Download Aberrant Circulating Levels of Purinergic Signaling Markers are

Aberrant Circulating Levels of Purinergic Signaling Markers are Associated with Several
Key Aspects of Peripheral Atherosclerosis and Thrombosis
Juho Jalkanen1, Gennady G. Yegutkin2, Maija Hollmén2, Kristiina Aalto2, Tuomas Kiviniemi3, Veikko
Salomaa4, Sirpa Jalkanen2 and Harri Hakovirta1,2
1
Department of Vascular Surgery, Turku University Hospital, Turku, Finland; 2Medicity Research
Laboratory, Department of Microbiology and Immunology, University of Turku, Turku, Finland; 3Heart
Center, Turku University Hospital, and; 4National Institute for Health and Welfare, Helsinki, Finland
Running title: Purinergic Signaling in Peripheral Artery Disease
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Subject code:
[17] Peripheral artery disease
[134] Atherosclerosis basic science pathophysiology
[135] Risk factors
[138] Cell signaling
Address correspondence to:
Dr. Sirpa Jalkanen
MediCity Research Laboratory
University of Turku
Tykistökatu 6 A
FIN-20520 Turku
Finland
Tel: +358 2 333 7007
Fax: +358 2 333 7000
[email protected]
In December 2014, the average time from submission to first decision for all original research papers
submitted to Circulation Research was 14.47 days.
DOI: 10.1161/CIRCRESAHA.116.305715
1
ABSTRACT
Rationale: Purinergic signaling plays an important role in inflammation and vascular integrity, but little is
known about purinergic mechanisms during the pathogenesis of atherosclerosis in humans.
Objective: To study markers of purinergic signaling in a cohort of patients with peripheral artery disease
(PAD).
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Methods and Results: Plasma ATP and ADP levels and serum nucleoside triphosphate
diphosphohydrolase-1 (NTPDase1/CD39) and ecto-5'-nucleotidase/CD73 activities were measured in 226
stable peripheral artery disease (PAD) patients admitted for non-urgent invasive imaging and/or treatment.
The major findings were that ATP, ADP, and CD73 values were higher in atherosclerotic patients than in
controls without clinically evident PAD (P < 0.0001). Low CD39 activity was associated with disease
progression (P = 0.01). In multivariable linear regression models high CD73 activity was associated with
chronic hypoxia (P = 0.001). Statin use was associated with lower ADP (P = 0.041) and tended to associate
with higher CD73 (P = 0.054), while lower ATP was associated with the use of angiotensin receptor
blockers (P = 0.015).
Conclusions: Purinergic signaling plays an important role in PAD progression. Elevated levels of
circulating ATP and ADP are especially associated with atherosclerotic diseases of younger age, and
smoking. The anti-thrombotic and anti-inflammatory effects of statins may partly be explained by their
ability to lower ADP. We suggest that the pro-thrombotic nature of smoking could be a cause of elevated
ADP, and this may explain why cardiovascular patients who smoke benefit from platelet P2Y12 receptor
antagonists more than their non-smoking peers.
Keywords:
ADP, ATP, atherosclerosis, CD39, CD73, purinergic signaling, thrombosis, smoking, adenosine,
adenosine receptor.
Nonstandard Abbreviations and Acronyms:
ARB
angiotensin receptor blocker
CAD
coronary artery disease
NT5E
ecto-5'-nucleotidase
NTPDase1
nucleoside triphosphate diphosphohydrolase-1
PAD
peripheral artery disease
P-AFOS
plasma alkaline phosphatase
DOI: 10.1161/CIRCRESAHA.116.305715
2
INTRODUCTION
Although atherosclerosis is a widely studied disease and extensive interventions are practiced, it
still remains a major cause of death in Western societies.1 The contemporary view of the disease as an
inflammatory process of the vascular wall is well-established and reviewed in several excellent articles.1-3
The importance of extracellular purinergic signaling or metabolic pathways (i.e., the conversion of
circulating ATP and ADP to AMP, and then to adenosine, by cell surface-associated and soluble
nucleotidases) in inflammation and cell trafficking is acknowledged.4-7 Moreover, based on preclinical
findings, evidence suggests that purinergic signaling plays an important role in atherosclerosis.8,9 The first
such findings from human samples were reported by Lecka et al. in 201010, and strikingly, recent human
genetic studies indicate that NT5E (ecto-5'-nucleotidase/CD73) mutations are an important factor
contributing to peripheral arterial calcification.11 Thus, there is compelling evidence for the role of
extracellular ATP metabolism in vascular inflammation, but in a broader spectrum, very little is known
about the contribution of purinergic signaling in atherosclerosis.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Endogenous nucleotides, ATP and ADP, are released into circulation from cells under stress and
injury. After release, they create an important signaling cascade to modulate vascular tone, platelet
aggregation, and inflammation.12,13 Firstly, the endothelial cell surface enzyme nucleoside triphosphate
diphosphohydrolase-1 (NTPDase1, also known as CD39) converts ATP to ADP, and further to AMP. ADP
is the most significant mediator of platelet aggregation via activation of the P2Y12 receptor expressed on
platelets, and the clearance of ADP via NTPDase1/CD39 is crucial for thromboregulation.14-16 Subsequent
breakdown of ATP/ADP-derived AMP to adenosine is mediated by another enzyme, ecto-5'nucleotidase/CD73.12 Adenosine itself acts as a powerful local anti-inflammatory agent regulating vascular
permeability and leukocyte trafficking via different adenosine receptor subtypes.5,17,18 This study was
designed to investigate whether plasma concentrations of ATP and ADP, and serum CD39 and CD73
activities, can discriminate different patient groups suffering from PAD and thus explain the mechanisms
underlying disease onset and progression.
METHODS
This study was approved by the local Ethical Committee of the Hospital District of Southwestern Finland.
By approval of the Ethical Committee, a register under the name The Role of Purinergic Signaling in
Atherosclerosis (PURE ASO) was formed and is held at the Department of Vascular Surgery, Turku
University Hospital.
PAD patient cohort.
For one year, from February 2012 to March 2013, we enrolled every patient diagnosed with a stable state
of PAD of the lower extremities who were admitted to the Department of Vascular Surgery at the Turku
University Hospital (Finland) for elective invasive treatment and imaging, i.e., open surgery or angiography
and endovascular treatment. Patients admitted for urgent treatment from the emergency unit and those with
acute ischemia, infections, or major tissue loss were not included in the study. During our enrollment period,
227 suitable patients were screened. Only one patient declined, and 226 gave written informed consent. The
study patients had earlier been seen and diagnosed with PAD by a vascular surgeon at the outpatient clinic
of the department.
Non-atherosclerotic control cohort.
For non-atherosclerotic controls, 64 middle-aged (25–50 years of age; 41 males and 23 females) volunteers
not on permanent medications and without clinically evident PAD were recruited. In addition, to match the
older age groups of the PURE ASO patient cohort, 89 frozen (˗70°C) samples were obtained from the
DOI: 10.1161/CIRCRESAHA.116.305715
3
general population cohort of FINRISK 199719 (see Online Supplement for further details). Controls had to
be free of chronic heart diseases, PAD, stroke, diabetes, cancer, chronic obstructive pulmonary disease
(COPD), and rheumatic illnesses both on the basis of a self-report and register linkage with national health
care databases.
Blood samples.
Samples were drawn in the morning after at least 4 hr of fasting. Two 9 mL samples of whole blood were
obtained, and each sample was placed in a serum sample tube or EDTA sample tube and handled as
described in the Online Supplement.
Quantification of ATP and ADP levels in human plasma.
Plasma ATP and ADP levels were determined using an ATPlite enzyme-coupled assay kit with a long-lived
luminescent signal (Perkin Elmer, Groningen, The Netherlands) as described previously20, with further
details in the Online Supplement.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Measurement of soluble nucleotidase activity in human serum.
Soluble ADPase/NTPDase and 5’-nucleotidase activities were assayed radiochemically, as described
previously21, with further details in the Online Supplement.
Statistical analyses.
Statistical analyses were performed in association with a professional statistical provider, 4Pharma Ltd.
(Turku, Finland). First, simple distributional characteristics of purinergic signaling markers were compared
between the controls and the patient cohort. Markers with skewed distributions were log-transformed to
better fit to normality, which was tested using a Shapiro-Wilk test. The patient cohort was analyzed
according to different subgroups and background variables, and compared against other subgroups or the
controls using a t-test for normally distributed variables (including log-transformed data). Pearson
correlations were used to study the associations of continuous variables. Linear regression models were
used to study the possible trends in the means of log-transformed marker values across PAD localization
and the association of markers to age as a continuous variable.
