ABSTRACT
Rationale
Untargeted lipidomics is commonly performed at analytical flow rates, which consume more solvent and may require higher on‐column sample loads when sensitivity is limited by analyte abundance. Nano flow separations use lower flow rates and sample loads, reducing solvent consumption and facilitating improved electrospray ionization. We present here practical considerations for implementing a routine nano flow lipidomics workflow.
Methods
Bovine liver total lipid extract was spiked with SPLASH Lipidomix internal standards and analyzed by nano flow and high flow liquid chromatography coupled to a high‐resolution accurate mass Orbitrap‐based mass spectrometer. Full‐scan polarity switching was used for untargeted profiling, and the AcquireX Deep Scan workflow was applied to support data‐dependent MS/MS acquisition in the complex matrix. A SPLASH dilution series was analyzed in triplicate injections to compare analytical response across on‐column loads between the two workflows.
Results
At 25 ng on‐column, nano flow yielded 1266 total lipid annotations and 835 high‐quality annotations, compared with 919 total and 518 high‐quality annotations for high flow at 100 ng. Unintentional fragmentation decreased under nano flow conditions by 21%–31% across the representative lipid species evaluated. The use of nano flow allowed for the detection of lower on‐column loads across several standards, extending the lower end of the response range by up to 40‐fold compared with high flow analyses.
Conclusions
Nano flow lipidomics improved sensitivity for untargeted analysis by increasing the number of lipid annotations, reducing unintentional fragmentation, and extending the analytical response to lower on‐column loads. Together with practical guidance around sample preparation, injection volume, washing, and equilibration, these results support nano flow chromatography as a sensitive and reliable approach for sample‐limited untargeted lipidomics.
1. Introduction
Lipids are ubiquitous in living systems and play a central role in biology as structural components of cellular membranes, mediators in cellular signaling, and stores of metabolic energy [1, 2]. Dysregulation of lipid metabolism has been linked to a broad range of diseases including neurodegeneration, diabetes, and multiple types of cancer [3, 4]. Over the last few decades, increasing interest in lipid biology has driven the development of numerous analytical methods for lipid characterization [5, 6].
Among these approaches, liquid chromatography coupled to mass spectrometry (LC–MS) has become a dominant platform because it provides broad coverage across lipid classes while maintaining sensitivity in complex biological matrices [6, 7]. Many advancements in the field have been enabled by the development and increased accessibility of high‐resolution accurate mass (HRAM) instrumentation, including orbital trapping (Orbitrap) and time‐of‐flight mass spectrometers [8, 9]. Together with the development of powerful bioinformatics tools, these advancements have enabled scientists to confidently pursue lipidomics analyses from an untargeted perspective, which has accelerated lipidomic discoveries [10, 11]. Despite this progress, challenges remain in the detection and annotation of low‐abundance species, particularly when sample availability is limited or extensive lipidome coverage is required [12, 13].
Nano flow LC coupled to electrospray ionization (ESI) mass spectrometry has been applied to lipid analysis since the mid‐2000s. Early studies demonstrated sensitive separation and structural characterization of phosphatidylcholines from plant and mammalian tissues [14], as well as broad profiling of phospholipids and glycosphingolipids from glioblastoma cells using nano flow LC coupled to high‐resolution FT‐ICR and ion‐trap MS/MS [15]. More recent work has extended capillary and nano flow lipidomics to high‐depth plasma profiling using Orbitrap Tribrid mass spectrometry [16], comprehensive lipid analysis of saliva [17], and profiling of intact glycosphingolipids from human tissue, serum, and cells with automated structural annotation [18]. Collectively, these studies demonstrate the versatility of nano flow LC–MS across lipid classes, biological matrices, and mass spectrometry platforms. The increased sensitivity observed at lower flow rates is supported by improved ESI efficiencies, as smaller droplets can increase charge density and facilitate faster desolvation, whereas fewer analyte molecules within each droplet can reduce ion competition and ion suppression [19]. Nano flow can also reduce in‐source fragmentation for some lipid classes, although the extent depends on ion transfer conditions and source parameters, including temperature, radiofrequency (RF) settings, spray configuration, and gas parameters [20, 21]. Together, these effects can increase analytical sensitivity and improve lipid annotation quality when nano flow LC is coupled directly to ESI‐MS [13, 19].
