Research ArticleAPPLIED SCIENCES AND ENGINEERING

Spatially resolved 3D metabolomic profiling in tissues

See allHide authors and affiliations

Science Advances  27 Jan 2021:
Vol. 7, no. 5, eabd0957
DOI: 10.1126/sciadv.abd0957
  • Fig. 1 3D-SMF.

    (A) The schematics of 3D metabolic profiling over 50 to 150 depth layers in a tonsil specimen is illustrated. (B) Calibration of the mass peak lists was performed using the negatively charged ions. (C) Postanalysis of the acquired TOF-SIMS data included extraction of the mass channels, their corresponding intensity counts, and duration of data collection. (D) To determine the exact value of a mass channel, we identified the peaks of the relevant mass channels during the image acquisition or after the measurements. (E) Mass channels were carefully chosen on the basis of the intensity and width of mass channels. (F) The acquired images from the mass channels were analyzed using log transforms and filtering, followed by the data export into an American Standard Code for Information Interchange (ASCII) text. (G) 3D metabolic distributions in the acquired TOF-SIMS datasets were processed by postanalysis methods. (H) The spatial clustering of all TOF-SIMS datasets into six distinct clusters was performed using six unique colors. (I) Overlaid images of isotope-labeled mass channels with those containing distinct 10 lipid fragments were shown. (J) TOF-SIMS signal difference between isotope-labeled and unlabeled dataset of the same GC was illustrated. (K) Prediction of isotope-labeled dataset amino acid mass channel was calculated from mass image differences of labeled and unlabeled channels.

  • Fig. 2 Anticorrelated and correlated spatially resolved metabolic distributions in tissues.

    (A) A 5-μm-thick unlabeled tonsil slice was imaged by a TOF-SIMS device, providing 3D metabolite images over 138 depth slices. (B) The calibration and identification of the 189 mass channels were carried out, and the resultant measurement data were exported into an ASCII text file. (C) The bar graphs for total ion 3D signals were used for the highest 50 correlated and anticorrelated metabolite pairs. (D) The t-SNE plot for the unlabeled tonsil dataset demonstrated the highest (blue) and lowest correlated (anticorrelated) metabolites (green). (E) Six clusters were obtained from a 3D K-means algorithm to visualize selected five metabolites of 50, 63, 74, 79, and 385 m/z (C23H45O4, cholesterol) on a heatmap of normalized z-scores. (F) Pairwise correlations of the 189 metabolites of the unlabeled tonsil tissue dataset were computed on a clustered heatmap. (G) Five metabolites of 50, 63, 74, 79, and 385 m/z (C23H45O4, cholesterol) were visualized for six clusters. The first column contains pixel-binned 2D images, the second column shows the 3D clusters, and the third column displays the 3D voxels of the five metabolites. The clusters and the original metabolites were quantified by F1 scores (0.661 to 0.982).

  • Fig. 3 Isotope-conjugated antibodies were labeled on tissue sections.

    (A) Cell-specific labeling of an unlabeled tonsil-sliced sample using an isotope-tagged antibody library. (B) Specific cellular features and metabolic compounds were identified. (C) After labeling, the tonsil tissue was analyzed by TOF-SIMS imaging to acquire a 3D dataset. (D) The t-SNE plot exhibited the highest correlated metabolites (blue), anticorrelated metabolites (green), and isotope-tagged antibody labels (black) in the labeled tissue dataset. (E) The calibrated mass spectra consisted of 189 compounds that contained 20 peaks for cell features and 169 peaks for metabolites. (F) Pairwise 3D correlations for only the antibody-labeled masses of the isotope antibody–labeled tonsil dataset across 20 binned slices were shown on a hierarchically clustered heatmap. (G) A clustered heatmap was used to represent the pairwise 3D correlations between a subset of metabolic compounds and cell-specific labels. (H) The hierarchically clustered heatmap showed pairwise 3D correlations of the five binned slices of the metabolite compounds and antibody labels. The dominant red/green colors in the heatmaps included a colormap for blue (low correlations), green (medium correlations), and red (high correlations).

  • Fig. 4 Cell type–specific and metabolic correlations for the GC in tonsil tissues.

    (A) An optical imaging technique was used to observe the 5-μm sliced tonsil section before TOF-SIMS. (B) The selected ROI was identified consistently between the optical imaging platform and the TOF-SIMS platform. In the TOF-SIMS device, an internal optical camera was used to align the ROIs. (C) The image showed the outside of the GC captured by the internal optical camera before the TOF-SIMS measurement process. (D) The images were shown for the border of the GC captured by the internal optical camera image before the TOF-SIMS measurement process started. (E) Comparative hierarchically clustered heatmaps for the 3D correlations were implemented for inside, outside, and at the border of the GC in the labeled tonsil datasets, yielding unique lipidomic features in each dataset. (F) Comparative t-SNE maps of the same GC of the labeled tonsil tissues was displayed. The t-SNE contained correlated (blue), anticorrelated (green), and antibody-labeled (black) channels. (G) Bar plots of the secondary ion signal levels were computed for the selected antibody labels and the metabolite compounds. (H) The bar graphs showed the 50 highest correlated and anticorrelated rankings of metabolic pairs from the 3D Pearson correlations.

