Poster Contest Participants for 2024

Participants: 25
Uploaded Poster PDFs: 20

Topic(s): Other -omics > Glycomics

Poster Presentation
Poster #7b
Attended on Wednesday at 14:30

Understanding Mammalian Metabolic Flux into N-glycans by Stable Isotopic Tracing Analysis

Idris Wazeerud-Din (Presenter)
Children's Hospital of Philadelphia


Background: Congenital disorders of glycosylation (CDG) are one of the largest groups of inborn errors of metabolism with more than 170 types identified, most of which are under-studied. Among all the CDGs, PMM2-CDG is the most common and affects more than 50% of the CDG patients in the country. Recently, significant progress has been made in developing novel therapies for PMM2-CDG, including aldolase reductase inhibitors and mannose-1-phosphate replacement therapies. It is hypothesized that the clinical improvement and developmental gains achieved in affected children on these therapies may be due to increased monosaccharide flux into N-linked glycoprotein biosynthesis. This study aims to investigate the metabolic flux of stable isotope labeled 15N2-glutamine and 13C6-glucose into the glycoproteins from cultured human fibroblasts or mouse tissues post the whole-body metabolic tracing using in-depth high-resolution mass spectrometry (HRMS) methods.

Method: Both normal and PMM2-CDG human skin fibroblast cells were cultured in DMEM and dialyzed FBS with glutamine replaced by 15N2-glutamine. Mouse tissues were obtained after being intravenously infused with 15N2-glutamine for 4 hours. N-glycans from cells or tissues were released by PNGase F digestion, N-tagged with a modified quinoline and extracted using a HILIC plate in preparation for N-glycan heavy isotopic tracing analysis by UPLC-ESI-QTOF (G2Si Waters) and UHPLC-Orbitrap (IQ-X Tribrid Thermo Fisher) systems.

Results: We initiated metabolic tracing with 13.5 mM 15N2-glutamine in cultured skin fibroblasts when the cells reach 90% confluency for 24, 48, and 72 hours. On average, more than 20% GlcNAc in N-glycans was replaced with 15N1GlcNAc after 24 hrs of 15N2-glutamine tracing and the total 15N flux further increases at 48 and 72 hours. Interestingly, the rate of metabolic flux from glutamine to N-glycan is slow in PMM2-CDG cells compared to two concurrent control lines in a glycan specific pattern. On average, the rate of glutamine flux into high mannose glycans, Man6-9GlcNAc2, in PMM2-CDG cells is 70-80% of the control mean. As the number of GlcNAc in the N-glycan increases in the complexed glycans, such as Fuc1Gal2Man3GlcNAc4, the rate of flux dropped to 20% of the control mean, which could be explained by a reduced metabolic flux into the hexosamine biosynthetic pathway in PMM2-CDG cells. Using HRMS technology, we are able to detect as low as 1% 15N incorporating into GlcNAc in Fuc1GlcNAc2Man3GlcNAc2 glycan. As we examined the liver N-glycans extracted from a mouse infused with 15N2-glutamine through tail vein for 4 hours, a significant amount of 15N was detected in the GlcNAc fragments of mouse N-glycans.

Conclusion: Our metabolic N-glycan flux analysis using HRMS method was able to identify less than 1% metabolic flux through 15N2-glutamine and we can quantify the amount of flux as well as the rate of flux in cultured cells when there is >1% heavy isotopic incorporation. Furthermore, we provide evidence here that the metabolic flux through hexosamine biosynthetic pathway is reduced in a PMM2-CDG cell line and detectable within 72 hours of tracing, building a paradigm to study mammalian metabolic flux into N-linked protein glycosylation both in vitro and in vivo.

Topic(s): Small Molecule > Emerging Technologies > Assays Leveraging Technology

Poster Presentation
Poster #9b
Attended on Wednesday at 14:30

High-Throughput Bioanalysis Using an Automated Ambient Mass Spectrometry Platform

Nicolás Morato (Presenter)
Purdue University


Introduction: Central laboratory facilities, large-scale biomarker identification campaigns, and drug discovery studies, all rely on the high-throughput analysis of complex biological samples, from non-volatile buffers with high concentrations of detergents or salts, to tissue biopsies, cell cultures, bacterial colonies, or diverse biofluids. Traditionally the use of mass spectrometry (MS) for the study of these classes of samples has heavily relied on sample preparation and purification approaches prior to analysis. This, although effective at dealing with the complexity of biological matrices, drastically limits the inherent throughput of MS analysis and generates multiple challenges when dealing with large-scale sample cohorts. Here we describe the operation and several applications of an automated robotic platform based on desorption electrospray ionization (DESI) MS for the rapid and direct (i.e. no sample workup needed) analysis of arrays of biological samples, with throughputs better than 1 sample per second. Relevant examples of the screening of biosamples for biomarker discovery, routine bacterial identification, as well as bioactivity assessment using purified targets or cell cultures will be overviewed.

Methods: High-density biological arrays (up to 6,144 samples/array) are rapidly generated on PTFE-coated glass slides using a fluid handling workstation (Beckman Biomek i7 or Hamilton Vantage) equipped with a slotted floating 384 pin-tool. Samples (50 nL) are spotted from 384-well plates either in replicates or from multiple sources, with the spotting being completely user-defined in each experiment. Prepared plates are automatically transferred to a DESI stage and analyzed by MS(/MS) in a spot-to-spot fashion with effective analysis times in the order of 500 ms per spot. The large data volumes produced by this approach are automatically processed in real time by a combination of custom Python- and MATLAB-based software. For biological assays samples are automatically prepared using the fluid handler capabilities or externally with Multidrop Combi setups. For cell studies, prostate cancer cell lines (PC3, LNCaP) were grown and enzymatic target expression was induced. Target inhibition was assessed after a 24-hour incubation followed by media removal and the addition of methanol spiked with an internal standard. All cell cultures were performed in 96-well plates before transfer to the final plate for spotting. For bacteria arrays, single colonies are picked, deposited in 384-well plates pre-filled with solvent (for inactivation of the bacteria and safe analysis in open air), and 50-nL aliquots are spotted onto the PTFE array. Alternatively, analysis of the bacterial colonies can proceed directly from liquid culture media or from solid agar plates after image recognition and route definition algorithms are applied and interface with the control system of the platform.

Results: This general DESI-MS methodology has been applied to a variety of complex samples, including biological assays, tissue biopsies, human cell cultures, and bacterial colonies, with no need for sample work-up. Sub-second analysis times, low sample consumption (typically sub-ng), remarkable matrix tolerance, and excellent quantitative performance are important features of this approach. Several examples including those used for the efficient generation of large spectral libraries, the characterization of enzymatic processes and the evaluation of binding affinities, as well as the profiling and classification of biological specimens will be overviewed. Examples include the label-free characterization of novel potential targets for drug discovery in cancer and neurodegenerative diseases, as well as screening campaigns against such targets for the identification of hits and their subsequent confirmation in cell cultures. Similarly, the use of tissue instead of purified targets to assess overall metabolic pathway dynamics as well as downstream inhibition effects will be overviewed. Additionally, we will showcase the generation of bacterial sample arrays for rapid lipid and metabolite profiling. The unique spectral signatures obtained allow for high-accuracy (>95%) classification of ca. 30 species of microorganisms based entirely on small molecules, in a complementary fashion to commercial protein-based MALDI biotyping but with a simplified workflow and potential expansion to simultaneous identification of enzyme-based antibiotic resistance. Finally, current efforts towards the use of image recognition techniques for the direct analysis of random arrays such as bacterial colony plates, will be discussed.

Discussion/Conclusion: The use of DESI allows for direct analysis of a wide range of biological samples, despite large concentrations of non-volatile salts or detergents, without the need for any sample work-up. This advantage comes from the intrinsic mechanism of DESI which involves a contactless online microextraction event on the surface of analysis. Additionally, the intrinsic control over the sampling event that DESI provides allows for facile and rapid introduction of discrete samples in a high-density array by precise movement of the surface of analysis. These advantages, added to both custom and commercial robotics, software, and analytical instrumentation, make for a powerful automated platform well suited to handle large sample sets, such as that commonplace in central laboratory facilities, drug discovery campaigns, or biomarker screening studies. Additionally, the depth of the data acquired is well suited to provide valuable information, for instance combined with AI approaches, to guide drug development or even clinical decision-making.

Topic(s): Proteomics > Cases of Unmet Clinical Needs > none

Poster Presentation
Poster #14b
Attended on Thursday at 12:15

Identification of the Protein Expression of Androgen Receptor Variants Using Targeted Proteomics in Clinical Castration Resistant Prostate Cancer Models

Zoi Sychev (Presenter)
University of Minnesota, Twin Cities



Castration Resistant Prostate Cancer (CRPC) is a treatment resistant form of prostate cancer (PCa). Currently, there is not a way to identify which patients will develop this resistance before full blown CRPC develops. Therefore, all PCa treatment approaches are similar yet this results in tumor regression in some cases and progression in others. Targeting the androgen receptor (AR) is still the main focus of current therapies even in CRPC. Emergence of AR splice variants (AR-vs) after initial treatment is thought to be one of the primary mechanisms of resistance. AR-vs lack the ligand binding domain rendering Androgen Deprivation Therapy (ADT) ineffective in tumors expressing these variants. Recent work has identified the DNA and RNA species of ARVs in CRPC but investigation into whether the protein is translated are unknown. One exception is the approval of an antibody test that detects a specific AR-v, AR-v7, from the blood of PC patients and predicts ADT response. However, there are instances where a patient may not express ARv7 and are still resistant to ADT. This may suggest that other AR-vs, not currently detected at the protein level, are important predictors for ADT response. The primary goal of this work is to expand the protein identification of known and unknown AR-Vs that may predict response to ADT using targeted liquid chromatography tandem mass spectrometry (LC-MS/MS). Targeted LC-MS/MS can provide accurate, precise, sensitive, and reproducible detection of a pre-determined set of peptides in specimens without the need for antibody enrichment.


The goal of this work is to generate a robust set of unique peptides that behave as surrogates for AR-v proteins and then evaluate them via targeted LC-MS/MS.