After comparing the controls and the patient cohort, further explorative statistical modeling was performed
using only the patient cohort. First, based on univariate analysis the associations of suspected
cardiovascular risk factors (for CD39: smoking, COPD, hypoxia, uremia, renal insufficiency, and
hypertension; for CD73: hypoxia, renal insufficiency, uremia; for ATP and ADP: smoking and
hypertension) to log-transformed markers were tested individually in linear regression models containing
gender, age group and PAD localization as fixed effects, and P-AFOS as a covariate. In addition, the
associations of used medications (aspirin, beta-blockers, angiotensin converting enzyme (ACE) blockers,
calcium channel blockers, furosemide, warfarin, nitroglycerin, metformin, oral or inhaled cortisone,
angiotensin receptor blockers (ARBs), bisphosphonates and gliptins) to log-transformed markers were
tested individually in linear regression models containing gender, age group, PAD classification, statin use,
and clopidogrel use as fixed effects, and P-AFOS as a covariate. A stepwise backward elimination method
was applied for all the above-mentioned models. The variable under question was retained in the model
even when fulfilling the elimination criterion to examine its significance in the final reduced model. Second,
the identified factors appearing to be related to the marker were then selected for the final multivariate
model incorporating all the selected factors, and gender, age group, PAD localization and P-AFOS. This
full combined model was then simplified to a reduced model using the backward elimination approach
leaving only statistically significant factors into the model. In summary, we first tested cardiovascular risk
factors and medications separately in the linear regression model. From these factors, a combined model
was formed encompassing all factors that had shown significance. Finally, a backward elimination method
was applied to the combined model to determine the most decisive factors affecting marker levels.
DOI: 10.1161/CIRCRESAHA.116.305715
4
No P-value adjustments were performed. A P-value of 0.05 was used as the threshold for statistical
significance. All statistical analyses were performed using SAS Software for Windows version 9.3.
RESULTS
Description of the patient cohort and controls.
For a comprehensive description of the patient cohort and controls, see the Online Supplement
including Tables I, II, and III.
Elevated P-AFOS levels can interfere with interpretations of purinergic signaling, especially CD73.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Before further analysis, the cohort was screened for increased P-AFOS levels. The reference values
of the local hospital laboratory were used with an upper threshold of 105 U/L (international units per liter).
As P-AFOS is elevated in acute hepatobiliar diseases, and bone-related disorders, it was our assumption
(and unpublished data) that these diseases, especially hepatobiliar diseases, may have a significant effect
on purinergic signaling without contributing significantly to atherosclerosis. Thus, the markers of
purinergic signaling could be misinterpreted as a result of another underlying illness. Altogether, 12 patients
with elevated P-AFOS were noted. In six of these patients, P-AFOS was only marginally over the threshold,
and no underlying condition explaining this could be seen in the patient files. All these six patients were
included in the final analysis. Six patients with clearly elevated P-AFOS levels (one patient with liver
cirrhosis, two patients with chronic pancreatitis, and three patients with prostate carcinoma and bone
metastases) were identified and excluded from the final statistical analyses. P-AFOS concentrations
positively correlated with CD39 levels (Figure 1A). An even stronger positive correlation was observed
with CD73 (Figure 1B). A slightly negative correlation was noted for P-AFOS with ATP and ADP (Figure
1C and D). The correlations of CD39, ATP, and ADP with P-AFOS were abolished after exclusion of the
six previously mentioned patients with marked background illnesses. A positive correlation persisted for
only P-AFOS and CD73 (Pearson r = 0.226, P < 0.001).
Markers of purinergic signaling are elevated in the patient cohort.
Normal serum CD39 activity is 12–20 nmol/mL/hr, and CD73 activity is 120–280 nmol/mL/hr.20The control group displayed consistently normal levels of CD39 and CD73 activity (Table 1). CD39
activity did not significantly differ between the PAD patients and the control group (P = 0.09), but CD73
activity was significantly higher in the patient cohort (P < 0.0001) than in the controls. Normal plasma ATP
and ADP values are generally within a 500–3000 nmol/L range.20,21 Again, the control group displayed
consistently normal levels of ATP and ADP (Table 1). Both ATP and ADP were clearly higher in the patient
cohort than in the controls (P < 0.0001). There were no significant differences between genders amongst
patients (data not shown). CD39 was rather normally distributed, but for CD73, ATP and ADP some
patients had extremely high values, which skewed the distribution; thus log-transformed values were used
to calculate statistical significance.
23
For further insight and analysis, the patient cohort was divided into previously derived age groups
(<60, 60–69, 70–79, and >80 years) and into distinct clinically relevant subgroups of PAD depending on
the localization of the disease (aorta-iliac, femoropopliteal, crural, or pedal). More proximal disease
localizations tended to associate with higher ATP and ADP levels, and higher CD39 activity (see Figure
2A). However, despite this prominent trend, using log-transformed values and linear regression only CD39
showed a declining trend towards distal PAD (P = 0.017). All PAD localization categories significantly
differed from the control group for ATP and ADP, but not for CD39 (Figure 2A). CD73 activity did not
DOI: 10.1161/CIRCRESAHA.116.305715
5
show any meaningful trend depending on PAD localization, but aorta-iliac (P = 0.0002), femoropopliteal
(P < 0.0001) and crural (P = 0.042) groups significantly differed when compared to healthy controls (data
not shown).
For patients of a younger age, ATP/ADP levels and CD39 activity tended to be higher than those
of older age groups (see Figure 2B). However, only CD39 showed declining linear trend towards older age
in patients (P = 0.055) and in controls (P < 0.0001). Different age groups were compared directly against
their age-matched controls, and all showed clear statistically significant differences (Figure 2B). CD73
activity did not show any meaningful trend across age groups, and only patients in age group 70–79 showed
a significant increase (P = 0.003) when compared to its corresponding control group (data not shown).
Correlation of markers of purinergic signaling with dyslipidemia, hypertension, and hyperglycemia.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Since cardiovascular risk factors are known to associate with different PAD localizations, it was
our assumption that certain risk factors could affect purinergic signaling and lead to an atherosclerotic
disease. Thus, we tested potential correlations of cholesterol levels, triglycerides, systolic blood pressure,
glycemia, and creatinine levels against markers of purinergic signaling and found that several known risk
factor related elements positively correlated with ATP and ADP concentrations in PAD patients, but not in
controls without clinically evident PAD.
The most pronounced results of these tests amongst patients are presented in Figure 3. Systolic
blood pressure (P = 0.043, Figure 3A) and triglycerides (P = 0.048, not shown) had a statistically significant
positive correlation with ADP concentration. Total cholesterol had a strong positive correlation with ATP
concentration (P = 0.013, Figure 3C) and CD39 activity (P = 0.003, Figure 3B). Also triglycerides had a
strong positive correlation with ATP (P = 0.005, not shown). To illustrate the effect of hyperglycemia, the
GHbA1c value was used because it reflects a more constant elevation in blood glucose levels than a single
fasting glucose value. ATP and ADP levels correlated positively with increased glucose levels as a measure
of GHbA1c, but did not reach statistical significance (Figure 3D). Creatinine levels did not positively
correlate with ATP or ADP in the patient cohort. We also performed the same testing with the control
group. Systolic blood pressure, LDL cholesterol, triglycerides, and creatinine were not significantly
correlated with markers of purinergic signaling (data not shown). Only total cholesterol had a slight positive
correlation with ATP (P = 0.026).
Active smoking is associated with high ATP and ADP levels and increased CD39 activity in PAD patients.
To examine the effect of smoking on purinergic signaling as a cardiovascular risk factor, we formed
subgroups from PAD patients with only one distinct cardiovascular risk factor: smoking, hypertension,
dyslipidemia, or diabetes (see Table IV in the Online Supplement for subgroup characteristics). Especially
ADP tended to be higher in PAD patients with only smoking as a cardiovascular risk factor when compared
to other PAD subgroups with one distinct risk factor (see Table 2A). Smoking did not elevate ATP and
ADP levels in control subjects without clinically evident PAD (data not shown). ATP (P < 0.0001) and
ADP (P < 0.0001) levels were clearly higher and CD39 to some extent higher (P = 0.043) in smoking PAD
patients when compared to smoking controls (Table 2B). Similarly, we also tested PAD patients who had
only dyslipidemia as a risk factor against control subjects with dyslipidemia. Control subjects with clear
dyslipidemia had significantly lower ATP (P = 0.015) and ADP (P = 0.0475) when compared to PAD
patients with only dyslipidemia as a risk factor (Table 2C).
