Plasma metabolomic profiling of hypertrophic cardiomyopathy patients before and after surgical myectomy suggests postoperative improvement in metabolic function

A total of 18 patients with clinically documented HCM referred and scheduled for surgical myectomy were approached for written informed consent to participate in the study and were included if they planned to follow up at our center. There were no other inclusion or exclusion criteria. Those who consented underwent a venous blood draw at their preoperative evaluation, within 4 weeks of their scheduled procedure. Follow up blood draws were performed at their HCM clinic postoperative visit, approximately 3 months after surgery, as was done in a prior study of plasma proteomics in this population [4]. The study was approved by the Tufts University/Medical Center Health Sciences Institutional Review Board under IRB protocol # 9487. All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki. Patient characteristics were obtained from the medical record and are shown in Table 1.

Blood samples were collected in K2 EDTA tubes and centrifuged at 2000 g for 15 min at 4 °C to separate cells from plasma. The supernatant plasma was then aliquoted and stored at − 80 °C.

Plasma samples gathered before and after surgical myectomy were sent for commercial metabolomic profiling (Metabolon, Morrisville, NC) using UPLC-MS/MS. The company uses a standardized sample preparation and analysis pipeline, as previously described [10]. Briefly, proteins were removed by methanol precipitation and centrifugation and split into 5 aliquots. Organic solvent was removed from the resulting extracts by brief placement on a TurboVAP (Zymark) and samples were then stored overnight under nitrogen prior to analysis. One aliquot was saved for backup while the other 4 were used for analysis. Aliquots were dried and then reconstituted in buffers compatible with subsequent UPLC-MS/MS. Two aliquots were analyzed using reverse-phase (RP)/UPLC-MS/MS with positive ion mode electrospray ionization (ESI), one was analyzed with RP/UPLC-MS/MS with negative ion mode ESI and one aliquot was analyzed by HILIC/UPLC-MS/MS with negative ion mode ESI. All methods utilized a Waters ACQUITY ultra-performance liquid chromatography (UPLC) and a Thermo Scientific Q-Exactive high resolution/accurate mass spectrometer interfaced with a heated electrospray ionization (HESI-II) source and Orbitrap mass analyzer operated at 35,000 mass resolution. Each reconstitution solvent contained a series of standards at fixed concentrations to ensure injection and chromatographic consistency. One aliquot was analyzed using acidic positive ion conditions, chromatographically optimized for more hydrophilic compounds. In this method, the extract was gradient eluted from a C18 column (Waters UPLC BEH C18-2.1 × 100 mm, 1.7 µm) using water and methanol, containing 0.05% perfluoropentanoic acid (PFPA) and 0.1% formic acid (FA). Another aliquot was also analyzed using acidic positive ion conditions; however it was chromatographically optimized for more hydrophobic compounds. In this method, the extract was gradient eluted from the same afore mentioned C18 column using methanol, acetonitrile, water, 0.05% PFPA and 0.01% FA and was operated at an overall higher organic content. Another aliquot was analyzed using basic negative ion optimized conditions using a separate dedicated C18 column. The basic extracts were gradient eluted from the column using methanol and water, however with 6.5 mM Ammonium Bicarbonate at pH 8. The fourth aliquot was analyzed via negative ionization following elution from a HILIC column (Waters UPLC BEH Amide 2.1 × 150 mm, 1.7 µm) using a gradient consisting of water and acetonitrile with 10 mM Ammonium Formate, pH 10.8. The MS analysis alternated between MS and data-dependent MSⁿ scans using dynamic exclusion. The scan range varied slighted between methods but covered 70–1000 m/z. Raw data was extracted, peak-identified and QC processed using Metabolon’s hardware and software. Compounds were identified by comparison to library entries of purified standards or recurrent unknown entities. Metabolon maintains a library based on authenticated standards that contains the retention time/index (RI), mass to charge ratio (m/z), and chromatographic data (including MS/MS spectral data) on all molecules present in the library. Furthermore, biochemical identifications are based on three criteria: retention index within a narrow RI window of the proposed identification, accurate mass match to the library ± 10 ppm, and the MS/MS forward and reverse scores between the experimental data and authentic standards. The MS/MS scores are based on a comparison of the ions present in the experimental spectrum to the ions present in the library spectrum. While there may be similarities between these molecules based on one of these factors, the use of all three data points can be utilized to distinguish and differentiate biochemicals.

