Dopamine-modified TiO2 monolith-assisted LDI MS imaging for simultaneous localization of small metabolites and lipids in mouse brain tissue with enhanced detection selectivity and sensitivity

A dopamine-modified TiO2 monolith was developed to assist LDI MS imaging for small metabolites and lipids simultaneously with enhanced sensitivity.


Tissue preparation and sectioning
In all cases, the rodent brains were surgically dissected, frozen in liquid nitrogen, and stored at -80 ⁰C until use. Coronal tissue sections, 14-µm thick, were prepared from frozen mouse cerebrum using a cryostat (3050S, Leica Biosystems Inc., Buffalo Grove, IL) at -19 ⁰C and thaw-mounted onto conductive indium-tin oxide (ITO)-coated glass slides (Delta Technologies, Loveland, CO).
While most samples were analyzed immediately following preparation, some sections were stored at -80 ⁰C for later use. Optical images of the tissues were taken using a flatbed scanner (Epson Perfection V300, Epson America, Inc., Long Beach, CA) with a resolution of 2400 dpi before MSI. The optical images shown in Figure 3 were adjusted to aid in visualization of hippocampal structures using Adobe Photoshop 2014.

Materials characterization
UV-vis absorption spectra detection was performed on an EPOCH TM microplate spectrophotometer (Biotek Instruments, Inc., Winooski, VT) with a scanning range of 200-700 nm. TiO 2 nanoparticle sol solution at the concentration of 0.05 M was measured directly by adding 200 µL solutions into a standard 96-well plate.
Diffuse reflectance UV-vis spectrometry was performed on a Cary 5000 UV-Vis spectrophotometer (Agilent, Santa Clara, CA) with a wavelength scanning range of 200-700 nm. The same amounts of different TiO 2 materials were first coated on quartz slides (1" × 1", Ted Pella, Inc., Redding, CA) with an airbrush (Paasche Airbrush Company, Chicago, IL), and then the slides were used for detection, with bare quartz slides evaluated as blanks.
An environmental scanning electron microscope (ESEM) (Philips XL30 ESEM-FEG, FEI, Hillsboro, OR) was used to investigate the micro-morphology of the TiO 2 materials coated on tissue slices. Before ESEM detection, the samples were coated with gold for 1 min with a sputter coater (Desk-1 TSC, Denton Vacuum, Moorestown, NJ).

Laboratory-constructed system for MALDI matrix sublimation
Sublimation was carried out using a laboratory-constructed system, similar to one previously described, with some modifications. 1,2 Briefly, an aluminum foil boat was affixed with double-sided conductive copper tape to the inner base of a sublimation chamber. Samples were attached to a copper plate affixed to the bottom face of an ice-filled cold finger. The MALDI matrix-to-sample distance was ~20 mm. For each sublimation deposition, 350 mg of powdered DHB was distributed evenly in the boat. The sublimation chamber and cold finger were assembled together per manufacturer's instructions, pumped to intermediate vacuum (~10 mTorr), and placed in a heating mantle (Glas-Col LLC, Terre Haute, IN) to equilibrate the vacuum and cool the sample plate. The optimized deposition conditions for derivatized samples included supplying 120 V to the heating mantle for 12 min. After matrix deposition, the chamber was removed from the mantle, vented with room temperature air (25 ⁰C), and the sample promptly removed from the cold finger.

