Introduction

Throughout history, model organisms have been used to study biological phenomena in non-human species. The transparent nematode Caenorhabditis elegans is routinely used as a model organism. They gained popularity in biological research in the 1970s after extensive research was conducted by Dr. Brenner1 to elucidate their properties and breeding techniques. C. elegans nematodes have a rapid life cycle and large brood size, making them suitable for research requiring large sample cohorts2,3. They also have a consistent number of somatic cells over the generations4, allowing for controlled studies using various altered cell lineages. In addition, the high transparency of the small nematode bodies enables clear optical observations5 and even whole nematode staining6,7. Therefore, C. elegans has been used in developmental biology1,8,9, neurology10,11, genetics1,12,13, and toxicological studies14,15,16. C. elegans has also attracted significant interest in studies of lipid metabolism and storage due to the evolutionary conservation of lipids and related pathways between the nematodes and mammals17. To trace the lipids in C. elegans, lipid-specific staining methods, such as Nile Red or Oil Red O (ORO) staining, are typically used in combination with chromatography-based mass spectrometry (MS)17,18. Staining can provide spatial information about lipids19, but cannot achieve exact molecular identification. On the other hand, liquid-chromatography MS (LC-MS) analysis enables the sensitive detection and exact identification of lipids but does not provide spatial information. LC-MS only provides an estimation of the entire population of nematodes17, and information regarding intrapopulation variation of different molecular species is not available. MS imaging (MSI) can be used to visualise the spatial distribution of various molecules in C. elegans. MSI has gained popularity over the past 20 years, owing to its ability to provide spatially resolved information for a wide range of molecules20,21,22,23 and trace the fate of pharmaceutical compounds24,25. In particular, matrix-assisted laser desorption/ionisation (MALDI), a powerful soft-ionization technique, offers non-destructive analysis. MSI is often paired with post-analysis histological staining methods to produce multimodal images26,27,28,29. MALDI has previously been used to image small animals such as zebrafish (Danio rerio)30,31,32 and fruit fly (Drosophila melanogaster)33,34,35. However, the application of MSI to C. elegans is limited by its small size (adult length of \(\sim 1\) mm) and the exterior cuticle layer. In particular, it is challenging to preserve the internal structures of nematodes during sectioning, which is required to remove the cuticle for MALDI analysis of C. elegans. While previous studies demonstrated MALDI-MSI of C. elegans, the presence of cuticle layers36 and the use of the freeze-cracking method37, or the alternative freeze-and-thaw method22 resulted in limited visualization of the internal structures. In addition, the small size of C. elegans limits the reliability of embedding and sectioning methods, which hinders the acquisition of consecutive sections necessary for three-dimensional (3D) reconstructions. Considering the impact of 3D visualisation and its significant advances in understanding tissues and organs24,38,39, 3D MSI observations of C. elegans are expected to facilitate lipidomics studies by providing better correlations between the anatomical structures and optical imaging information. In this study, we developed a microfluidics-based workflow for sectioning C. elegans to correlate lipid molecular information with anatomy. This protocol can be used to obtain consecutive sections of nematodes with well-conserved internal structures, including the pharynx, intestine, and embryos. These sections provide high visibility of internal structures to facilitate MSI analysis. Additionally, various imaging modalities can be applied to a single section to help visualise molecular distributions specific to internal structures. Moreover, well-preserved internal structures can be used as reference points between consecutive sections to achieve a 3D reconstruction of the nematode from optical and MSI images, allowing for a better understanding of target molecules in the nematode.

Results

Strategy to obtain well-conserved longitudinal sections of nematodes

Our sample preparation strategy (Fig.1a) relies on (1) selecting an appropriate support to keep the nematode body straight, (2) selecting an optimal embedding medium, and (3) achieving precise positioning before and during cryotome sectioning. Although a recent study reported that nematodes straightened into a line when frozen22, we collected anesthetised wild-type, \(N2\), C. elegans using a microfluidic device to minimise sample damage (Fig. 1b–d). We used a polydimethylsiloxane (PDMS) line-patterned mould with channel widths of 50 \({\upmu {\textrm{m}}}\). This geometry was designed for young adult nematodes as they are easier to handle compared with the larval phases (L1–L4). Furthermore, within a population, there tend to be fewer embryos and more mature adults. The embedding medium was selected and adapted based on previous publications focusing on MALDI-MSI compatibility and medium pliability30,31,40,41,42,43. The viscosity of the medium is important as it needs to be poured over the aligned nematodes and exhibit sufficient pliability to enable the removal of the microfluidic chip and mould without damaging the sample. After removal from the mould, the nematodes were well encapsulated in the gelatin (Fig. 1c), preserving their orientation and remaining straightened as during alignment on the microfluidic chip. By carefully selecting the alignment and embedding methods, and developing a two-part mould and lid set (Fig. 1d), we ensured the consistent placement of nematodes during sectioning (Fig. 1e) and the acquisition of well-preserved consecutive sections (Figs. 1f,g, 2).

