Today’s and tomorrow’s imaging and circulating biomarkers for pulmonary arterial hypertension

Abstract

The pathobiology of pulmonary arterial hypertension (PAH) involves a remodeling process in distal pulmonary arteries, as well as vasoconstriction and in situ thrombosis, leading to an increase in pulmonary vascular resistance, right heart failure and death. Its etiology may be idiopathic, but PAH is also frequently associated with underlying conditions such as connective tissue diseases. During the past decade, more than welcome novel therapies have been developed and are in development, including those increasingly targeting the remodeling process. These therapeutic options modestly increase the patients’ long-term survival, now approaching 60% at 5 years. However, non-invasive tools for confirming PAH diagnosis, and assessing disease severity and response to therapy, are tragically lacking and would help to select the best treatment. After exclusion of other causes of pulmonary hypertension, a final diagnosis still relies on right heart catheterization, an invasive technique which cannot be repeated as often as an optimal follow-up might require. Similarly, other techniques and biomarkers used for assessing disease severity and response to treatment generally lack specificity and have significant limitations. In this review, imaging as well as current and future circulating biomarkers for diagnosis, prognosis, and follow-up are discussed.

Introduction

Pulmonary hypertension (PH) includes several pathologies. A new classification of PH designated five distinct categories differing in their pathophysiology, clinical presentation, diagnostic findings, and response to treatment (Table 1) [1]. Amongst these, pulmonary arterial hypertension (PAH) is characterized by pulmonary vascular remodeling featured by vasoconstriction, remodeling of the small pulmonary artery (PA) wall, and in situ thrombosis [2]. PAH is clinically characterized by progressively increasing pulmonary vascular resistance eventually leading to right ventricle (RV) hypertrophy (RVH), failure and death [3, 4]. PAH may be idiopathic or associated with various conditions (Table 1) [5].

Table 1.

Classification of pulmonary hypertension (PH)

1) Pulmonary arterial hypertension (PAH)  Idiopathic (iPAH)  Heritable  Related with collagen vascular disease (e.g., systemic sclerosis), congenital left-to-right shunt, HIV infection, use of drugs or toxins, portal hypertension or other causes (e.g., scleroderma)  Pulmonary hypertension of the newborn  Pulmonary veino-occlusive disease  Pulmonary capillary hemangiomatosis

2) Pulmonary hypertension with left-sided heart disease  Systolic dysfunction  Diastolic dysfunction  Valvular disease

3) Pulmonary hypertension associated with lung diseases and/or hypoxemia  Chronic obstructive pulmonary disease (COPD)  Interstitial lung disease  Sleep-disordered breathing, alveolar hypoventilation  Chronic exposure to high altitude  Developmental lung abnormalities

4) Pulmonary hypertension due to chronic thrombotic and/or embolic disease  Chronic thromboembolic pulmonary hypertension (CTEPH) in proximal or distal pulmonary arteries  Embolization of other matter, such as tumor, parasites or foreign material

5) Pulmonary hypertension with unclear multifactorial mechanisms  Hematologic disorders (i.e. myeloproliferative disorders, splenectomy)  Systemic disorder (i.e. sarcoidosis)  Metabolic disorder (i.e. glycogen storage disease)  Other causes like compression of pulmonary vessels, chronic renal failure

HIV Human immunodeficiency virus

Biomarkers are obviously an essential missing element for diagnosing PAH [1, 6]. In fact, the average delay before diagnosing PAH patients is 2.8 years, a period during which the disease often worsens [7]. The long current delay is explained by the lack of specificity of PAH symptoms. In fact, dyspnea is the cardinal symptom of PAH and is generally attributed to other pulmonary diseases such as asthma and chronic obstructive pulmonary disease, or to depression and anxiety disorders, before being associated with PAH. Moreover, the prevalence of PAH is estimated at 60 per million, which makes it rare and even more difficult to diagnose, since some physicians are not familiar with PAH [8]. Thus, new diagnostic tools should ideally reduce this delay in order to treat this pathology at its earliest stages. Indeed, PAH is diagnosed by an exclusion principle to strike out the other PH groups. Among the PAH group (group 1), iPAH is a diagnostic of exclusion since no known cause is related to the disease, as it is idiopathic (Fig. 1).

Fig. 1.

Pulmonary hypertension diagnostic algorithm. This represents the usual algorithm followed by patients with suspected pulmonary hypertension to diagnose the class of pulmonary hypertension in order to better adapt the treatment

Fortunately, major improvements have been made in the treatment of PAH in the past decade. Despite these novel therapies, most patients still display persistent exercise intolerance, and the long-term prognosis remains poor [4, 9–12]. A goal-oriented therapy has been proposed to improve long-term outcomes of PAH patients [13]. However, this proposal has emphasized the lack of a consistent protocol for evaluating the course of the disease. Indeed, establishing such goals is hampered by the gap in the evidence-based assessment of PAH treatment efficacy. Only limited tools are currently available for predicting the long-term outcome in PAH. An accurate biomarker should guide the escalation of pharmacotherapy given the risk/benefit ratio and the timing of referral for lung/heart transplantation, i.e. the ultimate tool for saving patients from RV failure. This involves the capacity to evaluate the improvement or degradation of the patient’s condition. In such circumstances, referral for life-saving interventions such as lung transplantation would not be uselessly delayed in the significant portion of patients who are unresponsive to new therapies [14].

Current understanding of PAH pathophysiology

Pulmonary arterial hypertension is a vascular disease that is largely restricted to small PA. Many abnormalities contribute to this syndrome of obstructed, constricted small PA. This includes abnormalities in blood concentration of a few growth factors, neurotransmitters and cytokines, namely increases in platelet-derived growth factor (PDGF), serotonin, interleukin-6 (IL-6), and endothelin-1 (ET-1). Nonetheless, these factors may not be reliable biomarkers since they are increased by stress, inflammation, and other common assaults on body integrity, and therefore, may not be specific to PAH. The media in PAH also exhibits an increased activation of the NFAT (nuclear factor of activated T cells) transcription factor, leading to increased Ca²⁺-dependent PA smooth muscle cell (PASMC) proliferation and decreased mitochondrial-dependent apoptosis [15]. There is also a metabolic disorder due in part to an impairment of mitochondrial metabolism known as the Warburg effect, which is defined as a decrease in glucose oxidation and an increase in glycolysis [16]. This phenomenon is well described in cancer, and mounting evidence supports the fact that PAH also displays a Warburg effect [16]. Finally, the adventitia is infiltrated with inflammatory cells and exhibits metalloprotease activation [17]. Despite recent advances in the therapy of PAH, mortality rates remain high (40% at 5 years) [18]. Over the last 5 years, many investigators (including our laboratory) have endeavored to understand the mechanisms for enhanced PASMC proliferation and decreased apoptosis, which account for distal PA remodeling. Indeed, the lung peripheral vascular bed is increasingly the target of novel treatment strategies, which aim at restoring physiological vascular tone and at reducing the remodeling process [19]. Many pathways are implicated in PAH development since this pathology is a combination of many factors such as PASMC deregulation (enhanced proliferation and migration, resistance to apoptosis), abnormal endothelial cells (causing endothelial barrier disruption and plexiform lesions), and inflammation as well as increased elastolytic activity. Many factors have been shown to be implicated in PAH pathogenesis. Among others, RhoA/ROCK, BMPRII (bone morphgenetic protein receptor II) dysfunction, STAT3 (signal transducer and activator of transcription-3), and NFATc2 upregulation seem to play a critical role in PASMC proliferation. The issue is that neither STAT3, an enhancer of NFAT expression, nor NFAT itself can be targeted, because of their critical function in a plethora of physiological processes, such as the immune response. More recently, our group has demonstrated that NFAT expression is STAT3-dependent while its activation relies on the oncogene protein Pim1 (proviral integration site for Moloney murine leukemia) [20]. Inflammatory factors such as TGF-β, IL-6, and BMP can stimulate PASMC proliferation [21]. Modification in BMPRII levels can trigger a large amount of signaling pathways. For example, transcription factor PPARγ (peroxisome proliferator activated receptor gamma) is decreased by BMPRII dysfunction and Rabinovich et al. remarkably showed that PPARγ is implicated in PASMC proliferation and migration, but also in endothelial cells dysfunction contributing to PAH pathobiology [22–24]. MicroRNA (miRNA) are also implicated in the sustainability of the disease, Courboulin et al. have indeed identified miR-204 as implicated in PAH. Other microRNA, such as miR-21, miR-34c, and miR-133, are implicated in right ventricular failure and, thus, may also play a role in PAH [25, 26]. These factors, proteins, and pathways are not yet used as biomarkers nor therapeutic targets, but this shows that the past decade brought knowledge in understanding PAH in a way to potentially treat it.

In this review, currently available biomarkers will be presented together with their limitations and more recent improvements.

