In situ expression of Bcl-2 in pulmonary artery endothelial cells associates with pulmonary arterial hypertension relative to heart failure with preserved ejection fraction

Abstract Abstract

We have previously reported that pulmonary artery endothelial cells (PAECs) can be harvested from the tips of discarded Swan-Ganz catheters after right heart catheterization (RHC). In this study, we tested the hypothesis that the existence of an antiapoptotic phenotype in PAECs obtained during RHC is a distinctive feature of pulmonary arterial hypertension (PAH; World Health Organization group 1) and might be used to differentiate PAH from other etiologies of pulmonary hypertension. Specifically, we developed a flow cytometry-based measure of Bcl-2 activity, referred to as the normalized endothelial Bcl-2 index (NEBI). We report that higher NEBI values are associated with PAH to the exclusion of heart failure with preserved ejection fraction (HFpEF) and that this simple diagnostic measurement is capable of differentiating PAH from HFpEF without presenting addition risk to the patient. If validated in a larger, multicenter study, the NEBI has the potential to assist physicians in the selection of appropriate therapeutic interventions in the common and dangerous scenario wherein patients present a clinical and hemodynamic phenotype that makes it difficult to confidently differentiate between PAH and HFpEF.

Pulmonary hypertension (PH) is a clinical state characterized by increased blood pressure in the pulmonary vascular system, including the pulmonary artery, pulmonary vein, or pulmonary capillaries. PH includes a diverse variety of etiologies but is defined as a mean pulmonary artery pressure exceeding 25 mmHg at rest. Because of the multiple, sometimes unrelated etiologies that can lead to PH, the World Health Organization (WHO) has adopted the clinical classification system developed at the Dana Point 2008 World Symposium on Pulmonary Arterial Hypertension and updated in 2013 that classifies the disease into five distinct groups.¹

Pulmonary arterial hypertension (PAH) is a distinct and particularly ominous subset of PH, categorized and referred to as WHO group 1 PH. This disease state is characterized by intimal proliferation, medial thickening, and microvascular pruning. The distinct pathophysiology underlying group 1 disease has informed the emergence of three distinct classes of pharmacotherapies, providing hope in the face of what had previously been a rapidly fatal disease. While effective in group 1 disease, several of these pharmacotherapies are strongly contraindicated in several other classes of PH, particularly PH due to left heart disease (WHO group 2).

The management of all classes of PH involves periodic right heart catheterization (RHC), both for definitive diagnosis of the disease and for evaluating response to therapeutic measures. RHC provides a number of hemodynamic measurements, but these are often insufficient for categorizing a patient’s PH into a distinct WHO group in the absence of additional clinical information. In most cases, the composite clinical phenotype of a patient is sufficient to clearly categorize the etiology of PH.

Nevertheless, there are frequent exceptions to this principle, and in many instances the diagnosis of PAH cannot be made with reasonable certainty, raising the vexing clinical conundrum of whether to withhold PAH treatments and accept the risk of suboptimal therapy or to prescribe PAH treatment and risk the possibility of unmasking subclinical left heart failure and subacute pulmonary edema that can result from inappropriate use of PAH pharmacotherapies. One of the best examples of this challenge is PH associated with heart failure with preserved ejection fraction (HFpEF), a WHO group 2 entity, which often presents a clinical and hemodynamic syndrome indistinguishable from idiopathic PAH (IPAH) and has thus been described as the “great masquerader” of PAH.

In 2013, our group described a novel methodology for harvesting pulmonary artery endothelial cells (PAECs) from the elastic balloons of Swan-Ganz (SG) catheters after RHC.² Whereas our initial approach to purifying PAECs from contaminating leukocytes was selective culture, we have more recently refined a methodology for selectively enriching PAECs by magnetic beads decorated with antihuman CD146 antibodies. Combining this with improved catheter collection and processing techniques, we are now able to perform molecular testing on PAECs maintained at 4°C within 2 hours of their removal from the patient’s body. We submit that this workflow allows a virtual in situ snapshot of the cellular biology of the pulmonary vascular endothelium in patients undergoing RHC.

This brief report describes a novel diagnostic tool that we refer to as the normalized endothelial Bcl-2 index (NEBI), a flow cytometery–based analysis of the level of Bcl-2 protein expression in PAECs harvested from SG catheters after RHC. We show that in a limited clinical study, high NEBI values are associated with PAH, whereas lower values are associated with HFpEF. Further, we report that this test can be performed on SG catheters stored overnight on ice without an impact on the results, suggesting that a “clip-and-ship” workflow may be feasible to allow the performance of this test in a central facility.