Fig. 1
figure 1

Embedding and sectioning optimization. (a) Schematic of the sample preparation workflow. (b) Photographs of nematodes trapped in the PDMS chip and (c) embedded in gelatin after removal of the PDMS chip. Red arrows indicate nematodes. (d) 3D printed mould and lid. (e) Frozen sample block secured onto a cryotome specimen disc. (f) Glass slide with gelatin sections. A nematode (white line in the air pocket) is indicated by a yellow arrow. Inset: close-up view of the nematode. (g) Confirmation of the location and preservation of the sections using hematoxylin staining. (h) Representative microscopy images defining the fragmentation rankings. The rankings are based on the number of fragments produced: low = 1–2 fragments, low/moderate = 3–4 fragments, moderate = 5 fragments, moderate/severe = 6–7 fragments, severe = 7+ fragments. (i) Comparison of the fragmentation of nematode sections and j) ratio of nematode to air pocket for the two embedding media. Bars in (j) show the standard error (SE) and correspond to \(\pm 1 SE\).

Full size image
Fig. 2
figure 2

Comparison of intact nematodes and nematode sections. (a) brightfield image of an intact nematode. (b,c) brightfield image of consecutive nematode sections. Arrows indicate different parts of the C. elegans anatomy corresponding between the sections and intact nematode.

Full size image

Evaluation of embedding media

The selection of an appropriate embedding medium is critical for the successful conservation of anatomical structures. Gelatin is a commonly used medium for MSI40,41. However, when using only gelatin, air pockets of various sizes are typically observed around the nematodes and throughout the entire section, making it challenging to determine the nematode location and preserve the sections. Therefore, we tested several compositions of embedding media to determine the optimal conditions for minimising the air pocket size around the nematodes. Carboxymethyl cellulose (CMC)42 is another common MSI medium often used for hollow samples, including lungs and swim bladders of zebrafish31,43. Therefore, we prepared gelatin–CMC mixtures with 10% gelatin and 2% or 5% CMC for comparison with 10% gelatin. Figure S1 in the Supplementary Information shows optical images of nematodes embedded in these three media. Increasing the CMC content increased the viscosity of the media, and 10% gelatin with 5% CMC did not provide reproducible results (Fig. S1e). Therefore, the quality of the sections was evaluated only for 10% gelatin and 10% gelatin–2% CMC media. The level of nematode fragmentation was ranked as low, medium, or severe depending on the preservation of their shape and structure, as graphically depicted in Fig. 1h,i. The fraction of low-fragmentation sections was considerably lower in 10% gelatin–2% CMC than in 10% gelatin. Note that 10% gelatin–5% CMC was previously used for small biological samples30,31. Although the present study identified 10% gelatin–2% CMC as the optimal medium for the longitudinal sectioning of C. elegans, the optimal composition may differ for other biological samples. To further quantify the air pocket size, we analysed the images by manually tracing the region of interest (ROI) around the edges of the nematodes and inner edges of the air pockets (Fig. S2); a higher ratio of nematodes to air pockets indicates smaller air pockets. Figure 1j shows that the addition of 2% CMC to gelatin decreases the size of the air pockets surrounding the nematodes in the section. Furthermore, this medium improves the preservation of the nematode structure, indicating that it provides better support for the encapsulation and sectioning. Therefore, we determined that 10% gelatin–2% CMC was the optimal embedding medium for subsequent sectioning experiments.

Precise positioning during cryotome sectioning

To enable the precise positioning of the target nematode during sectioning, it is necessary to clearly mark the nematode position in the block. The nematodes themselves are too small to be easily seen and most of the sections do not contain any nematodes, which significantly decreases the collection efficiency of sections suitable for analysis. To overcome this problem, air pockets in the block were used as guides to locate the nematodes (Fig. 1f, and Fig. S3). Sections were collected from the first appearance of the air pockets. Once collected, the samples were observed by eye, without the use of a microscope, and estimations were made regarding the presence of nematodes in the sections (Fig. 1f,g) using the air pockets. For estimation accuracy, we used the following classifications. True positive: The nematode location was visually estimated based on the presence of an air pocket and confirmed by microscopic observation. False positive: The nematode location was estimated visually, but no nematode was present when inspected under a microscope. False negative: The nematode location was not observed during visual inspection but was noticed during microscopic observation. True negative: no air pockets or nematode were observed visually or by microscopy, so the section was excluded from statistical analysis. Across four slides, with 51 sections and 119 air pocket locations, the rate of true positives was 68.91%, while false negatives and false positives accounted for 14.29% and 16,81%, respectively. The false positives were likely caused by air pockets in the sections directly before or after the collected nematodes (Fig. S3b and c; see Figure S4). We also attempted to fully remove the air pockets for better nematode encapsulation; however, this resulted in a lower collection accuracy. The presence of some air pockets is inevitable and beneficial as a collection guide.