We developed the AR-v targeted LC-MS/MS assay by first designing AR splice variants from unique splice regions via RNA-guided sequences translated to amino acid sequences. The AR-variants full amino acid sequences were then in-silico trypsin digested using Expasy software tool that generated tryptic peptides. Blast analysis of the AR-v peptides was performed where no overlap with other human proteins was observed. These AR-v peptide sequences were commercially synthesized along with heavy isotopically analogs to be used as internal standards in the LC-MS/MS method. The LC-MS/MS system consisted of a XR HPLC system (Shimadzu) with a QTRAP 6500 triple quadrupole mass spectrometer (SCIEX) using electrospray ionization and multiple reaction monitoring (MRM) for analysis. Experiments were conducted under high-flow conditions with a flow rate of 0.700 mL/min and a 7-minute run time employing gradient chromatography. Mobile phase solutions consisted of 0.1% formic acid in water (mobile phase A) and 0.1% formic acid in acetonitrile (mobile phase B). A Kinetex (Phenomenex) 2.6 um C-18 100 A 50 x 3 mm column was used for all chromatographic runs. The method examined 9 peptides and included 2 to 3 MRM transitions per peptide. Calibration curves were constructed using unlabeled peptides over 0.10 pmol/uL to 1 pmol/uL. Specimens examined consisted of a panel of prostate cancer cell lines (n= 6), CRPC patient derived xenografts tumors (n=4) and a negative control cell line (no androgen receptor expression), Cos-1, which served with a dual function and was used as matrix for peptide MRM optimization experiments. Cell line and patient derived xenografts tumor specimens were processed using lysis buffer containing 7 M of urea, 2 M thiourea, 0.4 M Tris pH 8.0, 20% acetonitrile (ACN), 10 mM TCEP and 25 mM chloroacetamide. Protease inhibitor HALT was added to the lysis buffer at 1x concentration immediately before addition to the specimens. Viscous specimens, likely due to chromatin release, were sonicated using a probe sonicator set at 30% amplitude, for 5 seconds while samples were on ice. The specimens then underwent proteolysis using LysC at pH 8.2 and incubated at 37 C for 4 hours and then trypsin at pH 8.2 was added and incubated at 37C in a warm air incubator overnight (~16 hrs). After incubation, the samples were acidified with trifluoroacetic acid to pH 3 (final concentration of 0.3-0.6%) or less to stop further enzymatic activity. Specimens were desalted using reverse phase HLB columns and the elutions were dried down using a speed vacuum. Samples were re-suspended with 0.1% formic acid. Internal standard solution was spiked-in to each calibrator and cell line specimen and then added to a 2 mL autosampler vials with 300 µL glass inserts. Linearity and sensitivity for each peptide included in the method was examined using calibration curves. Data analysis was conducted using Multiquant (SCIEX) software.


The targeted MRM approached was successfully achieved for quantification of 9 peptides from 8 proteins in 6 prostate cancer cell lines and 4 patient derived xenografts tumor specimens. Out of the 8 proteins evaluated, we observed 3 known targets (AR-exon1, AR-v7 and AR-v12) and 4 novel AR-vs in 22Rv1 PCa cell lines. Additionally, CRPC patient derived xenografts were assessed where 3 novel AR-vs were detected.


These preliminary results using a LC-MS/MS platform show promising identification and quantification of novel AR-vs that have not been measured before at the protein level. Subsequent analysis of larger cohorts will further elucidate AR-vs role in PCa treatment resistance mechanisms and possibly assist in future treatment approaches.

Topic(s): Practical Training > Tox / TDM / Endocrine

Poster Presentation
Poster #15a
Attended on Wednesday at 12:15

Development of a Rapid and Sensitive LC-MS/MS Assay for Dolutegravir Quantitation in Breast Milk

Ashley Rackow (Presenter)
Johns Hopkins University


Introduction: The use of antiretroviral therapy (ART) in pregnant people living with HIV can prevent peri- and post-natal transmission of the virus. Dolutegravir (DTG) is an integrase strand transfer inhibitor that is utilized in several combinatorial ART regimens. DTG is a lipophilic drug, and may deposit in breast milk during the postpartum period. Previous work has shown transplacental and breastmilk transfer of DTG, resulting in neonatal and infant drug exposure. However, DTG pharmacokinetics (PK) in breast milk are incompletely understood. Bioanalytical tools are required to understand the multicompartment pharmacology and efficacy of DTG in preventing mother-to-child HIV transmission.

Objectives: To develop and validate a liquid chromatographic-mass spectrometric (LC-MS/MS) assay in breast milk to support clinical trials research.

Methods: Breast milk was purchased acquired from BioIVT (Hicksville, NY). DTG was acquired from Toronto Research Chemicals (Toronto, ON), and its isotopically labeled internal standard, 13C, 2H5-DTG, was acquired from Alsachim (Illkirch, FR). DTG was spiked into breast milk to prepare calibration standards and quality controls. Drug was isolated via protein precipitation, and reconstituted material was analyzed using an API 5500 mass spectrometer (SCIEX, Redwood City, CA) interfaced with a Shimadzu LC-40. Analytical separation occurred using a Waters Acquity C8 column, and the instrument was operated in selective reaction monitoring and positive ionization modes. Monitored transitions were 420.20/277.10 for DTG and 426.20/133.10 for 13C, 2H6-DTG. The assay was validated in accordance with FDA bioanalytical method validation recommendations.

Results: The analytical measuring range of the assay was 0.5-1000 ng/mL, with an analytical run time of 2.2 minutes. Inter-assay precision and accuracy ranged from 4.6% to 15.1% and -2.4% to 9.6%, respectively, for materials prepared at the lower limit of quantification (0.5 ng/mL), low (1.5 ng/mL), mid (75 ng/mL) and high (800 ng/mL) quality control levels. Matrix effects assessments showed 12% ion suppression for DTG and 13C, 2H5-DTG, and recovery efficiencies of 99%; relative matrix effects were negligible. DTG also demonstrated stability under freeze-thaw, re-injection, and room temperature conditions.

Conclusion: An LC-MS/MS method for DTG quantification in breast milk has been developed and validated. The assay met acceptable performance criteria and may be used in downstream applications to understand the multicompartment pharmacology and distribution of DTG into breast milk.

Topic(s): Small Molecule > Lipidomics > Cases of Unmet Clinical Needs

Poster Presentation
Poster #15b
Attended on Wednesday at 14:30

Let Me Talk Your Ears Off: Determination of Chemical Markers of Ménière's Disease in Earwax

Allix Coon (Presenter)
University at Albany, SUNY



An increasing trend in the field of medical diagnostics is to utilize non-traditional biological matrices that can be collected in a minimally invasive manner as reporters of disease. These can include tears, sweat, fingernails and cerumen, otherwise known as earwax. Earwax is a highly lipid-rich matrix that can potentially provide information on diseases associated with lipid dysregulation. One disorder that is being investigated is Ménière’s disease. It is a disease of the inner ear that results in extreme vertigo and eventual hearing loss. The current approach to disease detection involves eliminating other potential disorders though “diagnosis by exclusion”. As a result, its presence is extremely time consuming to determine, and thus a rapid and accurate method of diagnosis would be of high benefit.


Earwax was collected at a clinic from confirmed Ménière's disease and non-Ménière's disease donors. The individual plugs were suspended in ethyl acetate and the mass spectrometric technique, direct analysis in real time – high-resolution mass spectrometry (DART-HRMS) was used to acquire their chemical profiles. The acquired mass spectra were then subjected to multivariate statistical analysis to ascertain the chemical differences between the two groups.


Earwax is reported to contain hundreds of compounds and thus the DART-MS-derived data were screened against a compiled list of the known earwax constituents contained in the literature. A heatmap rendering of the mass spectral data for the subset of masses that matched those of earwax constituents reported in the literature was created, and analysis of variance was performed to reveal which masses were within a 95% confidence limit. A total of 100 masses were found to be important for differentiating the Ménière's disease and non-Ménière's disease samples. When inspecting the bar graph renderings of these masses, it was concluded that a subset of fifteen were presence/absence markers that were only found in one of the two groups including cholesteryl stearate and a cholesterol derivative with the formula C28H48O that we detected in previous studies. A partial least squares (PLS) classification model built on the basis of these markers exhibited an 83.08% leave-one-out cross validation, and an external validation prediction accuracy of 85% when screening the cerumen of unknowns.


The results revealed 15 potential compounds that could serve as reporters for Ménière's disease. This work demonstrates that: (1) earwax can serve as a viable and readily accessible matrix that can be analyzed to reveal a disease state; and (2) a rapid and accurate Ménière’s disease diagnosis approach could be developed through interrogation of earwax using ambient ionization mass spectrometry.

Topic(s): Small Molecule > Tox / TDM / Endocrine

Poster Presentation
Poster #16a
Attended on Thursday at 09:15

Evaluating Analytical Performance of Tacrolimus LC-MS/MS Assay Using Ascomycin Versus Tacrolimus-C13D2 Internal Standards.

Kwaku Twum (Presenter)
University of North Carolina - Chapel Hill


Objective: To investigate and compare imprecision, accuracy, linearity, and potential matrix interference in tacrolimus LC-MS/MS assay using ascomycin or tacrolimus-C13D2 as internal standards.

Introduction: Therapeutic monitoring of whole blood tacrolimus concentration via liquid chromatography-tandem mass spectrometry (LC-MS/MS) remains essential to reduce the risk of rejection of a transplant organ. Conventional LC-MS/MS approaches for tacrolimus measurement have utilized the internal standard ascomycin, a structural analog. Recently, deuterated isotopically labeled tacrolimus internal standards have become increasingly commercially available, which may provide improved performance over ascomycin. This study evaluated the method performance for measurement of tacrolimus in whole blood using ascomycin versus deuterated tacrolimus (tacrolimus-C13D2) internal standards.

Methods: Whole blood tacrolimus concentrations are measured by LC-MS/MS as part of routine patient care at UNC Health McLendon Clinical Laboratories (Chapel Hill, NC). In brief, EDTA whole blood tacrolimus is extracted using an automated TECAN auto-pipetting system, which involves pipetting (in the following order) 100 µL of zinc sulfate, 100 µL of calibrator/control/sample, 250 µL lysing solution containing acetonitrile and internal standard (ascomycin or tacrolimus-C13D2) into microcentrifuge tubes. Samples are vortexed and centrifuged at 13000g for 3 minutes. A 100 µL aliquot is transferred to a 96-well microtiter plate and analyzed using a Waters Xevo TQD Mass Spectrometer equipped with an Acquity UPLC BEH C18 Column (2.1x50 mm) with an Acquity UPLC BEH C18 guard column (2.1 x 5mm). An electrospray ionization interface is operated in positive ion [ES(+)] mode and MS/MS detection in multiple reaction monitoring mode. The performance in the linearity of calibrators, precision, accuracy, and interference of the matrix using the two internal standards were investigated. To assess imprecision, three levels of UTAK quality control materials [L1(3.65 ng/mL), L2(13.26 ng/mL), and L3(20.32 ng/mL)] spanning the analytical measuring range (AMR) were tested in over 20 analytical runs spanning ten days to calculate percent coefficient of variation (CV). Accuracy was assessed by comparing a) measured results in residual external quality assessment materials with peer results and calculating standard deviation index (SDI) and b) patient result comparisons using tacrolimus-C13D2 to a validated assay that uses ascomycin. Linearity and % recovery of six-point calibration results prepared using tacrolimus-C13D2 or ascomycin for each analytical run were compared. Statistical analysis of agreement between results was all performed using Microsoft Excel.

Results: Between-day imprecision was measured at the CV of 8.40% (L1), 2.66% (L2), 3.67% (L3), and 3.63% (L1), 3.87% (L2), 9.27% (L3) for QC measured with tacrolimus-C13D2 and ascomycin respectively. Using previously reported patient samples (n=17) with ascomycin as the internal standard, tacrolimus-C13D2 showed a mean bias of +0.55 ng/ml or 3.00%. Testing of residual CAP survey samples showed a better CAP survey challenge closer to the peer group mean for tacrolimus-C13D2 (SDI -1.48, -1.17, and -1.35) compared to ascomycin (SDI -2.1, -2.2, -2.1). Six calibrator levels measured each day of testing demonstrated linear response and showed a slightly better recovery of manufacturer-assigned concentrations using tacrolimus-c13d2 (mean bias -0.02%, R2 of 0.9997) compared to ascomycin (mean bias -0.58%, R2 of 0.9984).

