregate of rankings of common transcription factor binding sites prediction programs (OPPOSUM using JASPAR, JASPAR CORE, X2K using databases ENCODE, ChEA2015, JASPAR and TRANSFAC, ChEA2016, archs4, ENRICHR, ChEA3, PASTAA).
4. Discussion
This study has detected increased levels of the anti-inflammatory IL-10 and significantly higher IL-10/ IL-12 ratio in the PBMC supernatant of patients with IC who exhibited evidence of acute renal injury following exercise. We also conducted gene expression profiling of sampled PBMC which demonstrated a complex systemic inflammatory process in this group of patients. Our results suggest that some patients with IC may have an altered immune status which contributes to systemic inflammation and acute renal injury following exercise.
Our previous profiling of this cohort of patients suggested that those who exhibit exercise induced NAG/Cr rise also have a greater rise in oxidative stress (hydrogen peroxide; H2O2) and a greater increase in endothelial reactivity (flow mediated dilation of the brachial artery) following exercise as compared to those with no NAG/Cr rise [11]. Interestingly, this group also demonstrated a longer average walking distances on the treadmill (395m vs 195m; p = 0.05) despite comparable disease severity, as measured by the ankle-brachial pressure index (0.71 vs 0.83; p = 0.61), as compared to those with no NAG/Cr rise in this previous study [11].
Consistent with the literature, there was a small but non-significant increase in CD16+ cells post exercise, likely to be CD16+ monocytes [22, 23]. To avoid over-reliance on one type of bioinformatic analysis, several different approaches were taken, allowing themes common to the majority to emerge. Gene Ontology analysis pointed to an important molecular role for signal transduction. The signature of a cytokine pathway in this analysis could relate to the other signatures for IL6 as an important node, and STAT sites in the clusters of regulated genes. The signature indicating use of GPCRs could also be related to PTGS2 (which generates both pro-inflammatory and anti-inflammatory eicosanoids) and to CRE sites, as eicosanoids signal in part via GPCRs coupled to adenylate cyclase and cyclic-AMP.
Activation of the central pro-inflammatory NF-κB pathway involves multiple cell surface receptors, kinases, and ubiquitinylation of the inhibitory component IκBα. Thus, the ubiquitinylation, kinase and transmembrane signalling GO hits are likely to represent involvement of NF-κB.
IL-6 was prominent on some of the RNA-seq analysis, particularly in the network analysis as a highly connected gene. IL-6 may now be specifically inhibited by neutralising antibodies, that are licensed for inflammatory disease [24]. As IL-6 elevates so-called acute phase markers. Serum CRP (C-Reactive Protein) would be an interesting follow-up and has been linked to long-term risk of atherosclerotic cardiovascular disease [25].
The GO for biological process was also insightful. As a form of ischemia-reperfusion, claudication could be expected to activate cellular processes involved in cell energy deprivation. It is interesting that related signals could be detected in peripheral blood, indicating systemic activation. The PBMC sampled had GO signals for catabolism including protein catabolism. In addition, the GO signatures suggested a commitment to RNA synthesis. This would be in keeping with the other signatures for transcription factor activation and gene regulatory activity and is likely to be involved in a systemic response to claudication. It could reflect gene transcription for pro-inflammatory activity, pro-repair activity or metabolic adaptation.
Taken together, the modulated genes had signatures encompassing hypoxia, metabolic adaptation to starvation, inflammatory activation, renal protection, and oxidative stress. Some of these pathways were pointed to by a combination of induced transcription factor and induced target genes, such as NFκB driven inflammation (namely C-REL (p65 NFκB) and target genes IL6, CXCL3 (CXC-Motif Chemokine Ligand 3), NFκB, PTGS2, CXCL2 (CXC-Motif Chemokine Ligand 2), DUSP4 (Dual Specificity Phosphatase 4) and SOD2). Similarly, induction of hypoxia-inducible factor (HIF), and some of its target genes (VEGFA, HIF1A, SEMA7A (Semaphorin 7A), SLC2A3 (Solute Carrier Family 2 member A3), HK2 (Hexokinase 3), PFK (Phosphofructokinase)) pointed to a role for responses to hypoxia, as expected in ischemia [26]. The catabolic response, suggested involvement of the ischemia-activated signalling kinase AMPK (5’ Adenosine Monophosphate Activated Protein Kinase) as a driver as it drives catabolism [27]. Some of the genes and predicted transcription factors are related to oxidative stress, notably SOD2, BACH1, NFE2L2, MAFG. This ties in with our previous description of elevated oxidative stress in these patients. Of interest was the novel protective gene NAMPT (Nicotinamide Phosphoribosyl Transferase), which has been shown to be protective in renal ischemia [28]. CREB target genes were also prominent and statistically significant (NR4A2, PURB, RBKS, RPP25, PTPRK, DTX4, QPRT, EAF1, TRAF4, STYK1, TGIF1, JMJD1C, PDE4B, BANK1), consistent with a role for cyclic-AMP activators such as eicosanoids derived from PTGS2 (COX2).