In contrast to proteomics workflows, which are commonly performed in the nano flow regime, lipidomics workflows are most often carried out using analytical flow separations [6, 22]. These methods are well established, robust, and compatible with a wide range of commercially available column dimensions and stationary phase chemistries [6, 22]. Continued improvements in mass spectrometer performance have further increased the sensitivity and utility of analytical flow lipidomics workflows [6, 8, 9]. Nevertheless, certain applications may benefit from lower flow separations, particularly when sample amounts are limited, low‐abundance lipid species are of interest, or solvent consumption is an important consideration [12, 13]. At the same time, nano flow lipidomics remains more technically challenging to implement than established analytical flow workflows [12, 16, 21]. Nano flow separations require appropriate instrumentation, low dead volume connections, and stable electrospray emitters [12]. Method development is also constrained by the choice of commercially available capillary columns, which mainly feature C18 phases optimized for peptide separations, whereas a broader range of stationary phases, including C30, Shield RP18, and CSH Phenyl‐Hexyl, has been explored for lipid analysis at analytical flow rates [16, 22, 23, 24, 25]. Implementing a robust nano flow method also requires careful method setup to minimize carryover and gradual buildup, and maintain stable performance [18, 26]. Nano flow lipidomics can deliver sensitivity and solvent‐use advantages but is less forgiving from an implementation standpoint [12, 16]. Thus, routine nano flow lipidomics requires additional attention to wash procedures, re‐equilibration, low dead volume connections, and source stability [26, 27]. However, many of those practical barriers are becoming less significant as nano LC instrumentation continues to improve. Modern platforms, including the Thermo Scientific Vanquish Neo UHPLC system used in this study, incorporate automated system control and simplified operation that can reduce the level of operational complexity associated with nano flow separations. With appropriate training and standard procedures, nano flow workflows can therefore be implemented more readily in laboratories with existing LC–MS expertise.
In this work, an established high flow lipidomics method was compared with a nano flow LC–MS method using Orbitrap‐based HRAM mass spectrometry. The study evaluates the differences in lipid annotation depth, sensitivity, unintentional fragmentation, and analytical response, while also identifying practical considerations for routine nano flow lipidomics.
2. Materials and Methods
2.1. Sample Preparation
Bovine liver total lipid extract (25‐mg/mL stock solution) and SPLASH Lipidomix internal standards (Avanti Polar Lipids) were prepared at two different concentrations by combining fixed volumes of liver extract (2.0 μL) and SPLASH (19.8 μL) with isopropanol:acetonitrile (50:50, v/v) to different final volumes. The higher concentration was prepared by mixing 2.0‐μL liver extract, 19.8‐μL SPLASH, and 178.2‐μL isopropanol:acetonitrile (50:50, v/v) (final 200 μL). The lower concentration was prepared by mixing 2.0‐μL liver extract, 19.8‐μL SPLASH, and 978.2‐μL isopropanol:acetonitrile (50:50, v/v) with a final volume of 1000 μL.
A SPLASH Lipidomix dilution series was prepared in isopropanol:acetonitrile (50:50, v/v), with each level brought to a final volume of 200 μL. A 10% level was prepared by combining 20‐μL SPLASH stock with 180‐μL isopropanol:acetonitrile (50:50, v/v). The remaining levels at 5%, 1%, 0.5%, 0.1%, 0.05%, 0.01%, 0.005%, and 0.001% were prepared sequentially from the prior level using either a twofold dilution by mixing equal volumes of the previous level and diluent or a fivefold dilution by mixing 40 μL of the previous level with 160 μL of diluent. Each dilution level was analyzed in triplicate.
2.2. High‐Performance Liquid Chromatography
For the bovine liver total lipid extract comparison, nano flow analyses were performed at 5 and 25 ng on‐column, whereas high flow analyses were performed at 100 ng on‐column. These loading amounts were selected to reflect typical loading ranges reported in previous lipidomics studies [12, 13, 18]. All solvents used for sample preparation, mobile phase preparation, and wash solutions were Fisher Chemical Optima LC–MS grade.
Nano flow liquid chromatography (nLC) separations were performed using a Vanquish Neo UHPLC system in direct‐injection mode with pressure control set to 1500 bar, equipped with a Thermo Scientific EASY‐Spray C18 column (75 μm × 15 cm, 2 μm) maintained at 45°C. The instrument was configured to begin data acquisition during sample injection. The flow rate was set to 700 nL/min, and the injection volume was 0.1 μL. Mobile phase A consisted of 60:40 acetonitrile:water (v/v) containing 10‐mM ammonium formate and 0.1% formic acid, and mobile phase B consisted of isopropanol:acetonitrile:water (88:10:2, v/v/v) containing 10‐mM ammonium formate and 0.1% formic acid. The gradient (Figure S1) was held at 30% B from 0.0–9.5 min, increased to 43% B at 10.5 min, 55% B at 11.0 min, 65% B at 16.0 min, 85% B at 19.0 min, and 100% B at 20.0 min, then held at 100% B from 20.0–25.0 min for column washing. A “zebra wash” sequence was then applied by alternating low and high organic conditions in 0.5‐min segments: 5% B (25.5–26.0 min), 100% B (26.5–27.0 min), 5% B (27.5–28.0 min), 100% B (28.5–29.0 min), followed by 5% B (29.5–30.0 min) to return to low organic. The column was then equilibrated prior to the next injection using an equilibration factor of 5 column volumes at the initial conditions. Autosampler wash solutions consisted of a weak wash of acetonitrile:water:isopropanol (45:30:25, v/v/v) and a strong wash of hexanes:acetone:isopropanol (5:10:85, v/v/v). Both weak and strong wash steps were performed after sample draw for 5 s at 80 μL/s, corresponding to 400 μL per wash event for each solvent.