Conclusion: Whole blood tacrolimus measurement via LC-MS/MS demonstrated comparable analytical performance using tacrolimus-C13D2 or ascomycin as an internal standard. Linearity, imprecision, and accuracy in tacrolimus measurements showed acceptable performance for both internal standards, though a negative bias in measured tacrolimus compared with peers was observed. In evaluating the choice of internal standard, other factors such as cost and ease of obtaining pure isotopically labeled analog must also be considered.

Topic(s): Small Molecule > Cases in Clinical Analysis > Cases of Unmet Clinical Needs

Poster Presentation
Poster #20b
Attended on Thursday at 12:15

Desorption Electrospray Ionization Mass Spectrometry as a Sensitive Tool for Intraoperative Diagnosis of Isocitrate Dehydrogenase (IDH) Mutation in Glioma

Mahdiyeh Shahi (Presenter)
Purdue University


Introduction: Glioma is the most common neoplasm in the central nervous system with tumor resection as the primary treatment option and total tumor resection being associated with better patient outcomes. However, gross total resection is difficult to achieve due to the infiltrative nature of glioma. Thus, rapid detection of molecular features with high sensitivity and minimal to no sample preparation is crucial for intraoperative diagnosis. Isocitrate dehydrogenase (IDH) mutation status plays a critical role in the diagnosis and prognosis as patients with IDH-mutant glioma exhibit higher survival rates. However, IDH mutations are commonly determined by immunohistochemistry (IHC) and genetic testing, methods that are incompatible with intraoperative evaluation. Mutations in IDH enzymes lead to the accumulation of an oncometabolite, 2-hydroxyglutarate (2HG), which is the product of α-ketoglutaric acid reduction in a nicotinamide adenine dinucleotide phosphate (NADPH)-dependent manner. Consequently, elevated concentrations of 2HG are associated with IDH-mutant glioma, which can play a significant role in the assessment of IDH mutation status in glioma. Desorption electrospray ionization mass spectrometry (DESI-MS) enables rapid and sensitive detection of 2HG in glioma biopsies. Cystathionine, a key intermediate in the transsulfuration pathway, has been found in high concentrations in glioma tumors, and it too can be detected by DESI-MS and targeted as a potential biomarker in glioma samples.

Objectives: In this study, DESI-MS is used for intraoperative assessment of IDH mutation status in glioma patients by measuring relative 2HG concentration in less than two minutes.

Methods: Adult patients undergoing glioma resection were recruited and their biopsied tissues were analyzed during surgery by DESI-MS. Tumor biopsies were smeared on glass slides and analyzed by DESI-MS with no subsequent sample treatment. Measurements were done in MS/MS mode to improve specificity and minimize matrix interferences. Signals for 2HG fragment ions were measured relative to those of glutamate (Glu), as the endogenous reference molecule.

Results: The DESI-MS predictions of IDH mutation status made on 129 core samples from 34 patients agreed with IHC and/or polymerase chain reaction (PCR) results (sensitivity, specificity, and accuracy all 100%). In addition, 2HG was observed in 80% of the margin biopsies indicating that the tumor infiltrates beyond the surgical margins. Initial offline studies of cystathionine revealed high concentrations in IDH-mutant and IDH-wildtype tumors compared to the normal mouse brain.


IDH mutation status was assessed correctly through direct analysis of brain biopsies by DESI-MS in less than two minutes. IDH-mutant glioma samples were strongly associated with high concentrations of 2HG. Elevated concentrations of 2HG in margin biopsies show that tumor infiltration beyond surgical margins is common. The results show the value of DESI-MS as a rapid and sensitive tool for intraoperative evaluation of glioma with the potential for achieving better patient outcomes.

Topic(s): Proteomics > Cases of Unmet Clinical Needs

Poster Presentation
Poster #23b
Attended on Wednesday at 14:30

Detection of Patient Monoclonal Serum Free Light Chains by On-Probe Extraction coupled with High-Resolution Mass Spectrometry

Priscilla Yeung (Presenter)
Stanford University Department of Pathology


INTRODUCTION: Multiple myeloma is the second most common hematolymphoid malignancy in the United States and has a median survival of 6 years after diagnosis. It is part of a larger group of plasma cell neoplasms that range from monoclonal gammopathy of uncertain significance (MGUS), to smoldering myeloma (SM), to fulminant multiple myeloma (MM), and include extramedullary manifestations such as plasmacytomas or amyloidosis. One of the essential clinical markers of plasma cell neoplasms is serum free light chains (FLC), which are circulating antibody light chains that are unbound to heavy chain. The current widely used immunoassay method quantifies total serum FLCs, including the polyclonal background. Although a skewed kappa/lambda (K/L) FLC ratio is often used as a proxy for clonality, some patients, such as those with renal disease, can have ambiguous results that can benefit from a direct measurement of clonality.

OBJECTIVES: The purpose of this study was to develop an assay that couples an on-probe extraction immunocapture step with high-resolution mass spectrometry (OPEX-MS) to determine the clonality of serum free light chains.

METHODS: Remnant patient samples with FLC immunoassay results from the Stanford Clinical Chemistry Laboratory were collected and processed according to Institutional Review Board protocols approved by Stanford Health Care. Immunocapture with on-probe extraction was performed using streptavidin probes and biotinylated capture antibodies (anti-kappa light chain and anti-lambda light chain) on a Gator Plus analyzer. Captured proteins were eluted using 0.3% formic acid to break the antibody-antigen association and rinsed with phosphate buffered solution with Tween. The analyte capture and elution steps were repeated 30 times for optimal signal to noise ratio. Liquid chromatography-mass spectrometry was performed on a TLX-2 multi-channel HPLC coupled with a Q-Exactive Plus mass spectrometer. Proteins were separated on a reverse phase HPLC column (MAbPac, 2.1 mm X 5 mm). Mobile phase A was 0.1% formic acid in water and mobile phase B was 0.1% formic acid in acetonitrile. Data analysis was performed on BioPharma Finder 5.1 software using a sliding window ReSpect deconvolution algorithm.

RESULTS: Four cohorts of samples from unique patients were tested based on Binding Site Optilite FLC results: Negative (n = 50), Kappa Positive (n = 49), Lambda Positive (n = 46), and Dual Elevated (n = 100). In the Negative, Kappa Positive, and Lambda Positive cohorts, OPEX-MS showed good overall concordance (91%) with the immunoassay method. In samples with conflict results, majority of the samples in the Kappa Positive cohort that were negative by OPEX-MS had K/L FLC ratios that were only slightly elevated. OPEX-MS identified clonal FLCs in a subset of the Dual Elevated cohort with a normal K/L ratio.

CONCLUSIONS: The OPEX-MS method can serve as a complementary approach to determine clonality in patients with difficult-to-interpret FLC immunoassay results.

Topic(s): Small Molecule > Tox / TDM / Endocrine > Cases in Clinical Analysis

Poster Presentation
Poster #24b
Attended on Thursday at 12:15

Concordance of Newborn and Maternal Drug Screen Results by Immunoassay and Mass Spectrometry

Hannah Brown (Presenter)
Washington University School of Medicine in St. Louis



The interplay between maternal health and neonatal outcomes informs current guidelines that recommend universal screening for substance use disorders in obstetric patients. As part of this screening process, maternal and neonatal drug testing is frequently performed to detect in utero drug exposure. While maternal urine drug testing exhibits high sensitivity and specificity, neonatal urine drug testing faces preanalytical and analytical challenges due to difficult sample collection protocols, low drug concentrations, and drug metabolites differing from those targeted by immunoassay (IA) or mass spectrometry (MS)-based methods. Consequently, meconium is often the preferred specimen type to detect neonatal drug exposure due to a longer window of detection compared to urine. Most laboratories do not have the ability to analyze meconium and, therefore, send out meconium testing to specialized reference laboratories, which can delay results for several days or more. In response highly sensitive MS-based methods for rapid, in-house neonatal urine drug testing have been developed, although the concordance of these methods with gold-standard meconium testing is not well documented.


The primary objective of this study was to evaluate concordance of drug screening results of paired newborn urine and meconium samples with maternal urine samples using IA and liquid chromatography mass spectrometry (LC-MS/MS) methods, aiming to investigate maternal-fetal drug transfer.


The number of positive drug screen results for newborns born at Barnes Jewish Medical Center with paired urine and meconium samples from January 2021 to October 2022 were tabulated retrospectively for the de-identified dataset (n=1,424). Of the 1,424 newborns, 831 newborns had mothers with drug screens performed within three days of birth (+/- 3 days). Urine drug testing was performed on freshly collected urine by in-house LC-MS/MS methods. All paired meconium samples were sent to an outside reference laboratory where they were screened by IA and confirmed with MS. The menu of available meconium testing performed in our patient population allowed for a paired urine-meconium comparison for the following drugs: amphetamines, cocaine, cannabinoids, opiates, and cannabinoids. The cutoffs for the MS urine screening method (in ng/mL) were: amphetamines 5, cocaine 1, opiates 25-250 (depending on specific compound), cannabinoids 20 (as 11-nor-9-Carboxy-Δ9-tetrahydrocannabinol glucuronide). The cutoffs for meconium screening (in ng/g) were: amphetamines 100, cocaine 100, opiates 100, cannabinoids 20.


Of the 1,424 newborns, 120 screened positive for amphetamines, 65 screened positive for cocaine, and 2 screened positive for PCP in urine and/or meconium. Positivity rates were equivalent for amphetamines (7.4%), cocaine (3.9%), and PCP (0.1%). McNemar’s test p-values indicate there is no significant difference between urine and meconium for the detection of amphetamines and cocaine (p < 0.001). There was insufficient data to analyze concordance for PCP due to low positivity rate. For opiates, 57 newborns screened positive. The drug class with the highest positivity rate is cannabinoids, with 866 newborns screening positive. There was a significant difference in positivity rate (percent positive agreement (PPA)) for opiates and cannabinoids (20% and 2.5%, respectively). A cohort of 831 newborns with mothers having a urine drug screens performed within three days of birth were identified. Positivity rates were higher in maternal urine for all drug classes except cannabinoids (69.4% positivity rate in maternal urine; 96.9% positivity rate in meconium). Notably, in the case of amphetamines, cocaine, and opiates, an additional 7, 24, and 32 instances of in utero drug exposure were identified in maternal urine alone, accounting for 9%, 34%, and 42% of all positive screens, respectively. Specifically in the case of cannabinoid positivity, 99.8% of all newborns with in utero cannabinoid exposure screened positive in meconium and/or maternal urine.


In this retrospective analysis including paired data from meconium and urine collected from 1,424 neonates, we observed that a sensitive and rapid LC-MS/MS method for neonatal urine tests detected equivalent numbers of neonates exposed to cocaine and/or amphetamines in utero compared to common meconium drug testing methods. Further, given the poor analytical performance of screening for cannabinoids in urine, meconium only testing may be appropriate for this drug class, in combination with maternal urine drug testing. Higher positivity rates in maternal urine highlight the importance of incorporating maternal drug screening results into the interpretation of newborn urine drug screens, as it provides a comprehensive understanding of fetal exposure, especially in instances in which meconium is unavailable. This study highlights the potential value of sensitive and rapid LC-MS/MS methods for neonatal urine drug testing.