Primary tests

The first signs of PAH are observed by physical examination, including an increase in jugular venous distension, a holosystolic murmur of tricuspid valve regurgitation and a loud P2. Hepatomegaly, peripheral edema, ascites, and cool extremities are eventually observed when RV failure is more severe. Electrocardiography is usually performed at this stage, highlighting a right axis deviation, RV hypertrophy, and peaked P-waves. Moreover, nocturnal hypoxemia occurs in >75% of PAH patients, independently of the occurrence of apneas or hypopneas [27]. Isolated alteration in diffusing lung capacity for carbon monoxide is traditionally seen on pulmonary function tests, although pulmonary function may be normal or exhibits mild obstructive/restive disorder in PAH patients [28, 29]. When cardiopulmonary exercise testing is performed, PAH patients generally display altered cardiac response and gas exchange abnormalities [30]. However, these abnormalities are generally seen when right heart failure is already present and lack specificity.

New York Heart Association/World Health Organization functional class

The New York Heart Association and World Health Organization (WHO) functional classes (FC) are based on symptoms of dyspnea with degree of exercise intolerance [31], and help to classify PAH severity. PAH severity is most commonly assessed objectively with the use of WHO functional class combined with exercise capacity and RV function. Various studies have shown that FC and exercise capacity as assessed by the cardiopulmonary exercise test [32–34] or the 6-min walk test [34–37] have good discriminating properties and predict survival in PAH [35]. However, caution must be exerted when using the functional classification as an endpoint since the assignment of patients to different classes is subject to investigator and patient bias, even if they have held up in multivariate analysis [31, 35, 38, 39].

Right heart catheterization (RHC)

Right heart catheterization is the gold standard for the diagnosis of PH. It is to date the only direct proof of PH. PH is defined as a mean pulmonary artery pressure (PAP) at rest >25 mmHg (Fig. 1). The indication for RHC relies on the estimated systolic PAP on echocardiography, the presence of additional echocardiographic findings suggestive of PH and the presence of symptoms or risks factors for PH [40]. Crucial information such as PAP (systolic, diastolic and mean) values, right arterial and ventricular pressures, PA wedge pressure, pulmonary and systemic vascular resistance, cardiac output and cardiac index and mixed venous oxygen saturation can all be measured by catheterization [41]. RHC is essential for therapy selection since it provides important data on pulmonary circulation vasoreactivity [41], and provides important information about disease severity and response to therapy. Nonetheless, the technique is invasive, although the incidence of complications in experienced centers is small [42–44].

Imaging

Imaging represents a powerful tool for assessing lung peripheral vascular bed, as well as heart function and morphology. Indeed, although PAH patients suffer from a lung disease, functional capacity and prognosis is intimately related to heart function [45]. Thus, it is not only critical to be able to observe improvements at the lung level but also to visualize and analyze RV functions. Actually, the complex and unique shape and the contractile pattern of the RV make it more difficult to assess [46]. Different imaging techniques are now available to assess lungs as well as the RV. The development of new non-invasive biomarkers would be of great interest at diagnosis, but also for follow-up.

Echocardiography

Echocardiography is a key screening tool in the diagnostic algorithm of PAH (Fig. 1). In comparison with invasive measurements, it has the advantages of being safe, easily and rapidly performed, portable, widely available and thus well suited for screening, despite the fact that it is considered performer-dependent [47]. After PH is diagnosed by RHC, disease monitoring by echocardiography is done every 3–6 months.

Baseline measurements

Doppler echo is performed on patients with suspected PAH to assess several parameters such as heart function and shape, as well as pulmonary and cardiac hemodynamics.

Actually, echo allows the assessment of right and left heart chamber sizes, and PAH patients display an enlarged right atrial and RV size and RVH, instead of a crescent-shaped RV. Moreover, it is also possible to see pericardial effusions and diastolic septal shift [48], which have prognostic values for PAH patients [48, 49]. Echo also allows exclusion of other types of PH, including valvular dysfunction, congestive heart failure and congenital heart disease.

Furthermore, RVSP can be estimated by tricuspid regurgitation using Doppler, right arterial pressure and pulmonary artery acceleration time (PAAT). PAAT is decreased in PAH patients and known to correlate with disease severity. mPAP can also be approximated using Doppler function, with a correlation varying from 0.57 to 0.93 compared to PAP measured by RHC. PAP measured by echo also allows the calculation of other echo-derived parameters (e.g., aorta diameter, pulmonary velocity, heart rate), and the calculation of PVR, which is well correlated to disease severity [50]. LV eccentricity index has also been shown to have some predictive value on PAH patients outcome [8, 48].

Some findings have also highlighted the prognostic values of the Tei-Index and the tricuspid annular plane systolic excursion (TAPSE) [51]. The Tei-Index, already used to evaluate the performance of the myocardium, has proved useful in the evaluation of ventricular function. Tricuspid annular displacement has also shown good results in the prognosis of PH patients [52]. Furthermore, systolic displacement of the tricuspid annulus towards the RV apex is referred as TAPSE, correlates with RV ejection fraction, and powerfully reflects RV function. A decreased TAPSE portends a poor prognosis in patients with dilated cardiomyopathy [53–55], and Forfia et al. [56] showed that there is also a correlation between TAPSE and PAH prognosis (survival) in PAH patients. The latter authors have also demonstrated a good correlation between TAPSE values and mPAP, but an even better one with PVR. Furthermore, TAPSE is highly reproducible, unlike mPAP estimated by echocardiography. Nonetheless, further studies have to be performed to confirm these results.

Follow-up values

Echocardiography is also a good method to monitor therapeutic efficiency on RV function as it is a critical element of PAH prognosis and since it can be visualized on echo. Indeed, echo is also used to assess the efficiency of treatment as a follow-up technique. PAH deterioration is accelerated in PAH patients with RV dysfunction [4, 57]. Indeed, this imaging technique assesses parameters such as ventricular and septal morphologies, and maximal velocity of tricuspid regurgitation jets. All these parameters are significantly different between the intravenous epoprostenol (a prostacyclin analogous) and control groups [58]. Moreover, in another randomized controlled trial, a significant difference in changes in ventricular morphology, i.e. the minimum diameter of the inferior vena cava and Doppler measurement (RV ejection time and mitral valve peak velocity) in the patients taking bosentan (a endothelin receptor antagonist) was observed [8, 58], indicating that echo can be used as a follow-up technique.

Echo also represents a useful tool in the early diagnosis of PH. Indeed, first-degree relatives of PAH patients should undergo routine echo every 3–5 years to ensure that no hereditary PAH is developing, or if symptoms manifest themselves in order to get a diagnosis as quickly as possible if they develop PAH [41, 43, 59]. Echo is also useful in screening other at-risk population including scleroderma and HIV-positive patients [41, 60, 61].

Limitations and future developments

The main issue with this technique is its poor inter-observer agreement, and, due to its 2-D assessment, heart shape based on certain assumptions. Moreover, the ultrasound window can be limited not only because of high body mass index (BMI), lung fibrosis, or emphysema but also due to the sternum position of the patient [52]. Like other imaging techniques [magnetic resonance imaging (MRI), computed tomography (CT) scan], echo is constantly improving and has been shown to be accurate in assessing RV function [62]. Nonetheless, despite all its disadvantages, echo is likely to remain the most prominent tool for screening PAH and for follow-up, at least for the time being. This is explained by its availability and, despite its lack of precision, this imaging technique has nonetheless amply proven its reliability for screening. Furthermore, this imaging tool undergoes technical improvement.

The RV has a very complicated shape, which is hard to assess by echo due to 2-D limitations. Some authors believe that assessing LV compression yields more reliable results than RVH [52]. The 3-D echocardiography is in progress, which is of a great interest since it would decrease the requirements for geometric assumptions and provide more accurate estimations of LV and RV volumes and ejection fractions [63, 64]. New improvements are also expected concerning insights into RV and LV structure, function and interdependence [8, 65].

Recently, a novel application for Doppler echocardiography permits RV strain analysis, a very promising development [66]. Indeed, in a recent study in 30 PAH patients, Filusch et al. have shown that RV systolic strain and strain rate were significantly altered and correlated with levels of N-terminal pro-brain natriuretic peptide (NT-proBNP, a serum biomarker) and reduced 6MWD compared to matched control. Moreover, reduced strains also correlated with both mPAP and PVR in these PAH patients [51]. Thus, detection of the strains is able to discriminate between the different WHO FC and predict in which class patients will fall. Results from RV strain analysis are also related to invasive pulmonary hemodynamic parameters such as PVR, mPAP and cardiac outpout, emphasizing its prognostic value [47, 67].

Unfortunately, some of the parameters discussed above may sometimes show lack of specificity. This explains why RHC must be made on patient with normal echo-derived RVSP but with high suspicion of PAH, to ensure whether or not this patient suffers PAH. Indeed, RHC remains the only diagnostic measure for PAH, and is usually used to confirm the presence of PAH. Nonetheless, echo is a precious tool in screening patients with PAH-like symptoms, to at least exclude pulmonary pathologies, if not being a PAH diagnosis tool itself.