Methods

Protection of human subjects

All studies described in this article were approved by the Institutional Review Board of the Allegheny Health Network. All patients provided informed consent before the administration of sedative agents and were deidentified by being assigned a study number. The database containing patients’ names linked to study numbers was kept on a secure server and was accessible only by the clinical coordinator (GL). A separate database containing relevant clinical information was kept, indexed to study number, and was used to populate Table 1.

Table 1.

Patient demographic and clinical characteristics

Characteristic	PAH (n = 13)	HFpEF (n = 9)	P values	Test
NEBI	6.90 ± 4.2	3.05 ± 2.0	0.01	t test, unequal variance
Age, years	61.9 ± 14	68.8 ± 10	0.24	Log transformed, t test, equal variance
Women	12 (92.3)	7 (77.7)	0.54	χ2a
NYHA classb			0.99	χ2a
III	9 (69)	7 (78)
II	3 (23)	2 (22)
I	1 (8)	0
WHO group 1 PAH subgroup
Idiopathic PAH	6 (46)	NA
Associated PAH
CTD-scleroderma	3 (23)	NA
Portopulmonary	2 (15)	NA
Congenital	2 (15)	NA
Previously diagnosed	11 (85)	NA
Newly diagnosed	2 (15)	NA
BMI	28.5 (22–34)	35 (33–37)	0.039	Wilcoxon (Mann-Whitney) rank sum
Hemodynamics
Mean pulmonary arterial pressure, mmHg	32 (27–45)	36 (24–40)	0.9	Wilcoxon (Mann-Whitney) rank sum
Mean right arterial pressure, mmHg	9.08 ± 3.6	12.0 ± 7.7	0.3	t test, equal variance
Pulmonary wedge pressure, mmHg	11.7 ± 3.2	16.7 ± 7.3	0.08	t test, unequal variance
Pulmonary vascular resistance, dyn·s/cm5	312 ± 160	202 ± 62	0.043	t test, unequal variance
Cardiac output	3.66 (2.9–4.1)	2.96 (2.6–3.2)	0.038	Wilcoxon (Mann-Whitney) rank sum
Other functional parameters
REVEAL score at time of RHC	8.3 ± 2.5	NA
6-minute walk distance, m	328 (186–398)c	244 (148–301)c	0.25	Wilcoxon (Mann-Whitney) rank sum
Comorbidities
Diabetes	3 (23)	6 (67)	0.079	χ2a
Hypertension	3 (23)	8 (89)	0.008	χ2a
Sleep apnea	2 (15)	6 (66)	0.026	χ2a
Coronary artery disease	0	3 (33)	0.055	χ2a
Atrial fibrillation	1 (8)	5 (55)	0.023	χ2a
COPD	2 (15)	0	0.49	χ2a
Medications
PDE5i	8 (61)	0
ERA	5 (38)	0
Prostacyclin	5 (38)	0
Riociguat	1 (8)	0
Macitentan	1 (8)	0
Beta blocker	0	5 (55)
ACEI/ARB	0	3 (33)
Aldosterone antagonist	0	1 (11)
Digoxin	6 (46)	2 (22)
Continuous O2	3 (23)	0
Nighttime O2	1 (8)	2 (22)
Triple therapy	3 (23)	0
Dual therapy	4 (30)	0
Monotherapy	6 (46)	0
Imaging
LAVI	26 (20–36)d	36 (32–52)	0.11	Wilcoxon (Mann-Whitney) rank sum
Interventicular septal thickness, mm	11 (8–12)	11 (9.5–13)	0.58	Wilcoxon (Mann-Whitney) rank sum
Posterior wall thickness, mm	9.5 (7.5–12)	10.5 (9.5–13)	0.31	Wilcoxon (Mann-Whitney) rank sum
Left ventricle ejection fraction	67.5 (65–75)	60 (58–65)	0.026	Wilcoxon (Mann-Whitney) rank sum
TAPSE, mm	21 (20–30)	20 (15–23)	0.34	Wilcoxon (Mann-Whitney) rank sum
Mitral E/A	1 (0.7–1.1)	1.15 (0.86–1.5)	0.67	Wilcoxon (Mann-Whitney) rank sum
Mitral valve deceleration time, ms	219 (172–313)	209 (144–271)	0.48	Wilcoxon (Mann-Whitney) rank sum

