Plasma endothelin-1 and vascular endothelial growth factor levels and their relationship to hemodynamics in idiopathic pulmonary fibrosis

Abstract

Background

Pulmonary hypertension (PH) is associated with a poor prognosis in idiopathic pulmonary fibrosis (IPF). Endothelin-1 (ET-1) and vascular endothelial growth factor (VEGF) are important in both fibrosis and vascular remodeling.

Objectives

We sought to determine the relationship between ET-1 and VEGF levels and hemodynamics in patients with IPF. We hypothesized that higher levels of ET-1 and VEGF would be associated with higher pulmonary artery pressures (PAP) and pulmonary vascular resistance (PVR) in patients with IPF.

Methods

We performed a cross-sectional analysis of 52 adults with IPF enrolled in a prospective cohort with available clinical data, platelet-free plasma, and hemodynamics. ET-1 and VEGF levels were measured via immunoassay. The associations of ET-1 and VEGF with PAP and PVR were examined using generalized additive models adjusted for age, gender, race/ethnicity, and forced vital capacity (% predicted).

Results

Sixteen of 52 (30.8%) had PH (mean PAP ≥ 25 mmHg). After multivariable adjustment, higher ET-1 levels were significantly associated with higher systolic (p = 0.01), diastolic (p = 0.02), and mean (p = 0.01) PAP and possibly higher PVR (p = 0.09). There were no significant associations between VEGF levels and hemodynamics.

Conclusions

Higher levels of ET-1 were associated with higher PAP and possibly higher PVR in participants with IPF. In a sub-group of patients, ET-1 may be a contributor to pulmonary vascular disease burden in IPF.

INTRODUCTION

Pulmonary hypertension (PH) is a common complication of idiopathic pulmonary fibrosis (IPF) that is associated with a high mortality rate when present [1–7]. While most individuals with IPF experience mild-to-moderate elevations in pulmonary artery pressures (PAP), the development (and severity) of PH is not associated with the severity of lung function impairment [2,4,5]. It has been postulated that PH in IPF is mediated in large part by hypoxic pulmonary vasoconstriction, but this is unlikely to be the sole determinant of pulmonary vascular disease, given the lack of a clear relationship between lung function and PH [2,4,5].

Endothelin-1 (ET-1) is a potent vasoconstrictor and promoter of pulmonary arterial smooth muscle cell growth with a pivotal role in pulmonary arterial hypertension (PAH) [8–10]. Elevated levels of ET-1 have been demonstrated in the peripheral and pulmonary circulations of patients with PAH and secondary PH [8,11–15]. In addition to its impact on the pulmonary endothelium, ET-1 has mitogenic effects that might contribute to the pathogenesis of pulmonary fibrosis [16–23]. Adults with interstitial lung disease (ILD) have been shown to have elevated circulating ET-1 levels, however few studies have examined the association of ET-1 with pulmonary vascular disease in IPF [24–26].

As with ET-1, vascular endothelial growth factor (VEGF) is an important mediator in the development of PAH and may play a role in the pathogenesis of IPF (and IPF with PH) [27–29]. In animal models of IPF, VEGF appears to modulate both fibrosis and endothelial function, although its effects are somewhat controversial [29–32]. VEGF levels may correlate with lung function in IPF patients, but to our knowledge the relationship between VEGF levels and hemodyamics has not been studied [26,33].

We examined the cross-sectional associations between plasma ET-1 and VEGF levels and hemodynamics, specifically PAP and pulmonary vascular resistance (PVR), in a cohort of patients with IPF. We hypothesized that higher levels of ET-1 and VEGF would be associated with higher PAP and PVR.