Radionuclide imagery

This type of imaging technique includes single photon emission computed tomography (SPECT), positron emission tomography (PET), scintigraphy, and magnetic resonance spectroscopy imaging [67]. A few decades ago, before the emergence of CT and MRI, radionuclide imaging was commonly used [47]. The presence of perfusion (Q) abnormalities using perfusion scintigraphy has already been shown in late PAH. However, in addition to impairment in perfusion, there is also a concomitant impairment in ventilation (V) as seen by planar scintigraphy [68].

Using SPECT, which is more sensitive than planar scintigraphy for assessing focal V–Q disturbances, the V/Q SPECT-derived ratio can be measured and provides topographic V–Q distribution and an objective quantification of cross-sectional lung V–Q imbalance with a correlation of lung morphology on CT. Some groups have been able to find V impairment resulting in V–Q mismatch in patients with mild PAH. As indicated by the significant correlation observed between the standard deviation of the entire lung V/Q ratios and the mean PAP in these patients, the lung pathophysiology causing V–Q imbalance may be more advanced, as shown by persistently elevated PAP in iPAH [68]. However, the main indication for lung V/Q scan remains to date the exclusion of chronic thromboembolism as the cause of PH [69].

Using PET scan, Wong et al. demonstrated that RV myocardial blood flow is increased in iPAH, whereas high O₂ extraction fraction was associated with poorer NYHA class and more severe RV failure. This may explain why an additional demand to the heart may lead to cardiac ischemia and RV failure [70].

Another interesting aspect is the Warburg effect since it is a hallmark of both cancer and PAH [16]. This pathological phenomenon is characterized by an imbalance between oxidative phosphorylation and glycolysis, promoting glycolysis and lactate production. This also involves increased glucose uptake compared to non-pathological tissues. The PET scan is already well developed and commonly used in cancer to assess the development of a Warburg effect and response to treatment, by correlating that effect with tumor severity and aggressiveness. Thus, this technique may be thought as also being of some prognostic value in PAH management. Indeed, using PET, Abikhzer et al. [71] have reported that RV, which is usually invisible in PET, could be imaged in a PAH patient due to increased glucose uptake. This might represent a future approach to assess RVH severity. Moreover, Marshboom et al. have recently demonstrated that this technique may also be possible as assessed the PAH lungs, as it also displays Warburg effect [72].

New tools are also emerging, such as radioligand binding of phosphodiesterase 5 (PDE5) [73, 74] and ET-1 [75], which are currently being evaluated in animal models. Indeed, these receptors are already targeted in the pathophysiology of PAH, and the hypothesis that their level is related to the evolution of patients’ state is under scrutiny. Thus, it might be interesting to verify the relevance of such radioligands by quantifying the changes in the expression of PDE5 and ET-1 as biomarkers in patients. These endpoints will obviously require further studies and will need to be validated more extensively in PAH animal models before going through human trials.

Magnetic resonance spectroscopy imaging has much contributed to the diagnosis and prognosis of cancer through the study of the metabolic changes occurring in tumors. Thus, this technique might also bring a contribution to the field of PAH. Moreover, this imaging tool supplements the use of MRI in cancer in order to facilitate the diagnosis and to decide upon a proper therapeutic strategy. Hence, it might well be possible to apply such a combination of techniques to PAH [76].

CT scan

The gating system for the lung and heart (based on the electrocardiogram gating system) has been a major improvement in CT, allowing to overcome artifacts due to respiration and heartbeat. This allowed the functional assessment of the heart in addition to lung parameters. Indeed, it is believed that RV remodeling follows PA remodeling. Thus, it might be of crucial interest to assess PA remodeling before the effect can be visualized in the right heart as a result of disease progression. Moreover, assessing lung remodeling would allow assessing the response to treatment in distal PA where the pathophysiological process actually takes place.

Baseline measurements

Some investigators have studied the main PA diameter in relationship to mean PAP. Unfortunately, its correlation with mean PAP varies from weak to strong even when corrected for the size of the aorta, the thoracic vertebrae or the arterio-bronchial segment in more than three lobes [77, 78], and for the BMI [79]. Moreover, a progressive dilatation of the main PA overtime is frequently discrepant with PAH severity [80].

More importantly, images obtained by CT may allow specialists to visualize characteristics that help in the differential diagnosis of PAH. For example, the combination of mosaic lung attenuations, a marked variation in the size of segmental vessels, peripheral parenchymal opacities as well as dilatation of bronchial and non-bronchial systemic arteries is common in chronic thromboembolic pulmonary hypertension (CTEPH) (73%) but is rare in PAH (14%), which can provide a differential diagnosis between these two pathologies [77]. Similarly, the CT scan allows the exclusion of severe parenchymal abnormalities responsible for PH (e.g., pulmonary fibrosis), and can identify other possible causes of PH (congenital malformations, coronary heart disease or cardiac dysfunction) [81]. Finally, it may document pericardial effusions, which is directly related to PAH severity [77].

More recently, data from electrocardiographically-gated multidirectional CT studies have shown that functional parameters such as right PA dispensability, systolic–diastolic RV outflow tract dimensions, and diastolic wall thickness are measurable with good inter-observer agreement, and can reliably be used as criteria for PAH diagnosis [77]. In a recent study, Revel et al. [82] found that the most reliable parameter among all electrocardiographically-gated CT parameters evaluated for identifying PAH patients is right PA wall distensibility. This measure also correlates with mean PAP [71]. The right ventricle/left ventricle ratio, using a gated or non-gated multi-directed CT reflects RV function if acute pulmonary embolism. Whether these results can be extended to PAH patients remains unknown [83].

Devaraj et al. [78] have tested a composite index using the CT-derived ratio of the diameter of the ascending aorta and echo-derived RV systolic pressure. The combination of these two parameters from two different imaging techniques showed a relationship with mean PAP and was a better predictor of PH than either measurement alone. That index has been found to increase specificity to 100% with respect to RHC-derived diagnosis compared to the criterion of main PA diameter alone [84]. Indeed, a combination of the (CT-derived main PA diameter)/(ascending aorta) ratio and echo-derived RV systolic pressure has been found to better correlate with RHC-derived mean PAP than either single measurement [78]. A simultaneous use of parameters derived from CT and echo also represents a more informative combination since CT provides anatomical data whereas echo identifies their functional implications.

Limitations and future developments

A major limitation of CT-scan is X-ray exposure, which is problematic when multiple follow-up measures are necessary. Nevertheless, major improvements are currently made to limit X-ray levels and the exposure time required for a given scan quality [85]. Moreover, the contrast agents used may exhibit significant nephrotoxicity and be allergenic in some patients (e.g., iodine). Claustrophobia is a more mundane limitation that can be taken into account.

New investigations based on technological improvements are constantly being undertaken to circumvent limitations related to radionuclide imaging, the technique replaced by CT-scan, but also related to CT-scan itself. Dual-energy CT-perfusion imaging is a new technology that decreases X-ray exposure with results comparable to perfusion scintigraphy. That technique can provide CT pulmonary angiograms, high-resolution morphologic images, and spatially matched perfusion images in the very same scan, with the same radiation dose that is sustained in a classic CT pulmonary angiogram. Nevertheless, a few technical issues remain to be addressed [86, 87]. Another novel technique is area-detector CT, which will probably soon be developed in the lung, although radiation exposure remains significant. The CT-scan is important as it allows the exclusion of underlying pathologies related to PH. Furthermore, new types of image acquisition allow the collecting of increasing amounts of data in a single scanning session.

MRI

MRI has represented a major advance in the assessment of lung and cardiac anatomy and function. Compared to echo, MRI provides a higher spatial resolution and exhibits better inter-observer reproducibility [88], and one of its major advantages is that planes can be moved, and due to the 3-D view, no approximation is required. Thus, cardiac MRI is now considered as the gold standard for the detailed study of the RV and is currently used as an endpoint in the multinational European Union-funded Framework 6 EURO-MR project for PH, which will provide an important opportunity to more definitely establish its utilization as an endpoint [89]. Moreover, recent improvements of hardware and software currently enables semi-automated myocardial blood border definition, reducing intra- and inter-observer variability in ventricular volume measurements and decreasing the time spent on analysis [8, 52, 90–95]. Finally, technical improvements of cine acquisition have enabled the analysis of dynamic visualization of several heartbeats (as cine images) with a significant decrease in breath-holding [46], which is frequently problematic in PAH patients who usually complain of dyspnea.

Baseline measurements

Cardiac MRI is most commonly used to assess end-systolic and end-diastolic RV and LV volumes and the resulting RV ejection fraction, stroke volume and cardiac output. Cardiac output can also be assessed by cine phase contrast methods measuring volumetric flow in the main PA. Not surprisingly, the RV end-diastolic and end-systolic volumes have been shown to be significantly increased, whereas stroke volume and cardiac output are decreased in PAH compared to controls. Moreover, the RV ejection fraction has been described as being impaired by ~50% in PAH patients [52]. MRI also allows the assessment of RV hypertrophy, including (1) RV mass, which is increased two- to threefold in PAH [52], (2) the Fulton ratio (RV mass/LV mass with septum), which is increased by 80% in PH patients, and (3) end-systolic RV wall thickness. As RV failure progresses, LV end-diastolic volume and peak LV filling rate are impaired in PAH patients and may be assessed with cardiac MRI [52, 96]. In order to assess cardiac function under stress conditions, some authors have proposed cardiac MRI using dobutamine, a β₁-adrenergic agonist [97], or during exercise using a MR-compatible ergometer [98]. These experiments documented the incapacity of PAH patients to increase stroke volume from rest to exercise.