High flow separations were performed on a Thermo Scientific Vanquish Horizon UHPLC system equipped with a Thermo Scientific Accucore C30 column (2.1 × 150 mm, 2.6 μm) maintained at 45°C. The flow rate was 260 μL/min with a 2.0‐μL injection volume, and mobile phases A and B were the same as those used for the nano flow method. The gradient was held at 30% B from 0.0–2.0 min, increased to 43% B at 2.0 min and 55% B at 2.1 min, then ramped to 65% B at 12.0 min, 85% B at 18.0 min, and 100% B at 20.0 min, then held at 100% B from 20.0–25.0 min. The gradient returned to 30% B at 25.1 min and was held at 30% B from 25.1–30.0 min for re‐equilibration. The outer needle wash solvent was isopropanol:acetonitrile:water (50:45:5, v/v/v).
2.3. Mass Spectrometry
Mass spectrometry analyses were performed on a Thermo Scientific Orbitrap Excedion hybrid mass spectrometer. Full‐scan MS1 data were acquired with polarity switching over m/z 250–1700 at a resolution setting of 120 000 (FWHM) at m/z 200, using an automatic gain control (AGC) target of standard (1e6 absolute AGC), a maximum injection time of 50 ms, and an RF lens value of 50%. High flow analyses used a Thermo Scientific OptaMax Plus H‐ESI source with spray voltages of 3.5 kV (positive) and 2.5 kV (negative), sheath gas 40, auxiliary gas 8, ion transfer tube temperature 275°C, and vaporizer temperature 250°C. Nano flow analyses used a Thermo Scientific EASY‐Spray source with spray voltages of 1.9 kV (positive) and 1.5 kV (negative) and an ion transfer tube temperature of 275°C. Bovine liver total lipid extract samples were acquired both using full‐scan MS1 polarity‐switching runs and using an AcquireX Deep Scan workflow with data‐dependent acquisition in both positive and negative polarities (5 ID runs in each polarity). In the AcquireX Deep Scan workflow (both positive and negative polarities), MS1 scans were acquired at 120 000 resolution over m/z 250–1700 with an AGC target of standard and a 50‐ms maximum injection time, and MS2 scans were acquired at 30 000 resolution with an AGC target of standard (1e5 absolute AGC), a 1.2 m/z isolation window, stepped HCD NCE 20%/30%/40%, and a 54‐ms maximum injection time. Precursors were selected using a minimum intensity threshold of 20 000, with dynamic exclusion of 2 s, and MS2 scan range mode was set to auto‐extended.
2.4. Data Analysis
Bovine liver total lipid extract was processed using Thermo Scientific LipidSearch 5.2 software for lipid annotation and relative quantification. Raw files were imported and processed using a consistent workflow across nano flow and high flow datasets. During the alignment step, the lipid peak areas were class‐normalized to the corresponding SPLASH internal standard using the known standard amount. Initial output included lipid annotations with any adduct and grade assignments A, B, or C (D grades were pre‐filtered out because they lack specific head group and substituent fragment ions). Follow‐up filtering was applied using class‐specific criteria based on the preferred main ion for each lipid class and the LipidSearch annotation grade. To make the filtering criteria more adaptable to other software platforms, the preferred main ions used for filtering are summarized in Table S1, whereas the complete LipidSearch‐specific filter expression, including grade criteria, is provided as Data S1.
Unintentional fragmentation was evaluated for three deuterium‐labeled standards from SPLASH Lipidomix (phosphatidylinositol [PI] 15:0‐18:1(d7), cholesterol ester [ChE] 18:1(d7), and phosphatidylglycerol [PG] 15:0‐18:1(d7)) and three endogenous lipids (PI 18:0–20:4, ChE 18:2, and PG 18:2–18:2) under nano flow and high flow conditions. For each lipid, the signal‐to‐noise (S/N) ratio of the selected product ion and the S/N ratio of the corresponding precursor ion were obtained from the MS1 data. The precursor adducts and corresponding precursor and product ions used for the S/N calculations are summarized in Table S2. The percentage of unintentional fragmentation was calculated according to the method described by Criscuolo et al. [21] using the following equation:
For the dilution series, SPLASH standards were evaluated in Thermo Scientific TraceFinder 5.2 software. Integrated peak areas from triplicate injections were plotted against the on‐column load for each SPLASH standard on logarithmic axes (log10 scale) using GraphPad Prism 11.0. On‐column load was calculated from the concentration of each standard at each dilution level and the injection volume used for each workflow (0.1 μL for nano flow and 2.0 μL for high flow). The response range was selected from consecutive dilution levels that showed a consistent increase in peak area with increasing on‐column load and linear behavior on logarithmic axes. Low concentration points showing response flattening or increased replicate variability were excluded.