Reproducibility of sample preparation and MALDI-MSI data

To distinguish the potential variation artefacts caused by the sectioning process or MALDI-MSI sample preparation and measurement from the inherent variations between individuals or sections, MALDI analysis was performed on a set of multiple sections from different nematodes. The samples for assessing the variations due to preparation and analysis steps were collected from two separate gelatin blocks, which were sectioned, matrix-coated, and analysed on different days. Conversely, to visualise the differences within and between nematodes, consecutive sections of each nematode were acquired from the same glass slide and analysed simultaneously. All analysed nematodes were cultured in a single Petri dish, and the growth phases were synchronised. We compared the absolute intensity variations between individual nematodes and consecutive sections of eight nematodes. Specifically, the dataset for the variations between individual nematodes (designated as “N2 all”) were evaluated by analysing the MS intensities for the first top sections of eight nematodes collected on two glass slides, analysed on different days (Fig. 3a). Datasets for the variations between consecutive sections (“Nematode 1–4”) were obtained by analysing the MS intensity for the consecutive sections of a selected four of the eight nematodes (Fig. 3b). The number of sections for each nematode varied between two and six. The data are presented in the Supplementary Information. The other four nematodes were excluded from the consecutive section analysis because of section loss, mainly caused by curling of the gelatin section during thawing on the glass slides, in combination with alternating positive and negative ionisation mode analysis, where only half of the available sections could be used in this comparison (see “Methods” section). To obtain the MS intensities, we normalised the data to a matrix peak, and used the lipid mass at m/z 423.31 (Fig. 3d), which was assumed to be a lysophospholipid LPA (18:0) based on an exact mass database query (see Table S1). This peak was used as a reference because it was present throughout the nematode body, enabling the consistent comparison of localised lipids in specific regions across different sections. Figure 3c compares the absolute intensities at m/z 423.31 for the “N2 all” and “Nematode 1–4” datasets. As “N2 all” includes data collected on different days, the variations may have arisen from different instrumental effects and sample-preparation conditions, in addition to the intrinsic variations within and between the nematodes. In contrast, the “Nematode 1-4” datasets reflect only the differences between the sections from the same nematode, i.e., variations related to the analysis location in the body and intrapopulation variation. Additionally, any instrumental effects from matrix deposition and analysis were minimised, as all data were collected simultaneously. Statistical analyses were performed to compare datasets. Both one-way analysis of variance (ANOVA) and paired t-tests between all group combinations showed no significant differences in the variations between individual nematodes and consecutive sections (see Fig. S5). This indicates that the signal intensity variations introduced by sample preparation and measurement are smaller than the intrinsic variations within and between individual nematodes, thereby proving that the proposed sample preparation and MSI analysis methods are suitable for analysing C. elegans.

Fig. 3
figure 3

Average MS intensity variations in nematodes and consecutive sections. Schematics showing the differences between the data collection methods used to obtain the datasets a) “N2 all” and b) “Nematode 1–4”. (a) Twelve consecutive gelatin block sections containing 5 different nematodes. (b) Consecutive sections obtained from a single nematode. “Nematode 1-3” in (a) represent data points in “N2 all”, while “Consecutive section 1–3” in (b) correspond to separate data points in “Nematode 1-4”. (c) Absolute intensity of the MS peak at m/z 423.31 for the datasets “N2 all” and “Nematode 1–4”. Intensities were obtained from average nematode ROI spectra in arbitrary units. The bars correspond to \(\pm 1 SE\). (d) MS spectra of a representative nematode from the “N2 all” dataset and its gelatin background, with matrix peaks m/z 167.08 and m/z 333.13, and a lipid signal at m/z 423.31.

Full size image

Organ specificity: differences in lipid distribution observed by MSI

Longitudinal sectioning revealed nematodes with preserved internal structures (Fig. 4a) and neutral lipid distribution by ORO staining (Fig. 4b; Section Cross-validation by ORO staining for more details). The MSI data exhibit clear trends in which the lipid masses of m/z 466.35 and m/z 648.49 are distributed throughout the nematode body at varying intensities (Figs. 4c, e; S6b, d). In contrast, lipids with masses m/z 599.35 and m/z 939.71 are only present in the embryos and pharynx, respectively (Figs. 4d, f, g; S6c, e; S7). We also obtained the corresponding average MS spectra (over a range of lipid-derived masses) for the sections containing the entire nematode ROI and background (Fig. 4h, i), with the surrounding matrix-covered embedding medium taken as the background. Through this comparison, we confirmed that all lipids mapped in Fig. 4c–f originate from the nematodes and not the matrix or gelatin. In addition, we observed that lipids mainly localised within organs, such as that characterised by a peak m/z 939.71, showed high signal intensities in smaller, organ-specific ROIs but lower average signal intensities in the whole-section spectrum (Fig. 4h, i; see Fig. S8). For a more accurate identification of lipids, exact masses and tandem mass spectrometry (MS/MS) analysis were used (see Supplementary Information, Fig. S9–12). Exact mass-based identifications were acquired via high-mass-resolution analysis of the sections using a QExactive system with a query to the LipidMaps database44. MS/MS spectra were acquired using the timsTOF system with a query of the fragment masses to the Alex123 database45. The resulting identifications of the lipids are summarised in Table S1. Due to MS/MS analysis and identifications, the spatial distribution of some lipids can be reasonably understood. However, the MS/MS analysis of some masses was inhibited by the sample size and instrumental limitations which will be covered later in the “Discussion” section.

Fig. 4
figure 4

Multi-modal imaging of a single C. elegans section. (a) Optical image of a single longitudinal C. elegans cross-section, with (b) post-MALDI-MSI ORO staining. Distribution of several lipids at (c) m/z 466.35, (d) m/z 599.35, (e) m/z 648.49, and (f) m/z 939.71. (g) overlaid image of the data shown in (d)–(f). The peaks at m/z 167.08 and m/z 333.13 are from the matrix. (h) Average spectrum compared to a background of norharmane-sprayed embedding medium. (i) Expanded section of the light-blue shaded region in h). The nematode and background spectrum ROIs contain approximately 1500 spectra each.

