Diagnostic and prognostic significance of serum angiopoietin-1 and -2 concentrations in patients with pulmonary hypertension

Several biomarkers for detecting pulmonary hypertension (PH) have been reported. However, these biomarkers are deemed insufficient to detect PH in its early stages. We evaluated the utility of serum angiopoietin (ANGP), a glycoprotein related to angiogenesis, as a diagnostic and prognostic biomarker of PH. Patients with PH who underwent right-heart catheterization, were retrospectively studied. Serum concentrations of ANGP-1 and ANGP-2 were measured using an enzyme-linked immunosorbent assay in patients with PH (n = 32), those with idiopathic pulmonary fibrosis (IPF) without PH (as a disease control, n = 75), and age-matched healthy controls (HC, n = 60). Nineteen patients (59.4%) with PH had World Health Organization group 3 PH. Serum ANGP-2 concentration, but not ANGP-1, in patients with PH was significantly higher compared with that in HC (p = 0.025) and in patients with IPF without PH (p = 0.008). Serum ANGP-2 concentration in patients with PH positively and significantly correlated with N-terminal pro-B-type natriuretic peptide (r = 0.769, p < 0.001), right ventricular diameter on echocardiography (r = 0.565, p = 0.035), and mean pulmonary arterial pressure (r = 0.449, p = 0.032) and pulmonary vascular resistance (r = 0.451, p = 0.031) on right-heart catheterization. ANGP-1 and ANGP-2 were expressed on lung vascular endothelial cells, as shown by immunohistochemistry. Patients with PH with higher ANGP-2 concentration (≥ 2.48 ng/mL) had significantly worse survival (p = 0.022). Higher ANGP-2 concentration was a significant worse prognostic factor (hazard ratio = 6.063, p = 0.037), while serum ANGP-1 concentration was not. In conclusion, serum ANGP-2 may be a useful diagnostic and prognostic biomarker in patients with PH, especially in patients with group 3 PH.

. Patient characteristics and data of right heart catheterization in patients with PH, IPF without PH, and HC. Data are presented as median (range) or n (%). PH pulmonary hypertension, PAH pulmonary arterial hypertension, IPF idiopathic pulmonary fibrosis, HC healthy control, CTD connective tissue disease, ILD interstitial lung disease, SLE systemic lupus erythematosus, SSc systemic sclerosis, PM polymyositis, RA rheumatoid arthritis, CPFE combined pulmonary fibrosis with emphysema, IIP idiopathic interstitial pneumonia, UCIIP unclassifiable idiopathic interstitial pneumonia, COPD chronic obstructive pulmonary disease, DPB diffuse panbronchiolitis, WHO world health organization, PAP pulmonary arterial pressure, PVR pulmonary vascular resistance, PAWP pulmonary artery wedge pressure. and ANGP-2 in the 32 patients with PH, 60 age-matched HC, and 75 patients with IPF without PH are shown in Fig. 1. In HC, the serum ANGP-1 concentration was significantly and negatively correlated with age (Supplementary Figure E1a: r = − 0.721, p < 0.001), and the serum ANGP-2 concentration tended to be positively correlated with age (Supplementary Figure E1b: r = 0.227, p = 0.081). The serum ANGP-1 concentration in patients with PH tended to be higher but was not significantly higher compared with that in age-matched HC ( Fig. 1a; median 40.6 ng/mL vs. 31.9 ng/mL [p = 0.179], respectively). The serum ANGP-1 concentration in patients with PH was significantly higher than that in patients with IPF without PH ( Fig. 1a;  Relationships between ANGP concentrations and physiologic, echocardiographic, and RHC parameters. The relationships between the serum ANGP-1 concentration and several examination findings, which were simultaneously evaluated, are shown in Fig

Comparison between patients with low and high ANGP-2 concentration. Thirty-one patients
with PH, who were followed up, were divided into two groups according to median serum ANGP-2 concentration: the ANGP-2 high group (≥ 2.48 ng/mL) and the ANGP-2 low group (< 2.48 ng/mL) ( Table 2). The ANGP-2 high group included more patients with group 3 PH (p = 0.016), a higher mMRC dyspnea scale (p = 0.010), and had a higher NT-ProBNP concentration (median, 93 pg/mL vs. 608 pg/mL; p < 0.001) compared with the ANGP-2 low group. Furthermore, the ANGP-2 high group tended to have a lower %DLCO (44.8% vs. 26.6%; p = 0.072), and a higher proportion of patients treated with long-term oxygen therapy (p = 0.072) compared with the ANGP-2 low group.