3. Results and Discussion
Lipid annotations in bovine liver total lipid extract were compared across different on‐column loads using nano flow and high flow LC–MS methods (Figure 1). At a 25‐ng on‐column load, nano flow yielded 1266 total lipid annotations compared with 919 annotations obtained with high flow at 100 ng on‐column. After applying a class‐specific filter (see Data S1), which required the expected main ion and appropriate LipidSearch grades for each lipid class, 835 high‐quality annotations remained for nano flow at 25 ng compared with 518 for high flow at 100 ng. At a further reduced nano flow load of 5 ng, 579 high‐quality lipid annotations were reported. Therefore, the 5‐ng nano flow condition produced approximately 12% more high‐quality annotations than high flow while using 20‐fold less material on‐column, whereas the 25‐ng nano flow condition produced approximately 61% more high‐quality annotations while using fourfold less material. These results highlight the practical sensitivity advantage of nano flow under low loading conditions and are consistent with previous reports showing improved sensitivity at reduced flow rates [12, 13, 20].
FIGURE 1.
Total lipid annotations before and after class‐specific filtering across on‐column loads for nano flow and high flow LC–MS. The total height of each bar represents annotations before filtering, with values above the bars indicating the total count. Teal represents high‐quality annotations after filtering, where gray represents annotations removed by filtering.
The extent of this advantage, however, is not expected to scale directly with the theoretical sensitivity gain from reducing column diameter [28]. The increase in high‐quality lipid annotations reflects the performance of the complete LC–MS workflow and is not determined by column dimensions alone. The extent of sensitivity gain can also vary among lipid classes because electrospray response is analyte‐dependent [20]. For example, surface‐active lipid species that already ionize efficiently may show a smaller relative benefit from reduced flow, which has been previously reported for several lipid classes under similar conditions [13, 20]. To determine whether the overall increase in annotation depth observed in Figure 1 was broadly distributed across the lipidome or driven by specific classes, high‐quality lipid annotations were next summarized by class.
At 25 ng, nano flow produced more annotations than high flow in 19 of the 23 lipid classes evaluated (Figure 2). Several of the largest differences were observed for phosphatidylcholine (PC), fatty acid (FA), phosphatidylethanolamine (PE), diacylglycerol (DG), and sphingomyelin (SM), whereas triacylglycerol (TG), lysophosphatidylethanolamine (LPE), and cardiolipin (CL) showed a similar number of annotations between the two workflows. Nano flow also extended coverage to several additional lipid classes. Lysophosphatidylinositol (LPI), lysophosphatidylglycerol (LPG), monoacylglycerol (MG), sphingoid bases (SPH), and ceramide phosphate (CerP) were observed at 25 ng with nano flow but were not present in the filtered 100‐ng high flow dataset. In contrast, acylcarnitine (AcCa) was represented only in the high flow filtered dataset, with three high‐quality annotations. Even at 5 ng on‐column, nano flow reported annotations across 19 lipid classes and produced more annotations than high flow for several classes, including PC, FA, PE, SM, DG, ChE, and PG. These results show that the increased annotation depth with nano flow was distributed across multiple lipid classes and was largely maintained as the on‐column load was reduced.
FIGURE 2.
Number of high‐quality lipid annotations reported for each lipid class after class‐specific filtering for high flow LC–MS at 100 ng on‐column (gray), nano flow LC–MS at 25 ng on‐column (teal), and nano flow LC–MS at 5 ng on‐column (blue). The dashed box contains an expanded view of lipid classes with less than 30 annotations. The underlying annotation counts are provided in Table S3.
The class‐specific differences in Figure 2 should not, however, be attributed to flow rate alone. This study compared an established high flow lipidomics method, which uses an Accucore C30 column, with a nano flow method implemented using an EASY‐Spray C18 column. The C30 method was used as the high flow reference because it represents a well‐established approach for broad untargeted lipidomics, whereas the choice of commercially available nano flow column chemistries was limited for development of the nano flow method. C18 and C30 phases differ in chromatographic selectivity, and previous comparisons have shown that stationary phase choice can affect both the number and type of lipid species resolved [22, 29]. C30 phases can provide greater selectivity for structurally similar hydrophobic lipids, with particularly clear effects reported for TG and ChE isomers, whereas differences in PC, PI, and PE coverage have also been observed between reversed‐phase columns [29]. Therefore, differences among individual lipid classes in Figure 2 may reflect stationary phase selectivity in addition to flow regime, sample load, gradient conditions, and source configuration. A matched C18 comparison at both flow regimes would be needed to separate these effects and can be performed in future studies. The class‐level results should therefore be interpreted as a comparison of two complete LC–MS workflows rather than as an effect of flow rate alone.