Full size image

Cross-validation of neutral lipid distribution by ORO staining

ORO staining was performed on the sections following MALDI-MSI analysis to visualise the distributions of neutral lipids and correlate them with the MALDI results. A challenge of this post-analysis step is removing the matrix from the sections without affecting the lipids. For this purpose, we employed a vacuum-based physical method instead of a methanol-based chemical method (see “Methods”). Figure 4b shows an ORO-stained image of a section after MALDI analysis. Higher staining intensities are observed in areas corresponding to higher lipid signals and the presence of multiple lipids. This selectivity was also confirmed by the lack of staining in the cuticle, which is consistent with intact nematode staining (See Supplementary Information, Fig. S13). In addition, we found that ORO staining is useful for confirming MSI data that occasionally contains background signals. To further investigate the efficacy of ORO staining as a post-analysis method, we compared the data with the staining of intact nematodes and sections before MSI analysis (Fig. S13). These results show some slight changes in the colour intensity and distribution after analysis. In post-MSI analysis, the distribution of the lipid-bound dye is more uniform than that obtained without MSI analysis (Fig. S13b), which exhibits droplet-like structures similar to those observed for whole-nematode ORO staining (Fig. S13a).

3D consecutive sections for correlation with anatomical structures

After obtaining well-preserved sections for 2D MSI, we explored the potential of the proposed workflow for 3D MSI. We prepared 10 consecutive sections, starting from the section containing the top cuticle layer to the last section containing the bottom layer of the nematode. Figure 4 shows a stacked image of consecutive sections obtained using optical and MSI modalities. Because a single slice has a thickness of 10 \({\upmu {\textrm{m}}}\), the 10 sections span \(\sim\) 100 \({\upmu {\textrm{m}}}\), which corresponds to the nematode diameter, including the generally thinner top and bottom cuticle sections. As the MSI analysis was performed in dual-polarity mode to visualise a wider range of lipids and their interactions, the resultant stack was composed of odd sections obtained in negative-ion mode and even sections obtained in positive-ion mode (see Fig. S14 for separate ionisation modes). The negative-ion mode mapped the distribution of m/z 466.33 (Figs. 4c; S6c), whereas the positive-ion mode mapped the distribution of m/z 510.40, corresponding to LPC O-18:0 from the phosphatidylcholine (PC) lipid group (see S15 for MS/MS data). Variations in the MSI signal intensity are observed as the brighter coloured areas in the MSI images of the consecutive sections, and even within a single section. This is particularly noticeable in sections \(\sharp 4\)–6 of Fig. 5. Because the measured signals are related to membrane lipids, the differences could be related to changes in the cell density. Compared to the distribution of other lipids (Fig. S6), higher intensity spots are observed for both the embryo- and intestine-related areas. Embryo-related areas are visible as red-coloured spots on the right side of the MSI-analysed sections \(\sharp 4\)–8. The remaining high-intensity areas correspond to the intestine. Additionally, some areas show no detectable signals, such as \(\sharp 6\) and \(\sharp 10\), which is due to physical gaps in the nematode sections. These gaps were likely caused by the fragility of the sections, as inferred from the corresponding optical images. These sections may be thinner or contain smaller amounts of tissue compared to other sections. A possible method for correcting for these gaps is described in the “Discussion” section.

Fig. 5
figure 5

Optical and MSI images of consecutive sections of C. elegans. (Left stack) Optical images of ten consecutive sections obtained from a single C. elegans nematode. (Right stack) Corresponding MSI-derived lipid distributions. MSI images are alternating distributions of lipids found at m/z 466.33 in negative ionization mode, and m/z 510.40 in positive ionization mode.

Full size image

Discussion

Advantages of the proposed workflow

In this study, we aimed to develop a method for improving the correlation between lipidomic information and the spatial distribution of lipids in the nematodes. Although the C. elegans lipidome has been studied before using staining and chromatography-based methods, the correlation of lipids to different anatomical features has proven challenging due to the small size of the nematode bodies. Based on its ability to visualise single nematodes and differentiate lipid groups, Raman microscopy has been proposed as a potential method for bridging the gap between lipid location and identity46,47,48. However, accurate lipid identification based on Raman data is limited. MSI has proven to be an effective technology for lipid spatial distribution and identification. A previous study applied the so-called “popcorn” method in which the nematodes are frozen at \(-80^\circ {\textrm{C}}\) and later thawed for use22. Although this method successfully aligned the nematodes on gelatin blocks and retained their rod-like shape, the freeze–thaw process can disrupt the nematode cells49, resulting in the loss of distinguishable structures after subsequent sectioning (see Fig. S16). In contrast, the present method preserves the internal structures of the nematode in the sections (Fig. 2) , enabling the correlation between the structures and MSI images. As shown in Fig. 4, we observed different spatial distributions of lipids, such as lyso-phosphoinositol (LPI) 18:0, and monomethylated phosphoethanolamine glucosylceramide, mmPEGC, localization to the embryos and pharynx, respectively. The precise mass identification of lipids is expected to facilitate a deeper understanding of their biological functions, although the literature on lipidomic analysis of specific organs is scarce. For example, LPE (18:0) m/z 466.33 (Fig. 4c) is associated with cell membranes, which may explain its widespread presence in the nematode sections. Additionally, the mmPEGC lipid at m/z 939.71, specifically PE-NMe-HexCer 39:0;O450, has been indicated in the cholesterol metabolism of C. elegans51. Nematodes mainly obtain cholesterol from their diet, which is consistent with the observation of mmPEGC primarily in the pharynx and beginning of the intestine. Nonetheless, for the exact structure of the lipids, further analysis such as LC-MS/MS is necessary. However, that was not within the scope of the current study, as we focused our efforts on developing a reproducible method for C. elegans sectioning and analysis. Although the droplets could have been affected by the laser ablation, resulting in a more uniform distribution, a comparison with ORO staining nonetheless revealed that the areas with multiple lipid species identified by MALDI also exhibit higher stain intensities. This is reasonable considering that ORO stains all neutral lipids, so a higher colour intensity indicates a higher concentration of lipids. Additionally, to demonstrate the applicability of the present research to broader C. elegans lipid research, the results of the wild-type nematode data were compared to those of a known lipid mutant (See Supplementary Information; Fig. S17)52,53. Furthermore, these results highlight the value of our developed workflow for nematode sectioning and molecular analysis that avoids interference from the rigid cuticle.