Delayed contrast enhancement (DCE) has been found to be frequent in PAH patients, especially when associated with connective tissue disease [99]. DCE is mainly observed at the RV intersection points and in intraventricular septum area. Although DCE may be associated with myocardial ischemia, it is more consistent with myocardial fibrosis. The extent of DCE in the myocardium has been shown to be inversely related to measures of RV functions [100]. Finally, PAH patients have been shown to display an interventricular asynchrony caused by a longer RV systolic contraction using tagged MRI [46]. Cardiac MRI is also of value to exclude other causes of PH such as congenital heart disease of LV impairment (e.g., diastolic dysfunction, constrictive pericarditis).

MRI can also be used to assess the pulmonary circulation. MRI-measured proximal PA pulsatility, a marker of PA stiffness predicts mortality in PAH patients [101]. However, the correlation between PA distensibility and RHC-derived mean PAP remains controversial [101, 102]. Furthermore, 2-D MR angiograms have been reported to differentiate CTEPH from PAH with a sensitivity of 92% compared to a V/Q scan [103]. Similarly, Nikolaou et al. [104] have been able to differentiate CTEPH from PAH with an impressive accuracy of 90%, using contrast enhanced perfusion MRI and MR angiograms with parallel imaging techniques [104]. Moreover, PAH patients have significantly decreased pulmonary blood flow and prolonged mean time transit in the whole lung, as documented by 3-D dynamic contrast-enhanced perfusion MRI [104]. Nevertheless, CT angiograms are still believed to be superior to MRI angiograms for visualizing vascular abnormalities.

Some groups have also suggested MRI-guided catheterization and real-time MRI-guided catheterization allow the assessment of an impressive list of direct and calculated parameters, namely the construction of RV pressure volume loops, RV afterload, myocardial contractility, pump function, and RV–PA coupling, which are all useful to assess PAH severity and response to therapy [46, 105]. Using both RHC and cardiac MRI, some have estimated RV power. Indeed, RV mechanical efficiency is lower in NYHA FC III than II, thus correlating with PAH severity. Decreasing mechanical efficiency of the RV is a characteristic of deterioration of RV functions in iPAH. This team has also assessed glucose uptake by PET scan. However, the reduction of RV mechanical efficiency is not due to a metabolic shift to glucose oxidation, as assessed by PET-scan [106].

Unfortunately, the relationship between MRI measures and pulmonary hemodynamics is less straightforward. Previous studies have suggested that PAP correlates with RV mass [52], Fulton ratio [52], RV wall thickness, PAAT [52], pulmonary blood flow transit time [107–109], and septal curvature (r = 0.77). If leftward ventricular septal bowing is observed, systolic PAP may be expected to be >67 mmHg [46]. Among different MRI-derived parameters, namely PA areas, PA stain, average velocity, peak velocity, acceleration time, and ejection time, it has been determined that average velocity best correlates with mean PAP reference. This may imply that RHC might lose its prominence in the future [46]. However, significant RV and septal thickening (as seen in long-standing PAH), low left ventricular systolic pressure may also influence the importance of the septal curvature [52]. Similar correlations have been suggested for pulmonary vascular resistance [110]. However, these conclusions should be considered with caution as these correlations were not observed by other teams [111].

Followup values

The accuracy and reproducibility of cardiac MRI in assessing cardiac functions and morphology makes it a powerful tool for treatment follow-up [90]. Indeed, changes in MRI-derived measurements are relevant in PAH. For instance, a progressive dilation of RV, and a decrease in LV systolic volume and RV stroke volume as detected at a 1-year follow-up by MRI have been correlated with a worse prognosis in PAH [52, 96]. Changes in RV stroke volume overtime also correlated with changes in exercise capacity [112]. Changes in myocardial perfusion, RV mass, and interventricular septal shift were also observed with therapy [8, 46, 113, 114].

Limitations

The main limitation of MRI is imposed by the core physical principle of this technique, i.e. its incompatibility with metal compounds such as the delivery pump for continuous parenteral therapies commonly used in PAH [8]. Nevertheless, some teams have extended patients’ tubulure, allowing the pump to be left outside the MRI room [8]. Moreover, MRI remains expensive, not widely available, and time-consuming compared to CT scan [90]. MRI is also problematic for claustrophobic patients (i.e. 2.3% of patients subjected to MRI), although it can easily be circumvented by light sedation and the development of a short-bore machine reducing the noise and the close-space feeling [115, 116]. As for other imaging technics, body movements, as well as respiratory and cardiac motions may result in artifacts. Fortunately, technical improvements in acquisition speed have reduced breath-holding and scan times [52]. Finally, end-stage renal disease has been associated with cases of nephrogenic systemic fibrosis when gadolinium-containing contrast agents are used [90]. Due to its availability and ease of use, echo will probably remain the tool of choice for first-tier screening and detection, whereas MRI will be used subsequently to assess lung and heart functions [52].

Conclusion

Imaging techniques are improving very fast, and, alone or in combination, allow the measuring of an increasing number of clinically relevant parameters both at baseline and during follow-up. Although RHC is still the gold standard to establish a diagnosis of PAH, novel imaging tools are coming progressively closer to the same diagnostic quality. While some standardization of procedures is required to reduce inter-observer and inter-site variability, imaging techniques are very useful to support a diagnosis of PAH diagnosis, to exclude other causes of PH, and to objectively assess disease severity.

Circulating biomarkers

In parallel with imaging techniques, circulating biomarkers are intensively explored by several groups. Ideally, blood and urine biomarkers that (1) become abnormally elevated/decreased in early or severe/end-stage disease (depending on the biomarker), (2) are independent of indirect consequences of PAH (e.g., renal and left ventricle function); and (3) parallel the progression of the disease or the favorable response to therapeutic interventions. Unfortunately, no such biomarker has yet been identified. Nonetheless, some of those presented below show some interest and despite the fact that no perfect biomarker is yet available, a combination of several biomarkers might be sufficiently accurate for diagnostic or prognostic purposes.

Cardiac-derived biomarkers

Brain natriuretic peptides (BNP and NT-proBNP)

Plasma brain natriuretic peptide (BNP) and its biologically inactive N-terminal fragment, NT-proBNP [117], are cardiac hormones secreted by the cardiac myocytes and are used as ventricular dysfunction biomarkers [118]. In PAH, BNP and NT-proBNP, levels increase due to enhanced synthesis by the RV and reflect the RV structure and function [117, 119, 120]. Both forms appear to predict survival in idiopathic PAH (iPAH) [117, 119, 121, 122] as well as other types of PH [123, 124]. Furthermore, natriuretic peptide levels parallel changes in pulmonary hemodynamics overtime [253]. However, natriuretic peptides are not a specific PAH biomarker since they are largely influenced by renal function, fluid retention and diuretics, situations frequently encountered in PAH [125, 126].

Troponin T

Cardiac troponins are regulatory proteins that can be detected by highly sensitive assays in peripheral blood after their release, as a result of cardiac myocyte membrane disruption. Torbicki et al. have evaluated the prognostic value of that biochemical parameter in patients with severe PH [127]. Detectable troponin T was associated with a higher heart rate and NT-proBNP levels, a shorter 6-min walk distance, and a poor prognosis. Indeed, 63% of patients with detectable levels of cardiac troponin T died during the 2-year follow-up. Although troponin T is a marker of disease severity, and of prognostic value, it rises only in end-stage disease [38, 127] and may be confounded by left heart disease and renal impairment [128–130].

Osteopontin

Osteopontin (OPN), being a cytokine, is related to the inflammatory process. Circulating OPN has been shown as an independent predictor of survival in iPAH patients, and correlates with severity [131]. Indeed, OPN is significantly higher in iPAH patients compared to healthy subjects. In a retrospective study, baseline level of OPN has been shown to correlate with 6-min walk distance, mean PAP, and FC. However, the origin of OPN production is still unclear. It is hypothesized to be produced by the RV but this still has to be confirmed [131]. Moreover, its biomarker value has to be confirmed in larger studies.

Endothelium-derived biomarkers

cGMP

Cyclic guanosine 3′,5′-monophosphate (cGMP) is produced by the activation of guanylate cyclase and is considered as an intracellular second messenger of natriuretic peptides, bradykinin and nitric oxide (NO) [132]. cGMP levels are significantly elevated in PAH patients' urine compared to healthy controls, and inversely correlate with the hemodynamic severity of PAH [133]. Wiedemann et al. have measured cGMP levels in PAH and PH patients, and, besides a marked increase in cGMP and ANP levels in these patients, found that iloprost (a stable prostacyclin analog) inhalation caused a decrease in cGMP in parallel with pulmonary vasodilation and hemodynamic improvement [134]. However, due to sizable interindividual variations in cGMP levels, this biomarker has poor discriminative properties [135].