Despite this limitation, the filtered data showed that nano flow maintained broader lipid class coverage with less material on‐column. At 25 ng, most classes contained more high‐quality annotations than the 100‐ng high flow condition, and several additional lipid classes were reported only with nano flow after filtering. This pattern is consistent with previous lipidomics studies reporting improved lipidome coverage and sensitivity at reduced flow rates [16, 20]. Together, these results support another practical advantage of nano flow for sample‐limited lipidomics, while also showing that the extent of this benefit can vary across classes and chromatographic conditions.
Minimizing unintentional fragmentation is important in untargeted lipidomics because it can decrease sensitivity and compromise lipid detection [21]. Moreover, the fragment ions formed before intentional MS/MS can overlap in m/z with intact lipid ions, potentially increasing false‐positive annotations and reducing annotation confidence [21]. Unintentional fragmentation was evaluated by comparing percent fragmentation for six representative lipids from the PI, ChE, and PG classes under high flow and nano flow conditions (Figure 3). The percentage of unintentional fragmentation was calculated using the S/N of a selected product ion and its corresponding precursor ion, as described in Section 2. Lower unintentional fragmentation was observed with nano flow for all six lipid species evaluated. Fragmentation decreased by 21% for both PI species, 24%–31% for the ChE species, and 23%–30% for the PG species relative to their respective high flow measurements. Overall, the consistent decrease across the representative PI, ChE, and PG lipid species indicates less unintentional fragmentation with the nano flow workflow under the conditions evaluated.
FIGURE 3.
Comparison of unintentional fragmentation under high flow and nano flow LC conditions. Unintentional fragmentation (%) was determined for six lipid species. PI 15:0–18:1(d7), ChE 18:1(d7), and PG 15:0–18:1(d7) were labeled internal standards from SPLASH Lipidomix. PI 18:0–20:4, ChE 18:2, and PG 18:2–18:2 were unlabeled endogenous lipid species. Teal circles represent high flow measurements, and blue squares represent nano flow measurements. Connected points show measurements of the same lipid species under both LC flow conditions.
Analytical response was next evaluated using the SPLASH Lipidomix dilution series (Figure 4). Integrated peak areas were plotted against the amount loaded on‐column for representative PC, SM, and DG standards. Nano flow extended the response to lower on‐column loads for all standards evaluated. The lower end of the selected response range was extended to twofold lower on‐column load for SM, 10‐fold lower for PC, and 40‐fold lower for DG compared with high flow. Similar class‐dependent improvements in sensitivity and quantitative response have been reported for reduced‐flow lipidomics. Danne‐Rasche et al. showed that nLC extended the linear dynamic range for most lipid standards evaluated compared with narrow‐bore HPLC, whereas more recent work by Girel et al. also demonstrated increased response for lipid standards under reduced‐flow conditions [13, 20]. Higher on‐column loads were not evaluated with nano flow in the present study, so the upper ends of the response ranges should not be directly compared between the two workflows. Overall, these results showed another advantage of nano flow, allowing for an extension of the analytical response toward lower on‐column loads, further supporting its use, especially when sample amount or analyte abundance is limited.
FIGURE 4.
Comparison of nano flow and high flow LC–MS response for representative SPLASH Lipidomix internal standards: (A) PC 15:0–18:1(d7), (B) SM d18:1–18:1(d9), and (C) DG 15:0–18:1(d7). Integrated peak area is plotted against the on‐column load (pg) using a log10 scale. Points represent the mean ± SD of three replicate injections. Filled black circles represent nano flow measurements with a 0.1‐μL injection, and open black squares represent high flow measurements with a 2.0‐μL injection.
The improved performance at lower on‐column loads also introduces practical considerations when implementing a nano flow lipidomics workflow. Transitioning a sample preparation developed for high flow to nano flow may require consideration of the sample matrix in addition to injection volume and on‐column load.
However, a separate sample preparation workflow may not be required for every sample type. To evaluate this directly, mouse plasma was prepared using isopropanol protein precipitation, and the nano flow injection volume was reduced to 0.01 μL. Pooled plasma QC samples were analyzed over a sequence of approximately 100 injections, with a total of 12 QC injections used to evaluate analytical stability (Figure S2). Across the sequence, representative lipid standards showed consistent retention times and peak area responses, whereas mass errors remained within ±5 ppm. No evidence of emitter failure or clogging was observed. These results show that, under the conditions evaluated, the existing plasma preparation could be used reliably with nano flow by reducing the injection volume. For other sample types, additional sample preparation may still be needed when residual salts or other nonvolatile matrix components are present. These considerations have been discussed in more detail for other miniaturized LC–MS workflows [30, 31].
