|
Abstract INTRODUCTION:
The most common gynaecological condition and indication for hysterectomy worldwide is uterine leiomyomas (fibroids). Despite this, the pathophysiology of the condition remains poorly understood. Metabolomic analysis of the fibroid secretome offers novel mechanistic insight and potential identification of a future therapeutic target to reduce the heavy, irregular menstrual bleeding, chronic pelvic pain and infertility associated with the disease.
OBJECTIVES:
Identify significant differences in secreted metabolite abundances between fibroids, control myometrium and adjacent myometrium with future promise of better mechanistic understanding of disease and non-hormonal therapeutic targets.
METHODS:
Fresh surgical fibroid or myometrial samples were placed into basal DMEM medium with primocin and incubated for 24 hours. The resultant conditioned media was rendered acellular by centrifugation (4000 rpm, 4°C, 10 min). Deproteinised media containing deuterated internal standards were analysed using the 1290 Infinity II liquid chromatography system coupled to a 6550-quadrupole time-of-flight mass spectrometry (Agilent, UK). Prior to the analysis of study samples, optimal injection and reconstitution volumes were established. Data scaling by total protein concentration (BCA assay) and total metabolite abundance were compared. Ethical approval was obtained for the collection and usage of samples (LWRTB REC: ethical approval 19/WA/0271 and INTERPRET REC: approval 19/SC/0449).
Raw data was processed using the Agilent MassHunter software suite and Metaboanalyst (Version 6.0). Features were extracted using ±10 ppm theoretical accurate mass and ±0.3 mins retention time window and were filtered based on their frequency (>70%) and variability across quality control samples (CV <30%). An accurate mass, retention time database containing 469 intermediary metabolites (Mw 72-785) was used to facilitate compound identification. Subsequent multivariant analysis using ANOVA (p<0.05 adjusted for FDR), univariant analysis using fold change and two paired t-test (FC 1.2, p<0.05 adjusted for FDR), and analysis of clinicodemographic variables was conducted following data pre-processing. Temporal profiling of the secretome was achieved by comparing time intervals collected and significant metabolite abundances. Iterative MS/MS was performed to aid in metabolite identification.
RESULTS:
129 samples from 48 patients were included in the final study, 19 samples from 2 patients were used for temporal profiling over 17, 24 and 50 h. Optimisation experiments showed that a 5 µL injection and a 50 µL reconstitution volume were superior.
Covariate metadata analysis was performed to assess the contribution of fibroid subtype (by anatomical location), BMI (normal, overweight and obese), ethnicity (Caucasian, Black, Asian), pre-operative haemoglobin (<120 g/L Hb defined as anaemic), hormonal agents, menstrual phase, menopausal status (pre and post) and parity had on metabolite abundances. No demographic data demonstrated distinct metabolic profiles in either polarity.
For targeted secretome analysis, 141 and 121 features were matched in positive and negative polarities respectively using an in-house metabolite database (level 1 identification). Of these matched features, 42(25 positive and 29 negative respectively, with duplicates accounted for), showed a significant change in metabolite abundance across univariant t-test (p<0.05) and multivariant ANOVA (p<0.05) analysis. For univariate analysis, most of the differences in metabolite abundances were shown between fibroid and control myometrium ((13 significant metabolites in positive and 20 in negative (blank contribution<5%)), where metabolite abundance was increased in fibroids relative to control (FC >1.2, no scaling). Most significant metabolites included amino acids and derivatives (including L-proline, L-ornithine, D-alanine), energy metabolism intermediates (including succinic acid, isocitrate, and malic acid), carbohydrates and derivatives (including glycerol, gluconic acid, and raffinose), and nucleotide metabolites (including hypoxanthine, inosine, and uracil).
For untargeted secretome analysis, 5000 features in positive and negative polarities were matched by RT and accurate mass. 2950 positive and 2094 negative features remained after filtering by QC frequency and sample variation. Of these, 252 positive and 335 negative features showed a significant change in metabolite abundance across univariant t-test (p<0.05) and multivariant ANOVA (p<0.05) analysis. For univariate analysis, most of the differences in metabolite abundances were shown between fibroid and control myometrium ((166 significant metabolites in positive and 31 in negative (blank contribution<5%)), where metabolite abundance was increased in fibroid relative to control (FC >1.2, no scaling). Raw data was manually interrogated to assess metabolite peak quality and abundance of the precursor ion, and of which, 35 positive and 101 negative features were confirmed. Metabolite identification is ongoing, including targeted MS/MS experiments to identify unknown significant features.
CONCLUSION:
This is the first study to demonstrate metabolomic differences in the secretome of fibroids compared to both control myometrium and adjacent myometrium using LC-QTOF-MS. We have shown that there are 48 metabolites with significant differences in abundance, that have been named and identified using RT and accurate mass from a reliable database. We have shown that metadata did not correlate with these metabolite differences. Further work is required to name and identify the significant metabolites found using the untargeted workflow, and to perform metabolite enrichment/pathway analysis to understand the biological significance in terms of pathophysiology and as a therapeutic target.
|
= Discovery stage. (53.14%, 2025)
= Translation stage. (22.33%, 2025)
= Clinically available. (24.53%, 2025)