Microparticles

Circulating microparticles (MP) are submicron membrane fragments shed from activated or damaged vascular cells, and released during apoptosis and/or activation of various cell types [136]. Circulating endothelial MP have been described as markers of endothelial injury as well as systemic vascular remodeling [137]. Amabile et al. [138] have shown that levels of circulating endothelial platelet endothelial cell adhesion molecule (PECAM⁺), VE-cadherin⁺, E-selectin⁺, and leukocyte-derived MP levels were increased in PH patients compared to healthy individuals. More specifically, levels of PECAM⁺ and vascular endothelium cadherin in PAH were significantly correlated with the hemodynamic severity of PAH, suggesting that endothelial MP might be reliable predictors of disease severity in PH. Corroborating these results, Bakouboula et al. have shown that MP containing active platelet tissue factor (thrombokinase) and CD105 (endoglin) were increased in patients with PAH compared to the control group [139]. Pro-coagulant MP were also linked to PH severity. They hypothesized that circulating MP in the pulmonary vascular bed might also contribute to lung injury via diverse pathways: impaired perfusion, vascular remodeling, leukocyte recruitment, and inflammatory response. MP containing active platelet tissue factor and CD105 may possibly be considered as PAH early biomarkers in the near future provided that other studies confirm the above findings and further characterize these MP and their involvement in PAH.

Given these encouraging results, research is still ongoing on MP in animal models to better assess their role and their potential as biomarkers. In order to characterize circulating MP during hypoxic PAH and to study their effects on endothelial function, the study of Tual-Chalot et al. has provided evidence that hypoxic circulating MP induce endothelial dysfunction in rat aorta and PA by decreasing NO production [140]. Furthermore, they showed that MP display tissue specificity.

Platelet-derived MP may influence vascular function via their involvement in leukocyte adhesion, thrombus formation and interaction with endothelium. The role of platelet MP in thromboxane A₂ (TXA₂) production was examined in rabbit PA [141]. TXA₂ is well known to be raised in atherosclerosis, which is a chronic vasculature disease in which oxidative stress plays a key role [142]. That study provided evidence that platelet MP act as a cellular source of TXA₂ in rabbit aorta and PA. Yet, this experimental approach demonstrated the ability of platelet MP and endothelial cells to participate in the transcellular conversion of arachidonic acid to the important vasoconstrictor TXA₂. Because TXA₂ is an interesting mediator in hypertension known to be deregulated in human PAH [143], that work has provided an insight into the role of platelet MP in vascular tone control.

Plasma von Willebrand factor

Plasma von Willebrand factor (vWF) is a large glycoprotein synthesized mainly in endothelial cells. The presence of dysfunctional endothelial cells in PAH has been suggested to be due to increased proteolysis and vWF release [144]. Veyradier et al. demonstrated that baseline vWF levels and vWF proteolysis were increased in ten patients with severe PAH [144]. Moreover, these alterations were reversible upon initiation of continuous epoprostenol infusion, and paralleled the improvement of hemodynamic measurements. More recently, Kawut and coworkers showed that increased vWF levels at baseline and follow-up were associated with worse survival in a cohort of 66 PAH patients [145].

D-dimer

D-dimer is a specific marker of cross-linked fibrin and is associated with microvascular thrombosis, an important component of PAH pathophysiology. D-dimer, as measured with an ELISA method, is increased in PAH patients and correlates with patients’ functional capacity and hemodynamic severity in iPAH [146, 147]. On the other hand, such correlation was not found in PAH associated with systemic sclerosis [148]. These data may suggest that microvascular thrombosis might not play an important role in the pathogenesis of PAH in patients with systemic sclerosis, or that microvascular thrombosis occurs in the systemic circulation independently of the pulmonary vascular remodeling in these patients. Moreover, D-dimer elevation occurs in diverse clinical situations, limiting its accuracy in PAH.

Cancer-shared PAH biomarkers

Inflammation state

GDF-15

Growth differentiation factor (GDF)-15, which belongs to the transforming growth factor (TGF)-superfamily of cytokines, was identified as an independent predictor of long-term mortality in iPAH [149]. That biomarker displays a tight correlation with echo-measured RV systolic pressure NT-proBNP plasma levels in systemic sclerosis-associated PAH. These results suggest that GDF-15 may have value as a biomarker, but larger studies on PAH patients are obviously required at this stage [150, 151].

Endothelin-1/endothelin-3

Endothelin-1 (ET-1) is a proliferative cytokine and a potent endogenous vasoconstrictor with remodeling properties. Plasma ET-1 levels have been shown to be increased in PH patients [152–155], and ET-1 protein as well as mRNA expression are enhanced in endothelial cells of affected vessels [156]. More importantly, synthetic ET receptor antagonists (e.g., bosentan) were shown to have beneficial effects in patients with various forms of PAH [157–159].

Rubens et al. [153] have found that active ET-1 and its precursor, big ET-1, correlate with disease severity and are good prognostic markers for PAH patients. However, ET-1 measurement is not easily achieved and encounters technical difficulties such as the extremely short half-life of the peptides (a few minutes) and their sensitivity to physiological and pathological factors [160]. Endothelin-3 (ET-3) is produced in many organs by various cells, including endothelial cells. Montani et al. [161] have demonstrated that the ET-1/ET-3 ratio is increased, whereas ET-3 plasma concentrations are decreased in PAH. ET-1 and ET-3 levels were correlated with hemodynamic and clinical markers of disease severity, suggesting that the ET-1/ET-3 ratio might be a novel prognostic factor in PAH.

LIGHT

Otterdal et al. [162] investigated thrombus formation and inflammation in PAH pathogenesis. They demonstrated that lymphotoxin-like inducible protein that competes with glycoprotein D for herpes virus entry mediator on T lymphocytes (LIGHT) serum levels, a platelet-derived ligand of the tumor necrosis factor superfamily, were increased in PAH compared to control subjects. Moreover, LIGHT levels are significantly related to mortality in PAH patients. These findings suggest that, since the prothrombotic effects of LIGHT in PAH involve endothelium-related mechanisms, these effects might be part of a common terminal pathway involved in PAH pathogenesis. However, further evidence is needed to confirm these hypotheses, and larger trials must be carried out before LIGHT can be considered as a reliable biomarker [162].

C-reactive protein

C-reactive protein (CRP), a marker of inflammation and tissue damage, has been demonstrated to be an efficient outcome predictor in PAH. Indeed, its level is significantly higher in PAH patients compared to control, independently correlating with severity and predicting mortality and clinical worsening. Moreover, normalization of CRP levels with therapy was associated with improvement in functional capacity, cardiac index and long-term survival [163, 164].

Oncogenes/tumor suppressors

A growing number of cancer hallmarks turn out to be present in PAH. Our recent review [165] reports the therapeutic value of proto-oncogenes/tumor suppressor proteins. These therapeutic targets might also have biomarker potential. Examples of the activation of oncogenic pathways in PAH include the increase in cytokines and growth factors such as epidermal growth factor (EGF), PDGF, cytokines (e.g., IL-6) [166], or peptide agonists such as ET-1 [167–170]. These factors can activate biological pathways such as proliferation, migration, survival and metabolism switches. Some cytokines are already considered as cancer biomarkers [171–174]. One might surmise that a cytokine such as IL-6 may also have value as a PAH biomarker. The main issue to be addressed to assess that hypothesis is the lack of specificity of many of these factors. However, cancer and PAH are easy to differentiate, given their different symptoms, which would bring more value in using proto-oncogenes as PAH biomarkers.

Pim1 proto-oncogene

A preclinical trial is currently ongoing to assess the potential of the Pim1 protoconcogene as a biomarker in PAH. As explained earlier, our current understanding of PAH physiopathology relies on NFAT and STAT3 activation. Thus, our group has shown that STAT3 is responsible for enhancing not only NFAT but also Pim1 expression [20]. In fact, the latter oncogene is required for NFAT activation in PAH. This means that, in PAH, NFAT cannot be activated in the absence of Pim1 and cannot deregulate Ca²⁺ homeostasis and other mechanisms responsible for PASMC proliferation and resistance to apoptosis found in distal PA.

Interestingly, Pim1 is not normally expressed in human tissues. Although Pim1 is also found in diverse types of cancer, Paulin et al. showed that Pim1 exhibits high specificity for the pulmonary vascular remodeling that characterize PAH [20]. In PAH, Pim1 is markedly overexpressed in the pulmonary vascular bed. Moreover, the activation of these pathways is also observed in circulating immune cells, possibly due to the inflammatory state of the lung. Indeed, our group has shown that it is possible to measure Pim1 expression in the buffy coat of PAH patients. More importantly, preliminary data suggest Pim1 blood levels are elevated in early disease and strongly correlate with the severity of the disease. Finally, the simplicity of the methodology used to measure Pim1 (from blood sample and urine) supports Pim1 as a promising biomarker in PAH [20].