Wash solvents are also an important consideration when transitioning to nano flow. Lipidomics can be especially demanding with respect to carryover and buildup because strongly retained and poorly soluble material can accumulate in the LC flow path. For the nano flow method used here, a weak wash solvent matching the initial conditions of the gradient and a strong wash consisting of isopropanol/acetone/hexanes (85/10/5, v/v/v) were used to reduce carryover and buildup observed during method development. In addition, a zebra wash consisting of alternating 5% and 100% mobile phase B was included at the end of the gradient before re‐equilibration (Figure S1). The zebra wash should also be evaluated for the specific sample type and chromatographic gradient being used, as some optimization of this step may be needed. Column re‐equilibration follows the same general chromatographic principles at nano flow and analytical flow rates. For this workflow, five column volumes were used, consistent with published work supporting this as an appropriate equilibration volume under reversed‐phase conditions and can serve as a practical starting point when implementing this method [32]. This parameter can be further optimized by the user while monitoring chromatographic reproducibility and method robustness. Taken together, these wash and re‐equilibration considerations provide guidance for minimizing carryover and maintaining reproducible chromatography during routine nano flow lipidomics.
4. Conclusions
In conclusion, the nano flow LC–MS workflow provided greater sensitivity and deeper lipid annotation coverage while using less material on‐column compared with the established high flow workflow. Nano flow increased high‐quality lipid annotations across multiple lipid classes, reduced unintentional fragmentation for the representative lipids evaluated, and extended the analytical response to lower on‐column loads while also reducing solvent consumption. Although some of the observed differences may also reflect differences between the complete LC–MS workflows, the overall results demonstrate several advantages of the nano flow workflow compared with the highflow workflow particularly when sample amount or analyte abundance is limited. The analytical stability and practical considerations evaluated here further support the routine implementation of nano flow lipidomics with appropriate attention to sample preparation, loading, washing, and re‐equilibration. Overall, nano flow provides a sensitive and complementary approach for untargeted lipidomics.
Author Contributions
Ciara J. Sallowicz: writing – original draft, methodology, conceptualization, investigation, visualization, writing – review and editing, project administration, formal analysis, data curation. Thiago Mattos: methodology, investigation, writing – original draft, writing – review and editing, conceptualization. Rahul R. Deshpande: investigation, conceptualization, methodology. Bashar Amer: resources, investigation. Susan S. Bird: conceptualization, writing – review and editing, supervision.
Conflicts of Interest
C.J.S., T.M., R.R.D., B.A., and S.S.B. are full‐time employees at Thermo Fisher Scientific.
Supporting information
Figure S1: Analytical LC gradient, flow rate, and zebra wash used for the nano flow method.
Figure S2: Analytical stability of the nano flow LC–MS workflow. Pooled plasma QC samples were analyzed over a sequence of approximately 100 injections, with 12 QC injections evaluated. Representative SPLASH internal standards are shown for (A) peak area, (B) mass error (ppm), and (C) retention time (RT). Mass errors remained within ±5 ppm across all QC injections.
Table S1: Lipid classes and respective main ions used to filter results for high‐quality annotations.
Table S2: Precursor and product ions used to evaluate unintentional fragmentation of representative endogenous lipids and SPLASH Lipidomix internal standards, including precursor adducts and corresponding m/z values.
Table S3: Number of high‐quality lipid annotations by lipid class for nano flow and high flow LC–MS analyses at different sample loads.
Data S1: LipidSearch filter used to obtain high‐quality lipid annotations.
Acknowledgments
The authors would like to thank Craig Dufresne and Yang Jiao for discussions and optimization of wash solvents and gradient for nano flow applications.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. van Meer G., Voelker D. R., and Feigenson G. W., “Membrane Lipids: Where They Are and How They Behave,” Nature Reviews. Molecular Cell Biology 9, no. 2 (2008): 112–124, 10.1038/nrm2330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Simons K. and Ikonen E., “How Cells Handle Cholesterol,” Science 290, no. 5497 (2000): 1721–1726, 10.1126/science.290.5497.1721. [DOI] [PubMed] [Google Scholar]