3D reconstruction and future perspectives

The consecutive nematode sections with well-retained internal structures enable visual, feature-based correction and reconstruction of the sections, to produce 3D visualisations of the sample. Using the Landmark Correspondences transformation function in ImageJ (version 2.16.0)54,55, the optical images were transformed based on the similarities between physical features of the adjacent sections (Fig. 6a–c). The reconstruction was performed using the Rigid Moving Least Squares method. Although this method is considered to be suitable for transformations without causing large deformations, it is still known to cause some pixel deformation56. This could be more prominent in small grid sizes with a limited amount of corresponding pixels, such as the small nematode sections. Nonetheless, the effect of this transformation is shown in Fig. 6e–g and 3D reconstruction videos Supplementary Video S1 and S2 which are available in the Supplementary Information. Although the nematode sections were misaligned without transformation, it was possible to reconstruct the nematode from the consecutive sections. Using the transformed and aligned optical images as templates, we further transformed MSI images into a 3D visualisation of mass distributions in the nematode (see Fig. 6d, g; Supplementary Video S3 in Supplementary Information). For example, the intestinal line from the head to tail is clearer in the 3D reconstructed video than in separate consecutive section images. When viewing the 3D reconstructions of optical and MSI data together (Supplementary Video S4), distribution patterns in the structures of the optical reconstruction coincide with the mass distributions of the MSI reconstruction, as discussed previously in the Results section. In addition, this correlation can be used to visualise organ-specific or mass-of-interest distributions, and multi-mass overlays.

Fig. 6
figure 6

Image reconstruction and 3D visualisation. (a) Schematic overview of feature-based reconstruction. Yellow circles represent the selected feature and location used for reconstruction, while the numbers in the left and right image represent the matching features in the nematode graphics. (b) Optical image used as a template for feature-based reconstruction. Structures used as markers are labelled with yellow numbers (an example is circled in red). (c) Target image for reconstruction based on the template (b). Matching structures are labelled with yellow numbers with an example circled in red. Inset: corresponding MSI image. (d) Resulting reconstructed image. Inset: corresponding MSI image. (e) Misaligned stacked 3D images of consecutive sections. (f) Stacked 3D reconstructed images. (g) Stacked 3D MSI images after feature-based reconstruction.

Full size image

Technical limitations

A major limitation of the developed workflow is the multistep sample-preparation method that results in poor controllability of the process. The first step in this process is nematode handling and alignment. Freezing nematodes prevents movement and provides straight samples, but can damage their internal structures. Although protocols for minimising freezing damage have been developed, they are generally used for storage purposes57,58. Furthermore, the embedding medium affects the quality and stability of the sections, and thus, the imaging results59. However, the embedding medium should be optimised for each sample and study purpose, therefore, the medium presented in this study might not be optimal for other samples. The matrix deposition method and the resulting crystal size can also influence the results through their spatial resolution59,60 and the compatibility with the used laser spot size. The rapifleX system has a relatively small laser spot size of 5 \({\upmu {\textrm{m}}}\). However, adult C. elegans have a diameter of 60–70 \({\upmu {\textrm{m}}}\), which corresponds to only 12–14 pixels across the nematode body when analysed with a 5 \({\upmu {\textrm{m}}}\) laser spot (Fig. S18). Additionally, MSI analysis of such small samples limits the visibility of the images and thus the interpretation of the results, through insufficient spatial resolution. New methodological improvements, such as the tissue-expansion mass-spectrometry imaging (TEMI) approach could be considered for application to C. elegans research for helping overcome this limitation61. Furthermore, although the timsTOF system has the laser spot size of 5 \({\upmu {\textrm{m}}}\), using this setting did not provide sufficient material for MS/MS analysis. Instead, a larger M5 flat setting with a resulting spot size of 114 \({\upmu {\textrm{m}}}\) had to be used. By using such a laser spot size, MS/MS analysis of C. elegans can result in the fragmentation of higher intensity peaks arising from the surrounding matrix and embedding medium. Alternatively, performing LC-MS/MS analysis would provide lipid identification, but not the confirmation of its location in the nematode body. Finally, in future studies, the addition of internal standards to the matrix solution should be considered. Although we showed that the observed variations are more likely associated with the variations between samples than measurement artefacts (Fig. 3), internal standards would further improve the accuracy of data analysis. Internal standards could be used to normalise the spectral data to facilitate the quantitative analysis of lipids.