Other oncogenes/tumor supressors

The p27 tumor suppressor is often downregulated in cancers harboring p53 mutations. Interestingly, overexpression of p27 decreases PASMC proliferation and prevents hypoxia-induced PH in vivo [175, 176]. Conversely, p53 deficiency in a hypoxia mouse model is associated with a more severe PAH, an increase in HIF-1α, and a loss of p21 expression. Furthermore, that tumor suppressor has been shown to have a protective role in a left-to-right shunt model, which strengthens its protective effect [165].

Yang et al. aimed to investigate the role of apoptosis in the PA remodeling of PH secondary to hypoxia and to determine the relative expression of selected genes [177]. Compared to a control group, the apoptotic index of the hypoxic group decreased significantly. Through the methods of in situ hybridization and RT-PCR, this group found that Bcl-2 expression was increased, whereas Bax was decreased significantly in the hypoxic group. The alteration in Bcl-2 and Bax expression induced by hypoxia plays an important role in PA remodeling, which is the main pathological change found in PH secondary to hypoxia [177].

Current understanding of cancer pathophysiology in apoptosis resistance is fundamental to develop efficient treatments, which might also be of help in PAH treatment. However, a number of studies are required to evaluate whether these oncogenes/tumor suppressors are quantifiable and have biomarker values. Indeed, a compounding factor is that these proteins are not secreted. Actually, unlike with cancer where the tumor is usually removed and in situ hybridization may be performed, there is no direct way to extract these polypeptides with PAH. Nonetheless, it may be thought that, as for Pim1, some of these factors might be found in circulating immune cells and, thus, in patient’s buffy coat. Moreover, proteomic technology and future improvements in microsurgery might facilitate the investigation of oncogenes/tumor suppressors in PAH.

miRNA

Small noncoding miRNA (21–23 nt) are now recognized as important regulators of gene expression and are involved in most physiological and pathological processes. Several studies are currently assessing whether miRNA expression in blood samples might be used to identify PAH progression and to help in its diagnosis. Indeed, miRNA can be secreted and may thus be found in fluid samples. Furthermore, as already discussed, inflammatory mediators are important actors in PAH. Our laboratory has demonstrated that the pathways involving NFAT, STAT3 and miR-204 described in PASMC (see below) may also be found in inflammatory cells, since altered levels of mRNA and oncogenes have been demonstrated in PH patients' buffy coats [14, 20].

Tumor suppressor miR-204

Our group has characterized the crucial implication of miR-204 in PAH pathogenesis. In fact, STAT3 is responsible for the decreased expression of miR-204, which is itself a STAT3 inhibitor. In PAH, the increase in circulating factors activates STAT3, and, therefore, downregulates miR-204. Furthermore, as said earlier, STAT3 enhances NFAT expression. Thus, via many pathways previously described [15, 178], NFAT leads to the proliferative and anti-apoptotic phenotype that is responsible for the vascular remodeling caused by PAH. This feedback loop between STAT3 and miR-204 might explain the resistance to inhibitors of circulating factor receptors such as imatinib (a tyrosine kinase inhibitor) and bosentan (an endothelin receptor antagonist). We have described the therapeutic value of miR-204 restoration, since nebulization with a miR-204 mimic reverses the pathology in the monocrotaline-PAH rat model [19]. The latter study also highlights the biomarker potential of miR-204. Indeed, the decrease in miR-204 expression is found specifically in the lung and in PAH models, and correlates with PAH severity in human buffy coat. Thus, although these results must be replicated in larger patient cohorts, miR-204 represents a promising biomarker that also correlates with disease severity.

miR-21 oncogene

As previously mentioned, PAH is characterized by an increase in circulating blood factors. A noteworthy observation is that TGF and bone morphogenetic protein (BMP) stimulation have been shown to rapidly induce miRNA-21 (miR-21) [179, 180]. The latter miRNA is significantly increased and correlates with poor prognosis in lung cancer [181].

Wang et al. [182] described the profile of miRNA expression in human arteries with arteriosclerosis obliterans. Among them, miR-21 was mainly found in arterial smooth muscle cells and was increased by >sevenfold in arteriosclerosis obliterans, which was related to HIF-1α expression. In cultured human arterial smooth muscle cells, cell proliferation and migration were significantly decreased by the inhibition of miR-21. These authors were also able to confirm that tropomyosin-1 is a target of miR-21 that is involved in the intracellular effects of miR-21. The study of Sarkar et al. [183] has investigated the role of miR-21 in hypoxia-induced PASMC proliferation and migration. The latter group showed that miR-21 expression increased by threefold in human PASMC after a 6-h hypoxia (3% O₂) and remained high (twofold) even after a 24-h hypoxia. They further showed that miR-21 is essential for hypoxia-induced cell migration. Protein expression of miR-21 target genes, i.e. programmed cell death protein-4, Sprouty 2, and PPARα, was decreased by hypoxia and in PASMC overexpressing miR-21 under normoxia, and was increased in hypoxic cells in which miR-21 had been knocked down. Their findings suggest that miR-21 plays a significant role in hypoxia-induced PAMSC proliferation and migration through the regulation of multiple gene targets.

Using PAH rat models (hypoxic and monocrotaline), Caruso et al. demonstrated miR-22 and miR-30 were downregulated, whereas miR-322 and miR-451 were significantly upregulated in both PAH models. Interestingly, miR-21 was selectively and consistently downregulated after monocrotaline injection, but not under hypoxia. They also described the downregulation of miR-21 expression in iPAH human samples [184]. These findings suggest that a substantial loss of miR-21 is associated with vascular remodeling in monocrotaline-exposed lungs. The authors suggested that reduced BMP signaling is possibly related to miR-21 downregulation in the monocrotaline model, which might be involved in the alteration of smooth muscle cell phenotype found in PAH.

Taken together, these studies highlight the importance of ascertaining the role of miR-21 in the different PAH models as well as defining the potential effect of miR-21 modulation in disease prevention. Undoubtedly, these data emphasize the importance of investigating miR-21 in the pathology of PAH.

miR-210 oncogene

miR-210 is the main hypoxia-sensitive miRNA involved in the regulation of the hypoxic response in tumor cells as well as tumor growth [185]. Its expression is induced by HIF-1α and is known to protect cancer cells from hypoxia-induced apoptosis, and to participate in mitochondrial impairment and cell cycle regulation, all features that are also found in PAH [19, 184, 186–188]. Nonetheless, miR-210 expression has not been found to be altered either in human or PAH animal models.

miR-145 tumor suppressor

The biomarker potential of miR-145 has already been assessed in cancer [189]. miR-145 is the most abundant miRNA in normal vascular walls, and is significantly decreased in pathological conditions of vascular walls with neointimal lesions and dedifferentiated vascular smooth muscle cells. Cell dedifferentiation is well known to play a crucial role in the pathological vascular remodeling process that occurs in several vascular pathologies, including PAH. Differentiation can be affected by several stimuli, including growth factors typically increased in PAH. Furthermore, Cheng et al. [190] demonstrated that its restoration leads to the loss of the pathological phenotype in an in vitro model of vascular smooth muscle cells as well as in an in vivo model of carotid angioplasty in rats. Hence, it might be very interesting to assess the biomarker value of miR-145 in PAH as well to determine whether it is also reduced in PAH patients’ buffy coat, and whether it correlates with the severity of the disease.

Conclusion

Being already considered as biomarkers in cancer, miRNA might be potential biomarkers in PAH in the near future. Since miRNA are very short sequences that are present at low concentrations in the blood, extraction inaccuracies as well as the variance of replicate values will constitute a serious methodological challenge [191]. Standardization of sample conservation, technical dosage and extraction are of utmost importance before translating the profiling of serum miRNA to clinical practice, lest such a promising and powerful tool be misused.

Metabolic biomarkers

Serum lactate dehydrogenase (LDH)

The Warburg effect is increasingly correlated with aggressiveness and poor prognosis in cancers [192]. Given that a similar metabolic disorder is also found in PAH, serum LDH might be a biomarker, although not specific, of the Warburg effect in PAH.

Urine metabolomic derived biomarker

A few teams are currently working on the use of metabolomics-derived biomarkers able to differentiate cancer from healthy patients, and correlating with severity. These groups have measured various metabolites in patients’ urine such as quinolinate, 4-hydroxybenzoate, and gentisate [193]. It might be of major interest to investigate whether specific differences are observed in the urine of PAH patients.

Oxidative stress biomarkers and PAH

Reactive oxygen species (ROS) are predominantly implicated in cell damage via the inhibition of protein, lipid and DNA normal functions. These highly reactive species may initiate several peroxidative reactions damaging the genetic cell apparatus, as well as cell membranes, proteins and lipids. ROS may be produced in the lung tissue in severe PH patients as a result of tissue hypoxia [194], ischemia [195] or via inflammatory cascade activation and increased production of inflammatory cytokines [196, 197]. Oxidative stress occurs when repeated external insults result in excessive ROS formation, which overwhelms antioxidant systems, thus creating an altered redox state favors oxidation. Oxidative stress is also considered as a cancer hallmark, but the present discussion will focus on the role of ROS markers in PAH.