A potential limitation of this work is that enriched PBMC’s may include a proportion of other contaminating cells. In future studies, flow cytometric analysis may be beneficial to quantify this. However, all samples were treated similarly and so any potential contamination would not be expected to confound our findings.
5. Conclusions
Our results suggest that some patients with IC have an altered immune status making them ‘vulnerable’ to systemic inflammation and end organ injury following exercise. We have identified a panel of genes which are differentially expressed in this vulnerable group of patients and may be potentially useful for clinical screening. Further validation is necessary to evaluate the utility of this gene panel as a prognostic biomarker for patients with intermittent claudication.
Supporting information
S1 Fig. Forward scatter vs side scatter plot for cytometric analysis.
https://doi.org/10.1371/journal.pone.0265393.s001
(DOCX)
Acknowledgments
Infrastructure support for this research was provided by the NIHR Imperial Biomedical Research Centre (BRC).
References
- 1. Song P, Rudan D, Zhu Y, Fowkes FJI, Rahimi K, Fowkes FGR, et al. Global, regional, and national prevalence and risk factors for peripheral artery disease in 2015: an updated systematic review and analysis. Lancet Glob Heal. 2019 Aug;7(8):e1020–30.
- 2. Robless P, Mikhailidis DP, Stansby GP. Cilostazol for peripheral arterial disease. Cochrane database Syst Rev. 2007 Jan;(1):CD003748. pmid:17253494
- 3. Brevetti G, Silvestro A, Di Giacomo S, Bucur R, Di Donato A, Schiano V, et al. Endothelial dysfunction in peripheral arterial disease is related to increase in plasma markers of inflammation and severity of peripheral circulatory impairment but not to classic risk factors and atherosclerotic burden. J Vasc Surg. 2003 Aug;38(2):374–9. pmid:12891123
- 4. Gokce N, Keaney JF, Hunter LM, Watkins MT, Nedeljkovic ZS, Menzoian JO, et al. Predictive value of noninvasively determined endothelial dysfunction for long-term cardiovascular events in patients with peripheral vascular disease. J Am Coll Cardiol. 2003 May;41(10):1769–75. pmid:12767663
- 5. Silvestro A, Scopacasa F, Ruocco A, Oliva G, Schiano V, Zincarelli C, et al. Inflammatory status and endothelial function in asymptomatic and symptomatic peripheral arterial disease. Vasc Med. 2003 Nov;8(4):225–32. pmid:15125481
- 6. Brevetti G, Silvestro A, Schiano V, Chiariello M. Endothelial dysfunction and cardiovascular risk prediction in peripheral arterial disease: additive value of flow-mediated dilation to ankle-brachial pressure index. Circulation. 2003 Oct;108(17):2093–8. pmid:14530195
- 7. Gnasso A, Carallo C, Irace C, De Franceschi MS, Mattioli PL, Motti C, et al. Association between wall shear stress and flow-mediated vasodilation in healthy men. Atherosclerosis. 2001 May;156(1):171–6. pmid:11369011
- 8. Normahani P, Khosravi S, Sounderajah V, Aslam M, Standfield NJ, Jaffer U. The Effect of Lower Limb Revascularization on Flow, Perfusion, and Systemic Endothelial Function: A Systematic Review. Angiology. 2020 Nov;3319720969543.