The entire patient cohort was further tested for the effect of smoking on all markers of purinergic
signaling. CD39 activity steadily increased with the degree of smoking: patients who had never smoked
(mean, 16.4; SD, 6.4 nmol/mL/hr), had quit smoking (mean, 17.9; SD, 5.5 nmol/mL/hr) or were active
smokers (mean, 20.3; SD, 7.7 nmol/mL/hr). Active smokers had significantly higher CD39 activity than
DOI: 10.1161/CIRCRESAHA.116.305715
6
those who had smoked but quit (P = 0.041) or those who had never smoked (P = 0.001), using the Student’s
t-test.
Low CD39 activity is associated with disease progression.
CD39 activity was significantly lower in patients with critical ischemia (n = 118; mean, 17.3; SD,
6.5 nmol/mL/hr) than in patients with claudication (n = 97; mean, 19.3; SD, 6.9 nmol/mL/hr) using
Student’s t-test (P = 0.025). This suggests that higher CD39 activity could protect from disease progression
to critical ischemia. However, this could also be associated with an increasing severity of ischemia because
CD39 activity steadily decreased from Rutherford values of 1 to 6 (i.e., from mild claudication to major
tissue loss) giving a correlation (Spearman) of ˗0.174 (P = 0.01). Other markers did not show any
correlations.
Similarly, patients with severe coronary artery disease (CAD) (mean, 17.7; SD, 7.6 nmol/mL/hr)
had lower CD39 than patients with mild CAD (mean, 19.1; SD, 6.3 nmol/mL/hr), but this was not
statistically significant (P = 0.264).
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Explorative statistical modeling of multiple background variables related to high or low levels of ATP and
ADP, or CD39 and CD73 activity.
In accordance with the previously presented trends of increased ATP and ADP levels, and CD39
and CD73 activity in the patient cohort, explorative statistical modeling of multiple background variables
was performed (by linear regression) to gain more insight on which variables affect marker levels. For this,
only the patient cohort was used, with log-transformed values for purinergic markers. All of the statistically
significant findings are summarized in Table 3.
The first column, “Full combined model”, indicates all variables, which showed some significance
in their respective group analysis (demographic factors, cardiovascular risk factors, and medication) and
were used in the final combined reduction model. The second column, “Reduced combined model”, shows
variables that remained significant. For ATP, only high P-AFOS levels and the use of angiotensin receptor
blockers (ARBs) were associated with low levels of ATP. Similarly for ADP, high P-AFOS was associated
with low values, but the relationship was not as strong as that of ATP. For ADP, younger age and
hypertension were associated with high values, and the use of statin medication was associated with low
values. Despite this finding, no association between the use of statins and CD39 activity was detected. No
associations between CD39 and medications were detected at all. For CD39 only smoking and uremia were
significant in the final combined model. Out of all of the markers, CD73 had the strongest and most diverse
associations with different factors. The most pronounced phenomenon was the association of high CD73
activity with chronic hypoxia, i.e., severe COPD, but not to milder COPD or smoking. In addition, higher
P-AFOS levels strongly associated with increased CD73 activity, as already noted in Figure 1. Surprisingly,
the use of warfarin was associated with increased CD73 activity, as was female gender. When medications
alone were tested, the use of statins had the strongest association with increased CD73 activity (data not
shown), but this effect was to some extent diminished in the final combined model, in which warfarin
exhibited a stronger association.
DOI: 10.1161/CIRCRESAHA.116.305715
7
DISCUSSION
We hypothesized that malfunctions of CD73-derived extracellular adenosine production would
contribute to atherosclerosis and that this effect would be observed in our PAD patient cohort. Instead,
throughout our patient cohort, we observed high CD73 activity, and to some extent CD39 activity as well.
Elevated CD73 was especially associated with hypoxemic atherosclerotic disease, while elevated CD39
was associated with smoking and ATP/ADP-dependent entities. All markers of purinergic signaling were
normal in the controls, and values of ATP, ADP, and CD73 were significantly higher in atherosclerotic
patients. CD39 activity was not significantly increased in the patient cohort as a result of the high CD39
activity in the youngest control population (controls <40 years, Figure 2B). The youngest control group
with the highest CD39 activity also had the lowest ATP and ADP levels (see Figure 2B).
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Our findings are in line with previous studies of the role of CD39 and ATP/ADP interplay in
thrombosis and vascular inflammation.24,25 It is also well-documented from lung tissue and bronchial fluid
samples that purinergic markers, especially CD73, are elevated in the presence of COPD and hypoxia,26,27
and in in vivo animal studies demonstrating the effects of CD73 in hypoxia transduction.17 In the present
study, the same effects were observed in blood samples from PAD patients. However, CD73 activity was
clearly more associated with a chronically hypoxic state than CD39 activity, and CD39 activity had a
stronger relationship with active smoking, which could be interpreted as intermittent hypoxia. Both CD73
and CD39 expression and activity can be driven by hypoxia,28 but as seen here and before,29 CD39 is
somewhat unaffected by chronic hypoxia, but instead intermittent hypoxia can have an effect.
However, the increased ATP and ADP levels in systemic circulation of the studied patients may
reflect diminished endothelial bound ecto-nucleotidase levels in the vascular wall. Similar experimental
results were reported previously both in young atherosclerosis-prone apolipoprotein E-deficient mice26 and
in chronically hypoxic bovine vasa vasorum endothelial cells.29 A similar underlying pathological
mechanism may occur in the atherosclerotic hypoxic vessels observed in the patient cohort of this study.
Lost ecto-nucleotidases from the vascular wall could be reflected in the high circulating levels observed
herein.
Notably, NTPDase1/CD39 and other nucleotidases may circulate in the bloodstream either as "true"
soluble enzymes16,21 or in the form of microparticle-embedded enzymes30. Similar patterns of nucleotide
metabolism were observed in human serum and heparinized plasma, and also after additional
ultracentrifugation of serum samples for 1 hr at 100,000 g21 (and our unpublished data). These data exclude
the potential release of membrane-bound nucleotidases from the blood elements in the course of serum
preparation and provide further evidence for the predominant contribution of soluble rather than
microparticle-associated enzymes to the measured NTPDase/ADPase and ecto-5'-nucleotidase/CD73
activities.
The results of this study indicate that in a general population of patients with PAD, it is not the
impairment of CD39 and CD73 that contributes to atherosclerosis, but it is the burden of circulating ATP
and ADP that plays a more determinant role in the pathogenesis of the disease. This mechanism is especially
notable in young patients. Old age, which is a known major independent risk factor for PAD, does not seem
to play a significant role within this mechanism. Alternatively, age-related impairments (e.g., of CD39),
even in the absence of increased levels of ATP and ADP, could be a contributing factor. It is a well-known
clinical observation amongst practitioners of vascular surgery, as also demonstrated by Diehm et al.,31 that
smoking is associated with proximal PAD in young patients, while old age is often associated with a distal
disease. Now, for the first time, our findings demonstrate that this clinical phenomenon is connected to
changes at a molecular level.
DOI: 10.1161/CIRCRESAHA.116.305715
8
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
All atherosclerotic patients in this cohort displayed relatively high levels of ATP and ADP when
compared to those of the controls and known normal levels derived from past experiments.20,21 Distinct
cardiovascular risk factors were shown to drive the increase in ATP and ADP levels amongst subjects with
a clinically significant atherosclerotic disease, but not amongst the controls. However, in the multivariable
modeling, high ATP levels could not be clearly appointed to any specific cardiovascular risk factor, and
ADP could only be appointed to hypertension. ATP and ADP appeared to be especially high in proximal
PAD and younger age, but to some extent this phenomenon lacked strength after log-transformation. This
observation could, however, strongly associate with smoking. ADP was clearly higher in patients who had
only smoking as a cardiovascular risk factor when compared to smoking controls and other risk factorbased subgroups. PAD patients with only diabetes as a risk factor, on the other hand, had the lowest ATP
and ADP levels and amongst the cardiovascular risk factor-related laboratory values, hyperglycemia had
the weakest correlation with ATP and ADP. An explanation to this weak association could be that, in
practice, diabetes is mostly associated with distal PAD, while smoking, dyslipidemia, and hypertension are
associated with proximal PAD.31 Thus, ATP and ADP could be a common denominator of molecular
pathophysiology amongst several cardiovascular risk factors, although smoking may be the strongest driver
for both of these factors, especially ADP.