Conclusion

The workflow presented in this study achieved the successful consecutive sectioning of C. elegans with conserved internal structures and provided 3D MALDI-MSI images with features correlating with these structures. Consecutive sections that span the entire nematode with preserved internal structures enable the improved visualisation of lipids in the nematodes via correlating MSI and ORO staining images. Additionally, the reproducibility of the protocol and its ability to control the sample orientation are beneficial for studies using bigger amounts of samples. In addition, we confirmed that the sample-preparation process is compatible with MALDI-MSI, which requires the preparation of samples that do not interfere with later analysis. The proposed method enables the visualisation of the lipid profile within the sample including organ-specific distributions. Finally, staining protocols before and after MALDI-MSI analysis can be included in the workflow to add another layer of information that is comparable to that obtained by staining whole nematodes. The present technique could be applied to various C. elegans studies requiring a detailed correlation between the anatomical features and lipid distributions.

Methods

Chemicals

Porcine skin gelatin powder (Type A, gel strength \(\tilde{3}\)00 g Bloom), ultra-high viscosity CMC, agar powder, and DPX slide mounting medium were acquired from Sigma Aldrich. Japan. Crystalline norharmane (\(\ge\)98% by thin-layer chromatography), ORO solution (0.5% in propylene glycol), and red phosphorus (97%) were purchased from Sigma Aldrich, The Netherlands. Sodium azide (\(\geqq\)97.0%, extra-pure reagent) and xylene (JIS reagent grade 1, \(\geqq\)80% by gas chromatography) were purchased from Nacalai Tesque Japan. Tissue-Tek\(^{\circledR }\) Optimal Cutting Temperature (OCT) Compound was purchased from Funakoshi Japan. Isopropanol (suitable for ultra-liquid chromatography mass spectrometry and convergence chromatography/supercritical fluid chromatography), methanol (suitable for ultra-performance liquid chromatography and ultra-high-performance liquid chromatography), and chloroform (suitable for high-performance liquid chromatography) were obtained from Biosolve, The Netherlands. Imsol Mounting Medium and ethanol (70% (v/v), TechniSolv) have been obtained from VWR, The Netherlands. Isopropanol and ORO powder were purchased from Fujifilm Wako, Japan. Hematoxylin and Eosin (H&E) Stain Kit was purchased from Cosmo Bio, Japan.

Nematodes and culturing

The wild-type N2 (Bristol) strain of C. elegans was used in the experiments, which was provided by the Caenorhabditis Genetics Center (CGC) at the University of Minnesota (funded by The National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440)). The nematodes were cultured on nematode growth medium (NGM) plates consisting of 2% agar seeded with Escherichia coli OP50, provided by CGC. Once many gravid adults were present on the plate, egg bleaching was performed according to an established protocol62 to achieve phase synchronisation. As the final step of the bleaching protocol, the nematode batches were left to incubate at \(20^\circ {\textrm{C}}\) until grown to the young-adult stage, at which time they were collected for further preparation.

Medium preparation

Three different embedding media were tested: 10% gelatin, 10% gelatin with 5% CMC, and 10% gelatin with 2% CMC. All media were prepared by dissolving gelatin and/or CMC powders in deionised water using 50 mL Falcon tubes, followed by briefly vortexing the ingredients for about 30 seconds and heating until a homogenous mixture was formed. In the case of the gelatin/CMC media, the powders were mixed before adding deionised water. The CMC containing media required heating in a water bath for minimally 2 hours after vortexing to remove air bubbles. During water-bath heating, the water level was maintained to ensure that it covered the entire tube containing the embedding medium. Before embedding the samples, the medium was reheated in a water bath until it liquefied. To maintain the heat for as long as necessary during sample preparation and prevent solidification of the medium, the glass beaker used for the water bath was wrapped in thick aluminium foil, and another piece of foil was used to cover the beaker.

Nematode embedding

Once the young-adult stage was reached, 30 nematodes were collected in batches of 10. Each batch of 10 nematodes was placed in a 10 \(\upmu {\textrm{L}}\) droplet of 30 mM sodium azide for anesthetising the nematodes and removing the E. coli food source from the cuticle. After washing, the nematodes were transferred onto a Wormsheet PDMS microfluidics chip63,64 (Biocosm, Hyogo, Japan), to separate them into two groups of 5. Each group was placed in a 3.5 \(\upmu {\textrm{L}}\) droplet of M9 buffer before being pressed with the plastic cover provided with the Wormsheet. This step was performed to ensure an even dispersion of the nematodes on the PDMS chip to facilitate sectioning and analysis. The nematodes were set in etched channels with a width of 50 \({\upmu {\textrm{m}}}\), appropriate for young adult nematodes according to the product specifications. Then, the plastic cover was removed by sliding it off in the direction of the channels. A total of 7 \(\upmu {\textrm{L}}\) of M9 buffer was used to prevent the PDMS chip from drying too quickly or becoming too moist during cover removal. If the chip is too dry, it is not possible to slide the cover, whereas lifting the cover dislodges the nematodes from the channels. When the chip is too moist, the nematodes may slide out of the channels as the cover is removed by sliding. Once the cover has been successfully removed, excess liquid was removed from the edges of the chip using a lint-free wipe. A mould was designed in Fusion360 software (V.2.0.15299) and printed in-house using an Ultimaker S3 3D printer. The printed mould was placed around the PDMS chip and filled with the embedding medium. Because of the high viscosity of the solution, even when heated, a modified glass pipette with the tip removed was used instead of a semi-automatic pipette. To ensure a flat surface for the following step, a lid was placed on top of the embedding medium and lightly pressed to prevent the medium from drying in the ambient air. Additionally, to avoid trapping air in the medium, the mould was slightly overfilled, and the lid was slid from the side before being pressed. The medium was left to stand for 12 hours at 4 \(^\circ {\textrm{C}}\) to set. Then, the sample preparation was continued by removing the filled mould and its lid from the PDMS chip and glass slide. The mould lid was then removed, and the transfer and retention of nematodes in the block was confirmed using a stereomicroscope after placing the mould upside down on a glass slide. Once the presence and orientation of the nematodes were confirmed, the block was gently removed from the mould by slowly detaching the edges using the flat end of a clean, metal weighing spoon. Then, a smaller mould was placed around the areas of the block containing the nematodes and lightly pressed to remove excess embedding medium, enabling easier sample handling during sectioning and the placement of additional sections on a single glass slide. A small piece of tape was placed at the corner of the block, ensuring that it did not cover the nematodes. Another layer of the prepared embedding medium was then added on top of the trimmed block and covered with the lid, following the same procedure described previously for the first layer of the block. This process effectively produces a gelatin–nematode “sandwich structures”, with C. elegans placed in the centre of the block. The tape acted as a guide for later cryo-sectioning, as it remained visible after freezing. The gelatin was again cooled at 4 \(^\circ {\textrm{C}}\) for 12 hours. Then, the gelatin blocks were fully removed from the mould using the same procedure used to remove the first layer of gelatin. The blocks were placed in aluminium boats and then flash-frozen using liquid nitrogen. The boats prevented direct contact between the samples and liquid nitrogen. Once frozen, the samples were stored in aluminium foil and Falcon tubes in a \(-80 ^\circ {\textrm{C}}\) freezer until required for cryosectioning. A schematic and visual overview of the workflow are presented in Fig. 1a.