Matched pairs t-tests were performed on log transformed paired patient samples pre- and post-myectomy to obtain an initial list of potentially important metabolites. The matched pairs t-test is equivalent to the one-sample t-test performed on the differences of the observations taken on each subject (i.e., calculate (x₁ – x₂) for each subject; test whether the mean difference is zero or not). The test statistic is given by \({\text{t}} = { }\left( {{\overline{\text{x}}}_{1} - {\overline{\text{x}}}_{2} } \right)/{\text{n}}\), with n – 1 degrees of freedom, where \({\overline{\text{x}}}_{1}\), \({\overline{\text{x}}}_{2}\) are the sample means for groups 1 and 2, respectively, s_d is the standard deviation of the differences, n is the number of subjects (so there are 2n observations). A volcano plot was generated using RStudio, with BioConductor package EnhancedVolcano (https://​github.​com/​kevinblighe/​EnhancedVolcano). Hierarchical clustering was performed on log transformed data using ArrayStudio (Qiagen Digital Insights, Redwood City, CA).

The false discovery rate (FDR) was also calculated to account for multiple comparisons between patients. The FDR for a given set of compounds can be estimated using the q-value. In order to interpret the q-value, the data must first be sorted by the p-value then choose the cutoff for significance (typically p < 0.05). The q-value gives the false discovery rate for the selected list (i.e., an estimate of the proportion of false discoveries for the list of compounds whose p-value is below the cutoff for significance) [11].

18 patients were enrolled into this pilot study. Patient characteristics are summarized in Table 1 and listed more extensively in Additional file 1: Table S1. HCM patients referred for surgical myectomy at Tufts Medical Center were approached consecutively and those that gave consent and planned to return for a postoperative visit were enrolled. The patients varied in age from 37 to 76. Twelve of eighteen were female (67%) and fifteen of eighteen had NYHA heart failure classification of 3 or greater (83%). Two of the patients carried pathogenic Mybpc3 mutations, one carried a pathogenic Myh7 mutation, and 15 patients had no pathogenic mutations found during screening (83%). Two out of eighteen patients had a history of atrial fibrillation (11%) and two of eighteen had a history of ventricular tachycardia or ventricular fibrillation leading to ICD placement (11%). Seventeen of eighteen had medical comorbidities in addition to HCM. Sixteen out of eighteen were taking beta blockers. LVOT gradients were documented for all patients, either at rest or with provocation, ranging from 60 to 160 mm Hg, with a mean maximal gradient of 102.4 ± 30.6. Thirteen of eighteen (72%) had at least mild mitral regurgitation. No patients had midventricular obstruction. All 18 patients underwent surgical myectomy, while five had concurrent mitral valve surgery, two had concurrent coronary artery bypass grafting and two had aortic valve replacement for concurrent aortic stenosis. The two patients with atrial fibrillation had concurrent MAZE procedures. All had no residual LVOT gradient on follow up echocardiogram done around the time of the postoperative visit.

Metabolomic analysis was performed on paired plasma samples from 18 patients. This identified a total of 1,340 metabolites. We wanted to first understand in more detail the proteome profiles of these samples and the relationships of the individual pre- and post-surgery samples based on relative concentration of all 1,340 metabolites identified. Consequently, we performed hierarchical clustering using all samples across all metabolites (Fig. 1). Hierarchical clustering sorts samples by similarity of metabolite concentrations in an unbiased fashion and has the potential to reveal dominant biomarkers that could distinguish the preoperative, LVOT obstruction state from the postoperative state where LVOT obstruction is no longer present. Samples with a more comparable expression pattern cluster together and separate from samples with a more dissimilar expression pattern. This hierarchical cluster analysis of all samples with all metabolites demonstrated that each paired Pre/Post patient sample clustered together and separated from all other patients, with one exception, subject 2875 (Fig. 1). This patient had concurrent aortic stenosis and also underwent aortic valve replacement, and thus may be expected to have a more divergent shift in metabolomic profile. This result indicates that, in general, the overall expression profile of all metabolites is more closely related within a patient than between Pre- and Post-surgery, suggesting that each person has a unique overall plasma metabolite fingerprint distinct from any other person, and is consistent with previous plasma proteomic profiling in HCM patients before and after surgical myectomy [4]. No dominant metabolic biomarkers of the postoperative state that would drive clustering into the preoperative and postoperative states were found.