Topic(s): Proteomics > Emerging Technologies > Cases of Unmet Clinical Needs

Poster Presentation
Poster #25b
Attended on Wednesday at 14:30

From the Non-invasive Breast Microenvironment to Metastatic Breast Cancer: Pathological Variation in Collagen Proteomic Signatures by Mass Spectrometry Imaging

Taylor Hulahan (Presenter)
Medical University of South Carolina


Introduction: While ductal carcinoma in situ (DCIS) does not directly pose a significant mortality risk, the subsequent development of invasive breast cancer (IBC) doubles a patient’s breast cancer-specific mortality rate compared to the general population. Once diagnosed as IBC, metastasis in regional areas decreases the survival rate by 13%. Notably, the progression to IBC is characterized by the disruption of an intact basement membrane, in which collagens are a key constituent. Prolyl hydroxylases are understood to be essential in breast cancer invasion and metastasis, but hydroxylation sites of collagen critical to the primary tumor and metastatic sites have yet to be defined. We hypothesize that the hydroxylation status of discrete collagen proline residues contributes to pathologies of breast cancer progression spanning non-invasive to metastatic disease.

Methods: Extracellular matrix (ECM)-targeted spatial proteomics using Matrix-Assisted Laser Desorption/Ionization-Quadrupole Time-Of-Flight (MALDI-QTOF) imaging followed by high-resolution mass-accuracy proteomics for peptide identification was used to assess ECM proteomic changes within non-invasive, invasive, and metastatic breast cancers. To investigate ECM peptide signatures of DCIS and IBC pathologies, a comparative analysis was performed using a 22-sample cohort derived from 17 different patients defined as pure DCIS (n=9), mixed DCIS-IBC (n=8), and pure IBC (n=4). To spatially define ECM alterations with breast cancer metastases, lymph node metastases were compared to patient-matched primary tumors and normal lymph nodes using specimens from triple-negative breast cancer (TNBC) patients (n=5) and tissue microarrays from 31 generational South Carolina women with IBC (n=21).

Results: Comparative analysis between DCIS and IBC pathologies revealed over 1,000 putatively identified peaks that were linked to pathological annotations or adjacent regions. Forty-three peaks had significantly different intensity profiles between DCIS and IBC pathologies by Mann-Whitney U-test (p-value<0.05). Eight of these differentially expressed peptides were identified as fibrillar collagen sequences within annotated triple helical regions and many contained hydroxylated proline sites. These fibrillar collagen domains could discriminate between pathologies by area under the receiver operating curve (AUROC)≥0.75 and Wilson/Brown t-test (p-value<0.05). To determine if proximity to IBC influenced the proteomic profile of DCIS pathologies, mixed DCIS-IBC specimens (n=4) with DCIS lesions varying distances from the invasive cancer field were examined for ECM proteomic field effects. Notably, DCIS lesions within invasive regions had a more similar proteomic profile to IBC than DCIS lesions located more distally from the IBC region via principal component analysis and hierarchical clustering. A four-sample subset used for higher-resolution analysis demonstrated DCIS intra-tumoral heterogeneity at the individual lesion level. To assess collagen domain variation with metastases, a study on the TNBC metastatic niche of axillary lymph nodes has been initiated. Spatial segmentation analysis of 187 putatively identified peptides and 680,216 pixels revealed 10 uniquely localized proteomic groups with groups shared between the primary tumor and the metastatic niche. Of these 187 putatively identified peptides, 158 peptides were similarly expressed between the primary tumor and metastatic lymph nodes by Wilcoxon rank-sum test (p>0.05). Seven peptides could discriminate between metastatic and normal lymph node specimens, while 22 peptides could discriminate between metastatic lymph nodes and the primary tumor (AUROC>0.70; p-value < 0.05). Within the tissue microarray set, 10% of peptides from fibrillar collagens with variable sites of proline hydroxylation could discriminate between metastatic and normal lymph nodes (AUROC>70%; p-value< 0.01).

Conclusions: Our preliminary interrogation highlights emerging differences in collagen post-translational site modifications between DCIS, IBC, and breast cancer metastases. These signatures could be useful for understanding breast cancer recurrence and progression. Further investigation is being pursued in an expanded, racially diverse cohort across all stages of the breast cancer spectrum.

Topic(s): Small Molecule > Tox / TDM / Endocrine > Metabolomics

Poster Presentation
Poster #28a
Attended on Thursday at 09:15

Development of Simple LC-FAIMS-MS/MS Method for the Quantification of Nicotine and Its Metabolites in Urine

Danting Liu (Presenter)
Mayo Clinic



The use of tobacco products, particularly smoking, is the main preventable cause of lung cancer in the U.S. Nicotine is the active component in tobacco, responsible for addiction. The clinical test for nicotine and its metabolites in urine is a widely accepted method to evaluate nicotine exposure. In addition, anabasine, an analog of nicotine present in trace amounts in tobacco products, is also used as an indicator for monitoring compliance in tobacco cessation and effectiveness in nicotine replacement therapy. Currently, the conventional methods used to detect these molecules in the clinical laboratory are direct dilution or solid phase extraction (SPE), followed by LC-MS/MS analysis. However, direct dilution and injection methods may introduce bias in quantification and are prone to interferences, especially with low concentration samples. SPE purification is effective at overcoming these limitations, but this technique adds cost and labor to the analysis. Here, we present a novel LC-FAIMS-MS/MS method capable of detecting nicotine and its metabolites in urine with better sensitivity and specificity, without adding labor and cost.


Residual urine samples from the nicotine positive and negative patients were used in this study. In the test, 50 μL of calibrator, QC, blank, or patient urine samples were mixed with 50 μL of internal standard working solution (200 ng/mL Nicotine-D4, Nonicotine-D4, Anabasine-D4 and Cotinine-D3) and 800 μL of Mobile phase A (5 mM Ammonium Bicarbonate in 0.1% Formic Acid) in a 96 well-plate at room temperature. The sample was then separated by the LC column (C18, 3.0 x 50 mm, 2.6 um PS C-18 Kinetex) using a 10 min gradient (mobile phase B, 0.1% Formic Acid in ACN) and analyzed by TSQ Altis Plus Triple Quadrupole MS with or without FAIMS installed. The solid phase extraction (SPE) method was performed to compare with the dilution method. Nicotine and its metabolites in 50 μL of the urine sample were extracted with a strong cation Bond Elut PlexaPCX 30 mg plate (Agilent), followed by the same LC-MS/MS method. The four analytes were quantitated using internal standards and calibration curves.

Result and Conclusion:

The results obtained from the LC-FAIMS-MS/MS method showed that, compared with the regular dilute-and-shoot method, using the FAIMS cleaned up the background noise up to 80% and largely increased the signal-to-noise ratio (S/N). The chemical noise signals that interfere with the quantification of anabasine and cotinine, especially at low concentrations, were completely eliminated. The Compensation Voltage (CV) values used in FAIMS for each analyte were optimized to produce the highest quantitative accuracy and precision. The agreement between the quantifier and qualifier ions was excellent, with linear regression analysis yielding R2 greater than 0.99 for all four analytes. The total imprecision of the assay was less than 10% across the analytical measuring range. Overall, our new method is simple and utilizes low cost sample preparation, requires minimal instrument maintenance, while providing accurate, precise, and specific measurements, which makes this method an ideal, high throughput analysis technique for the clinical laboratory.

Topic(s): Small Molecule > Tox / TDM / Endocrine > Identifying High Value Tests

Poster Presentation
Poster #29a
Attended on Wednesday at 12:15

Evaluation of the Roche Benzodiazepines II Immunoassay for Urine Drug Testing in Clinical Specimens

Mengyuan Ge (Presenter)
UC San Diego Health


Background: Benzodiazepines are one of the most commonly prescribed medications in the United States and are frequently linked to instances of abuse and overdose. Historically, FDA-cleared benzodiazepine urine immunoassays cross-react poorly with glucuronidated metabolites excreted in urine. False negative results are especially prevalent with lorazepam which is almost exclusively excreted at lorazepam-glucuronide. Some clinical laboratories have addressed this problem with the addition of beta-glucuronidase to enhance assay sensitivity as a laboratory developed test (LDT). Roche Diagnostics recently received FDA clearance to offer a benzodiazepine immunoassay that includes beta-glucuronidase.

Methods: Performance characteristics of two FDA-cleared benzodiazepine urine immunoassays (Benzodiazepine Plus, no glucuronidase and Benzodiazepines II, with glucuronidase; Roche Diagnostics) and a benzodiazepine immunoassay LDT (HS-BENZ, with glucuronidase) were evaluated using 258 urine specimens. These immunoassays were directly compared to an LC-MS/MS benzodiazepine LDT to determine clinical sensitivity and specificity. Cross-reactivity of all three immunoassays were compared and evaluated based on the measured benzodiazepine concentrations determined by the LC-MS/MS LDT. Cross-reactivity for 7-aminoclonazepam and lorazepam was assessed using drug-free urine spiked with reference materials (Cerilliant) at concentrations ranging from 100 to 1000 ng/mL.

Results: The Benzodiazepines II and LDT immunoassays exhibited greater clinical sensitivity (100% and 95.2%) compared to the Benzodiazepines Plus assay (66.7%). Clinical specificity of 100% was observed for all three assays. Cross-reactivity of the Benzodiazepines II assay was greater across the range of benzodiazepine concentrations tested in comparison to two other immunoassays, and in particular, the 7-amino-clonazepam metabolite as it was the only immunoassay out of the three that was sensitive enough to detect the presence of specimens containing this metabolite alone. Cross-reactivity analysis shows that the Benzodiazepine II assay detected 7-aminoclonazepam or lorazepam at 300 ng/mL, while the other two assays required 800 ng/mL and 600 ng/mL for a positive result.

Conclusions: A comprehensive evaluation of these three immunoassays demonstrates that the Benzodiazepines II immunoassay has increased clinical and analytical sensitivity compared to the Benzodiazepines Plus and HS-BENZ immunoassays. The inclusion of a beta-glucuronidase greatly improved the sensitivity of the Benzodiazepines II and HS-BENZ immunoassays for lorazepam, which is primarily excreted as a glucuronide metabolite in urine. As one of the first commercially available FDA-cleared benzodiazepine urine immunoassays which incorporates a beta-glucuronidase, clinical laboratories should consider implementing this assay to detect benzodiazepines more robustly in their patient populations.

Topic(s): Small Molecule > Tox / TDM / Endocrine

Poster Presentation
Poster #30b
Attended on Thursday at 12:15

Analytical Validation of an Automated Dried Blood Spot Desorption LC-MS/MS Method for Levetiracetam and Lamotrigine

Joshua Miller (Presenter)
Mayo Clinic


Introduction: Levetiracetam and lamotrigine are antiepileptic drugs used in the treatment of tonic-clonic seizures but require monitoring due to their narrow therapeutic ranges. Traditional monitoring methods utilize serum samples, requiring frequent venipunctures for patients. Dried blood spots (DBS) are an alternative specimen type which can easily be collected at home. However, DBS require laborious sample preparation before measurements using LC-MS/MS, limiting their use in high test volume laboratories. Herein we report a novel LC-MS/MS method that allows for the coincidental quantification of levetiracetam and lamotrigine from DBS utilizing direct inline flow through desorption technology which requires no sample preparation.