In addition to high levels of ATP and ADP, smoking
patients exhibited significant CD39 activity. This could indicate that CD39 activity is required to clear the
elevated levels of ATP and ADP caused by smoking. It would be logical to assume that without an elevated
level of CD39 activity, ATP and ADP levels would be even higher in smoking patients. Lower CD39
activity could contribute to disease progression amongst PAD patients. Such an effect was observed when
comparing patients with claudication to those with critical ischemia. Lower CD39 activity was observed in
patients with critical ischemia, and CD39 activity decreased steadily as the disease progressed according to
the Rutherford classification. In previous studies, soluble CD39 exhibited significant antithrombotic
effects.15,16,32
Building on clinical knowledge and past findings, it is known that smoking has a highly damaging
effect on lower extremity vascular bypass grafts. Smoking independently increases the risk of graft failure
by up to 4-fold.33 It was recently shown that smokers with atherosclerotic diseases benefit more from
antiplatelet therapy targeting the P2Y12-receptor than non-smoking atherosclerotic patients. However, the
mechanism remains unknown.34-36 The effect of smoking on cytochrome P450 (CYP)1A2, which converts
clopidogrel into its active metabolite, was provided as a rationale,37 but this does not explain a similar
smoker’s paradox observed with prasugrel and ticagrelor, which have different mechanisms of action.38 On
the basis of our findings, we suggest that the beneficial effect is specifically due to the inhibition of ADP
via the P2Y12 receptor expressed on platelets because circulating ADP levels are elevated in smokers.
Moreover, statins are known to have anti-inflammatory and antithrombotic effects. A suggested
mechanism of action could be an improvement in CD39 activity.39 In the present study, the use of statins
was not associated with CD39 activity, but the use of statins was clearly associated with lower ADP levels.
We already observed a similar phenomenon in a prospective setting in which young diabetic patients
exhibited low ADP values after statin therapy, but no change in CD39 activity.40 As Kaneider et al. showed
in an in vitro study, statins restored impaired CD39 function.39 They also did not report increased levels of
expression. Thus, it could be that the beneficial effects of statins are directed to CD39 on the vascular wall
and are only observed as changes in nucleotide, but not NTPDase, levels when measuring circulating
values. However, as noted in the present study, and as shown earlier, statins improve CD73 expression,
although the effect is only transient.40,41 An explanation could be that statins inhibit Rho-GTPase-dependent
endocytosis of ecto-nucleotidases41 but lack the ability to induce new protein synthesis.
The use of antihypertensive ARBs, however, was clearly associated with low ATP values amongst
the patient cohort. This could to some extent explain the beneficial effects in cardiovascular events
associated with the use of ARBs.42,43 However, on the basis of our results, this is currently an entirely
theoretical interpretation.
DOI: 10.1161/CIRCRESAHA.116.305715
9
In conclusion, we showed, for the first time, changes in intravascular nucleotide turnover in
different subgroups of patients suffering from PAD and related cardiovascular risk factors. We believe that
our findings indicate an important role for purinergic signaling during the atherosclerotic process,
particularly contributing to atherosclerotic diseases associated with a younger age, tobacco smoking, and
thrombosis.
ACKNOWLEDGMENTS
We thank Dr. Jan-Erik Wickström for helping to recruit patients, Tommi Pesonen, M.Sc., for professional
assistance and guidance in statistical analyses, and Sari Mäki, Anne Meyer, and Teija Kanasuo for technical
assistance.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
SOURCES OF FUNDING
The study was supported by the Academy of Finland, the Sigrid Juselius Foundation, and the Clinical
Research Fund (EVO) of Turku University Hospital.
DISCLOSURES
The authors have nothing to disclose.
REFERENCES
1.
Weber C, Noels H. Atherosclerosis: current pathogenesis and therapeutic options. Nat Med.
2011;17:1410–1422. doi:10.1038/nm.2538.
2.
Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med.
2005;352:1685–1695. doi:10.1056/NEJMra043430.
3.
Libby P. Inflammation in atherosclerosis. Arterioscler Thromb Vasc Biol. 2012;32:2045–2051.
doi:10.1161/ATVBAHA.108.179705.
4.
Colgan SP, Eltzschig HK, Eckle T, Thompson LF. Physiological roles for ecto-5'-nucleotidase
(CD73). Purinergic Signal. 2006;2:351–360. doi:10.1007/s11302-005-5302-5.
5.
Grünewald JK, Ridley AJ. CD73 represses pro-inflammatory responses in human endothelial
cells. J Inflamm (Lond). 2010;7:10. doi:10.1186/1476-9255-7-10.
6.
Antonioli L, Pacher P, Vizi ES, Haskó G. CD39 and CD73 in immunity and inflammation. Trends
Mol Med. 2013;19:355–367. doi:10.1016/j.molmed.2013.03.005.
7.
Eltzschig HK, Sitkovsky MV, Robson SC. Purinergic signaling during inflammation. N Engl J
Med. 2012;367:2322–2333. doi:10.1056/NEJMra1205750.
8.
Zernecke A, Bidzhekov K, Ozüyaman B, Fraemohs L, Liehn EA, Lüscher-Firzlaff JM, Lüscher B,
Schrader J, Weber C. CD73/ecto-5'-nucleotidase protects against vascular inflammation and
neointima formation. Circulation. 2006;113:2120–2127.
doi:10.1161/CIRCULATIONAHA.105.595249.
DOI: 10.1161/CIRCRESAHA.116.305715
10
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
9.
Buchheiser A, Ebner A, Burghoff S, Ding Z, Romio M, Viethen C, Lindecke A, Köhrer K, Fischer
JW, Schrader J. Inactivation of CD73 promotes atherogenesis in apolipoprotein E-deficient mice.
Cardiovasc Res. 2011;92:338–347. doi:10.1093/cvr/cvr218.
10.
Lecka J, Bloch-Boguslawska E, Molski S, Komoszynski M. Extracellular purine metabolism in
blood vessels (Part II): Activity of ecto-enzymes in blood vessels of patients with abdominal aortic
aneurysm. Clin Appl Thromb Hemost. 2010;16:650–657. doi:10.1177/1076029609354329.
11.
St Hilaire C, Ziegler SG, Markello TC, Brusco A, Groden C, Gill F, Carlson-Donohoe H,
Lederman RJ, Chen MY, Yang D, Siegenthaler MP, Arduino C, Mancini C, Freudenthal B,
Stanescu HC, Zdebik AA, Chaganti RK, Nussbaum RL, Kleta R, Gahl WA, Boehm M. NT5E
mutations and arterial calcifications. N Engl J Med. 2011;364:432–442.
doi:10.1056/NEJMoa0912923.
12.
Gordon JL. Extracellular ATP: effects, sources and fate. Biochem J. 1986;233:309–319.
13.
Zimmermann H, Zebisch M, Sträter N. Cellular function and molecular structure of ectonucleotidases. Purinergic Signal. 2012;8:437–502. doi:10.1007/s11302-012-9309-4.
14.
Robson SC, Wu Y, Sun X, Knosalla C, Dwyer K, Enjyoji K. Ectonucleotidases of CD39 family
modulate vascular inflammation and thrombosis in transplantation. Semin Thromb Hemost.
2005;31:217–233. doi:10.1055/s-2005-869527.
15.
Huttinger ZM, Milks MW, Nickoli MS, Aurand WL, Long LC, Wheeler DG, Dwyer KM, d'Apice
AJF, Robson SC, Cowan PJ, Gumina RJ. Ectonucleotide triphosphate diphosphohydrolase-1
(CD39) mediates resistance to occlusive arterial thrombus formation after vascular injury in mice.
Am J Pathol. 2012;181:322–333. doi:10.1016/j.ajpath.2012.03.024.
16.
Yegutkin GG, Wieringa B, Robson SC, Jalkanen S. Metabolism of circulating ADP in the
bloodstream is mediated via integrated actions of soluble adenylate kinase-1 and NTPDase1/CD39
activities. FASEB J. 2012;26:3875–3883. doi:10.1096/fj.12-205658.
17.
Thompson LF, Eltzschig HK, Ibla JC, Van De Wiele CJ, Resta R, Morote-Garcia JC, Colgan SP.