Cryosectioning

Sectioning of the samples was performed using a Leica Biosystems CM1850 (Leica Biosystems, Tokyo, Japan) cryotome set at \(-20 ^\circ {\textrm{C}}\). Before sectioning, all the equipment and samples were left inside the cryostat for about 5 minutes for temperature adjustment. Frozen sample blocks were secured onto the specimen disk using Tissue-Tek\(^{\circledR }\) O.C.T. Compound (Funakoshi, Japan). The orientation and expected locations of the nematodes were confirmed using the embedded tape as a guide. Sections with a thickness of 10 \({\upmu {\textrm{m}}}\) were transferred onto the glass slide by gently placing the slide on the section to prevent displacement of the nematodes. During sectioning, sections obtained near the marking tape were visually inspected for the presence of air pockets within the gelatin sheets, indicating the presence of C. elegans. Once air pockets were observed, the sections were collected by thaw-mounting on the conductive side of indium-tin oxide (ITO) glass slides (Matsunami Glass, Japan). A total of 12 sections were collected on each slide. The sections collected for staining and placement tests were thaw-mounted on regular glass slides (Matsunami Glass). Using the PDMS chip channel width as a guide, it was estimated that each nematode resulted in approximately five to six sections, as shown in Fig. 3b. However, owing to external factors such as the tilt of the gelatin block caused by the OCT layer and cryotome sample holder, additional sections were acquired to ensure complete retrieval of the nematode material.

Matrix application (sublimation and spraying)

To ensure sufficiently small norharmane crystals necessary for 5 \({\upmu {\textrm{m}}}\) high-spatial-resolution imaging, an HTX SublimatorTM system (HTX Technologies, LLC, Chapel Hill, NC, USA) was used for matrix deposition60. Norharmane was used as the matrix in this study. Approximately 50 \(\hbox {mg}\) of norharmane was dissolved in 1.5 \(\hbox {mL}\) of methanol, to coat one slide in a single sublimation run. For each sublimation run, the system was preheated to 60 \(^\circ {\textrm{C}}\), and then the matrix solution was poured into a matrix tray and left to dry. Afterwards, the sample was inserted into the sample holder and left to reach a vacuum pressure of 40 mPa. The tray temperature was set to 140 \(^\circ {\textrm{C}}\) and sublimation was performed for 220 seconds. Finally, the corners of the slides were cleaned using methanol to allow for contact with the slide holder and conductivity during the analysis. To enhance the lipid extraction needed for MS/MS analysis, M3+ Sprayer (HTX Technologies) was used for matrix deposition as it provides a more humid environment. The matrix was prepared at a concentration of 7 mg/mL of norharmane in 2:1 \(CHCl_3\):methanol. Criss-cross pattern was used over 15 spray cycles with 30 seconds drying time between cycles. The nozzle flow rate was set to 0.12 mL/min and a temperature of 30 \(^\circ {\textrm{C}}\).