Objective: The goal of this study was to validate an LC-MS/MS method utilizing inline flow-through desorption technology to quantify levetiracetam and lamotrigine from DBS samples.

Methods: DBS were prepared by spiking bovine whole blood with levetiracetam and lamotrigine and depositing these mixtures onto dried matrix cards. Direct inline matrix desorption and chromatographic separation were achieved using a Thermo Scientific DSX-1 equipped with a Vanquish HPLC system. Samples were purified online using a Cyclone-P (1 x 50 mm) TurboFlow column and then eluted onto an Accucore™ aQ (3 x 150 mm) analytical column for further chemical separation. The HPLC was coupled to a Thermo Scientific TSQ Altis Plus triple quadrupole mass spectrometer for ion selection and detection. Sample quantification was performed using isotopically labeled internal standards and by comparison to an external calibration curve. Each batch contained five calibrators with concentrations across the linear range. Twenty runs over the course of ten days were used to determine accuracy and precision.

Results: Calibration curves for both anticonvulsants maintained a linear response (R2>0.99) which encompassed their respective therapeutic windows. Within-run precision varied with %CVs of 3.6-7.8% and between-run precision varied with %CVs of 4.5-9.7% for both drugs. Passing-Bablok regression analysis of accuracy studies revealed excellent correlations (R2=0.99) for both drugs. Our method yielded a 14-day mean difference in concentration <±20% and <±10% compared to their originally measured concentration for levetiracetam and lamotrigine, respectively, indicating that the DBS are stable for at least 14 days.

Conclusions: Automatic flow through desorption technology in tandem with LC-MS/MS can accurately and precisely quantify levetiracetam and lamotrigine from DBS samples. This investigation bolsters the number of analytes that are amenable to DBS desorption LC-MS/MS analysis, highlighting the versatility of this technology for high throughput laboratories. Furthermore, the utilization of DBS versus serum has the potential to benefit patients undergoing long-term therapeutic drug monitoring.

Topic(s): Other -omics > Metabolomics > Precision Medicine

Poster Presentation
Poster #31a
Attended on Wednesday at 12:15

Innovations and Applications of Quantitative Metabolomics in Human Health

Rupasri Mandal (Presenter)
University of Alberta


INTRODUCTION: Metabolomics involves the comprehensive characterization of small molecule metabolites in specific biological specimens, such as blood, urine, feces and saliva. Metabolomics has traditionally been divided into two complementary but somewhat different approaches: untargeted and targeted metabolomics. Untargeted metabolomics is a discovery-based approach that aims to “qualitatively” measure all detectable spectral features over a large number of biospecimens. Targeted metabolomics studies are aimed at quantitatively measuring a small number of targeted compounds that can be readily used in clinical chemistry, medical diagnoses, and disease biomarker discovery studies. While untargeted methods can generate data on up to 1000 compounds, targeted methods usually generate data only on 100-200 compounds.

OBJECTIVE: In an effort to allow targeted metabolomics to achieve the same coverage as untargeted metabolomics, we have developed a custom made, comprehensive, quantitative LC-MS/MS-based assay for targeted metabolomics analysis of biospeimens such as serum, urine and fecal extracts. This assay allows for the identification of up to 1000 metabolites along with another 400 metabolite sums and ratios.

METHODS: Our method uses a combination of direct injection (DI) mass spectrometry with a reverse-phase LC-MS/MS. It combines the derivatization and extraction of analytes, and the selective mass-spectrometric detection using multiple reaction monitoring (MRM) pairs. Two separate panels involving two different precolumn derivatization reactions were developed for this assay: Panel A - Phenylisothiocyanate (PITC) derivatization targeting amine-containing compounds and Panel B - 3-nitrophenylhydrazine (3-NPH) derivatization targeting keto- and carboxyl-containing compounds. Isotopically-labeled internal standards are used for metabolite quantification. Calibration of metabolite concentration ranges in both panels was adjusted for different biofluid types.

For panel A, a 96 deep-well plate with a filter plate attached via sealing tape, containing the required reagents and solvents, was used to prepare the plate assay. The first 14 wells of each plate are used for calibration and quality control purposes. For all metabolites, except organic acids, samples are first thawed on ice and then vortexed and centrifuged at 13,000x g. 10 µL of each sample are loaded onto the center of the filter on the upper 96-well plate and dried in a stream of nitrogen. Subsequently, PITC is added for derivatization. After incubation, the filter spots are dried again using an evaporator. Extraction of the metabolites is then achieved by adding 300 µL of extraction solvent. The extracts are obtained by centrifugation into the lower 96-deep well plate, followed by a dilution step with MS running solvent.

For organic acid analysis (panel B), 90 µL of ice-cold methanol is added to 30 µL of each sample for overnight protein precipitation. The sample is centrifuged at 13000x g for 20 min. 50 µL of supernatant was loaded into the center of wells of a 96-deep well plate, followed by the addition of NPH. After incubation for 2h, isotope-labeled internal standards, BHT stabilizer and water are added before LC-MS injection.

Mass spectrometric analysis was performed on Sciex 5500 QTrap® tandem MS instrument equipped with an Agilent 1290 series UHPLC system. The samples are delivered to the MS by a standard LC method followed by a direct injection (DI) method. Data analysis was done using Analyst 1.6.2. Calibration regression, accuracy and precision of QC standards, and spiked recovery of each targeted metabolite were used for method validation.

RESULTS: This custom assay can be used for the targeted identification and quantification of up to 1000 metabolites across 21 chemical classes including amino acids and derivatives, biogenic amines, organic acids, nucleotides/nucleosides, ketone and keto acids, indole derivatives, vitamins and derivatives, sulfates, dipeptides and 11 different classes of lipids. In addition, 400 biomedically related metabolite sums and ratios can be calculated from concentrations of these 1000 targeted metabolites. In total, this custom assay can measure up to 1400 metabolites and ratios/sums. The accuracy of QC standards with 3 different concentration levels are in the range of 80% to 120% with satisfactory precision values of less than 20%. The recovery rates of spiked serum, urine and fecal extract samples with three different concentration levels are in the range of 80% to 120% with satisfactory precision values of less than 20%. More than 1500 biological samples such as serum, urine and fecal extracts have been analyzed in our laboratory variations of this assay since 2021. The analysis time is approximately 1 sample/hour on a single LC-MS/MS instrument or about 100 samples over 4 days.

CONCLUSIONS: We have developed a comprehensive, sensitive, high-throughput, low-volume, quantitative targeted LC-MS/MS assay for the analysis of up to 1400 metabolites and ratios, across 21 chemical classes for serum, urine, and fecal extracts. Only 40 µL of a given sample are required for the entire analysis. We have adapted this assay into a 96-well plate format to enable high-throughput analysis. This targeted assay will be able to deliver the same metabolite coverage with far more precise quantification, over much shorter periods of time at much lower costs than the most comprehensive untargeted assays. We believe that targeted metabolomics assays can now exceed the performance of untargeted assays and will continue to pull away from untargeted methods in terms of both performance and utility over the coming years.

Topic(s): Proteomics > Various OTHER > none

Poster Presentation
Poster #35a
Attended on Wednesday at 12:15

Surface Proteomics Identifies Distinct Therapeutic Targets for ALK-Positive Lung Cancer

Jacelyn Greenwald (Presenter)
Ohio State University


<font color='red'><b>Admin Note:</b> This abstract was has changed from the original submission of "Using Surfaceomics to Assess the Role of Adipose Tissue-Derived Extracellular Vesicles in Alzheimer’s Disease" on March 9, 2024.</font>


We hypothesize that aberrant expression of the ALK gene in ALK+ lung cancer induces the expression of yet unidentified surface-accessible proteins, which can be targeted for therapeutics. Our objective is to isolate these surface proteins, with the intent to better understand the downstream consequences of ALK rearrangement and fusion. These proteins can then be targeted with antibody-based therapeutics to improve ALK cancer survival outcomes.


For this study, we will use the H3122 and H2228 cell lines which have naturally occurring EML4-ALK translocations. Cells will be biotinylated using NHS esters of biotin (Thermo). A second surface technique will involve using trypsin in small quantities (2-3ug/mL) and incubating cells for 30 minutes. Both the surface and "total minus surface" fractions will be analyzed. A third cohort will be whole intact cells.

Protein digestion and surface protein isolation: For surface proteomics (surfaceomics) analysis, we apply two techniques: first, sulfo-N-hydroxysulfosuccinimide (NHS) esters of biotin (Thermo) are used to label primary amine groups on the cell surface. These tags form an amide bond with primary amines on surface-exposed lysines and N-termini. Surface proteins are then captured using neutravidin agarose (Thermo). The sulfo-NHS-SS-biotin tag includes a disulfide bond that can be reduced to elute the proteins from the neutravidin column.

A second surfaceomics approach involves light surface trypsinization to “shave” the cell and release proteins from its surface. Surface shaving uses trypsin in small quantities (around 2-3 μg/ml) and incubates cells for 15-30 minutes with shaking. Cells are then centrifuged to separate the cleaved surface peptides from the remaining shaved cells.

Mass spectrometry-based proteomics: All samples are subject to lysis with detergent, followed by reduction and alkylation. Digestion is performed using MS grade trypsin/LysC on a Protifi S-trap micro spin column. Peptides are eluted, then dried down and resuspended in water with 0.1% formic acid. LC-MS/MS is performed using a commercial timsTOF Pro mass spectrometer (Bruker Scientific) coupled to a nanoElute 2 high-performance liquid chromatography (HPLC) column. Samples are run on the column using a 120-minute linear reverse phase gradient. The column is 25 cm in length with 1.7 μm particles (IonOpticks, Australia). For each LC-MS/MS experiment, 200 ng of protein is loaded.

Data are collected using data-dependent acquisition (DDA) mode. The timsTOF Pro also has ion mobility, which adds an additional level of separation to the HPLC. The parallel-accumulation serial fragmentation (PASEF) capability of the timsTOF Pro allows us to collect more mass spectrometry information, leading to greater quantification and more sensitive protein assignment.

Data analysis: TimsTOF DDA-PASEF datasets are processed using FragPipe in conjunction with the MS/MS search engine MSFragger and IonQuant. Proteome FragPipe outputs are further processed using FunRich software and Reactome pathway analysis.


Preliminary studies have identified distinctly different protein expression profiles between surface and whole cell populations, with 2,887 proteins significantly differentially expressed between the two (adjusted p-value 0.05). Pathways which are upregulated in the surface fraction include toll-like and other receptor signaling, while cells without their surface intact expressed a greater proportion of metabolic and biosynthetic processes. Interestingly, many of the proteins implicated in cancer were upregulated in the surface fraction, including proteins mapping to non-small cell lung cancer pathways.

Comparing the control cell lines to the EML4-ALK variants reveals several proteins of interest, including CHI3L2 chitinase-3-like protein-2. This has already been identified as a biomarker in gliomas and is correlated with immune infiltration. Both EML and ALK are upregulated in the mutant populations, which validates the model.