Increases in oxidative stress and in lipid peroxidation products are associated with elevated PA pressure in different animal models of PH [198–200]. Altered ROS generation is believed to play a crucial role in the vascular responses observed under hypoxic conditions, such as deregulation of pulmonary vasomotor tone [201–203]. Moreover, some animal models studies have shown that ROS play a role in pathological RV remodeling. It is well established that increased PAH causes pressure overload of the RV, leading to RVH [204, 205]. Increased ROS generation and altered redox state are some of the critical features of the transition from hypertrophy to heart failure. Among various causes of hemodynamic and structural abnormalities of the decompensating RV are neurohormonal signaling (angiotensin-II, ET-1, aldosterone), natriuretic peptides, and inflammation, but also oxidative stress (including ROS and reactive nitrogen species) [206].

Uric acid

Uric acid is an endogenous free radical scavenger protecting cells from ROS and reactive nitrogen species damage. It has been demonstrated that uric acid levels are increased in PAH [207]. Voelkel et al. hypothesized that the site of uric acid production in PAH may be either the ischemic lung tissue or the ischemic RV, or both [207]. Uric acid serum levels correlate with hemodynamic and functional parameters [208, 209], and are related to survival in adult [33, 208] and pediatric [209] PAH. Uric acid is thus a biological serum marker of disease severity in PAH. However, uric acid serum levels are influenced by allopurinol, diuretics and hypoxemia and display significant interindividual variability, limiting its accuracy as a biomarker in PAH.

Isoprostanes

Isoprostanes are products of the peroxidative attack of membrane lipids and are present at nanomolar concentrations in the blood of normal individuals [210, 211]. In several disease states, and particularly pulmonary diseases such as PAH, isoprostane levels are increased [199, 212]. Moreover, given their potent and multiple biological activities, isoprostanes have been thought to account for many of the pathophysiological processes found in PAH (e.g., decrease in endothelial relaxing factors) [213]. A key question about isoprostane analysis is which isoprostane should be analyzed among the 64 distinct isoprostanes that can be generated from arachidonic acid [214]. The 5- and 15-series F₂-isoprostanes are produced in closely equal amounts in vivo, whereas the 8- and 12-series F₂-isoprostanes are produced in lower yields [215]. Considering their relative stability, the availability of inexpensive assays for quantification, the easy collection of blood and/or urine samples, and the direct relationship between isoprostane accumulation and oxidative stress intensity has made these “markers” of oxidative stress optimal biomarkers of lung disease states [213]. Hence, isoprostanes might also be potentially useful markers for PAH.

Serotonin (5-HT)

An elevation of circulating peripheral serotonin (5-HT) has been shown to occur upon PAH development both experimentally and clinically [216, 217]. This vasoactive amine may also produce ROS [218–220] and can induce protein carbonylation in PASMC [221]. Moreover, 5-HT stimulates PASMC proliferation, PA vasoconstriction and local microthrombosis [222]. In addition, studies have demonstrated that inhibition of 5-HT receptors and 5-HT transporters is able to delay PAH development and extend rat survival [223, 224]. The therapeutic value of 5-HT also highlights its interest as a biomarker. Thus, given the evidence that PAH patients have increased plasma 5-HT levels, and that this hormone may be considered as a pro-oxidant, it represents an attractive and easily assayed biomarker for PAH diagnosis and follow-up.

Derived nitric oxide products

NO may act as a signaling molecule, a toxin, and a pro-oxidant as well as a potential antioxidant [225]. It has been proposed to act as a pro-oxidant when reacting with superoxide anion (O₂ ⁻), thereby generating peroxynitrite [226], or by itself at high concentrations [227]. NO is a pulmonary vasodilator produced and released by the endothelium. Among its main functions are vascular tone regulation [160], platelet aggregation and inhibition of vascular smooth muscle cells proliferation. A reduction in available NO occurs in ROS-overproducing cells. The alteration of eNOS expression has also been associated with systemic and pulmonary hypertension [228–230]. In theory, low plasma NO might be a marker of endothelial dysfunction and possibly PAH. However, NO is too unstable to be measured in gaseous form in the blood [160]. However, NO treatment has already been tested, without much success, but NO-products and NO itself may help PAH diagnostic and follow-up.

HIF-1

HIF-1, which contributes to oxidative stress, is involved in the PAH pathogenesis, raising a genuine interest in this protein. HIF-1 is a transcription factor responsible for mediating physiological responses to hypoxia, and is found in a variety of organisms. The HIF-1α subunit, also called “oxygen sensor”, may be accumulated under normoxic conditions as a result of various stimuli such as ROS, growth factors, oncogenic activation and NO [231, 232]. The effect of NO on HIF-1α expression is reciprocal and depends on the chemical structure and concentration of the NO donor [233]. HIF-1 alteration may also indicate oxidant damage [198], and, thus, both in vitro and in vivo studies have demonstrated that HIF-1 protein levels are inversely proportional to O₂ levels in cell media [234, 235]. Furthermore, HIF-1 has long been believed to increase “early-response gene” transcription, which modulates vascular remodeling in the lung as well as cardiovascular function [236, 237]. Thus, HIF-1 is a promising candidate as a novel biomarker and/or as a target for PAH treatment. Moreover, its pathway actors should be carefully studied in order to find their biomarker potential.

NADPH oxidase/superoxide dismutase (SOD)

Using monocrotaline PAH model, Redout et al. have demonstrated that NADPH oxidase and mitochondria-derived ROS production increase RV failure [204]. Their findings indicate that mitochondrial SOD-1 and SOD-2 mRNA levels were decreased, and nitrotyrosine staining confirmed the presence of oxidative stress. Importantly, Archer et al. have identified mitochondrial SOD-2 methylation as a potential epigenetic mechanism for PAH in the Fawn-Hooded rat PAH model [238]. It may be relevant to investigate whether SOD-2 is also downregulated in human PAH. Interestingly, SOD-2 epigenetic downregulation impaired H₂O₂-mediated redox signaling, activating HIF-1α and creating an apoptosis-resistant state. Importantly, SOD-mimetic therapy reversed PAH in vivo, reinforcing the involvement of HIF-1. The study by Archer et al. also demonstrated that an increase in SOD-2 can restore mitochondrial function, inhibit proliferation and increase PASMC apoptosis in vitro. This study was the first demonstration of an epigenetic basis for PAH. These studies demonstrate that SOD-2 might be a powerful tool for PAH treatment and might have potential diagnostic value in PAH in the future, directly or via proteins functionally related to SOD-2.

Conclusion

There is an urgent need for alternative biomarkers that are safer and more reliable than those that are currently available. Because ROS may promote vasoconstriction, vascular smooth muscle cells proliferation and vascular remodeling in many PH forms, research should be pursued towards a thorough understanding of its critical role. There is mounting evidence of a cross-talk between PAH pathogenesis and oxidative stress, and monitoring oxidative stress signs and damage may help to improve PAH diagnostic and prognosis. Additionally, these data may provide a rationale for developing therapeutic strategies in RV failure due to PAH.

Other biomarkers

A renal-derived biomarker: creatinine

Renal dysfunction is well known to predict mortality in many cardiovascular diseases. Not surprisingly, Shah et al. demonstrated that higher serum creatinine levels and decreased glomerular filtration were associated with hemodynamic severity and worse survival in PAH, results that were also reported by other groups [126, 164].

CNS-derived: sympathetic nerve activity

Velez-Roa et al. [239] showed that PAH patients had increased muscle sympathetic nerve activity compared to controls, especially in advanced PAH. Moreover, they demonstrated that sympathetic hyperactivity was partially chemoreflex-mediated. Similarly, McGowan et al. documented muscle sympathetic nerve activity burst frequency was higher in PAH patients and was inversely correlated to the low-frequency spectral component of heart rate variability [240]. The authors suggested that reduced heart rate variability and sympathetic nervous system excitation in PAH patients might increase the likelihood of sudden death or accelerate progression to RV failure. This hypothesis was supported by recent experiments documenting the relationship between muscle sympathetic nerve activity and disease severity and prognosis [254]. This deserves prospective evaluation in larger cohorts in order to consolidate the status of heart rate variability spectral power and sympathetic hyperactivity as useful noninvasive surrogate markers of disease severity in PAH.

Proteomic-derived biomarkers

Improvement in proteomic tools has led to the assessment of proteomic-derived PAH biomarkers. Indeed, Zhang et al. have suggested that simultaneous analysis of the expression of ten protein spots may distinguish patients with PAH from normal controls [241]. Nine proteins and their isoforms were significantly different in the serum of PAH patients relative to controls, including leucine-rich α-2-glycoprotein, haptoglobin precursor, albumin isoform-2, transferrin variant, C3 complement, hydroxypyruvate reductase isoform-1, RAF1, fibrinogen isoform γ-A, and fibrinogen isoform γ-B. In addition, significant and important associations between LRG levels and functional class FC (r = 0.71, p < 0.01) and cardiac output (r = −0.65, p < 0.01) were shown.