METHODS AND MATERIALS

Study Design

Participants were prospectively evaluated at the New York Presbyterian/Columbia University Medical Center Interstitial Lung Disease (ILD) and Lung Transplantation Programs between June 2007 and December 2009. Patients meeting criteria for the diagnosis of IPF based on the 2000 American Thoracic Society/European Respiratory Society consensus statement were included [34]. Patients with clinical evidence of collagen vascular disease and lung disease due to environmental or drug exposure were excluded. Patients referred to the ILD and Lung Transplantation Programs undergo transthoracic echocardiogram as part of their routine work-up. Patients with a suggestion of PH (estimated systolic PAP > 40 mmHg or evidence of right ventricular dysfunction when an adequate tricuspid regurgitant jet is not seen) and those being considered for lung transplantation go on to have hemodynamics performed. Demographic and clinical data were collected by research staff. Additional physiologic data including pulmonary function testing and six-minute walk distance (6MWD) were performed according to standard clinical protocols [35,36]. The study was approved by the Columbia University Medical Center Institutional Review Board.

We included patients with available hemodynamics and banked plasma. We excluded those with hemodynamics and blood draw that were more than 6 months apart, and those without available echocardiography to confirm the absence of left ventricular systolic dysfunction (ejection fraction < 50%) and/or more than moderate aortic or mitral valve disease.

Hemodynamics

Right heart catheterizations were performed at the Cardiac Catheterization Laboratory of New York Presbyterian hospital using standard clinical procedures. All catheter tracings were reviewed by two primary investigators (CEV, DJL); patients were excluded if pulmonary capillary wedge pressure (PCWP) was determined to be ≥ 15 mmHg by either investigator.

Biomarkers

Platelet-free plasma by venipuncture for each participant was collected by a trained phlebotomist, processed, and stored at −80°C. Plasma ET-1 and VEGF were measured by quantitative sandwich enzyme immunoassays (QuantiGlo, R&D Systems, Inc, Minneapolis, MN) at the Laboratory for Clinical Biochemical Research, University of Vermont. Reported manufacturer intra-assay coefficients of variation (CV) are between 2.6–3.4% and 2.8–7.9% and inter-assay CV are between 4.6–8.9% and 4.2–8.8% for ET-1 and VEGF, respectively.

Statistical Analysis

Continuous variables were expressed as means and standard deviations. Categorical variables were expressed as percentages. Independent sample t tests were used to compare continuous variables and chi-square or Fisher’s exact tests were used to compare categorical variables in study participants and those excluded. As they are not normally distributed, ET-1 and VEGF were natural logarithmically transformed. Generalized additive modeling was used to assess the relationship between continuous ET-1 and VEGF levels and hemodynamics. Models were adjusted for age, gender, race/ethnicity, and forced vital capacity (FVC) (percent predicted). Statistical significance was defined as P < 0.05. Analyses were performed using STATA 10.0 (StataCorp, College Station, TX).

RESULTS

A total of 129 patients were enrolled in the IPF cohort during the study period (Figure 1). Of these, 74 (57%) had clinical data and right heart catheterization with available catheter tracings at the time of the study. We excluded 10 (14%) with PCWP ≥15 mmHg, 3 (4%) without available echocardiograms, and 9 (12%) patients whose blood draw and right heart catheterization were performed more than 6 months apart. A total of 52 participants were included in the study. Characteristics of the study sample and those excluded (N = 77) are shown in Table 1. Participants were more likely to be male, to have had systemic hypertension, to have received corticosteroids, and had more pronounced impairments in lung function and functional class than those excluded. Among participants, the mean age was 63 ± 6 years old, 83% were male, 94% were non-Hispanic white, and 61% were former smokers. None were receiving targeted PH therapy, such as endothelial receptor antagonists (ERAs) or phosphodiesterase-5 inhibitors. Mean FVC was 56 ± 15% predicted, mean 6MWD was 407 ± 126 m, and most participants were classified as New York Heart Association (NYHA) functional Class II or III. The average time between blood draw and hemodynamic assessment was 60 ± 49 days. Forty (77%) of the sample had study measurements performed less than 3 months apart.

Figure 1. Study flow.

IPF=idiopathic pulmonary fibrosis; PCWP=pulmonary capillary wedge pressure

Table 1.