  • Fig. 5 3D clustering of metabolic fragments in spatially distinct cell distributions in tissues.

    (A) The prominent mass channels were identified in the inside GC dataset that contained 50 and 385 m/z (C23H45O4, cholesterol) and 122 and 98 m/z (C5H12NO+, lipid). The first column contains pixel-binned images, the second column shows the 3D clusters, and the third column displays the 3D visualization of voxels of the four metabolic channels. The 3D overlaps between metabolites and clusters were represented using colocalization in the fourth column and an F1 score in the range of 0.661 to 0.982 in the fifth column. (B) The prominent mass channels in the outside GC are 75, 80, 229, and 42 m/z. The F1 scores were 0.889 to 0.999. (C) The prominent mass channels in the border GC are 75, 42, 50, and 144 m/z. The F1 scores were 0.683 to 0.995. (D) A comparison of the detected metabolic compounds was shown in the form of a bar graph. Distinct distributions of lipidomics were obtained for inside, outside, and at the border of the same GC. (E) A comparison of TOF-SIMS mass channel images that were overlaid using the identified immune labels and lipid fragments. In each row, immune markers and lipid fragment metabolic images were assigned to unique colors.

  • Fig. 6 Spatially resolved clustering and coloring of isotope-labeled and unlabeled GC in serial tonsil sections.

    (A) A comparison of TOF-SIMS mass channel images overlaid on microscopy images with lipid fragments from the labeled and unlabeled GC was shown. Two rows of lipid fragment metabolic images were assigned a unique color and then overlaid together with the morphological image. (B) A comparison of TOF-SIMS mass channel images overlaid from the isotope-labeled mass data of immune targeted channels and lipid fragments for labeled and unlabeled GC was shown. Each row immune marker and lipid fragment metabolic image were assigned a unique color and then overlaid together.

  • Table 1 Library of isotope-tagged antibodies for cell labeling in TOF-SIMS.

    NumberMassLabelDescription
    1141PrSMA
    2143NdVimentin
    3148NdPanKeratin
    4149SmH3K9me3
    5150NdPDL1
    6155GdFoxp3
    7156GdCD4
    8158GdE-cadherin
    9159TbCD68
    10161DyCD20
    11162DyCD8A
    12165HoPD1
    13167ErGranzyme B
    14168ErKi-67
    15169TMCollagen I
    16170ErCD3
    17171YbHistone 3
    18173YbCD45RO
    19191IrIntercalator
    (DNA)
    20193IrIntercalator
    (DNA)
  • Table 2 Unique compound distributions inside, outside, and at the center of GCs in tonsil tissues.

    Masses (m/z)
    Inside germinalCommon correlated
    compounds
    194, 35, 66, 100, 201,
    42, 75, 76, 107, 16, 84,
    88, 185, 91, and 60
    Common
    anticorrelated
    compounds
    1, 35, 42, 50, 185, 58,
    57, 63, 64, 194, 66,
    196, 201, 74, 75, 84,
    85, 86, 88, 91, and 107
    Top correlation
    compounds pairs
    75-42, 91-42, 75-91,
    84-91, and 84-42
    Top anticorrelation
    compounds pairs
    50-16, 50-179, 50-76,
    16-122, and 359-431
    Outside germinalCommon correlated
    compounds
    128, 66, 35, 100, 42,
    75, 107, 79, 84, 86, 91,
    and 63
    Common
    anticorrelated
    compounds
    1, 128, 133, 35, 42, 64,
    66, 72, 73, 75, 84, 86,
    91, 92, 95, 100, 102,
    107, and 124
    Top correlation
    compounds pairs
    75-84, 60-76, 75-91,
    60-16, and 16-17
    Top anticorrelation
    compounds pairs
    73-75, 73-42, 73-91,
    73-84, and 73-100
    Border germinalCommon correlated
    compounds
    97, 98, 35, 66, 72, 73,
    42, 75, 74, 50, 84, 86,
    185, and 60
    Common
    anticorrelated
    compounds
    1, 35, 42, 185, 58, 63,
    64, 66, 67, 68, 194,
    201, 75, 82, 84, 85, 86,
    87, 88, 91, and 124
    Top correlation
    compounds pairs
    75-42, 50-185, 50-74,
    75-91, and 50-73
    Top anticorrelation
    compounds pairs
    1-158, 1-155, 1-98,
    208-1, and 267-1

Supplementary Materials

Stay Connected to Science Advances

Navigate This Article