^NoteUnless otherwise noted, data are displayed as mean ± SD, number (%) of patients, or median (25th–75th percentile). ACEI/ARB: angiotensin-converting enzyme inhibitor/angiotensin-receptor blocker; BMI: body mass index; COPD: chronic obstructive pulmonary disease; CTD: connective-tissue disease; ERA: endothelin receptor antagonist; HFpEF: heart failure with preserved ejection fraction; LAVI: left atrial volume index (calculated as mL/m²); mitral E/A: mitral ratio of peak early to late diastolic filling velocity; NA: not applicable; NEBI: normalized endothelial Bcl-2 index; NYHA: New York Heart Association; PAH: pulmonary arterial hypertension; PDE5i: phosphodiesterase 5 inhibitors; REVEAL: Registry to Evaluate Early and Long-term PAH Disease Management; RHC: right heart catheterization; TAPSE: tricuspid annular plane systolic excursion; WHO: World Health Organization.

^aThe Fisher exact P value was used.

^bIndividual classes were not evaluated, because there was no overall difference; P = 0.99.

^cn = 12 for PAH and n = 3 for HFpEF; not all patients had 6-minute walk tests completed.

^dn = 11; PAH patients with the subtype “congenital” were removed from the LA volume index category.

Study design

The study was designed to screen all patients undergoing RHC in the cardiac catheter laboratory of Allegheny General Hospital between August 20 and October 28, 2015, and to recruit those patients with an established diagnosis of PAH or HFpEF. Every Friday, the catheter lab schedule was provided to MJP and RLB, and patients meeting the specified criteria were identified. A clinical coordinator (GL) served as the interface between the research and clinical laboratories, coordinating consent with the operating physician and performing sample collection. During this time period, 203 patients were screened, catheters were collected from 50 patients, and 22 patients ultimately met the study criteria; data for those patients are shown in Table 1.

Inclusion/exclusion criteria

Only patients with a previous diagnosis of PAH or PH related to HFpEF were included. All patients were enrolled from the advanced heart failure and pulmonary vascular disease clinic directed by RLB. Patients were originally identified as having PAH with the standard criteria identified in the European guidelines.³ Patients with HFpEF were also identified with standard clinical criteria.⁴

After enrollment was completed, an expert on PAH (RLB) who was blinded to NEBI measurements reviewed the electronic health record to independently to confirm each enrolled patient’s diagnostic group. The blinded reviewer was asked to identify enrolled patients as having either PAH or HFpEF on the basis of clinical characteristics, as displayed in Table 1. Reviewers were then asked to further characterize patients if they had at least 3 risk factors for diastolic dysfunction based on their echocardiograms and RHCs completed at time of original diagnosis and also those performed closest to the time of PAEC isolation. The respective imaging and catheterizations were evaluated for evidence of precapillary, postcapillary, or mixed physiology. Diastology, presence of left ventricular hypertrophy, and left atrial volume index (LAVI) were assessed on echocardiography, and the pulmonary capillary wedge pressure (PCWP), transpulmonary pressure gradient, and diastolic pressure gradient, as well as the absence or presence of a positive fluid challenge at time of catheterization, were conjointly used to confirm the diagnosis of PAH or HFpEF. In those instances where the blinded reviewer disagreed with the diagnosis on record, a third blinded reviewer (AR, SM) rendered an opinion. A simple majority was then used to place the patient into their respective category. This occurred in only 1 patient with a previous diagnosis of PAH.

Catheter collection and processing

At the conclusion of the RHC procedure, the operating physician extracted the catheter from the transcutaneous sheath and placed it in ice-cold saline in a 50-mL plastic tube, swirling briefly to remove rinse residual blood. The catheter was then transferred to a 15-mL plastic tube containing 5 mL of ice-cold endothelial cell (EC) growth media (Cell Applications, San Diego, CA), cut 7–10 cm from its distal end, and allowed to drop into the media, and the tube was sealed. The tube containing the catheter tip was immediately transferred to the laboratory in an ice bucket for further processing.

PAECs were removed from the SG catheter via a proprietary method, adapted from that previously reported,² that uses gentle agitation and scraping of the balloon in the presence of Accutase (Innovative Cell Technologies, San Diego, CA) and were resuspended in 300 μL of EC growth media. The cells collected were incubated for 15 minutes at 4°C with antihuman CD146 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). After incubation, cells were spun down once again, the supernatant was discarded, and the cells were resuspended in 1 mL of growth media. The sample was then applied to a Miltenyi LS positive selection column under gravity flow. Once the flow-through was fully exuded from the column, the column was washed 3 times with 3 mL of EC growth media under gravity flow. After the final wash, the column was removed from the magnet, and 5 mL of growth media was plunged through column to exude the sample.