Crucial role for ecto-5'-nucleotidase (CD73) in vascular leakage during hypoxia. J Exp Med.
2004;200:1395–1405. doi:10.1084/jem.20040915.
18.
Thompson LF, Takedachi M, Ebisuno Y, Tanaka T, Miyasaka M, Mills JH, Bynoe MS.
Regulation of leukocyte migration across endothelial barriers by ECTO-5'-nucleotidase-generated
adenosine. Nucleosides Nucleotides Nucleic Acids. 2008;27:755–760.
doi:10.1080/15257770802145678.
19.
Vartiainen E, Jousilahti P, Alfthan G, Sundvall J, Pietinen P, Puska P. Cardiovascular risk factor
changes in Finland, 1972-1997. Int J Epidemiol. 2000;29:49–56.
20.
Helenius M, Jalkanen S, Yegutkin G. Enzyme-coupled assays for simultaneous detection of
nanomolar ATP, ADP, AMP, adenosine, inosine and pyrophosphate concentrations in
extracellular fluids. Biochim Biophys Acta. 2012;1823:1967–1975.
doi:10.1016/j.bbamcr.2012.08.001.
21.
Yegutkin GG, Samburski SS, Mortensen SP, Jalkanen S, González-Alonso J. Intravascular ADP
and soluble nucleotidases contribute to acute prothrombotic state during vigorous exercise in
DOI: 10.1161/CIRCRESAHA.116.305715
11
humans. J Physiol (Lond). 2007;579:553–564. doi:10.1113/jphysiol.2006.119453.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
22.
Niemelä J, Ifergan I, Yegutkin GG, Jalkanen S, Prat A, Airas L. IFN‐β regulates CD73 and
adenosine expression at the blood–brain barrier. Eur J Immunol. 2008;38:2718–2726.
doi:10.1002/eji.200838437.
23.
Bellingan G, Maksimow M, Howell DC, Stotz M, Beale R, Beatty M, Walsh T, Binning A,
Davidson A, Kuper M, Shah S, Cooper J, Waris M, Yegutkin GG, Jalkanen J, Salmi M, Piippo I,
Jalkanen M, Montgomery H, Jalkanen S. The effect of intravenous interferon-beta-1a (FP-1201)
on lung CD73 expression and on acute respiratory distress syndrome mortality: an open-label
study. The Lancet Respiratory Medicine. 2014;2:98–107. doi:10.1016/S2213-2600(13)70259-5.
24.
Mercier N, Kiviniemi TO, Saraste A, Miiluniemi M, Silvola J, Jalkanen S, Yegutkin GG. Impaired
ATP-induced coronary blood flow and diminished aortic NTPDase activity precede lesion
formation in apolipoprotein E-deficient mice. Am J Pathol. 2012;180:419–428.
doi:10.1016/j.ajpath.2011.10.002.
25.
Burnstock G, Ralevic V. Purinergic signaling and blood vessels in health and disease. Pharmacol
Rev. 2014;66:102–192. doi:10.1124/pr.113.008029.
26.
Zhou Y, Murthy JN, Zeng D, Belardinelli L, Blackburn MR. Alterations in adenosine metabolism
and signaling in patients with chronic obstructive pulmonary disease and idiopathic pulmonary
fibrosis. PLoS ONE. 2010;5:e9224. doi:10.1371/journal.pone.0009224.
27.
Esther CR, Lazaar AL, Bordonali E, Qaqish B, Boucher RC. Elevated airway purines in COPD.
Chest. 2011;140:954–960. doi:10.1378/chest.10-2471.
28.
Eltzschig HK, Ibla JC, Furuta GT, Leonard MO, Jacobson KA, Enjyoji K, Robson SC, Colgan SP.
Coordinated adenine nucleotide phosphohydrolysis and nucleoside signaling in posthypoxic
endothelium: role of ectonucleotidases and adenosine A2B receptors. J Exp Med. 2003;198:783–
796. doi:10.1084/jem.20030891.
29.
Yegutkin GG, Helenius M, Kaczmarek E, Burns N, Jalkanen S, Stenmark K, Gerasimovskaya EV.
Chronic hypoxia impairs extracellular nucleotide metabolism and barrier function in pulmonary
artery vasa vasorum endothelial cells. Angiogenesis. 2011;14:503–513. doi:10.1007/s10456-0119234-0.
30.
Banz Y, Beldi G, Wu Y, Atkinson B, Usheva A, Robson SC. CD39 is incorporated into plasma
microparticles where it maintains functional properties and impacts endothelial activation. Br J
Haematol. 2008;142:627–637. doi:10.1111/j.1365-2141.2008.07230.x.
31.
Diehm N, Shang A, Silvestro A, Do D-D, Dick F, Schmidli J, Mahler F, Baumgartner I.
Association of cardiovascular risk factors with pattern of lower limb atherosclerosis in 2659
patients undergoing angioplasty. Eur J Vasc Endovasc Surg. 2006;31:59–63.
doi:10.1016/j.ejvs.2005.09.006.
32.
Buergler JM, Maliszewski CR, Broekman MJ, Kaluza GL, Schulz DG, Marcus AJ, Raizner AE,
Kleiman NS, Ali NM. Effects of SolCD39, a novel inhibitor of Platelet Aggregation, on Platelet
Deposition and Aggregation after PTCA in a Porcine Model. J Thromb Thrombolysis.
2005;19:115–122. doi:10.1007/s11239-005-1381-y.
DOI: 10.1161/CIRCRESAHA.116.305715
12
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
33.
Willigendael EM, Teijink JAW, Bartelink M-L, Peters RJG, Büller HR, Prins MH. Smoking and
the patency of lower extremity bypass grafts: a meta-analysis. J Vasc Surg. 2005;42:67–74.
doi:10.1016/j.jvs.2005.03.024.
34.
Dong A, Caicedo J, Han SG, Mueller P, Saha S, Smyth SS, Gairola CG. Enhanced platelet
reactivity and thrombosis in Apoe-/- mice exposed to cigarette smoke is attenuated by P2Y12
antagonism. Thromb Res. 2010;126:e312–7. doi:10.1016/j.thromres.2010.03.010.
35.
Rollini F, Franchi F, Cho JR, Degroat C, Bhatti M, Ferrante E, Patel R, Darlington A, TelloMontoliu A, Desai B, Ferreiro J, Muniz-Lozano A, Zenni MM, Guzman LA, Bass TA, Angiolillo
DJ. Cigarette smoking and antiplatelet effects of aspirin monotherapy versus clopidogrel
monotherapy in patients with atherosclerotic disease: results of a prospective pharmacodynamic
study. J Cardiovasc Transl Res. 2014;7:53–63. doi:10.1007/s12265-013-9535-3.
36.
Gurbel PA, Bliden KP, Logan DK, Kereiakes DJ, Lasseter KC, White A, Angiolillo DJ, Nolin TD,
Maa J-F, Bailey WL, Jakubowski JA, Ojeh CK, Jeong Y-H, Tantry US, Baker BA. The influence
of smoking status on the pharmacokinetics and pharmacodynamics of clopidogrel and prasugrel:
the PARADOX study. J Am Coll Cardiol. 2013;62:505–512. doi:10.1016/j.jacc.2013.03.037.
37.
Desai NR, Mega JL, Jiang S, Cannon CP, Sabatine MS. Interaction between cigarette smoking and
clinical benefit of clopidogrel. J Am Coll Cardiol. 2009;53:1273–1278.
doi:10.1016/j.jacc.2008.12.044.
38.
Gagne JJ, Bykov K, Choudhry NK, Toomey TJ, Connolly JG, Avorn J. Effect of smoking on
comparative efficacy of antiplatelet agents: systematic review, meta-analysis, and indirect
comparison. BMJ. 2013;347:f5307.
39.
Kaneider NC, Egger P, Dunzendorfer S, Noris P, Balduini CL, Gritti D, Ricevuti G, Wiedermann
CJ. Reversal of thrombin-induced deactivation of CD39/ATPDase in endothelial cells by HMGCoA reductase inhibition: effects on Rho-GTPase and adenosine nucleotide metabolism.
Arterioscler Thromb Vasc Biol. 2002;22:894–900.
40.
Kiviniemi TO, Yegutkin GG, Toikka JO, Paul S, Aittokallio T, Janatuinen T, Knuuti J, Rönnemaa
T, Koskenvuo JW, Hartiala JJ, Jalkanen S, Raitakari OT. Pravastatin-induced improvement in
coronary reactivity and circulating ATP and ADP levels in young adults with type 1 diabetes.