Study participants

	Study Sample	Excluded	P value
Number	52	77
Age, years	63 ± 6	65 ± 8	0.12
Male gender, %	83	69	0.08
Race/ethnicity, %*
Non-Hispanic white	94	90	0.65
Non-Hispanic black	2	3
Hispanic	0	0
Non-Hispanic other	4	7
Ever-smoker, %	61	69	0.28
Height, cm	175 ± 10	169 ± 9	0.00
Weight, kg	86 ± 15	80 ± 15	0.03
Body mass index, kg/m2	28 ± 4	28 ± 5	0.72
Systemic hypertension, %	41	33	0.04
Corticosteroid use, %	38	33	0.07
PH therapy, %
Endothelin Receptor Antagonists	0	1	1.00
Phosphodiesterase-5 Inhibitors	0	4	0.16
FVC, % predicted	56 ± 15	66 ± 20	0.01
No.*	50	63
DLCO, % predicted	31 ± 10	36 ± 13	0.05
No.*	49	57
6MWD, m	407 ± 126	395 ± 142	0.57
No.*	47	36
NYHA Functional Class, %			0.04
I	4	22
II	56	36
III	38	36
IV	2	6
Hemodynamics
Right atrial pressure, mmHg	3.1 ± 2.7
Systolic PAP, mmHg	38.4 ± 12.6
Diastolic PAP mmHg	13.3 ± 6.6
Mean PAP, mmHg	22.3 ± 8.5
PCWP, mmHg	7.3 ± 3.2
PVR, dynes·s·cm−5	269.7 ± 195.8
Cardic index, L/min/m2	2.6 ± 0.6

Data shown as mean ± standard deviation or %.

^*Subset with available pulmonary function tests and 6MWD

Definition of abbreviations: FVC=forced vital capacity; DLCO=diffusion capacity of the lung for carbon monoxide; 6MWD=six-minute-walk distance; NYHA=New York Heart Association; PAP=pulmonary arterial pressures; PCWP=pulmonary capillary wedge pressure; PVR=pulmonary vascular resistance

The mean ET-1 level was 2.1 ± 0.6 pg/mL and the mean VEGF level was 54.9 ± 25.5 pg/mL. The mean right atrial pressure was 3.1 ± 2.7 mmHg, mean PAP was 22.3 ± 8.5 mmHg, PCWP was 7.3 ± 3.2 mmHg, PVR was 269.7 ± 195.8 dynes·s·cm⁻⁵, cardiac index (CI) was 2.6 ± 0.6 L/min/m² (Table 1). Sixteen of 52 (30.8%) had PH (mean PAP ≥25 mmHg). Among those with PH, the mean ET-1 level was 2.3 ± 0.6 pg/mL and the mean VEGF level was 55.5 ± 25.9 pg/mL, as compared to 2.1 ± 0.6 pg/mL and 54.6 ± 25.6 pg/mL, respectively, in those without PH.

The continuous relationship between ET-1 levels and hemodynamics are depicted in Figure 2. After adjustment for age, gender, race/ethnicity, and percent predicted FVC, there was a blunted J-shaped relationship between ET-1 levels and systolic PAP (P for nonlinearity = 0.01), diastolic PAP (P for nonlinearity = 0.02), mean PAP (P for nonlinearity = 0.01), and possibly PVR (P for nonlinearity = 0.09). There was no association between ET-1 levels and RAP, PCWP, or CI. There were no significant associations between VEGF levels and hemodynamics (Figure E1). Results were unchanged when adjusted for diffusing capacity of the lung for carbon monoxide (DLCO) (data not shown). ET-1 and VEGF levels were not associated with FVC, DLCO, 6MWD, or NYHA functional class (data not shown).

Figure 2. Relationship between plasma ln(endothelin-1) and hemodynamics.

PAP=pulmonary artery pressures; PCWP=pulmonary capillary wedge pressure; PVR=pulmonary vascular resistance

DISCUSSION

We have shown that higher plasma ET-1 levels are associated with higher PAP and perhaps PVR in a cohort of patients with IPF after adjustment for demographics and lung function. This relationship is non-linear and J-shaped, with only a sub-group of patients demonstrating correlations between higher levels of ET-1 and higher PAPs. Conversely, there were no significant associations between VEGF levels and hemodynamics. To our knowledge, this is the largest observational study of these circulating biomarkers in IPF.