Geraci et al. have also hypothesized that the gene expression pattern found in PAH patient lung tissue has a characteristic profile compared to healthy lungs [242]. Using oligonucleotide microarray technology, they characterized gene expression patterns in the lung tissue obtained from six PAH patients—including two patients with the familial PAH form—as compared to six matched healthy patients. All PAH lung tissue samples had a decreased expression of several genes encoding various kinases and phosphatases, whereas the expression of several oncogenes and genes encoding ion channel proteins was increased. Alterations in the expression of TGF-β receptor III, BMP2, MAP kinase kinase-5, RACK-1, apolipoprotein C-III, and laminin receptor-1 were found exclusively in PAH patients. The microarray gene expression technique is a very useful molecular tool providing novel pertinent data that may bring a better understanding of the pathobiology of different PH clinical phenotypes. Unfortunately, the methodology used in such studies cannot yet be used in the clinical practice due to its high cost. Nevertheless, it remains a promising new approach that should be kept in mind with the incessant technical improvements in proteomic analysis, and might also lead to the discovery of more easily assessable serum or urine biomarkers. Taken together, these results indicate that proteomic analysis might be helpful in PAH diagnosis.

Red blood cells distribution width

Red blood cell distribution width (RDW) reflects red blood cell size variability and is commonly reported in automated complete blood counts. The relationship between RDW and PAH was investigated in a prospective study of 162 patients [243]. The authors found RDW to be related to disease severity and demonstrated that it is independently associated with mortality in PAH. Moreover, RDW performed better as a prognostic indicator than NT-proBNP and added important prognostic value over NT-proBNP measurement and exercise capacity. The authors concluded that RDW and NT-proBNP might be used in combination to improve PAH prognosis.

Immune system-derived biomarkers: antibodies

Specific antibodies such as anti-Scl-70 may also be used as biomarkers to assess the implications of other pathologies [45]. Around 40% of PAH patients have positive but low antinuclear antibody titers [244]. Other antibodies seem to be promising, such as anti-endothelial cell antibodies [245] or even antifibroblast antibodies [246]. Indeed, considering the ease with which antibodies can be tested, it might be of great interest to assess the presence and level of such antibodies in the serum of PAH patients. This might represent easily measurable clinical parameters of potential usefulness for future diagnosis and prognosis.

Hyponatremia

Growing evidences demonstrate that hyponatremia is strongly associated with advanced RV dysfunction, poor survival and WHO FC in PAH [247]. Since it is already used and it is easy to assess, hyponatremia could represent a powerful and independent biomarker. However, the mechanism of hyponatremia remains elusive. Thus, even though renal dysfunction is frequently seen in PAH patients often treated with diuretics, Forfia et al. hypothesized that hyponatremia may be a result of neurohormonal activation (increased in PAH) possibly caused by more advanced RV dysfunction [247].

High-density lipoprotein cholesterol (HDL-C)

It is well known that HDL-C is associated with poor prognostic in cardiovascular diseases, but, interestingly, decrease in this factor is also associated with higher mortality and clinical worsening and outcome in PAH. Furthermore, Heresi et al., have demonstrated that a decrease in HDL-C does not appear to be caused by underlying cardiovascular risk such as diabete mellitus, BMI or other cardiovascular risk factors. Moreover, HDL-C level is even lower in PAH patients than in a “high-risk population” (patients with coronary disease, systemic hypertension and diabetes, conditions associated with lower HDL-C) [248].

Conclusion

All the biomarkers discussed in this part have been summarized in Fig. 2. As shown above, current circulating biomarkers may be helpful in the diagnosis of PAH, but most of them lack specificity, and, unfortunately, many appear only at severe stages of the disease. For these reasons, PAH experts now consider the identification and validation of a clinically relevant biomarker that is representative of PAH physiopathology as a high priority objective [89, 233, 249–252]. Cancer-related biomarkers are promising tools but may lead to misdiagnosis. However, it should be kept in mind that, since symptoms of cancer and PAH are very different, the use of cancer biomarkers as PAH biomarkers might well be feasible in practice. Moreover, the effectiveness of cancer diagnosis is constantly improving. Likewise, oxidative stress-derived biomarkers might be of some value for PAH diagnosis and prognosis purposes. Finally, technological improvements of proteomic tools may highlight the implication of new actors in PAH physiopathology and lead to the discovery of novel biomarkers and therapeutical targets. However, many of the promising biomarkers already described must still be assessed in larger patient cohorts to be validated. A coordination between large PAH treatment centers should be expected to include the standardization of study protocols and tools in order to increase the number of patient in such studies, and might significantly contribute to accelerate the validation of these biomarkers. In addition, the combination of different biomarkers might be highly relevant.

Fig. 2.

Pulmonary endothelium, circulating cells and microparticles derived biomarkers. A large number of disorders are accountable for PAH development and sustainability, which are also accountable for distinct potential biomarkers. Among these, pulmonary artery smooth muscle cells have a cancer like phenotype (miRNA, Pim1), metabolic disorders (lactacte dehydrogenase, LDH) and increased oxidative stress (isoprostane, uric acid). Furthermore, inflammation is enhanced, inducing production of biomarkers, such as Endothelin (ET), Interleukin, Englodin (Eng), GDF-15 and LIGHT. These markers could be assessed in blood samples

General conclusion

As of today, PAH diagnosis is still an exclusion diagnosis involving a long period between the discovery of symptoms and diagnosis, resulting in a long delay before initiating any treatment. Unfortunately, current imaging and circulating biomarkers are only reliable in late stage disease, limiting their utility for screening at-risk populations. Hence, improvements are much needed to improve diagnostic strategies. Nevertheless, imaging biomarkers represent crucial tools in determining the etiology of PAH, excluding other types of PH and providing information regarding RV function, disease severity and response to therapy. The very rapid improvement in imaging techniques seen in the previous decades strongly suggest that imaging by CT, MRI and even echo will be further improved, thereby decreasing the requirement for invasive techniques such as RHC and allowing an earlier and more accurate diagnosis. In addition, circulating biomarkers show promises in PAH. Technical improvements are also implemented in biochemical engineering, and the detection of very small amounts of protein is now possible, thus opening new avenues for circulating biomarkers. More importantly, studies on biomarkers including oncogenes and miRNA as well as oxidative stress-related proteins may increase the understanding of PAH pathophysiology and identify novel therapeutic targets.

Despite the technical improvements, the prospective validation of new imaging and circulating biomarkers relies on multicenter large-scale trials and standardization of the techniques. While no “perfect biomarker” is likely to be discovered any time soon, one might surmise that a combination of biomarkers will together allow earlier diagnosis, more accurate prognostication and improved long-term outcomes in PAH.

Abbreviations

5-HT

Serotonin or 5-hydroxytryptamine

BMP

Bone morphogenetic protein

BMPRII

BMP receptor 2

BNP

Brain natriuretic peptide

cGMP

Cyclic guanosine 3′5′-monophosphate

CT

Computed tomography

CTEPH

Chronic thromboembolic pulmonary hypertension

DCE

Delayed contrast enhancement

Echo

Echocardiography

EGF

Epidermal growth factor

eNOS

Endothelial nitric oxide synthase

ET-1/3

Endothelin 1 or 3

FC

Functional class

HDL-C

High-density lipoprotein cholesterol

IL-6

Interleukin 6

iPAH

Idiopathic PAH

LIGHT

Lymphotoxin-like inducible protein

LV

Left ventricle

miR-X

MicroRNA-X

miRNA

Micro RNA

MP

Microparticles

MRI

Magnetic resonance imaging

NFAT

Nuclear factor of activated T cells

NO

Nitric oxide

NT-proBNP

N-terminal pro-brain natriuretic peptide

OPN

Osteopontin

PA

Pulmonary arteries

PAAT

Pulmonary artery acceleration time

PAH

Pulmonary arterial hypertension

PAP

Pulmonary arterial pressure

PASMC

Pulmonary artery smooth muscle cell

PDE5

Phosphodiesterase 5

PDGF

Platelet-derived growth factor

PECAM

Platelet endithelial cell adhesion molecule

PET

Position emission tomography

PH

Pulmonary hypertension

Pim1

Proviral integration site for Moloney murine leukemia 1

PPAR

Peroxisome proliferator activated receptor (gamma γ or alpha α)

Q

Pefusion

RDW

Red blood cell distribution width

RHC

Right heart catheterization

ROS

Reactive oxygen species

RV

Right ventricle

RVH

Right ventricle hypertrophy

SPECT

Single photon emission computed tomography

SOD

Superoxyde dismutase

STAT3

Signal transducer and activator of transcription 3

TAPSE

Tricuspid annular plane systolic excursion

TGFβ

Tranforming growth factor beta

TXA₂

Thromboxane A₂

V

Ventilation

vWF

Plasma von Willebrand factor

WHO

World Health Organization