Front Physiol. 2012;3:338. doi:10.3389/fphys.2012.00338.
41.
Ledoux S. Lovastatin Enhances Ecto-5'-Nucleotidase Activity and Cell Surface Expression in
Endothelial Cells: Implication of Rho-Family GTPases. Circ Res. 2002;90:420–427.
doi:10.1161/hh0402.105668.
42.
Molnar MZ, Kalantar-Zadeh K, Lott EH, Lu JL, Malakauskas SM, Ma JZ, Quarles DL, Kovesdy
CP. Angiotensin-converting enzyme inhibitor, angiotensin receptor blocker use, and mortality in
patients with chronic kidney disease. J Am Coll Cardiol. 2014;63:650–658.
doi:10.1016/j.jacc.2013.10.050.
43.
Smink PA, Miao Y, Eijkemans MJC, Bakker SJL, Raz I, Parving H-H, Hoekman J, Grobbee DE,
de Zeeuw D, Lambers Heerspink HJ. The importance of short-term off-target effects in estimating
the long-term renal and cardiovascular protection of angiotensin receptor blockers. Clin
Pharmacol Ther. 2014;95:208–215. doi:10.1038/clpt.2013.191.
DOI: 10.1161/CIRCRESAHA.116.305715
13
TABLE 1. Differences in Purinergic Signaling Markers: Patient Cohort vs. Controls
Patients
Variable
N
CD39
220 18.01
CD73
Controls
Mean SE
Median
N
0.45
17.0
220 253
10.8
ATP
201 5165
ADP
190 4654
Mean SE
Median
P-value*
123 17.63 0.82
15.0
0.09
212.0
123 195
9.36
182
<0.0001
256
4088
153 1805
137
1104
<0.0001
288
3627
151 886
85
508
<0.0001
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
CD39 and CD73 are expressed as enzymatic activity in nmol/mL/hr. ATP and ADP are expressed as
nmol/L. *Student’s t-test for log-transformed values.
DOI: 10.1161/CIRCRESAHA.116.305715
14
TABLE 2. Markers of Purinergic Signaling in Subgroups of Patients with Only One
Cardiovascular Risk Factor and Comparison to the Corresponding Group of Controls
ATP (nmol/L)
A
ADP (nmol/L)
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Patient subgroup
Mean
SE
Mean
SE
Smoking (N = 17)
6615
1107
6686
1220
Hypertension (N = 12)
6648
1350
5190
1249
Dyslipidemia (N = 9)
5718
1059
4869
1553
Diabetes (N = 8)
5109
844
4144
971
B
Patients: smoking
Variable
N
Mean
SE
N
Mean
SE
P-value
ATP (nmol/L)
16
6615
1107
29
2044
335
<0.0001*
ADP (nmol/L)
15
6686
1220
29
846
122
<0.0001*
CD39 (nmol/mL/hr) 17
18.9
1.9
16
13
1.4
0.043**
CD73 (nmol/mL/hr) 17
229
31
16
191
27
0.103*
Controls: smoking
C
Patients with dyslipidemia Controls with
Variable
N
Mean
SE
N
Mean SE
P-value
ATP (nmol/L)
9
5718
1059
14
2404
378
0.014**
ADP (nmol/L)
8
4869
1553
14
1133
242
0.048**
CD39 (nmol/mL/hr) 9
18.7
1.9
12
16.3
1.6
0.343**
CD73 (nmol/mL/hr) 9
190
25
12
238
34
0.759*
dyslipidemia
* Student’s t-test for log-transformed values
** Student’s t-test for normally distributed values
DOI: 10.1161/CIRCRESAHA.116.305715
15
TABLE 3. Results of the Linear Regression Model on Factors Affecting Markers of Purinergic
Signaling Using the Patient Cohort
Full combined model
ATP
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
ADP
CD39
CD73
Proximal PAD 0.79 (0.20; SE, 0.23)
Male gender 0.44 (˗0.08; SE, 0.11)
Young age 0.32 (0.36; SE, 0.19)
Statins 0.21 (˗0.13; SE, 0.11)
ARBs 0.032 (˗0.29; SE, 0.13)
P-AFOS 0.031 (˗0.006; SE, 0.003)
Proximal PAD 0.92 (0.19; SE, 0.32)
Male gender 0.20 (˗0.19; SE, 0.15)
Young age 0.038 (0.64; SE, 0.29)
Hypertension 0.023 (0.40; SE, 0.18)
ACE blockers 0.32 (0.15; SE, 0.15)
Gliptins 0.21 (0.29; SE, 0.23)
Statins 0.018 (˗0.36; SE, 0.15)
P-AFOS 0.031 (˗0.006; SE, 0.003)
Proximal PAD 0.44 (0.14; SE, 0.12)
Male gender 0.57 (˗0.03; SE, 0.05)
Young age 0.97 (0.03; SE, 0.10)
Clopidogrel 0.13 (0.14; SE, 0.10)
Cortisone 0.25 (˗0.08; SE, 0.07)
Smoking 0.13 (0.12; SE, 0.08)
COPD 0.34 (0.07; SE, 0.08)
Hypoxia 0.37 (0.11; SE, 0.12)
Uremia 0.03 (0.26; SE, 0.12)
P-AFOS 0.17 (0.002; SE, 0.001)
Male gender 0.04 (˗0.17; SE, 0.08)
Statins 0.05 (0.16; SE, 0.08)
Nitroglycerin 0.18 (˗0.14; SE, 0.10)
Warfarin 0.35 (0.20; SE, 0.09)
Hypoxia 0.002 (0.55; SE, 0.17)
P-AFOS 0.0001 (0.007; SE, 0.002)
Reduced combined model
ARBs 0.015 (˗0.317; SE, 0.129)
P-AFOS 0.027 (˗0.006; SE, 0.003)
Young age 0.042 (0.62, SE, 0.28)
Hypertension 0.007 (0.47; SE, 0.17)
Statins 0.041 (˗0.03; SE, 0.15)
P-AFOS 0.07 (˗0.007; SE, 0.004)
Smoking 0.045 (0.17; SE, 0.07)
Uremia 0.023 (0.26; SE, 0.12)
Male gender 0.03 (˗0.17; SE, 0.08)
Hypoxia 0.001 (0.56; SE, 0.17)
Warfarin 0.046 (0.19; SE, 0.09)
Statins 0.054 (0.15; SE, 0.08)
P-AFOS 0.0001 (0.007; SE, 0.002)
The subsequent numerical value represents P-values with beta coefficients and standard errors in brackets
for log-transformed variables.
DOI: 10.1161/CIRCRESAHA.116.305715
16
FIGURE LEGENDS
Figure 1. Correlations of P-AFOS and markers of purinergic signaling. (A) CD39, (B) CD73, (C) ATP,
and (D) ADP. P-AFOS is expressed as U/L, CD39 and CD73 as nmol/mL/hr, and ATP and ADP as nmol/L.
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Figure 2. ATP, ADP, and CD39 values vary according to the localization of atherosclerotic lesions in
PAD (A) and by age group (B). A, ATP, ADP, and CD39 values were high in proximal PAD and decline
in distal PAD. The effect was strongest with CD39 (P = 0.017), calculated as a linear trend between groups
by linear regression (linear trends between groups for ATP and ADP were not statistically significant).
Bottom row: P-values represent a comparison to the control cohort using the Student’s t-test for logtransformed values. B, Similarly, ATP, ADP, and CD39 values were high in young age groups and steadily
declined in the older age groups. Amongst patients the linear trend was strongest again with CD39 (P =
0.055) but does not reach statistical significance with CD39, nor with ATP and ADP. Amongst controls
CD39 had a strong declining trend towards older age (P < 0.0001). Age groups are also compared directly
with their age-matched controls using the Student’s t-test for log-transformed values (bottom row). CD39
is expressed as nmol/mL/hr, and ATP and ADP as nmol/L.
Figure 3. Correlation of purinergic signaling markers with measurable numeric values relating to
hypertension, hypercholesterolemia, and hyperglycemia. A, ADP positively correlates with systolic
blood pressure. B and C, CD39 and ATP positively correlate with cholesterol levels. D, ATP tends to
positively correlate with blood glucose levels measured by GHbA1c, but this did not reach statistical
significance. CD39 is expressed as nmol/mL/hr, and ATP and ADP as nmol/L.
DOI: 10.1161/CIRCRESAHA.116.305715
17
Novelty and Significance
What Is Known?

