Metabolic reprogramming of the urea cycle pathway in experimental pulmonary arterial hypertension rats induced by monocrotaline

MCT-induced animal model was used to assess PAH development in rats. All experiments were conducted in accordance with the Guideline for Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication 85–23, revised 1996) and approved by the Institutional Committee for Use and Care of Laboratory Animals of FuWai Hospital (Beijing, China).

Sprague–Dawley rats (180–220 g, 6 weeks old) were provided by Vital River Laboratories Co., Ltd. (Beijing, China). A total of 15 male rats were housed under specific pathogen-free conditions (12 h light/12 h dark photoperiod, 25 ± 2 °C, 50% ± 5% relative humidity) and were allowed to acclimate for 2 weeks before experiments. Rats were divided randomly into two groups: the PAH model group received a single subcutaneous injection of MCT (60 mg/kg; Sigma, St. Louis, MO, USA, n = 7), whereas the control group (n = 8) was treated with saline. After 3 weeks, all the rats were weighed and anesthetized (chloral hydrate, 60 ml/kg, n = 15).

To examine the development of PAH, we measured the mean pulmonary artery pressure (mPAP), right ventricular systolic pressure (RVSP), and RVH. For right heart catheterization, a polyethylene catheter was inserted into the right external jugular vein and threaded into the RV and pulmonary artery to measure the mPAP and RVSP. All data were analyzed using the PowerLab data acquisition system (Power Lab 8/30; AD Instruments, Sydney, Australia). The RV free wall was removed from the left ventricle (LV) and septum. RVH was accessed by the weight ratio of the RV to the LV plus septum weight (RV/(LV + S)).

The rats were euthanized and dissected after catheterization. Following PBS perfusion, lung tissues were embedded in 4% formaldehyde for immunofluorescence staining or in 10% formalin for histological analyses. The tissues were cut into 5 μm-thick slices. Anti-α-smooth muscle actin (α-SMA, 1:300, Abcam) was incubated at 4 °C overnight and then with Alexa 488 conjugated anti-rat IgG at room temperature for 1 h. Slides were viewed with a fluorescence microscope (LSM 780, Carl Zeiss, Oberkochen, Germany). Double-blind quantitative analysis was adopted to evaluate both vascular thickness and muscularization level. To analyze the degree of pulmonary vascular remodeling, ten random visual fields of wall area/total vessel area and relative fluorescence intensity were analyzed per lung section at a magnification of 200 using ImageJ software (http://​rsbweb.​nih.​gov/​ij).

Rigorous method validation of polar metabolites was established before metabolomics analysis to ensure the accurate and reliable of the analytical method, such as linearity and lower limit of quantification, precision and accuracy, stability, replaceable matrix and carryover (published in our previous work) [15]. To ensure the accuracy of the analysis, pool sample and pool standard solution were used as quality control in the whole analytical batches. The metabolites with compound relative standard deviation less than 30% between pool sample and pool standard sample were further analysis.

The parameters for AJS electrospray ionization MS/MS in positive/negative ion mode were as follows: dry gas: nitrogen; dry gas temperature, 200 °C; dry gas flow rate, 14 l/min; nebulizer, 20 psi; sheath gas: nitrogen; sheath gas temperature, 250 °C; sheath gas rate, 11 l/min; capillary voltage, ± 3000 V and nozzle voltage, ± 1.5 kV. Multiple reaction monitoring was performed using the characteristic precursor-to-product ion transitions, fragmentor voltage (380 V), and collision energies. The polar metabolites were identified based on retention time by using authentic standards and quantified through standard curve samples.

A t-test was used to compare between two groups for normal distribution data or Mann–Whitney test for non-normal distribution data by using SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). A p value of less than 0.05 was considered significant. To identify the most significant metabolites involved in the pathophysiology of PAH, we used MetaboAnalyst 3.0, a useful online website, to explore the potential metabolite and the involved pathway [16]. For further data analysis, partial least squares discriminant analysis (PLS-DA) was used to visually discriminate between groups by using the SIMCA-P 14.1 software (Umetrics, Umeå, Sweden). To reduce the noises and artifacts of the metabolomic data, all measured concentrations were mean-centered and auto-scaled. The quality and predictability of the PLS-DA model was then evaluated by R2Y (cum) and Q2 (cum) values, respectively. Metabolite Set Enrichment Analysis was conducted to identify biologically meaningful patterns significantly enriched in the quantitative metabolomic data.

PAH is characterized by a sustained increase in pulmonary artery pressure and vascular remolding associated with pulmonary arteriole obliteration [17]. In the present study, the MCT-treated rats (n = 7) exhibited dramatically elevated mPAP (35.22 ± 5.75 vs. 17.45 ± 4.41, p < 0.001) and RVSP (39.97 ± 3.96 vs. 21.11 ± 4.53, p < 0.001) than those of the control group (n = 8) (Fig. 1a, b). MCT-treated rats also developed pronounced RVH evident by the drastic increase in RV/LV + S (31.01% ± 3.65% vs. 22.61% ± 5.34%, p < 0.05) (Fig. 1c). In addition, histological assessment demonstrated increased proliferation of the pulmonary vascular and the immunostaining of MCT-treated lung tissue showed increased α-SMA expression in the distal pulmonary arteries in the PAH model group relative to that in the control rats (Fig. 2a, b). These results indicated the successful establishment of the PAH model in our analysis.