The initial data demonstrates the surface protocol has great potential to identify novel therapeutic targets which can be attributed to ALK activation. As we continue to build upon this initial dataset, we anticipate the identification of strong targets for ALK+ lung cancer treatment.

Topic(s): Small Molecule > Emerging Technologies > Metabolomics

Poster Presentation
Poster #36b
Attended on Thursday at 12:15

High Throughput DESI Tissue Analysis of Angiotensin II Metabolism in Mice

Kenneth Virgin (Presenter)
Purdue University, Indiana University School of Medicine


INTRODUCTION: Angiotensin metabolism is essential for blood pressure regulation. Many antihypertensive drugs target the production and action of angiotensin II. However, angiotensin II is only part of an intricate cascade that is known to induce adverse effects such as dry cough or angioedema when disrupted. A technique is needed to rapidly analyze multiple metabolites simultaneously for new drug design and potentially patient-specific treatment. High throughput desorption electrospray ionization (HT-DESI) can analyze complex solutions of small metabolites at a rate of less than 1 second per sample with no additional sample preparation necessary. We hypothesized that HT-DESI could quantitatively monitor the production of angiotensin metabolites in mouse tissue.

METHODS: We prepared mouse tissue homogenates of liver, brain, jejunum, and kidney. We added angiotensin II to each sample to saturate tissue enzymes and quenched aliquots of the reaction over the space of one hour. A 50 nL aliquot of each solution was transferred to a PTFE-coated glass slide and analyzed under ambient conditions using a Waters DESI source and QTOF Synapt G2 Si mass spectrometer. No internal standard was used.

We then calibrated angiotensin metabolites as well as bradykinin and bradykinin fragment in the range of 10 to 100 uM using DADLE as an internal standard. Calibration samples were then analyzed with HT-DESI as stated previously.

RESULTS: Angiotensin II was almost completely converted into angiotensin III, IV, and (1-7) within an hour of incubation in tissue homogenate. The relative intensity of angiotensin II metabolites was markedly different between different tissue types suggesting tissue-specific metabolism consistent with current literature. The data showed very little noise between samples, but we did not include an internal standard to quantify the concentration changes.

Subsequent calibration of angiotensin metabolites with DADLE as an internal standard showed linearity in the range of 10 to 100 uM for bradykinin R^2 = 0.994, bradykinin fragment R^2 = 0.992, angiotensin II R^2 = 0.993, angiotensin III R^2 = 0.996, angiotensin IV R^2 = 0.991, angiotensin (1-7) R^2 = 0.998. Angiotensin converting enzyme II has a Km of 10-15 uM which is in the range of our calibration.

Tissue homogenates will be reanalyzed with DADLE as an internal standard to quantify tissue enzyme kinetics with and without ACE inhibitors enalapril and lisinopril and the ARB losartan.

DISCUSSION: The present data demonstrate the ability of HT-DESI to rapidly monitor angiotensin metabolism without additional sample preparation. Additionally, HT-DESI can quantify metabolite concentration with the use of an internal standard. By analyzing the substrate metabolism in vivo, one could rapidly assess the metabolic impact of ACE inhibitors or other drugs on multiple metabolic pathways.

This technology could accelerate the design of new antihypertensive compounds as well as open the path to design patient-specific drugs based on individual metabolic variations.

Topic(s): Small Molecule > Various OTHER > none

Poster Presentation
Poster #37a
Attended on Wednesday at 12:15

Establishment of an LCMS/MS Assay to Measure 11-Oxygenated Androgens

Tatiana Coverdell (Presenter)
National Institutes of Health


INTRODUCTION: 11-oxygenated androgens are important biomarkers for many disorders, including polycystic ovary syndrome (PCOS) and congenital adrenal hyperplasia (CAH). Unfortunately, testing of these compounds is limited due to a lack of assay availability, especially in a clinical setting. Future studies utilizing an LCMS/MS method for measuring these compounds could allow for the characterization of these biomarkers that accurately reflect disease control and facilitate treatment monitoring.

OBJECTIVES: The primary objective of this study was to establish an in house assay using liquid chromatography tandem mass spectrometry (LCMS/MS) to quantitatively measure the following 11-oxygenated androgens in serum: 11-ketotestosterone, 11-hydroxytestosterone, 11-hydroxyandrostenedione, and 11-ketoandrostenedione.

METHODS: Serum was obtained from healthy and affected patient populations. Serum was then combined with internal standards, diluted with deionized water, and then loaded on a supported liquid extraction cartridge. Methyl-tert-butyl-ether (MTBE) was used to elute the steroids from the column. The eluate was then concentrated under nitrogen and reconstituted with methanol and water. Samples were then injected via an autosampler and resolved with a C18 column on an Agilent 1290 binary pump and Agilent 6490 Triple Quadrupole LCMS/MS equipped with an electrospray ionization (ESI) source. Mobile phases consisted of aqueous ammonium fluoride and ammonium fluoride in methanol. Steroids were quantified using dynamic multiple reaction monitoring mode.

RESULTS: We determined retention times and qualifier ions for each compound. We also built calibration curves for each compound with R squared values greater than 0.98. For each compound, we determined an analytical measurement range that was relevant to values that would be expected for affected and healthy individuals. We also determined lower limits of quantification.

CONCLUSION: We established an in house LCMS/MS method for quantification of the following four 11-oxygenated androgens in serum: 11-ketotestosterone, 11-hydroxytestosterone, 11-hydroxyandrostenedione, and 11-ketoandrostenedione. This assay can be further used to diagnose and aid in treatment for individuals with PCOS, CAH, and other disorders resulting from excess adrenal androgens.

Topic(s): Proteomics > Proteomics > none

Poster Presentation
Poster #38a
Attended on Thursday at 09:15

Accurate Quantification of Viral Protein to Support Reliable Viral Load Calculations

Yan Yan Beer (Presenter)
Physikalisch-Technische Bundesanstalt


INTRODUCTION: The monitoring of viral load is important to inform on disease progression and successful medical treatment. However, variabilities in quantitative measurement results may reduce the effectiveness of clinical intervention and disease management.

The ability to determine the number concentration of viral particles present in a biological sample, known as viral load, is essential in our continual quest to reduce the burden of viral infections on society. However, viral load is indirectly measured via the quantification of sequence specific viral nucleic acids, or proteins, in a representative sample. Quantitative PCR is the preferred analytical method for most viral infections. The quantification of viral DNA in a sample is nevertheless a non-specific incomplete indicator for active replication of the virus, while the quantification of viral proteins complements the information about the replication of the virus.

Combining DNA and protein measurements will provide a more complete picture of virus replication, while protein quantification will improve the measurement accuracy of infectious viral particles.

OBJECTIVES: We intend to assess the comparability of viral load measurements using DNA and protein measurement results to reduce the measurement uncertainty of viral load measurements and improving overall measurement comparability.

METHODS: The Human Cytomegalovirus (HCMV) served as a model organism. HCMV virions were produced in human fibroblasts and purified through a tartrate-glycerol gradient.

HCMV DNA was extracted using a viral DNA extraction kit. Digital droplet PCR of the HCMV gene UL54 was used to quantify the DNA and to calculate the genome copy number concentration within a sample.

Bottom-up proteomics was applied for the quantification of viral proteins. To identify suitable proteins and signature tryptic peptides for quantification, a peptide mapping approach was used. Time course experiments were run for selected signature peptides to confirm complete digestion of targeted proteins, confirmation of completeness was assumed when the ratio to isotopically labelled to natural protein plateaued and equimolar release of the peptides was observed.

For quantification, isotope dilution mass spectrometry (IDMS) was performed. To this end, synthetic labelled peptides served as internal standards and natural synthetic peptides were used as references for calibration. To ensure SI-traceability, amino acids analysis was performed on the synthetic peptides, using an exact matching IDMS calibration protocol and purified certified reference material (CRM) amino acids to prepare the calibration standards. The measurement uncertainty of the protein quantification was estimated according to the “Guide to the Expression of Uncertainty of Measurement” (GUM).


Three viral capsid proteins, Major capsid protein (MCP), Triplex capsid protein 1 (Tri1), and Triplex capsid protein 2 (Tri2), were chosen for viral protein quantification. For each protein, 2-4 signature peptides were selected. Calibration curves ranging from 1-200 pmol/ml were established for all peptides to estimate the limit of detection. The limit of detection is below 88 pmol/ml. For all the selected signature peptides, the time course experiments revealed that a digestion plateau was reached within 8-10 hours of initiating the tryptic digestion.

CONCLUSION: Protein measurements provide complementary information about replication of infectious viral particles. Amino acid analysis, complete digestion, equimolar release of the target peptides and calibration using isotope dilution approaches ensure measurement results are traceable to the SI. This overall reference measurement improves viral load measurements and should be adaptable to different viruses. Therefore, it could play an important role in providing information and/or reference measurement procedures to monitor viral load for disease progression and successful medical treatment.

Topic(s): Small Molecule > Microbiology > Metabolomics

Poster Presentation
Poster #43b
Attended on Wednesday at 14:30

Changes in Infant Health-Related Metabolites During Pregnancy After a Nutritional Intervention

Emma Guiberson (Presenter)
Stanford University


Introduction: Pre-term birth is a leading cause of child mortality, and is particularly present in undernourished populations. Diet is a powerful lever for manipulating the gut microbiome, which has been linked to a wide variety of health conditions in humans including cardiovascular health, asthma, food allergies, and others. As the gut microbiome plays a crucial role in maternal health, and the infant microbiome is acquired in part from the mother, understanding changes in the maternal microbiome during pregnancy can give potential impact into infant health and preterm birth. In particular, understanding the activity of the gut microbiome through signaling molecules such as microbiota-dependent metabolites (MDMs) gives a unique perspective into the host-microbiota interactions. These metabolites can travel throughout the body via biofluids. Previous data shows wide-scale changes in microbial communities and metabolites resulting from dietary interventions. As a portion of the infant microbiome is acquired from the mother, understanding changes to the microbiome and MDMs of pregnant mothers is paramount. In this study, we investigated longitudinal changes in MDMs in pregnant women living in Burkina Faso as a result of nutritional intervention during pregnancy, to further elucidate the impact of diet on the microbiome.

Methods: Previously the lab has developed an MDM-focused library for analysis of untargeted metabolomics data from biological samples. Study participants in Burkina Faso either received a nutritional intervention during their pregnancy, or continued on an existing dietary program. Health data analysis conducted on these individuals showed overall positive impacts on the maternal and infant health of participants receiving a nutritional intervention. Fecal samples were collected prior to the interventions and periodically throughout the pregnancy, for a total of 240 individuals over 3 time points. 300 samples were analyzed using metagenomics for strain identification and using LC-MS for untargeted metabolomics analysis. Extracted metabolites were then analyzed via LC-MS using an Agilent 1290 Infinity II LC coupled to an Agilent QTOF 6545 in both positive and negative mode. Data were then converted to a vender-neutral format using MS-Dial and compared against previously-collected library spectra for identification using exact mass and retention time.