- 3. Quehenberger O., Armando A. M., Brown A. H., et al., “Lipidomics Reveals a Remarkable Diversity of Lipids in Human Plasma,” Journal of Lipid Research 51, no. 11 (2010): 3299–3305, 10.1194/jlr.M009449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Kuhajda F. P., “Fatty‐Acid Synthase and Human Cancer: New Perspectives on Its Role in Tumor Biology,” Nutrition 16, no. 3 (2000): 202–208, 10.1016/S0899-9007(99)00266-X. [DOI] [PubMed] [Google Scholar]
- 5. Han X. and Gross R. W., “Global Analyses of Cellular Lipidomes Directly From Crude Extracts of Biological Samples by ESI Mass Spectrometry: A Bridge to Lipidomics,” Journal of Lipid Research 44, no. 6 (2003): 1071–1079, 10.1194/jlr.R300004-JLR200. [DOI] [PubMed] [Google Scholar]
- 6. Cajka T. and Fiehn O., “Comprehensive Analysis of Lipids in Biological Systems by Liquid Chromatography‐Mass Spectrometry,” TrAC Trends in Analytical Chemistry 61 (2014): 192–206, 10.1016/j.trac.2014.04.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Watson A. D., “Thematic Review Series: Systems Biology Approaches to Metabolic and Cardiovascular Disorders. Lipidomics: A Global Approach to Lipid Analysis in Biological Systems,” Journal of Lipid Research 47, no. 10 (2006): 2101–2111, 10.1194/jlr.R600022-JLR200. [DOI] [PubMed] [Google Scholar]
- 8. Makarov A., “Electrostatic Axially Harmonic Orbital Trapping: A High‐Performance Technique of Mass Analysis,” Analytical Chemistry 72, no. 6 (2000): 1156–1162, 10.1021/ac991131p. [DOI] [PubMed] [Google Scholar]
- 9. Guilhaus M., Selby D., and Mlynski V., “Orthogonal Acceleration Time‐Of‐Flight Mass Spectrometry,” Mass Spectrometry Reviews 19, no. 2 (2000): 65–107, 10.1002/(SICI)1098-2787(2000)19:2<65::AID-MAS1>3.0.CO;2-E. [DOI] [PubMed] [Google Scholar]
- 10. Herzog R., Schuhmann K., Schwudke D., et al., “LipidXplorer: A Software for Consensual Cross‐Platform Lipidomics,” PLoS ONE 7, no. 1 (2012): e29851, 10.1371/journal.pone.0029851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Takeda H., Matsuzawa Y., Takeuchi M., et al., “MS‐DIAL 5 Multimodal Mass Spectrometry Data Mining Unveils Lipidome Complexities,” Nature Communications 15 (2024): 9903, 10.1038/s41467-024-54137-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Zardini Buzatto A., Kwon B. K., and Li L., “Development of a NanoLC‐MS Workflow for High‐Sensitivity Global Lipidomic Analysis,” Analytica Chimica Acta 1139 (2020): 88–99, 10.1016/j.aca.2020.09.001. [DOI] [PubMed] [Google Scholar]
- 13. Danne‐Rasche N., Coman C., and Ahrends R., “Nano‐LC/NSI MS Refines Lipidomics by Enhancing Lipid Coverage, Measurement Sensitivity, and Linear Dynamic Range,” Analytical Chemistry 90, no. 13 (2018): 8093–8101, 10.1021/acs.analchem.8b01275. [DOI] [PubMed] [Google Scholar]
- 14. Bang D. Y., Kang D., and Moon M. H., “Nanoflow Liquid Chromatography‐Tandem Mass Spectrometry for the Characterization of Intact Phosphatidylcholines From Soybean, Bovine Brain, and Liver,” Journal of Chromatography. A 1104, no. 1–2 (2006): 222–229, 10.1016/j.chroma.2005.12.005. [DOI] [PubMed] [Google Scholar]
- 15. He H., Conrad C. A., Nilsson C. L., et al., “Method for Lipidomic Analysis: p53 Expression Modulation of Sulfatide, Ganglioside, and Phospholipid Composition of U87 MG Glioblastoma Cells,” Analytical Chemistry 79, no. 22 (2007): 8423–8430, 10.1021/ac071413m. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. He Y., Brademan D. R., Hutchins P. D., Overmyer K. A., and Coon J. J., “Maximizing MS/MS Acquisition for Lipidomics Using Capillary Separation and Orbitrap Tribrid Mass Spectrometer,” Analytical Chemistry 94, no. 7 (2022): 3394–3399, 10.1021/acs.analchem.1c05552. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lee G. B., Caner A., and Moon M. H., “Optimisation of Saliva Volumes for Lipidomic Analysis by Nanoflow Ultrahigh Performance Liquid Chromatography‐Tandem Mass Spectrometry,” Analytica Chimica Acta 1193 (2022): 339318, 10.1016/j.aca.2021.339318. [DOI] [PubMed] [Google Scholar]
- 18. Schindler R. L., Oloumi A., Tena J., et al., “Profiling Intact Glycosphingolipids With Automated Structural Annotation and Quantitation From Human Samples With Nanoflow Liquid Chromatography Mass Spectrometry,” Analytical Chemistry 96, no. 15 (2024): 5951–5959, 10.1021/acs.analchem.4c00077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Wilm M. and Mann M., “Analytical Properties of the Nanoelectrospray Ion Source,” Analytical Chemistry 68, no. 1 (1996): 1–8, 10.1021/ac9509519. [DOI] [PubMed] [Google Scholar]