- 9. Joras M, Poredos P. The association of acute exercise-induced ischaemia with systemic vasodilator function in patients with peripheral arterial disease. Vasc Med. 2008 Nov;13(4):255–62. pmid:18940901
- 10. Wijnen MHWA, Coolen SAJ, Vader HL, Reijenga JC, Huf FA, Roumen RMH. Antioxidants Reduce Oxidative Stress in Claudicants. J Surg Res [Internet]. 2001;96(2):183–7. Available from: http://www.sciencedirect.com/science/article/pii/S0022480400960785 pmid:11266271
- 11. Jaffer U, Singh P, Aslam M, IITam FWK, Standfield NJ. Potential identification of a claudicant group vulnerable to acute kidney injury. Perfusion. 2016 Oct;31(7):560–7. pmid:26980004
- 12. Masud R, Shameer K, Dhar A, Ding K, Kullo IJ. Gene expression profiling of peripheral blood mononuclear cells in the setting of peripheral arterial disease. J Clin Bioinforma [Internet]. 2012 Mar 12;2:6. Available from: https://pubmed.ncbi.nlm.nih.gov/22409835 pmid:22409835
- 13. O’Garra A, Murphy KM. From IL-10 to IL-12: how pathogens and their products stimulate APCs to induce T(H)1 development. Nat Immunol. 2009 Sep;10(9):929–32. pmid:19692989
- 14. Girn HRS, Orsi NM, Homer-Vanniasinkam S. An overview of cytokine interactions in atherosclerosis and implications for peripheral arterial disease. Vasc Med. 2007 Nov;12(4):299–309. pmid:18048466
- 15. Signorelli SS, Mazzarino MC, Di Pino L, Malaponte G, Porto C, Pennisi G, et al. High circulating levels of cytokines (IL-6 and TNFalpha), adhesion molecules (VCAM-1 and ICAM-1) and selectins in patients with peripheral arterial disease at rest and after a treadmill test. Vasc Med. 2003;8(1):15–9. pmid:12866607
- 16. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics [Internet]. 2012/10/25. 2013 Jan 1;29(1):15–21. Available from: https://pubmed.ncbi.nlm.nih.gov/23104886 pmid:23104886
- 17. Thomas PD, Campbell MJ, Kejariwal A, Mi H, Karlak B, Daverman R, et al. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 2003 Sep;13(9):2129–41. pmid:12952881
- 18. Mi H, Muruganujan A, Ebert D, Huang X, Thomas PD. PANTHER version 14: more genomes, a new PANTHER GO-slim and improvements in enrichment analysis tools. Nucleic Acids Res. 2019 Jan;47(D1):D419–26. pmid:30407594
- 19. Mi H, Muruganujan A, Huang X, Ebert D, Mills C, Guo X, et al. Protocol Update for large-scale genome and gene function analysis with the PANTHER classification system (v.14.0). Nat Protoc. 2019 Mar;14(3):703–21. pmid:30804569
- 20. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019 Jan;47(D1):D607–13. pmid:30476243
- 21. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003 Nov;13(11):2498–504. pmid:14597658
- 22. Slusher AL, Zúñiga TM, Acevedo EO. Maximal Exercise Alters the Inflammatory Phenotype and Response of Mononuclear Cells. Med Sci Sports Exerc. 2018 Apr;50(4):675–83. pmid:29112629
- 23. O’Carroll L, Wardrop B, Murphy RP, Ross MD, Harrison M. Circulating angiogenic cell response to sprint interval and continuous exercise. Eur J Appl Physiol. 2019 Mar;119(3):743–52. pmid:30673849
- 24. Collison J. Tocilizumab improves quality of life in GCA. Nat Rev Rheumatol. 2019 Apr;15(4):188. pmid:30837703
- 25. Lagrand WK, Visser CA, Hermens WT, Niessen HW, Verheugt FW, Wolbink GJ, et al. C-reactive protein as a cardiovascular risk factor: more than an epiphenomenon? Circulation. 1999 Jul;100(1):96–102. pmid:10393687
- 26. Schödel J, Ratcliffe PJ. Mechanisms of hypoxia signalling: new implications for nephrology. Nat Rev Nephrol. 2019 Oct;15(10):641–59. pmid:31488900
- 27. Hardie DG, Schaffer BE, Brunet A. AMPK: An Energy-Sensing Pathway with Multiple Inputs and Outputs. Trends Cell Biol. 2016 Mar;26(3):190–201. pmid:26616193
- 28. Tran MT, Zsengeller ZK, Berg AH, Khankin E V, Bhasin MK, Kim W, et al. PGC1α drives NAD biosynthesis linking oxidative metabolism to renal protection. Nature. 2016 Mar;531(7595):528–32. pmid:26982719