Relationship between ANGP concentrations and prognosis in patients with PH. Thirty-one
patients with PH, who were followed up, were divided into two groups according to the median serum ANGP concentrations: the ANGP-1 high group (≥ 40.62 ng/mL), the ANGP-1 low group (< 40.62 ng/mL), the ANGP-2 high group (≥ 2.48 ng/mL), and the ANGP-2 low group (< 2.48 ng/mL). The Kaplan-Meier curves of survival probability from the time of ANGP measurement are shown in Fig. 5. Patients with a high ANGP-2 concentration had significantly lower survival than those with a low ANGP-2 concentration ( Fig. 5b; log-rank, p = 0.022), while a significant difference was not observed between patients with high and low ANGP-1 concentrations ( Fig. 5a; log-rank, p = 0.407). The 2-year survival rate was 44.7% in the ANGP-2 high group and 92.9% in the ANGP-2 low group (Fig. 5b). When limited to patients with IPF, those with PH had significantly poorer survival compared with those without PH (Supplementary Figure E4a; log-rank, p = 0.004). Moreover, patients with IPF and PH with a high ANGP-2 concentration had significantly lower survival than those with a low ANGP-2 concentration (Supplementary Figure E4c; log-rank, p = 0.018), while a significant difference was not observed between patients with IPF and PH with high and low ANGP-1 concentrations ( Supplementary Figure E4b; p = 0.330). Next, prognostic factors from ANGP measurements were evaluated in all patients with PH. Univariable Cox proportional hazards models were used to identify factors that predict mortality in all patients with PH (Table 3). Serum ANGP-2 concentration (ANGP-2 high: hazard ratio [HR] 7.716, p = 0.015), but not the ANGP-1 concentration (p = 0.404), was a significant prognostic factor. Furthermore, WHO functional class 3-4 (HR 4.752, p = 0.017), %DLCO (HR 0.897, p = 0.036), and long-term oxygen therapy (LTOT; HR 7.135, p = 0.005) were also significant prognostic factors. Age-adjusted multivariable Cox proportional hazards models showed that serum ANGP-2 concentration was still a significant prognostic factor (Table 3; ANGP-2 high: HR 6.063, p = 0.037).

Discussion
In the present study, 32 patients with PH, mainly group 3 PH (19 patients [59.4%]), and 75 patients with IPF without PH were retrospectively studied. Serum ANGP-2 concentration, but not ANGP-1, in patients with PH was significantly higher compared with that in age-matched HC, and that in IPF patients without PH. Serum ANGP-2 concentration in patients with PH positively and significantly correlated with mMRC dyspnea scale, NT-ProBNP/BNP concentration, RVD on echocardiography, and mean PAP and PVR on RHC. Patients with PH with a higher serum ANGP-2 concentration showed significantly worse survival. A high ANGP-2 concentration was a predictor of a poor prognosis, while serum ANGP-1 concentration was not. To our knowledge, this is the first study that showed the significance of ANGP-2 as a diagnostic and prognostic biomarker for mainly group 3 PH. The gold standard of PH diagnosis is RHC 1 , although RHC is relatively invasive and is not easy to conduct in clinical practice. NT-ProBNP/BNP, %DLCO, and echocardiography are less invasive, but their diagnostic and prognostic significance are not sufficient in patients with PH. Therefore, a novel and reliable biomarker that is easy to measure is desired for PH. Previous studies have reported that plasma endothelin-1 and cyclic guanosine monophosphate (cGMP)/NT-proBNP ratio were related to 6MWD 14 , and increased serum interleukin (IL)-6/ IL-8/IL-10/IL-12 predicted worse survival 15 in patients with PAH. In the current study, although cGMP and cytokines were not evaluated, serum ANGP-2 was identified as a useful biomarker for the diagnosis of PH. Moreover, ANGP-2 predicted survival in patients with PH.
