MSACL 2024 Abstract

INTRODUCTION
Neurofibromatosis type 1 (NF1) is a common autosomal dominant inherited human disorder. A hallmark of NF1 is the development of plexiform neurofibromas (PNFs) in 30-50% of NF1 patients. PNFs are complex peripheral nerve sheath tumors associated with nerve trunks, which can cause substantial morbidity including pain, neurologic deficit, and motor dysfunction. Studies showed that PNFs grow most rapidly in young children. Currently there is no cure for PNF apart from surgical removal. Furthermore, surgery is often impossible due to the invasive nature of PNFs, their large size, and their association with critical anatomic structures. There are no validated biomarkers to identify and follow progression of PNFs in individuals with NF1 and/or other NF1 related tumors. Mass spectrometry (MS)-based lipidomics/metabolomics is a systems approach that seeks to comprehensively profile lipids and metabolites for disease specific biomarker discovery.

OBJECTIVES
To aid in the tumor burden biomarker discovery, we use untargeted lipidomics on plasma from robust neurofibroma enabled mouse models, with samples collected at three longitudinal time points. Using the temporal metabolic profiles from neurofibroma-bearing DhhCre;Nf1fl/fl mice and machine learning model of mass spectrometry (MS)-based lipidomics data, we aim to discover metabolic biomarkers of PNFs progression in plasma and validate our findings with samples from NF1 patients.

METHODS
Plasma samples from DhhCre;Nf1fl/fl tumor bearing mice were collected at different time points, i.e. 1, 2, and 7 months of age. Samples were aliquoted in multiple vials (50 µL) to avoid freeze-thaw cycles that will interfere with detection of biomarker. Samples were stored at -80oC before analysis. In all animal experiments, we used mice of both sexes. All mouse MRI image data was acquired with a 7T Bruker Biospec system equipped with 400 G/cm gradients. Reference compounds of glucosylceramides and lactosylceramides, and internal standards were obtained from Avanti Polar Lipids and Cayman. Delipidized human plasma was used to prepare the calibrators and quality controls.

Untargeted lipidomics analysis was conducted on a UHPLC-HRMS system with Q ExactiveTM plus hybrid quadrupole-OrbitrapTM mass spectrometer interfaced with Vanquish ultra-high performance liquid chromatography (UHPLC) system (Thermo Scientific, Waltham, MA). An Acquity CSH C18 UPLC column (2.1 × 100 mm, 1.7 µm, Waters, Milford, MA) was used in separation. Untargeted data were pre-processed by Progenesis QI (Waters Corp.) for peak picking, alignment, deconvolution, and preliminary metabolite/lipid annotation by in-house and on-line available databases, i.e., HMDB and LipidMaps, and identity of potential biomarkers was confirmed by both accurate mass and retention time as well as fragmentation patterns if available.

Separation and quantification of discovered biomarkers, i.e., glucosylceramides (GC) and lactosylceramides (LC) were performed on a UPLC coupled with Xevo TQ-S triple-quadruple mass spectrometry (Waters, Milford, MA). The precursor and product ion pairs for MRM monitoring were either selected by MS/MS spectra of standards or calculated theoretically when standards for a given ceramide length were not available. The optimal signal for the ion pair of glucosylceramides was achieved in positive ion mode. The commercially available GC and LC species, were spiked in the blank human serum (charcoal stripped) to construct calibration curves and QCs.

RESULTS
The DhhCre;Nf1fl/fl mouse model of neurofibromatosis has a rapid disease course and controlled genetic background, providing an opportunity for biomarker discovery and longitudinal evaluation in the context of neurofibroma formation, progression, and tumor burden. We performed untargeted lipidomics analysis in plasma from DhhCre;Nf1fl/fl mouse model (n=19) at age 1 and 2 mo. (before tumors form) and 7 mo. (in mice with tumor of known volumes). At 1 month of age, peripheral nerves and dorsal root ganglion (DRG) were generally normal even with NF1 loss. Gene expression is not significantly different from controls. By the 2 month time point, macrophages increase and rare CD11c+, CD11b-, DC and T cells start to present in paraspinal nerve roots and ganglia. Small tumors can be detected by 4 months of age. Glucosylceramides (largest fold change) were increased in the Nf1 mouse model only at 7 mo. (tumor-bearing mice). We confirmed elevation of several GC and LC species in tumor bearing mice with targeted LC-MS/MS analysis. We correlated GC and LC measurements with tumor burden using the volumetric measurements of PNFs in the animal model at 7 mo. GC (d18:1/20:0), GC (d18:1/16:0), and LC(d18:1/24:1) had Spearman correlations of 0.48, 0.42, and 0.34 respectively. We then performed same targeted analysis of GC and LC in plasma samples from NF1 individuals (n=11) and age and sex matched healthy controls (n=11). PNF tumor burden analysis was performed using whole body MRI and patients were stratified into the following categories 0 - none, 1 - small, 2 - intermediate, and 3-large PNFs. This pilot analysis was under powered due to small sample size. Preliminary comparison of large (3) and intermediate (2) to small (1) and no tumors (0) showed striking differences and high Spearman correlation for LC(d18:1/26:0) (rho=0.81), total GC species (rho=0.73), LC(d18:1/24:1) (rho=0.73), and GC (d18:1/20:0) (rho=0.73).

CONCLUSIONS
This MS-based lipidomics analysis has provided valuable insights into the biochemical changes in PNF progresssion. Our results identified glycosphingolipids including glucosylceramide and lactosylceramide species as potential prognostic biomarkers for PNFs development from a longitudinal mouse model. We validated the same biomarkers in the NF1 patient cohort using targeted LC-MS/MS analysis and correlated to tumor burden.