Measurement of NEBI

After processing, PAECs were fixed in 4% paraformaldehyde for 30 minutes, spun down, and permeabilized by placing them in 50 μL rinsing buffer (phosphate-buffered saline containing 0.1% Triton and 0.05% bovine serum albumin) for 5 minutes. Cells were blocked with 5 μL of purified mouse immunoglobulin G (eBioscience, San Diego, CA) for 10 minutes. After blocking, 5 μL of Alexa Fluor 647 anti-human CD146 antibody (Clone P1H12, BD Biosciences) and 5 μL of FITC-anti-Bcl-2 antibody (Clone 100, Life Technologies, Carlsbad, CA) were added to the cells, and incubation was performed in the dark for 30 minutes. Cells were then centrifuged at 500 g for 5 minutes and washed 3 times with Hank’s balanced salt solution. MACSquant running buffer (200 μL; Miltenyi Biotec) was added, and the sample was transferred to flow cytometry tubes for MACSquant analysis. Unstained cells were used as negative control and single antibody-stained cells as positive control. The individual performing the NEBI measurement was blinded to the patient’s medical condition and diagnostic category, using only the study number to identify the sample and report results.

Echocardiographic measurements

Standard 2-dimensional and Doppler transthoracic echocardiograms were analyzed offline with digital analysis software (Syngo Dynamics, Siemens) by a single research echocardiographer (AR), who was blinded to the clinical classification of the patient and the NEBI. Left and right ventricular size and function, as well as left atrial volumes, were measured according to American Society of Echocardiography guidelines.

Statistical analysis

Univariate comparisons of NEBI, mean pulmonary artery pressure (mPAP), pulmonary vascular resistance (PVR), and systemic vascular resistance (SVR) between PAH and HFpEF patients were conducted with t tests or Mann-Whitney U tests. Spearman correlations were used to assess associations between NEBI, mPAP, PVR, and SVR. Univariate logistic regression analyses were used to determine the association of NEBI, mPAP, PVR, and SVR with the likelihood of PAH diagnosis. Multivariate logistic-regression analyses were separately adjusted to determine whether the association of NEBI with PAH was attenuated by mPAP, PVR, or SVR. A nonparametric analysis of variance (Kruskal-Wallis) test was used on NEBI from all groups in Figure 2 (including PAH subgroups), indicating significant differences. Post hoc tests failed to show any differences between any of the groups after P values were adjusted for 10 multiple comparisons. All analyses were performed in Stata (ver.12.1; StataCorp, College Station, TX). The level of statistical significance was set at a 2-sided P value of <0.05.

Figure 2.

Average normalized endothelial Bcl-2 index values for each subcategory of World Health Organization group 1 pulmonary hypertension versus heart failure with preserved ejection fraction (HFpEF). Error bars show SEM.

Results

Table 1 lists characteristics of the research subjects as well as NEBI measurements. Specifically, it should be noted that despite equivalent mPAPs, those with HFpEF were older, had more comorbidities associated with left heart disease, and tended to have higher PCWPs and imaging features consistent with impaired diastology and to have the higher left atrial volume index (LAVI) consistent with pulmonary venous congestion. This provides strong support for two central concepts important for this analysis: (1) that our clinical phenotyping was accurate and (2) that the two cohorts were well representative of the disease states we are trying to differentiate with the NEBI.

Figure 1 shows a scatter plot of NEBI values for the PAH and HFpEF groups. NEBI values were normally distributed, with equal variance; a t test was used to compare NEBIs for PAH (mean: 6.90) and HFpEF (mean: 3.05) patients, and the means were significantly different (P = 0.01). Logistic regression on clinical state (PAH vs. HFpEF) indicated an odds ratio of 1.71 (95% confidence interval: 1.0–2.9), meaning that for every unit increase of NEBI, the odds of having PAH increased by 71% (P = 0.04). When mPAP was adjusted for, NEBI (odds ratio: 1.76 [1.0–3.0], P = 0.04) was independently associated with PAH, but the association was attenuated when PVR was adjusted for (P = 0.09). Greater power is needed to fully adjudicate these potential associations.

Figure 1.

Scatter plot of normalized endothelial Bcl-2 index (NEBI) values for patients classified as having pulmonary arterial hypertension (PAH; i.e., World Health Organization group 1 pulmonary hypertension) or heart failure with preserved ejection fraction (HFpEF). Each symbol represents an individual patient. To the right of each scatter plot, the average value for NEBI ± SD is plotted.