Plasma samples (100 μl) were analyzed using the targeted metabolomic profiling platform. In total, 126 polar metabolites were quantified from the MCT-treated and control rat plasma. Unpaired t test and Mann–Whitney test were performed to determine the metabolite variations between the two groups. Thirteen plasma metabolites related to PAH were tentatively identified through the targeted metabolomic pattern analysis to be significantly altered between the MCT-treated and control groups (p < 0.05). The detailed information of the distinguished metabolites was summarized in Table 1. The metabolites were ranked by significance on the basis of the p values. Our results demonstrated that many metabolites involved in different metabolic pathways were altered in rat plasma after MCT treatment.

Thirteen differential metabolites were divided into five categories: organic acids (n = 7), nucleotides (n = 2), lipid (n = 1), organic compounds (n = 1) and “others” (n = 2), which comprised the materials that cannot clearly be classified into any of the other four categories. The organic acids accounted for the largest proportion of the metabolites. Among the 13 differential metabolites, only adenosine monophosphate (AMP) was significantly decreased in the PAH group than in the control group. The AMP concentration in the PAH group was only 0.03-fold of the control group. The rest of the differential metabolites (92.3%) in the PAH group were all elevated relative to those in the control group. In particular, phenylacetylglycine increased by 3.23 folds that in the control group (Table 1).

PLS-DA, a supervised method based on a partial least squares algorithm, shows a high sensitivity for biomarker detection [18]. In this study, PLS-DA was conducted to investigate the metabolite patterns of PAH model and control group. The score plot obtained though PLS-DA revealed that the PAH model aggregated to the right side, whereas the control group clustered to the left (Fig. 3a). There was a distinguished classification between the clustering of the PAH model and control groups in the plasma with R2Y and Q2 greater than 0.5, which suggested that the PLS-DA models showed good stability and predictability. Those results indicated that the differentially expressed metabolites can be used to separate the plasma samples into two distinct groups.

We then identified differential metabolites for class discrimination between the groups based on the variable importance in projection (VIP) score obtained from PLS-DA. A total of 15 differential metabolites features identified by PLS-DA were presented in the Fig. 3b (VIP score > 1.5). The VIP score and relative concentrations of the corresponding metabolite in each group were also presented. The distinguished metabolic features were ranked by significance on the basis of their specific VIP values. Most of the (84.6%, 11/13) metabolites obtained from unpaired t test were included in the 15 differential metabolites. These multiple metabolic changes reflected an important metabolic distinction of PAH in the heat map based on non-supervised hierarchical clustering (VIP score top 36, Fig. 4). Overall, the PAH plasma exhibited a distinct metabolic signature relative to that in the control group.

Over representation analysis is a method that uses a hypergeometric test to evaluate whether a particular metabolite set is represented more than expected by chance within a given compound list. Differential metabolites and their concentrations were imported to MetaboAnalyst 3.0 to exploit the most disturbed metabolic pathways via over representation analysis. The metabolites that discriminate PAH were involved in 17 pathways (Fig. 5). After the results were adjusted for multiple testing by using one-paired p value, only the urea cycle pathways were enriched with the metabolites of interest (p = 0.02).

Figure 6 shows the related urea cycle pathway from the KEGG and SMPDB. Urea cycle pathway, playing a major role in PAH severity and treatment response [19, 20], connected five major distinguished metabolites in this study. These metabolites were AMP, 4-hydroxy-proline, ornithine, urea, and N-acetylornithine which demonstrated great potential in differentiating the PAH group from the control group (p < 0.05, VIP score > 1). The corresponding metabolite profiles are shown in Fig. 7. Citrulline and aspartic acid are synthesized to AMP and arginosuccinic acid, which is then converted to arginine by argininosuccinate lyase. Arginine is the precursor of nitric oxide (NO); nitric oxide synthase (NOS) converts arginine to citrulline while simultaneously producing NO and water. Arginine can also be converted to ornithine and urea by arginase. N-acetylornithine can be converted to ornithine by the aminoacylase-1. Ornithine is then converted to polyamines and proline, which are involved in the proliferation of pulmonary arterial smooth muscle cells and collagen synthesis and contribute to the pathogenesis of PAH. Proline then can be converted to 4-hydroxy-proline by Prolyl 4-hydroxylase. These compounds are considered as candidate biomarkers because of their significant ability to differentiate the PAH model from the control, as demonstrated in this study. These results suggest that the disruption of the urea cycle pathway may contribute to PAH onset.