Purinergic signaling plays an important role in inflammation and vascular wall integrity.

Impairment of the signaling cascade may lead to atherosclerosis.

Smoking is the leading cause of lower extremity vascular bypass graft failure.
What New Information Does This Article Contribute?
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017

Purinergic signaling, i.e., ATP metabolism, plays an important role in the development of
peripheral artery disease at a young age and affects large arteries (aorta-iliac or femoropopliteal
arteries).

Smoking is associated with elevated ATP/ADP levels and disease burden.

Cardiovascular patients who continue to smoke could benefit from permanent ADP/P2Y12
antiplatelet therapy, especially if they have a vascular bypass.
In general, little is known about purinergic signaling in atherosclerosis, especially in peripheral artery
disease (PAD). In this study, we provide evidence that the pro-thrombotic effect of tobacco smoke is due
to elevated plasma ADP levels. Our findings provide evidence for a mechanism underlying the greater
beneficial effects of clopidogrel-like antiplatelet therapy in smokers than non-smoking patients. The use
of P2Y12 antiplatelet therapy for vascular bypass graft protection in PAD patients who smoke should be
tested in a prospective randomized study.
DOI: 10.1161/CIRCRESAHA.116.305715
18
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Downloaded from http://circres.ahajournals.org/ by guest on June 14, 2017
Aberrant Circulating Levels of Purinergic Signaling Markers are Associated with Several Key
Aspects of Peripheral Atherosclerosis and Thrombosis
Juho Jalkanen, Gennady G Yegutkin, Maija Hollmén, Kristiina Aalto, Tuomas O Kiviniemi, Veikko
Salomaa, Sirpa Jalkanen and Harri H Hakovirta
Circ Res. published online February 2, 2015;
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2015 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/early/2015/02/02/CIRCRESAHA.116.305715
Data Supplement (unedited) at:
http://circres.ahajournals.org/content/suppl/2015/02/02/CIRCRESAHA.116.305715.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/
ONLINE SUPPLEMENT
Methods
Patient Cohort
For the 226 subjects forming the patient cohort, all major demographic factors,
baseline characteristics, cardiovascular risk factors, and medications were recorded,
in addition to specific data on the severity and distribution of PAD. Disease severity
was classified according to the Rutherford classification and then converted to the
usual clinically relevant three categories: no symptoms (Rutherford 0), claudication
(Rutherford 1–3), and critical ischemia (Rutherford 4–6). Elective patients with minor
ulcers or dry peripheral gangrene of a small area, e.g., one to two toes, were
included in the study, but patients with infections or tissue loss leading to transmetatarsal amputation were not included.
Baseline characteristics and cardiovascular or risk factors were determined
on a yes/no basis, if recorded in the patient file. Type I and type II diabetic patients
were evaluated separately, but also together as a single variable of three categories:
no diabetes, diabetics on oral medication, and diabetic patients treated with insulin.
Patients with coronary artery disease (CAD) were also assigned in a variable of three
categories: i) no diagnosis or symptoms, ii) diagnosed and only mildly symptomatic
(no invasive treatment or history of acute myocardial infarction, AMI), and iii)
symptomatic (previous AMI or invasive treatment received). All autoimmune and
rheumatic diseases were reported (e.g., systemic lupus erythematosus, psoriasis,
inflammatory bowel disease, etc.), but the prevalence was so small that an overall
“rheumatic disease” variable was formed. Patients with dyslipidemia were either
already diagnosed and on statin therapy, or were newly diagnosed according to the
current national guidelines (based on cholesterol and triglyceride levels) and were
prescribed statin therapy during the index hospital stay.
Patients were divided into the following age groups (<60, 60–69, 70–79, and
>80 years) and categorized according to the distribution of disease burden in the
lower extremities (aorta-iliac, femoropopliteal crural, and pedal). Patients could
exhibit calcification at several levels, but a distinct category was assigned if one area
was proven to be clearly affected according to the Inter-Society Consensus on the
Management of Peripheral Artery Disease (TASC II) -classification and the practicing
vascular surgeon. Only one patient was excluded from categorization as a result of
several levels of significant atherosclerotic lesions.
Non-atherosclerotic Control Cohort
Sixty-four newly recruited volunteers were all screened by a practicing medical
doctor of the research group for symptoms and signs of PAD. A palpable pedal pulse
was used to determine whether subjects were without evidence of clinically evident
PAD. Eighty-nine subjects (73 males, 16 females) were selected from the FINRISK
1997 study based on age (35–74 years), health status, and residence in either
Southern Finland (Helsinki and Vantaa) or Eastern Finland (Province of North
Karelia). A member of the research group did not see these subjects. The subjects
had to be free of CHD, PAD, stroke, diabetes, cancer, COPD, and rheumatic
illnesses on the basis of both a self-report and register linkage and national health
care databases. These 89 subjects from FINRISK 1997 and the 64 new volunteers
without clinically evident PAD formed the control pool (n = 153) for the PURE ASO
patient cohort and were divided into age groups to match the patient cohort: <40 (n =
81), 40–59 (n = 16), 60–69 (n = 40), and 70–79 (n = 16) years. The oldest age group
(>80) could not be matched.
Blood Samples
Samples were drawn in the morning after at least 4 hr of fasting. Nine mL of whole
blood was drawn to a serum sample tube and 9 mL to an EDTA sample tube. The
serum sample tube was left to clot at room temperature while it was transported to
the Medicity Research Laboratory, University of Turku. On arrival, it was centrifuged
at 2000 g for 10 min, after which the serum was extracted for analysis. For EDTA
plasma samples, the accredited University Hospital Central Laboratory conducted
standard creatinine, lipid, glucose, and alkaline phosphatase (P-AFOS) analyses
according to a photometric International Federation of Clinical Chemistry and
Laboratory Medicine (IFCC) procedure (values expressed as international units per
liter, U/L), and the remaining plasma was transported to the Medicity Research
Laboratory for ATP and ADP analysis. Venous blood samples from the FINRISK
1997 cohort were drawn in 1997. All blood sample aliquots were stored at ˗70°C prior
to analysis.
Quantification of ATP and ADP Levels in Human Plasma
Briefly, 10 µL aliquots of EDTA plasma were transferred into two parallel wells of a
white non-phosphorescent 96-well microplate containing 100 µL of PBS with (A) or
without (B) a solution of 200 µM UTP and 5 U/mL NDP kinase from baker’s yeast S.
cerevisiae (Sigma). Following the addition of 50 µL of ATP-monitoring reagent,
sample luminescence was measured using a Tecan Infinite M200 microplate reader
(Salzburg, Austria). The difference in luminescence signals between well “A” (ATP +
ADP) and “B” (only ATP) enabled the quantification of ADP concentration, which was
converted into ATP through an NDP kinase-mediated reaction in the presence of
exogenous UTP. This approach allows simultaneous measurement of both ATP and
ADP content within the same sample.
Measurement of Soluble Nucleotidase Activities in Human Serum
For ADPase/NTPDase activity, serum (10 µL) was incubated for 60 min at 37°C in 80
µL of RPMI-1640 medium containing 5 mM β-glycerophosphate, 80 µM adenylate
kinase inhibitor Ap5A, and 50 µM ADP with a [2,8-3H]ADP tracer (Perkin Elmer,
Boston, USA). Likewise, 5’-nucleotidase activity was assayed by incubating 10 µL of
serum for 60 min with 300 µM [2-3H]AMP (Quotient Bioresearch, GE Healthcare,
Rushden, UK). Radiolabelled substrates and their dephosphorylated products were
separated by thin-layer chromatography and quantified by scintillation β-counting.
Enzymatic activities were expressed as nanomoles of 3H-substrate metabolized per
hr by 1 mL of serum.
Results
Description of the Patient Cohort
The patient cohort consisted of 226 participants of which 128 (56.6%) were male. All
were of Caucasian origin. Overall mean age was 69.8 years (SD ± 11.44, 46–93
years) with a small tendency for males to be younger (males, 67.9 years [SD ± 11.3],
and females, 72.3 years [SD ± 11.2]). Registered illnesses and cardiovascular risk
factors together with their prevalence are presented in Online Table I, along with
basic laboratory values in Online Table II, and PAD-related descriptive variables and
subgroup characteristics are presented in Online Table III.