ANGP-1 and ANGP-2 play a role in vascular development, angiogenesis, and vascular permeability. ANGP-1 and ANGP-2 competitively bind the Tie-2 receptor, which is primarily found on endothelial cells [4][5][6]16 . Binding of ANGP-1 to Tie-2 on endothelial cells promotes vascular integrity and leads to an angiostatic "non-leaky" condition 7 , while ANGP-2 binding to Tie-2 promotes angiogenesis and vascular permeability 17 . In the current study, serum ANGP-2 concentration in patients with PH was significantly higher compared with that in patients with IPF without PH and that in age-matched HC. In the early stage of group 3 PH, hypoxic vasoconstriction and loss of vascular bed lead to the redistribution of pulmonary circulation 18 . In such a circumstance, the recruitment capacity and distensibility of pulmonary vessels play a role to keep a functional pulmonary circulation     19,20 . ANGP-2 is mainly produced from endothelial cells 21 , and ANGP-2-induced angiogenesis may keep such a functional gas-exchange at the distended pulmonary vessels. In patients with PAH, ANGP-2 concentration reportedly showed positive correlation with PVR, and an elevated ANGP-2 concentration was associated with worse survival 13 . According to our study, patients with PH with a higher serum ANGP-2 concentration showed significantly worse survival, and a higher ANGP-2 concentration predicted a worse prognosis. A previous study reported that hypoxia, vascular endothelial growth factor, and basic fibroblast growth factor increased ANGP-2 release from endothelial cells 21 . Furthermore, ANGP-2 was expressed only at sites of vascular remodeling in a rodent model 8 . Therefore, excessive production of ANGP-2, which promotes angiogenesis, may be related to PH disease progression and may be a novel therapeutic target in patients with PH. In the present study, patients with PH mainly presented with group 3 PH (59.4%). Fifteen percent of patients with IIP and 27% of those with combined pulmonary fibrosis and emphysema had group 3 PH 22 . Furthermore, in terms of IPF, 14% of patients with IPF with a mild-to-moderate restriction in lung volume 23 and 31.6% of patients with advanced IPF had PH 24 . Many previous studies focused on group 1 PH, especially PAH. Major differences between group 1 PH and group 3 PH are that group 3 PH demonstrates poorer RV function, a lower mean PAP and PVR, and worse survival 25 . Although vasodilator treatment is less efficient or worsens the respiratory condition in patients with group 3 PH 25 , a mean PAP of ≥ 35 mmHg (severe PH) warrants consideration of vasodilator therapy. Moreover, a recent study showed that inhalational treprostinil in patients with group 3 PH could increase 6MWD and decrease NT-proBNP concentration 26 . In the present study, serum concentrations of ANGP-2 were significantly higher than in age-matched HC or patients with IPF (Fig. 1a). Therefore, the serum concentration of ANGP-2 may be a useful diagnostic biomarker in patients with PH, and the regular measurement of ANGP-2, in combination with other biomarkers, in patients with pulmonary diseases may be practical for the early detection of PH. Furthermore, serum ANGP-2 concentration was positively and significantly correlated with mean PAP and PVR. Therefore, less invasive biomarkers, such as serum ANGP-2, may be useful for clinical decision making in terms of PH treatment in patients with group 3 PH.