Several small studies have previously shown elevated levels of ET-1 in patients with ILD (but not necessarily IPF) and that ET-1 levels correlate with mean PAP [25,37–39]. Trakada and colleagues demonstrated that ET-1 concentrations were higher among those with ILD and concurrent PH compared to those without PH in 38 patients [38]. In IPF specifically, the relationship with hemodynamics has not been studied, but ET-1 has been shown to be elevated peripherally and to be overexpressed in the lung [24]. ET-1 appears to localize to small vessel endothelial cells but diffuse into surrounding tissue, suggesting that in certain individuals activation of the ET-1 pathway may result in uncontrolled fibrosis and perhaps widespread endothelial dysfunction via paracrine effects, resulting in IPF with PH [21,24,40,41]. Randomized clinical trials have studied the effect of ERAs on disease progression in IPF and IPF with PH (clinicaltrials.gov registry numbers NCT00879229 and NCT00768300) [42,43]. While the results of these trials have been somewhat disappointing, it is possible that a subgroup of patients with IPF (e.g., those with a particularly active endothelin pathway and therefore a large pulmonary vascular disease burden) may benefit from targeted PH therapy [44]. In fact, while most participants in our study clustered around mid-range ET-1 levels and had mild to modest elevations in PAP and PVR, a small group had strong linear associations between ET-1 and hemodynamics (as shown in Figure 2).

In PH and PAH, endothelial VEGF and VEGF receptor production is upregulated in response to vascular insult [27,28,45,46]. Evidence in humans is limited, but VEGF may also modulate angiogenesis and fibrosis in IPF [26,33]. While some studies have shown VEGF inhibition attenuates pulmonary fibrosis, areas of spared lung appear to have increased vascular density and VEGF upregulation [29–31]. Farkas et al. demonstrated reduced VEGF expression in fibrotic lung, an inverse correlation of VEGF levels with PAP, and improved endothelial apoptosis, vascular remodeling, and PH with VEGF augmentation, suggesting that decreased vascularity may underpin the development of PH in IPF [32]. While patients with IPF seem to have similar serum levels of VEGF compared with controls, in patients with severe disease there is significant correlation between VEGF levels and progression of disease [26,33]. These studies do not separately identify patients with concurrent PH, and our study suggests VEGF may not be a useful biomarker in IPF. The relationship between VEGF, angiogenesis/fibrosis and PH remains poorly described in humans.

Our study has several limitations. First, while the majority had blood draw and hemodynamics measured over a short interval (58% within 8 weeks and 39% within 4 weeks), ideally these would have been performed simultaneously and this may have introduced confounding. No patients were hospitalized between tests, implying relative clinical stability of the cohort. Second, patients were selected for hemodynamic evaluation based on screening echocardiography (or as part of lung transplantation evaluation) and therefore those included in the study tended to be sicker (e.g., had more impaired lung function) than those excluded, possibly resulting in bias. Third, as this is a cross-sectional study, no conclusions can be drawn about causality, i.e. whether elevated ET-1 precedes the development of PH complicating IPF or is simply a marker of pulmonary vascular disease burden. As is commonly seen in IPF, our study sample was predominantly white men, limiting the generalizability of these findings to women and individuals of other races. Last, because of the small number of patients with significant pulmonary vascular disease (31% with mean PAP ≥ 25 mmHg), we were underpowered to assess whether ET-1 may be useful to predict PH.

We have shown that higher ET-1 levels are associated with higher PAP and may be associated with higher PVR. Studies will be needed to evaluate ET-1 as a diagnostic screening tool for IPF with PH, and it remains to be seen whether subsets of individuals may benefit from targeted PH therapy.

Supplementary Material

Suppl Leg. Figure E1. Relationship between ln(vascular endothelial growth factor) levels and hemodynamics.

PAP=pulmonary artery pressures; PCWP=pulmonary capillary wedge pressure; PVR=pulmonary vascular resistance