- 20. Girel S., Galmiche M., Fiault M., et al., “Microflow Liquid Chromatography Coupled to Multinozzle Electrospray Ionization for Improved Lipidomics Coverage of 3D Clear Cell Renal Cell Carcinoma,” Analytical Chemistry 97, no. 9 (2025): 5109–5117, 10.1021/acs.analchem.4c06337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Criscuolo A., Zeller M., and Fedorova M., “Evaluation of Lipid In‐Source Fragmentation on Different Orbitrap‐Based Mass Spectrometers,” Journal of the American Society for Mass Spectrometry 31, no. 2 (2020): 463–466, 10.1021/jasms.9b00061. [DOI] [PubMed] [Google Scholar]
- 22. Narváez‐Rivas M. and Zhang Q., “Comprehensive Untargeted Lipidomic Analysis Using a Core‐Shell C30 Particle Column and High‐Field Orbitrap Mass Spectrometer,” Journal of Chromatography. A 1440 (2016): 123–134, 10.1016/j.chroma.2016.02.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lee J. C., Kim Y. B., and Moon M. H., “Enhancement of Acidic Lipid Analysis by Nanoflow Ultrahigh Performance Liquid Chromatography‐Mass Spectrometry,” Analytica Chimica Acta 1166 (2021): 338573, 10.1016/j.aca.2021.338573. [DOI] [PubMed] [Google Scholar]
- 24. Wilson I. D., Broeckling C., Gethings L. A., et al., “Development of a Single Mobile Phase for LC‐IM‐MS‐Based Discovery Lipidomics and Metabolic Phenotyping: Application to Methapyrilene Hepatotoxicity in the Rat,” Journal of Chromatography. A 1714 (2024): 464552, 10.1016/j.chroma.2023.464552. [DOI] [PubMed] [Google Scholar]
- 25. Lísa M., Holčapek M., and Sovová H., “Comparison of Various Types of Stationary Phases in Non‐Aqueous Reversed‐Phase High‐Performance Liquid Chromatography‐Mass Spectrometry of Glycerolipids in Blackcurrant Oil and Its Enzymatic Hydrolysis Mixture,” Journal of Chromatography. A 1216, no. 48 (2009): 8371–8378, 10.1016/j.chroma.2009.09.060. [DOI] [PubMed] [Google Scholar]
- 26. Mitulović G., Stingl C., Steinmacher I., et al., “Preventing Carryover of Peptides and Proteins in Nano LC‐MS Separations,” Analytical Chemistry 81, no. 14 (2009): 5955–5960, 10.1021/ac900696m. [DOI] [PubMed] [Google Scholar]
- 27. Cortés‐Bautista S., Molins‐Legua C., and Campíns‐Falcó P., “Miniaturized Liquid Chromatography in Environmental Analysis. A Review,” Journal of Chromatography. A 1730 (2024): 465101, 10.1016/j.chroma.2024.465101. [DOI] [PubMed] [Google Scholar]
- 28. Shen Y., Zhao R., Berger S. J., Anderson G. A., Rodriguez N., and Smith R. D., “High‐Efficiency Nanoscale Liquid Chromatography Coupled On‐Line With Mass Spectrometry Using Nanoelectrospray Ionization for Proteomics,” Analytical Chemistry 74, no. 16 (2002): 4235–4249, 10.1021/ac0202280. [DOI] [PubMed] [Google Scholar]
- 29. Criscuolo A., Zeller M., Cook K., Angelidou G., and Fedorova M., “Rational Selection of Reverse Phase Columns for High Throughput LC‐MS Lipidomics,” Chemistry and Physics of Lipids 221 (2019): 120–127, 10.1016/j.chemphyslip.2019.03.006. [DOI] [PubMed] [Google Scholar]
- 30. Lin W., Mousavi F., Blum B. C., et al., “PANAMA‐Enabled High‐Sensitivity Dual Nanoflow LC‐MS Metabolomics and Proteomics Analysis,” Cell Reports Methods 4, no. 7 (2024): 100803, 10.1016/j.crmeth.2024.100803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Girel S., Meister I., Glauser G., and Rudaz S., “Hyphenation of Microflow Chromatography With Electrospray Ionization Mass Spectrometry for Bioanalytical Applications Focusing on Low Molecular Weight Compounds: A Tutorial Review,” Mass Spectrometry Reviews 44 (2025): 491–512, 10.1002/mas.21898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Schellinger A. P., Stoll D. R., and Carr P. W., “High Speed Gradient Elution Reversed Phase Liquid Chromatography of Bases in Buffered Eluents. Part II. Full Equilibrium,” Journal of Chromatography. A 1192, no. 1 (2008): 54–61, 10.1016/j.chroma.2008.02.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Analytical LC gradient, flow rate, and zebra wash used for the nano flow method.
Figure S2: Analytical stability of the nano flow LC–MS workflow. Pooled plasma QC samples were analyzed over a sequence of approximately 100 injections, with 12 QC injections evaluated. Representative SPLASH internal standards are shown for (A) peak area, (B) mass error (ppm), and (C) retention time (RT). Mass errors remained within ±5 ppm across all QC injections.
Table S1: Lipid classes and respective main ions used to filter results for high‐quality annotations.
Table S2: Precursor and product ions used to evaluate unintentional fragmentation of representative endogenous lipids and SPLASH Lipidomix internal standards, including precursor adducts and corresponding m/z values.
Table S3: Number of high‐quality lipid annotations by lipid class for nano flow and high flow LC–MS analyses at different sample loads.
Data S1: LipidSearch filter used to obtain high‐quality lipid annotations.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.