The present study has several limitations. First, this study adopted a retrospective design, and the number of patients enrolled was relatively small, which was also more group 3 PH predominant compared to the common distribution of PH prevalence. In addition, immunohistochemical analysis was conducted in only two due to the limited number of patients whose lung autopsy specimens were available. Second, longitudinal ANGP values were not evaluated. Third, treatments were not uniform across PH groups when serum ANGPs were measured, and may have affected serum concentrations of ANGPs. Therefore, larger prospective studies are needed to confirm the significance of ANGPs in patients with PH. In conclusion, the capacity of serum ANGP-1 and ANGP-2 as diagnostic and prognostic markers was retrospectively evaluated in patients with PH. Serum ANGP-2 concentration, but not ANGP-1, in patients with PH was significantly higher compared with that of age-matched HC and that of patients with IPF without PH. Serum ANGP-2 concentration in patients with PH positively and significantly correlated with mean PAP and PVR on RHC. Patients with PH with a higher serum ANGP-2 concentration showed significantly worse survival. A higher ANGP-2 concentration was a significant predictor of a worse prognosis. Collectively, these results suggest that serum ANGP-2 is easy to evaluate and may be useful as a diagnostic and prognostic biomarker for PH. Future studies should prospectively confirm the significance of ANGP-2 as a practicable biomarker in patients with PH.

Methods
Study design and patients. Thirty-two patients, who were diagnosed with PH from 2000 to 2020, were retrospectively studied. All patients were diagnosed with PH by RHC using a mean PAP of ≥ 25 mmHg. Seventyfive patients with IPF without PH were also studied. These patients met the IPF consensus criteria of the American Thoracic Society/European Respiratory Society/Japanese Respiratory Society/Latin American Thoracic Association 27 . All procedures in this study were performed in accordance with the study protocol and the 1964 Helsinki Declaration, as amended. The need for patient approval and informed consent was waived by Ethics Committee of Hamamatsu University School of Medicine due to the retrospective nature of the study. However, informed consent was obtained from all patients still visiting our clinics. The study protocol was approved by the Ethics Committee of Hamamatsu University School of Medicine (approval number: 17-232).
Data collection. Clinical, laboratory, and physiological data were obtained from medical records, which included FVC, DLCO, and partial pressure of oxygen at rest, minimum oxygen saturation, and walk-distance during the 6MWT. The extent of pulmonary emphysematous lesions on high-resolution computed tomography was evaluated by two observers. Echocardiography and RHC findings were also obtained from medical records. These data, which were measured on the date closest to the date on which the serum ANGP was evaluated, were used. Survival was analyzed from the date on which serum ANGP was measured.
Measurement of serum ANGP. Blood samples were collected from enrolled patients with PH. Serum concentrations of ANGP-1 and ANGP-2 were measured using an enzyme-linked immunosorbent assay (R&D Systems, Inc., Minneapolis, MN, USA). ANGP concentrations were compared with those of patients with IPF without PH (as a disease control) and age-matched HC.
Immunohistochemical staining of ANGPs. ANGPs were immunohistochemically stained in autopsy specimens. Briefly, deparaffinized sections were steeped in 0.3% hydrogen peroxide to inactivate endogenous peroxidase activity and then blocked with 10% normal goat serum. Sections were incubated with rabbit polyclonal antibody against ANGP-1 and ANGP-2 (anti-ANGPT1 or anti-ANGPT2 (C-term), Sigma-Aldrich, Saint Louis, MO, USA). After rinsing with phosphate-buffered saline, sections were incubated with biotin-conjugated goat anti-rabbit immunoglobulin G polyclonal antibody (Nichirey, Tokyo, Japan). Sections were then incubated with streptoavidin-peroxidase complex (Nichirey, Tokyo, Japan). The antigen-antibody complex was visualized with 3,3′-diaminobenzidine (DAB; Nichirey, Tokyo, Japan) and counterstained with hematoxylin. Statistical analysis. Statistical analyses were performed using JMP 13.1.0 (SAS Institute Inc., Cary, NC, USA). Categorical data were compared using the χ 2 test or Fisher's exact probability test for independence. Continuous data were analyzed using the Wilcoxon rank-sum test. Relationships between the serum ANGP concentrations and serial data were analyzed using Pearson's correlation coefficient. Relationships between the serum ANGP concentrations and categorical data were analyzed using Spearman's rank correlation coefficient. Overall survival of patient groups was estimated using the Kaplan-Meier method and Kaplan-Meier curves were compared using the log-rank test. The relationships between variables and mortality were evaluated using the Cox proportional hazards regression analysis. All tests were two-sided, and a p value of < 0.05 was considered statistically significant.