Figure 2 shows NEBI values for the subclassifications of PAH. We noted that NEBI values were highest in the 2 portopulmonary hypertension patients, and in these 2 patients the trend was for NEBI to increase with mPAP. The same score-pressure relationship was observed in the 2 congenital heart disease patients. The relationship of NEBI and mPAP can be addressed only in a much larger study, and it is notable that there was no discernible relationship between NEBI and mPAP in HFpEF patients.

We used a nonparametric Kruskal-Wallis rank test to determine whether the NEBI values were significantly different between any of the groups (P = 0.049). We then ran a post hoc test for multiple-group comparisons, using a Bonferroni-adjusted P value (P = 0.005) to account for all possible comparisons (10) of the groups in Figure 2. We found that NEBI values for IPAH were greater than those for HFpEF (P = 0.0085) and that those for portopulmonary hypertension were greater than those for HFpEF (P = 0.006). Although not significant when compared to the Bonferroni-adjusted P value, these trends are very promising, and our approach to the multiple-comparisons issue was very conservative. Therefore, a larger study based on a prospective design is needed to definitively show that NEBI can specifically differentiate IPAH from HFpEF.

Discussion

Our results demonstrate the potential utility of our NEBI as a rapid diagnostic measurement that may be capable of differentiating PAH from HFpEF through harvesting of cells from SG catheters that would otherwise be discarded after RHC. The NEBI process is performed on catheters only after the RHC procedure is complete, at the time that the catheter would otherwise be discarded. Thus, it presents no additional risk to the patient while adding the opportunity to undertake molecular profiling of the pulmonary vascular endothelium in situ.

We describe our molecular profiling test as an index of aberrant vascular remodeling specific to PAH based on the clear association between endothelial Bcl-2 expression and pulmonary vascular remodeling demonstrated in both cellular models and patients with PAH.^5-7 In 2007, Lévy et al.⁸ showed that the advent of Bcl-2 expression in the pulmonary vascular endothelium differentiated congenital heart disease–related PAH from surgically reversible, flow-mediated PH. Similar to Geraci et al.,⁷ Masri et al.⁹ showed that IPAH is characterized by the emergence of hyperproliferative, apoptosis-resistant ECs in the pulmonary vasculature that express, among other markers, Bcl-2. It has been suggested that these cells may represent the emergence of a dominant colony-forming EC¹⁰ due to environmental pressures. While anecdotal, the apparent agreement between lung biopsy specimens and our cell harvest technique is exciting and will require further confirmation. In addition, while the role of Bcl-2 in the overall pathophysiology of PAH is still not clear, it is now apparent that Bcl-2 expression in PAECs is a useful marker of PAH, at least to the exclusion of HFpEF.

Perhaps the most surprising finding of our this study is the degree to which the proximal pulmonary vasculature, where the SG catheter is lodged, seems to reflect the biology of the distal pulmonary vasculature, where PAH pathobiology is thought to originate and principally manifest itself. As mentioned above, the development of an acquired antiapoptotic phenotype in the pulmonary vascular endothelium is not entirely unexpected, but how and why this phenotype would propagate retrograde to the secondary and tertiary branches of the pulmonary artery are unknown. Since RHC is practiced periodically in most patients, it is intriguing to speculate that PAECs obtained during RHC might similarly be reflective of other distal biological processes, providing a new method for studying the pathobiology of PAH in situ at multiple time points along the natural history of the disease and in the context of therapeutic interventions.

We note several limitations of this study. First, the small sample size, owing to the relative rarity of pure WHO group 1 disease, makes it impossible to determine whether the magnitude of the NEBI has diagnostic value beyond simply a binary categorization of PAH versus HFpEF. Second, the association of NEBI magnitude with PAH versus HFpEF is limited by the very diagnostic imprecision that NEBI is intended to mitigate. Thus, although our efforts at clinical phenotyping were vigorous, it is possible that patients enrolled in this study were misclassified. Finally, our study groups consisted primarily of patients with previous diagnoses who were receiving concurrent therapy for PAH or HFpEF, respectively. This poses two potential problems. First, it is unclear at this point whether drug treatment, particularly PAH-specific therapy, could influence the expression of Bcl-2, given the positive effects of these pharmacotherapies on vascular remodeling. This could have resulted in an underestimation of the NEBI for PAH. Second, it is also not entirely clear at this point whether the NEBI may also be an index of irreversible remodeling, such that the index may reflect not only PAH but also a less responsive phenotype within PAH. A much larger, multicenter study with both therapy-naïve and previously treated patients will be needed to overcome these limitations.

Source of Support: Nil.

Conflict of Interest: The authors declare a patent pending for some techniques and diagnostic tools described in this article.