MALDI-MSI and MS/MS analysis

The prepared sections were analysed using a rapifleX Tissuetyper system (Bruker, Bremen, Germany)65. MALDI-MSI data were acquired using both positive and negative polarities, using a 355 nm Nd:YAG laser with a spot size of 5 \({\upmu {\textrm{m}}}\). A mass range of 100–1000 Da was used in reflectron mode to enhance the mass resolution in the lower mass range. High-mass-resolution MS was performed using a Q Exactive Plus system (Thermo Fisher Scientific GmbH, Bremen, Germany) coupled with a reduced-pressure electrospray ionisation/MALDI ion source (Spectroglyph LLC, Kennewick, WA, USA). The instrument was operated in both positive and negative ion modes with a mass resolution of 120,000 (full width at half maximum at m/z 400) with a laser pulse energy of 1.9 \({\upmu {\textrm{J}}}\) (1 kHz). Tandem mass spectrometry (MS/MS) analysis was performed using the trapped ion-mobility time-of-flight timsTOF system (Bruker) equipped with a MALDI source. To acquire sufficient material, the laser beam was set to 25 x 25 \({\upmu {\textrm{m}}}\) in M5 flat mode with a resulting field size of 114 x 114 \({\upmu {\textrm{m}}}\). Consequently, it was not possible to perform MS/MS imaging. Isolation window of 0.5 Da was used for m/z 423.31, 466.33 and 599.35, while 1 Da isolation window was used for m/z 648.49 and 939.71 in negative, and m/z 510.40 in positive ionisation mode. However, due to the low abundance of peaks at m/z 423.31 and 648.49, it was not possible to obtain their MS/MS spectra (Fig. S19). The collision energy range was between 30 and 50 \(\hbox {V}\). After matrix deposition and before imaging, red phosphorus was spotted onto the glass slide outside the area with the gelatin sections. This was then used to calibrate the positive and negative ionisation modes, over a measurement range up to 1000 Da66. Following data acquisition, images were prepared using flexImaging software (version 4.1, Bruker Daltonics GmbH, Bremen, Germany). A mass window of 332.9–333.8 Da, corresponding to the [2M-H-2H]\(^-\) ion of the norharmane matrix67 was used to normalise the negative ionisation mode datasets. For normalisation of positive ionisation mode data, the mass window was set to 334.8–335.7 Da corresponding to the [2M-H]\(^+\) norharmane ion. The peak at m/z 466.33, which was observed for all nematode samples in negative ionisation mode, was used as a guide for the delineation of ROIs containing only information pertaining to C. elegans. An example of this guided delineation is shown in Fig  4c. This was also confirmed by comparing multiple single-spot spectra from nematodes with the single-pixel spectra obtained for gelatin and ITO slides. The averaged spectra from the nematode ROIs were loaded into Mmass software (version 5.5.0)68. Peak picking was performed automatically with a signal-to-noise ratio (S/N) threshold of 3.0. Three norharmane peaks corresponding to [M-H]\(^-\) at m/z 167.0604, [2M-H-2H]\(^-\) at m/z 333.1135, and [3M-H]\(^-\) at m/z 499.1666 were used for spectral calibration. Information related to the absolute intensity, resolution, and S/N, was retrieved for the peaks of interest using Mmass with the resolution set to four decimal places. Data were visualised using OriginPro 2025 software (Learning Edition, version 10.2, OriginLab, Northampton, MA, USA). MS/MS data was converted to an Mmass compatible format using Data Analysis (Bruker Daltonics). Preliminary identification using exact masses was performed using the LipidMaps database, with a mass tolerance of \(\pm 0.05\). The mass tolerance was selected based on the mass resolution of the instrument and the mass shift observed for matrix related peaks across the spectrum67. Further identification using MS/MS analysis datasets were obtained using Alex123 database using the obtained masses of parent and fragment ions with a mass tolerance of \(\pm 0.01\) for both. For example, 1-octadecyl-sn-glycero-3-phosphoethanolamine was identified by the exact mass of m/z 466.3303, and the masses of its main fragments, m/z 405.2718, 269.2431 and 196.0340 (Fig. S9). Data visualisation in the present study focused on the negative ionisation mode data, as it showed a larger number of nematode-related masses.

Statistical analysis of method reproducibility

The statistical significance of the variance in the data obtained from MALDI-MSI was analysed by one-way ANOVA using the Origin software. ANOVA was performed on the absolute intensity values of the whole nematode ROIs for the peak at m/z 423, using data from eight separate nematodes with a single ROI per nematode. Furthermore, where applicable, four separate nematodes were analysed for each consecutive section, and two to six ROIs were acquired per nematode.

Histological staining

ORO staining

After MSI analysis, the sample slides were stained for neutral lipids using ORO197. First, the matrix was removed and the sections were fixed by fully submerging the slides in 10% neutral buffered formalin solution for 10 min. Subsequently, the slides were quickly rinsed by dipping in 60% isopropanol and then submerged in 0.5% ORO in propylene glycol solution for 1 hour. Excess ORO solution was removed by dipping the slides three times in 60% isopropanol followed by demineralised water. To preserve the stained slides, three drops of Imsol mounting medium were placed on the slide at a distance of approximately 1.5 cm from each other, covered with a coverslip, and left to set before scanning using a Leica Aperio CS2 digital pathology slide scanner (Leica Biosystems, Wetzlar, Germany). A similar protocol was used to stain sections that have not been analysed with MSI. However, the matrix removal step was omitted, and sections were submerged in 0.3% ORO in 60% isopropanol solution for 30 minutes. Whole nematode ORO staining was performed according to the published protocol7 .

Hematoxylin and eosin staining (H&E)

Slides were stained using the hematoxylin and bluing reagent to confirm the presence of C. elegans in the sections. Eosin staining was omitted as it was unnecessary for locating the nematodes and to avoid staining the embedding medium. The slides were stained with hematoxylin for 5 minutes. The dye was removed using demineralised water for 3 minutes before applying the bluing reagent for 20 to 30 seconds. Staining was completed by removing the reagent using demineralised water and dehydrating it in 100% ethanol. Finally, it was secured with DPX mounting medium, covered with a coverslip and left to dry before microscopic observation.