In order to measure metabolomic changes in post-operative patient plasma, metabolomic profiling was conducted. Briefly, patient plasma samples obtained before and after surgical myectomy were sent to Metabolon to be analyzed by mass spectrometry. Plasma metabolite levels were then compared by fold-change between the pre- and postoperative states for each patient. 1070 of these biochemicals were successfully identified while 270 remain structurally unidentified. All identified metabolites could be assigned to specific metabolic pathways and subpathways as listed in Additional file 2: Table S2. Of these, 215 metabolites exhibited a statistically significant fold-change post-myectomy, with a p-value < 0.05 by matched pairs t-test. These metabolites were further broken down into 139 metabolites upregulated and 76 metabolites downregulated post-myectomy, as shown in Fig. 2.

Review of the various metabolite changes across multiple metabolic pathways (Additional file 1: Table S1, Additional file 2: Table S2) revealed notable changes in metabolites for pathways such as Heme metabolism (Fig. 3). Heme metabolites include biliverdin and the bilirubins (Z,Z; E,E; E,Z or Z,E), which were significantly decreased. In addition, L-urobilin was decreased following surgery. When heme is broken down, it is first converted into biliverdin and bilirubin (via heme oxygenase (HO) and biliverdin reductase (BR) activities). After transport to the liver and excretion into the bile, bilirubin can be converted into urobilinogen via the gut microbiota. Subsequently, urobilinogen can be converted to either d-urobilin or l-urobilin.

Arginine metabolites were also altered after myectomy surgery, suggesting alterations in arginine metabolism and the urea cycle (Fig. 4). Homoarginine was significantly decreased while homocitrulline trended higher following surgery (Fig. 4). In addition, dimethylarginine (ADMA + SDMA) was also decreased following myectomy. Phenylacetylglutamine, an acetylated peptide that is a biomarker of urea cycle disorders, was also decreased following surgery.

Plasma phospholipids and sphingomyelins also showed a variety of changes after myectomy surgery (Fig. 5). Phospholipids are synthesized from diacylglycerols and polar head groups such as choline or ethanolamine and circulate in plasma as constituents of lipoproteins synthesized mostly in the liver. Phosphatidylcholine (PC) is the major phospholipid found in lipoproteins and is required for lipoprotein assembly and secretion. PCs (e.g. 1-palmitoyl-2-stearoyl-GPC (16:0/18:0) and 1,2-dilinoleoyl-GPC (18:2/18:2)), some phosphatidylinositol (PI) species and several phosphatidylethanolamine (PE) species were increased following surgery (Fig. 5 and Additional file 1: Table S1, Additional file 2: Table S2). There were also decreases in plasma for many sphingomyelins (e.g. sphingomyelin (d18:1/18:1, d18:2/18:0) and sphingomyelin (d18:2/18:1)) following myectomy. Sphingomyelins are phospholipids derivatives lacking glycerol backbones; they are composed of ceramide (a lipid made up of sphingosine and a fatty acid) and a polar head group (phosphocholine in most cases). Sphingomyelins are present in high levels in lipoproteins and these data suggest that the composition of lipoproteins change following myectomy.

Another metabolite that showed a significant change was 3-hydroxybutyrate (BHBA), which was decreased following myectomy (Fig. 6). BHBA (a ketone body) is produced in the liver, secreted and used by other tissues, including the heart where it can be converted to acetyl-CoA to support energetic needs. Biochemicals associated with food intake, xanthine/caffeine and benzoate metabolites (which are also the products of metabolism of plant polyphenols) were increased following surgery. This may reflect an increase in food intake as the patients return to normal following surgery and/or reflect improved circulation in the patients. Unnamed compounds are discrete biochemicals which do not correspond to a standard in the Metabolon library. Of the 270 unnamed compounds that were identified, 52 were significantly different (p ≤ 0.05) following myectomy.

Although many metabolites show significant alterations in paired t-testing, the relatively small number of patients and the multiple metabolites being compared cannot exclude type 1 error. To account for multiple statistical comparisons between metabolites, the false discovery rate (q-value) was also calculated for each metabolite. When applying a more stringent significance threshold of q-value < 0.05, 12 metabolites were identified to meet this threshold (Table 2). Six of these metabolites were structurally identified. These were perfluorooctanesulfonate (PFOS), perfluorooctanoate (PFOA), 3,5-dichloro-2,6-dihydroxybenzoic acid, 2-hydroxylaurate and bilirubin, which were all downregulated with fold-changes of 0.82, 0.77, 0.79, 0.85, and 0.63, respectively, and trigonelline, which was upregulated with a fold-change of 2.88. Five metabolites that were unable to be structurally unidentified were significantly downregulated in the postoperative state, while one was upregulated.