Results: This semi-targeted metabolomics pipeline method has routinely enabled the detection of many microbial dependent metabolites from biological samples, many of which change as a result of dietary interventions. In this study, we were able to detect dozens of MDMs that differ significantly from our library within fecal samples from pregnant subjects. In particular, we discovered a variety of metabolites that differ during the third trimester based on diet status. Of note within these changing metabolites are those that are commonly screened for in newborn metabolic testing as markers of metabolic diseases such as methymalonic acidosis (MMA). Many metabolites with negative associations with infant health showed decreased abundances in women receiving the nutritional intervention compared to the standard diet counterparts. This suggests that diet changes may reduce risk of these metabolic diseases in infants. Overall, these data suggest that a dietary intervention during pregnancy can impact gut microbiome diversity, resulting in changes to the metabolites present during and after pregnancy. These changes can potentially be passed onto infants during birth, leading to lifelong microbial implications. An important next step of this work is to determine how such diet-induced microbiome modification during pregnancy impacts infant microbiome, metabolome, and overall health.

Topic(s): Small Molecule > Metabolomics

Poster Presentation
Poster #49b
Attended on Wednesday at 14:30

The Investigation of Tryptophan Metabolites and Vitamin B in Myalgic Encephalomyelitis via High-Resolution Mass Spectrometry

Sandy Abujrais (Presenter)
Uppsala University



Tryptophan is mainly metabolised by two biological pathways, the kynurenine pathway (KP) and the serotonin pathway. KP is important for producing both neuroprotective compounds, such as kynurenic acid and neurotoxic compounds like 3-hydroxy kynurenine, 3-hydroxy anthranilinic acid and quinolinic acid. Kynurenine has neurotoxic potential and has been linked to depression and anxiety. Enzymatic cofactors involved in the tryptophan kynurenine pathway include the B vitamins, particularly pyridoxal 5′-phosphate and riboflavin.

Myalgic encephalomyelitis (ME) is a chronic multisystem disease that affects the nervous, immune system and energy metabolism. The cause of the disease remains unknown with no diagnostic test available. The investigation of these metabolites in ME will help increase our understanding of the disease pathology.


Plasma samples from 38 ME patients and 24 healthy participants were extracted using a protein precipitation method. The isotope dilution method was used to quantify 12 analytes of tryptophan metabolites (quinolinic acid , 3-hydroxykynurenine , kynurenine , 3-hydroxyanthranilic acid , tryptophan , anthranilic acid , xanthurenic acid , kynurenic acid , nicotinamide , 5-hydroxytryptophan, melatonin and serotonin) and five metabolites of vitamin B (riboflavin, pantothenic acid, pyridoxine, biotin and pyridoxal 5′-phosphate). A parallel reaction monitoring (PRM) is employed in a positive electrospray mode. The analyses were performed using a Waters Acquity UHPLC (Waters™) coupled to a high-resolution Q Exactive™ hybrid quadrupole-Orbitrap mass spectrometer (Thermo Scientific™).

Preliminary Data

Using direct infusion, mass spectrometric detection parameters, including MS/MS mass transitions were optimized for each analyte . Thereafter, the method was validated according to the European Medicines Agency (EMA) guidelines for bioanalytical method development. Precision, accuracy, freeze-thaw stability, recovery and matrix effect were evaluated by preparing quality control samples at four levels; lower limit of quantification, low, medium and high quality control. These validation parameters were found to be within the limits outline in the EMA guidelines.

There was a wide dynamic range for the analytes investigated ranging from 0.02 ng/ml for melatonin and 60000 ng/ml for tryptophan.


This rapid and sensitive method allows for the simultaneous analysis and absolute quantification of 15 analytes related to tryptophan metabolism focusing on KP and its enzyme cofactors.

This method can be employed in ME to increase our understanding of the pathological changes associated of the disease.

Topic(s): Proteomics > Precision Medicine > none

Poster Presentation
Poster #54b
Attended on Thursday at 12:15

Transferability and Degree of Harmonization of an LC-MS Based Reference Measurement Procedure for Apolipoproteins in a Network of Calibration Laboratories

Fred Romijn (Presenter)



For apolipoprotein standardization, a network of three calibration laboratories has been established for the value assignment of matrix-based reference and EQA materials. A multiplexed LC-MS based reference measurement procedure (RMP) has been developed for serum apolipoproteins apo(a), apoA-I, apoB, apoC-I, apoC-II, apoC-III and apoE. This study evaluates the transferability of the RMP among 3 calibration labs and assesses whether the current level of harmonization between the calibration labs fulfils predefined analytical performance specifications.


Six periodic ringtrial surveys were held among three calibration laboratories. The study protocol, calibrators, internal standards, quality controls (QC) and clinical samples (CS) were shared among the labs prior to the study. Precision and inter-peptide comparisons of the CS were used to evaluate intra-laboratory performance, while Inter-laboratory comparisons and the overall level of harmonization of the CS were defined to evaluate equivalence between the calibration labs. Precision of the common bilevel QC was assessed to monitor improvement in the level of harmonization over time.


Intra-laboratory precision generally fulfilled predefined performance, which was defined as CVa < 50% of minimal Total allowable Error. Additional inter-laboratory variation was observed. Specifically, median interlab variation for the quantifying peptides was 13.5%, 4.7%, 4.4%, 9.9%, 8.5%, 6.6% and 6.8% for apo(a), apoA-I, apoB, apoC-I, apoC-II, apoC-III and apoE, respectively. Median interlaboratory biases were 4.45%, 3.15%, 1.3%, 8.4%, 4.65%, 3.45% and 2.55% for these same proteins, respectively. For apo(a) specifically, the average interlaboratory CV for four samples around the cut-off value of 90 nmol/L was 13.7%. In QC samples, the average imprecision for all apos decreased from 6.0% and 18.1% for IQC1 and IQC2, respectively, to 5.2% and 9.5% over the course of six months, indicating improvement of analytical performance of the network of calibration labs over time.


This study shows the feasibility of transferring an LC-MS based multiplex apolipoprotein reference measurement procedure between laboratories, an essential requirement for establishing a network. Most CS fulfilled predefined performance specifications (<50% of minimal Total allowable error); Ongoing ringtrials will monitor the performance of the network of three apolipoprotein calibration labs required to maintain an accurate value-base for apolipoprotein certification of commercial reagents.

Topic(s): Small Molecule > Metabolomics > Precision Medicine

Poster Presentation
Poster #59a
Attended on Wednesday at 12:15

Propagation of Chemical Families from High-Confidence Level Metabolite Identification through Molecular Networking in the Context of Microbiome Research

Romina Pacheco (Presenter)
Université de Lille



Host-associated samples subjected to untargeted metabolomics have provided valuable insights into how microbes influence health in a bidirectional way[1,2]. However, accurate metabolite annotation and identification remain to be a challenge [3] along with ensuring analytical reproducibility and feature coverage for large cohorts of data [4].


This study has the objective to improve the confidence level of the metabolite annotation by using chemical pure standards and in silico prediction tools in order to propagate their chemical classes to the unannotated metabolites.


Human serum samples were analyzed using a UHPLC coupled to a high resolution Orbitrap Exploris 240 mass spectrometer with two optimized methods for polar and non-polar metabolites in negative and positive electrospray ionization mode. Intelligent data acquisition workflow was implemented in addition to the Data-Dependent Acquisition method in order to increase spectral data required for metabolite annotation. Ion Identity Molecular networking (IIMN) approach was applied using the GNPS on-line platform to expand the chemical class starting from the known metabolites, annotated with both public spectral reference library (GNPS) and an in-house spectral library, to the unknowns.


The metabolome profiling of the samples provided a total of 4840 linear and reproducible features (m/z-rt pairs) detected in positive and negative mode with both LC-MS methods. Focusing on the 410 unique metabolites that were annotated, 20% correspond to a high confidence level annotation. Among them, we identified lipids (45%), organic acids and derivatives (27%), organoheterocyclic compounds (10%), benzenoids (9%), organic oxygen compounds (5%) and other chemical superclasses, 38 classes and 76 subclasses. Propagation and in silico fragmentation tools allowed an increase of 20% of chemical categories assignment.


IIMN allowed us to reduce redundancies of ion species and to expand the chemical information of the unannotated metabolites. Metabolome mining tools, such as in silico approaches5, harness advanced machine learning and predict fragmentation spectra from known structures to complement our results. This will be essential for the implementation of a reproducible workflow for untargeted LCMS analysis of biofluids in the context of metabolomics in microbiome research. It will also help to increase and improve the identification of metabolites of interest to provide an appropriate biological interpretation.


1. Bauermeister, A., Mannochio-Russo, H., Costa-Lotufo, L.V., Jarmusch, A.K., Dorrestein, P.C., 2022. Nat. Rev. Microbiol. 20, 143&ndash;160.

2. Dekkers, K.F., Sayols-Baixeras, S., Baldanzi, G. et al., 2022. Nat Commun 13, 5370.

3. Plumb, R.S., Gethings, L.A., Rainville, P.D., Isaac, G., Trengove, R., King, A.M., Wilson, I.D., 2023. TrAC Trends Anal. Chem. 160, 116954.

4. Zhou, Z., Luo, M., Zhang, H., Yin, Y., Cai, Y., Zhu, Z.-J., 2022. Nat. Commun. 13, 6656.

5. Ernst, M.; Kang, K.B.; Caraballo-Rodr&iacute;guez, A.M.; Nothias, L.-F.; Wandy, J.; Chen, C.; Wang, M.; Rogers, S.; Medema, M.H.; Dorrestein, P.C.; et al. 2019. Metabolites. 9, 144.

Topic(s): Small Molecule > Tox / TDM / Endocrine

Poster Presentation
Poster #63a
Attended on Wednesday at 12:15

Development of LC-MS/MS Xylazine Assay for Investigation of Xylazine Prevalence in New Haven, CT Region

Leah Militello (Presenter)
Yale University


Introduction: Xylazine is a veterinary tranquilizer not approved for human use that has been implicated in an increasing number of overdose deaths nationwide. It is used as an adulterant in recreational drugs of abuse, including heroin, cocaine and fentanyl, and many users are exposed unknowingly. As an α2-adrenergic agonist, its physiologic effects include sedation, muscle relaxation and decreased perception of pain. Taken in combination with other drugs, especially other central nervous system depressants like alcohol or benzodiazepines, the effects can be lethal. Xylazine was involved in 10% of all drug overdose deaths in Connecticut in 2020. Typical drug abuse screening panels do not include an assay capable of xylazine detection, as no immunoassay designed to run on an automated analyzer has been FDA-approved thus far. There is increasing demand from providers in the emergency department for an assay to determine the presence of xylazine in intoxicated patients with unclear clinical presentation. Here we describe a laboratory-developed xylazine assay and investigate the presence of xylazine in patient samples that screened positive for other drugs of abuse within the geographic region served by our hospital system. This method also examined the impact of measuring the xylazine metabolite, 2,6-dimethylaniline, on assay sensitivity.

Objectives: Develop a liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay for detection of xylazine and use that assay to examine the prevalence of xylazine in the New Haven, CT region.