High blood pressure was very common in the patient cohort, up to 75%
(Online Table I). The prevalence of rheumatic diseases was also somewhat high
(16%). From these 35 rheumatic patients, only one was diagnosed with vasculitis,
and all others had PAD accompanied by rheumatoid arthritis (20/35), psoriasis
(8/35), systemic lupus erythematosus (3/35), or inflammatory bowel disease (3/35).
Smoking, diabetes, and dyslipidemia were more prevalent in
young age groups, while hypertension, renal failure, and rheumatic diseases were
more abundant in older age groups. In a similar fashion, smoking and dyslipidemia
were more prevalent in proximal PAD, and renal failure, rheumatic diseases, but also
diabetes, were more prevalent in distal PAD. Amongst all cardiovascular risk factors,
renal insufficiency and diabetes were also more prevalent in critical ischemia than in
claudication (see Online Table III).
Statin therapy was the most commonly used medication upon index
hospitalization and blood sampling (64%). Following statins, the most commonly
used medications were as follows: aspirin (59%), beta-blockers (56%), angiotensin
converting enzyme blockers (44%), calcium channel blockers (31%), diuretic
medications (27%, mostly furosemide), warfarin (22%), nitroglycerin (20%),
metformin (19%), oral or inhaled cortisone (19%, topical use excluded), and
angiotensin receptor blockers (ARBs) (18%), gliptins (11%), bisphosphonates (7%).
Only 7% of patients were on clopidogrel at index hospitalization, and only 1% were
on dipyridamole.
Description of the Control Cohort
The non-atherosclerotic control cohort consisted of 153 subjects of which 70.6%
were male. The mean age was 45.9 years (SD ± 18.5, 20–74 years). The percentage
of active smokers amongst the controls was 19%, of which 76% were male. In the
laboratory values of control subjects, we identified three subjects with different
degrees of renal insufficiency (creatinine, 212, 220, and 433 µmol/L). All had normal
levels of purinergic signaling markers and remained in the study. We also identified
14 control subjects with clear dyslipidemia (total cholesterol >6.5 and LDL >3.5).
They remained in the study and performed as controls to PAD patients with only
dyslipidemia as a cardiovascular risk factor (see Online Table IV).
Online Table II presents laboratory values for the patient cohort and controls.
An interesting finding was that total cholesterol and LDL levels were statistically
higher in the controls than in the patient cohort reflecting the fact that the patient
cohort was under medication. Triglycerides, however, were higher in the patient
cohort, as was systolic blood pressure. No difference was observed with HDL
cholesterol or creatinine.
Division of Subjects to Subgroups with One Distinct Cardiovascular Risk
Factor
For further insight and analysis, both patients and control subjects were divided into
subgroups according to one single cardiovascular risk factor. No other risk factor was
allowed. The groups formed, and their descriptions are presented in Online Table IV.
We were able to form groups for smoking, hypertension, dyslipidemia, and diabetes
for PAD patients, and groups for smoking and dyslipidemia for control subjects. No
groups could be formed for uremia or renal insufficiency and rheumatic diseases,
because they were always accompanied by another risk factor. There were also no
PAD patients without at least one cardiovascular risk factor.
Online Table I. Prevalence of Background Illnesses and Cardiovascular Risk
Factors in the Patient Cohort (n = 226) and Control Cohort (n = 153)
Prevalence in Patient
Cohort
Prevalence in Control
Cohort
COPD
21.2% (67% male)
0%
Chronic hypoxia
5.8% (69% male)
0%
Hypertension
74.8% (56% male)
0%
Sleep apnea
4.9% (82% male)
0%
Type I diabetes
6.2% (50% male)
0%
Type II diabetes
28.8% (71% male)
0%
Dyslipidemia
31.9% (57% male)
13% (81% male)
Renal insufficiency
23.9% (65% male)
2% (100% male)
Uremia
5.3% (83% male)
0%
Rheumatic disease
15.9% (39% male)
0%
Smoking
Patient cohort
Control cohort
Coronary artery
disease
Patient cohort
Control cohort
Diabetes
Patient cohort
Control cohort
Never smoked
Quit smoking
Current smoker
41.2%
27.4%
31.4%
(41% male)
(71% male)
(65% male)
81%
0%
No
Mild
19%
(76% male)
Severe
58.0%
12.4%
29.6%
(51% male)
(57% male)
(65% male)
100%
0%
0%
No diabetes
Oral
treatment
Insulin treatment
65%
14.2%
20.8%
(51% male)
(72% male)
(64% male)
100%
0%
0%
Online Table II. Basic Laboratory Values for the Patient Cohort (n = 226) and
Control Cohort (n = 153)
Patient Cohort
Control Cohort
P-value
Systolic blood pressure
(mmHg)
149 (SD, 27)
138 (SD, 22)
<0.0001*
Total cholesterol
(mmol/L)
4.3 (SD, 1.15)
5.1 (SD, 1.1)
<0.0001*
LDL cholesterol
(mmol/L)
2.23 (SD, 0.93)
3.1 (0.96)
<0.0001*
HDL cholesterol
(mmol/L)
1.41 (SD, 0.51)
1.46 (SD, 0.42)
ns
Triglycerides (mmol/L)
1.4 (SD, 0.82)
0.87 (SD, 0.40)
<0.0001*
95 (SD, 65)
88 (SD, 42)
ns
Creatinine (umol/L)
GHbA1c (%)
6.89 (SD, 1.3)
Fasting plasma glucose
(mmol/L)
*Student’s t-test for normally distributed values
NA
5.2 (SD, 0.67)
NA
Online Table III. Characteristics of the PAD Patient Cohort According to Age,
Disease Localization, and Severity (n = 226)
<60 years
60–69 years
70–79 years
>80 years
10.6%
32.7%
30.2%
26.5%
63%/37%
67%/33%
59%/41%
35%/65%
History of smoking
75%
78%
64%
23%
Hypertension
50%
74%
78%
86%
Dyslipidemia
38%
40%
21%
33%
Diabetes
42%
40%
38%
22%
Renal insufficiency
13%
14%
25%
37%
Rheumatic disease
8%
10%
24%
18%
Aorta-iliac
Femoropopliteal
Crural
Pedal
Localization
26.1%
43.8%
23.5%
6.2%
Male/female
56%/44%
58%/42%
(53%/47%)
50%/50%
Mean age (years)
66 (SD, 9.3)
71 (SD, 9.3)
78 (SD, 11.8)
74 (SD, 12.6)
History of smoking
86%
69%
26%
7%
Hypertension
63%
81%
83%
64%
Dyslipidemia
44%
34%
25%
7%
Diabetes
33%
43%
49%
50%
Renal insufficiency
9%
16%
42%
64%
Rheumatic disease
7%
13%
25%
43%
No
symptoms
/missing
Claudication
Critical ischemia
2.3%
44.5%
53.2%
55%/45%
58%/42%
68 (SD, 9.3)
74 (SD, 11.5)
Smoking
72%
50%
Hypertension
76%
77%
Dyslipidemia
42%
25%
Diabetes
25%
45%
Renal insufficiency
10%
35%
Rheumatic disease
11%
20%
Age group
Male/female
Disease severity
Male/female
Mean age (years)
Online Table IV. Descriptive Information on Patient Subgroups with Only One
Distinct Cardiovascular Risk Factor
N
Age
Male
%
Syst
BP
Chol
LDL
HDL
Trig
Smoking
(PAD)
17
57
(10)
48
144
(15)
4.5
(1.2)
2.4
(0.9)
1.5
(0.4)
1.4
(0.8)
Hypertension
(PAD)
12
77
(10)
33
164
(30)
3.9
(0.7)
2.3
(0.7)
1.2
(0.3)
1.1
(0.2)
Dyslipidemia
(PAD)
9
82
(5)
33
171
(25)
5.0
(1.2)
3.0
(1.2)
1.5
(0.2)
1.1
(0.3)
Diabetes
(PAD)
8
72
(15)
63
173
(24)
4.2
(0.9)
2.1
(0.7)
1.4
(0.5)
1.4
(0.6)
Smoking
(control)
29
50
(14)
76
135
(21)
5.3
(0.9)
3.3
(0.9)
1.4
(0.4)
1.5
(0.5)
Dyslipidemia
(control)
14
67
(5)
81
149
(22)
7.0
(0.6)
4.7
(0.6)
1.5
(0.6)
2.0
(0.7)
Standard deviation in parentheses. Lipid levels are expressed as mmol/L, systolic
blood pressure (Syst BP) in mmHg and age in years. Chol, total cholesterol; Trig,
triglycerides.