Methods: A total of 37 patient urine samples that screened positive by immunoassay for cocaine, opiates and/or fentanyl between October and November 2023 were selected for the study and subject to retrospective analysis. Sample preparation consisted of a centrifugation step, addition of the internal standard, methadone-d9, followed by vortexing and another centrifugation step. Mobile phase A was 0.1 % formic acid in water and mobile phase B was 0.1 % formic acid in acetonitrile. Liquid chromatography (LC) was performed using an Acquity UPLC HSS column (T31.8 µm, 2.1 x 100 mm) at 50°C. MS/MS used positive electrospray ionization and monitored m/z transitions of 221 > 164 for xylazine (primary), 221 > 90 for xylazine (qualifier), and 319 > 105 for methadone-d9 (IS). The measurement of the xylazine metabolite 2,6-dimethylaniline had m/z transitions of 122 > 105 (primary) and 122 > 107 (qualifier). Retention time was 1.94 minutes for xylazine, 2.26 minutes for 2,6-dimethylaniline, and 2.16 minutes for the methadone-d9 internal standard. Linearity was established between 1-1000 ng/mL.

Results: Xylazine was detected a total of 9 out of 37 patient samples, or 24%. Of these samples, one was positive exclusively for opiates and three were positive exclusively for cocaine. Two samples were positive for both cocaine and fentanyl, two were positive for both cocaine and opiates, and one was positive for both fentanyl and opiates. All but one of the opiate positive specimens confirmed positive for 6-MAM, indicating heroin use. The quantity of xylazine ranged from 6 ng/mL to greater than 1000 ng/mL, with a mean value of 215 ng/mL and a median of 24 ng/mL. The results of 2,6-dimethylaniline were all significantly lower than xylazine and it was not found to add value in addition to xylazine.

Conclusions: Analysis by LC-MS/MS showed a high prevalence (24%) of xylazine in patient urine samples that screened positive for cocaine, opiates, fentanyl or a combination of these drugs. This is consistent with the Center for Disease Control and Prevention (CDC) data reporting that around 30% of fentanyl overdoses have also been found to contain xylazine. This may indicate a high level of adulteration by xylazine in the illegal drug supply in New Haven. Measurement of the metabolite 2,6-dimethylaniline did not add value, so laboratories looking to develop assays for xylazine do not need to measure this metabolite.

Topic(s): Other -omics > Cases of Unmet Clinical Needs > Emerging Technologies

Poster Presentation
Poster #64b
Attended on Thursday at 12:15

Development of Technology for High-throughput and Passive Monitoring for Earlier Detection of Colorectal Cancer

Lauren Ford (Presenter)
Imperial College London



Colorectal cancer (CRC) is the third most common cancer globally. Patients with bowel cancer often have unspecific symptoms such as changes in bowel habits1, which can make diagnosis difficult. Bowel cancer screening is usually performed using Faecal Immunochemical Testing (FIT) which has a good sensitivity for advanced CRC but performs poorer in earlier detection cases.2 Furthermore, there is rising incidence of CRC in younger adults who often will not be included in the CRC population surveillance.3 Herein, we have identified an unmet need to develop smarter testing for detection of colorectal cancer using less invasive and patient friendly approaches. There has been a rising interest in digital health surveillance and integration of technology, including toilet habit monitoring.4


Herein, it is hypothesised that the surface chemistry of bespoke ceramics can be used to capture metabolite information in urine and faecal samples. By coupling to Laser Desorption – Rapid Evaporative Ionisation Mass Spectrometry (LD-REIMS) high throughput direct sample analysis can be carried out with the long term plan of incorporating into a toilet setting.


LD-REIMS data was collected on a Waters Xevo QTof G2XS mass spectrometer in the range of 50 to 1200 m/z at a rate of 1 scan/second. An Opotek Q-switched optical parametric oscillator (OPO) laser source set to a wavelength of 2940 nm was used for sample ablation. 24 Ceramic materials with differing hydrophobicity, porosity and pore size were tested for their coupling capability with LD-REIMS. Splash LipidoMix mass spectrometry standard (Avanti Lipids) was used as an internal standard spiked into urine for feasibility analysis. Spiked samples were applied to ceramic surfaces (100uL) and washed with water, ammonium acetate (aa) or acidified aa. Faecal samples were collected from the colorectal cancer clinics at Imperial College Healthcare Trust (ICHT) hospital sites (REC: 14/EE/0024). In total 10 CRC, 8 Adenoma and 15 healthy stool samples were collected. Faecal samples were applied to ceramics and washed. In house Peak Picking was performed using in house scripts in R and Matlab. Statistical analysis was performed using Python Jupyter Notebook.


Urine spiked with deuterated standards (Splash Lipidomix), LipidoMix contains 14 standards of deuterium labelled phospholipids, lysolipids, triacylglygerides, diacylglycerides and cholesterol, providing a good coverage of lipid species for analysis. Application of spiked biofluid solutions to ceramic surfaces resulted in an increased recovery of observed lipids. Overall CeO2 ceramic with a porosity of 38.4% and ZrB2 with a porosity of 22.0% resulted in the highest relative abundance of deuterated lipids detected within urine when washed with ammonium acetate solution (150mM, 100uL). Pooled faecal samples were applied to the ceramic surfaces and analysed by LD-REIMS these were compared to native faecal sample aerosolised directly without application to ceramics. For the application of faecal samples to ceramics ZrB2 performed most optimally. Individual faecal samples were also analysed after application to ceramic surfaces, support vector machine learning (SVM) models were built and leave one patient out cross validation (LOPO-CV) done to reveal clinical benchmark performance indicators such as Diagnostic Accuracy (81%), Sensitivity (76%) and Specificity (87%). Scanning Electron Microscopy (SEM) also revealed cell adhesion to ceramic surfaces.


The use of ceramic surfaces with specific porosity resulted in increased sensitivity of interesting features in complex biofluids. It is hypothesised that this increase in sensitivity is due to the specific cavity sizes of the ceramics and the hydrophobic surface chemistry enabling adhesion to the surface whilst salts from the biofluid matrix can be washed away reducing the ion suppression observed in this high throughput analysis. This increased sensitivity resulted in good diagnostic accuracy between Healthy, Adenoma and CRC patients showing promise for a passive testing capability. In adenoma and cancer patient cohorts, tentative annotation using metabolite databases revealed increased relative abundance of oxidised ceramides, oxidised fatty acids, and diacylglycerides are observed.


1. D’souza, N., Georgiou Delisle, T., Chen, M., Benton, S. & Abulafi, M. Faecal immunochemical test is superior to symptoms in predicting pathology in patients with suspected colorectal cancer symptoms referred on a 2WW pathway: A diagnostic accuracy study. Gut 70, 1130–1138 (2021).

2. Niedermaier, T., Weigl, K., Hoffmeister, M. & Brenner, H. Diagnostic performance of flexible sigmoidoscopy combined with fecal immunochemical test in colorectal cancer screening: meta-analysis and modeling. Eur J Epidemiol 32, 481–493 (2017).

3. Vuik, F. E. R. et al. Increasing incidence of colorectal cancer in young adults in Europe over the last 25 years. Gut (2019) doi:10.1136/gutjnl-2018-317592.

4. Park, S. min et al. A mountable toilet system for personalized health monitoring via the analysis of excreta. Nat Biomed Eng 4, 624–635 (2020).

Topic(s): Other -omics > Emerging Technologies > Metabolomics

Poster Presentation
Poster #67a
Attended on Wednesday at 12:15

Using Ambient Ionisation Mass Spectrometry on Polymer Modified Surfaces to Improve Detection of Biomarkers Related to Colorectal Cancer

Elmeri Latvanen (Presenter)
Imperial College London



Colorectal cancer (CRC) is the second leading cause of cancer mortality globally and this is largely due to the late diagnosis of the cancer. Current population screening tests such as the faecal immunochemical test (FIT) which work by detecting blood in stool, have a high sensitivity (83%) for the detection of advanced cancer, however when performed on patients with early-stage cancer, or adenomas the sensitivity of FIT tests drops to 40% and 16-34% respectively(1). To improve patient outcome through earlier detection of CRC alternative tests must be developed, which allow for high throughput population screening capable of detecting alternate biomarkers. The research presented proposes the use of surface modified Ambient Ionisation Mass Spectrometry (AIMS) to resolve current population screening shortcomings, specifically using Laser Desorption Rapid Evaporative Ionisation Mass Spectrometry (LD-REIMS).

The ability to do direct from sample analysis using AIMS has greatly improved the clinical adoption of mass spectrometry allowing for rapid analysis of samples without the need for sample preparation. However, the complexity of faecal samples brings up complications; with salts causing ion suppression, and the number of compounds present making the detection of biomarkers challenging. The introduction of a polymer coating to the sampling surface, allows for the preferential binding of lipid metabolites which when analysed using LD-REIMS can be used to separate among healthy, adenoma, and cancer cohorts.


Faecal samples were collected from the colorectal cancer clinics at Imperial College NHS Trust hospitals (REC: 14/EE/0024). In total 10 CRC, 7 adenoma, and 15 healthy samples were collected. Knitted polyester swabs were plasma treated and subsequently immersed in a polymer solution. Faecal water was prepared from the samples and added to the swabs. LD-REIMS data was collected on a Waters Xevo QTOF G2S Mass spectrometer in the range of 50 to 1200 m/z at a rate of 1 scan/s, an Opotek Q-switched optical parametric oscillator (OPO) laser source was used for aerosol generation. RP-LC-ESI-MS was performed by the national phenome centre. Data processing was done using AMX (Waters corporation) and in house peak picking pipelines, multivariate statistical analysis was done using in house python scripts.


Standard knitted polyester swabs were compared to modified surfaces with different physical and chemical properties to assess the effects on the resulting metabolite profile using LD-REIMS. Polymers chosen displayed a range of properties including hydrophobicity which was measured using water contact angle measurements (between 0° and 70.1°). Multiple machine learning models were generated to assess the key clinical benchmarks such as Diagnostic Accuracy, Sensitivity and Specificity. PEI modified surfaces displayed the best diagnostic accuracy for detecting cancer samples (Acc (93%), sens (87%), spec(99%), and ODTMS modified surfaces separate between healthy (Acc (87%), Sens (89%) and Spec (84%)) and adenoma (Acc (77%), Sens (62%) and Spec (91%)) samples. Metabolite features driving classification were identified from the LD-REIMS spectra and tentatively annotated using the Lipids Maps database and the Human metabolome database (HMDB). Annotations were confirmed using RP-LC-ESI-MS data. For the best performing surfaces (PEI & ODTMS) univariate analysis was carried out. Volcano plots were used to identify 672 features from the PEI data that were statistically significantly different in abundance between the cancer and normal samples compared to 1 feature using unmodified surfaces.


Surface modification had an impact on the observed metabolites using LD-REIMS when analysing faecal samples. Modification of the surface resulted in better classification of disease when analysed using LD-REIMS and increased relative abundance of metabolites important for disease classification. Once developed this technique can be used for high throughput population screening capable of earlier detection of CRC and therefore improved patient prognosis.


(1). Niedermaier, T., Weigl, K., Hoffmeister, M. & Brenner, H. Diagnostic performance of flexible sigmoidoscopy combined with fecal immunochemical test in colorectal cancer screening: meta-analysis and modeling. Eur J Epidemiol 32, 481–493 (2017).