Data analysis
The molecular ion distribution images of the tissue sections were visualized using flexImaging software 4.1 (Bruker Daltonics, Billerica, MA). MALDI MSI data acquired from triplicate or duplicate brain slices from the same mouse, thaw-mounted on three or two separate slides, were used in the statistical analysis.
For method development, samples from the same mouse were used with different sample preparation conditions on different slides.
In the method application experiments, optimized sample preparation conditions were used in the analysis of specimen from 8-month-old (young, n = 3) and 24-month-old (old, n = 4) mice. Slices from different animals were thaw-mounted on the same ITO slide. Statistical comparisons of peak intensities, areas, or S/Ns acquired from different brain regions of different animals, or samples from the same animal but prepared with different sample preparation conditions, were performed by exporting data from manuallydefined regions of interest (ROIs). Mass spectra from each ROI were imported into ClinProTools software 3.0 (Bruker Daltonics) with automatic baseline subtraction and total ion count normalization, except for the data presented in Figure 3 and Figure S5. Peaks were picked with an S/N threshold greater than 3 on average spectra. Picked peak parameters were exported as m/z value-peak area tables. To analyze changes in the levels of molecular signals in different subregions of mouse brain, ROIs were chosen. ROIs were outlined using ion maps and optic images aligned to the appropriate Mouse Brain Atlas schematics (http://mouse.brain-map.org/static/atlas). Four ROIs were selected: the CA1, CA3, and dentate gyrus of the hippocampus region, and corpus callosum (shown in Figure 6A). The mass spectra acquired at chosen regions were imported into ClinProTools, set to pick the peaks with an S/N >3. Peak areas were calculated and exported creating a single data set. For the multivariate analysis (principal component analysis, (PCA)) presented in Figure 6, the dataset was imported into OriginPro 8.5 (OriginLab Corporation, Northampton, MA). The mean values of peak areas for signals acquired in selected ROIs were calculated, and the values of technical triplicates or duplicates were used to calculate the relative standard deviation of the replicate measurements. For comparison of the averaged peak areas acquired from different animal groups (old vs. young) described in Figure 5, two way analysis of variance (ANOVA) was performed with OriginPro 8.5 to determine the significance of age effect on compound intensity.    Figure S7. Comparison of mass spectra from mouse brain tissue obtained using (blue) TiO 2 sub-micron particle-, (yellow) TiO 2 -DA sub-micron particle-, and (light blue) TiO 2 -DA monolith-assisted LDI, and mass spectra of blank samples with (dark blue) TiO 2 sub-micron particle-and (orange) TiO 2 -DA monolith-assisted LDI. Figure S8. Effect of water content in reaction solutions for further hydrolysis/condensation on TiO 2 -DA structure morphology and TiO 2 -DA materials-assisted LDI MSI measurements. (Left) Representative mass spectra acquired from mouse brain hemisphere sections coated with TiO2-DA materials generated in the presence of (A) 1%, (B) 2.5%, (C) 5%, and (D) 10% water in ethanol solution (reaction time is 60 min each). (Insets) SEM images of corresponding samples coated with TiO2-DA materials. (Right) Ion maps of representative lipid (m/z 844.5) distributions in the sections. Acquisitions were performed in the low spatial resolution mode with a 100-µm raster step size using settings producing a laser footprint ~100-µm in diameter. Figure S9. Ratios of an intact lipid to its fragment peak areas as determined using different TiO 2 -based LDI MS methods. The intact lipid (m/z 844.5) to fragment (m/z 141.0) ratios were calculated using data acquired from different subregions of mouse brain using LDI MS assisted with TiO 2 sub-micron particles (DA0), TiO 2 -DA sub-micron particles (DA1), and TiO 2 -DA monoliths (DA4). CA1, region I of hippocampus proper; CA3, region III of hippocampus proper; CC, corpus callosum.

S-12
Figure S10. Scanning electron microscopy image of the TiO2-DA monolith surface with low magnification (×5000). Figure S11. Repeatability of the unmodified TiO 2 sub-micron particle (DA0)-, TiO 2 -DA sub-micron particle (DA1)-, and TiO 2 -DA monolith (DA4)-assisted LDI MS measurements. Histograms depict average peak S/Ns of different molecular signals: (A) m/z 100-500, (B) m/z 500-700, and (C) m/z 700-900, acquired from region I of hippocampus (CA1), region III of hippocampus (CA3), and corpus callosum (CC) of different animals prepared and analyzed in different weeks. Average peak S/Ns with standard deviation error bars calculated using data obtained from measurements of adjacent brain slice triplicates collected from the same animals. The slices were deposited on different ITO glass slides and coated with the TiO 2 materials. The relative standard deviations of the triplicate data acquired using TiO 2 -DA monoliths are labeled on the corresponding bar graphs. Columns are positioned and boxed according to the week when samples were collected and analyzed.

S-17
Figure S14. Loading plots of (top) PC1 and (bottom) PC2 of data collected from young and old mice samples using the optimized TiO 2 -DA-assisted LDI method. Table S1. Tukey test results for all of the data sets presented in Figure S1 and Figure S3.  Table S2. Peak list collected from averaged mass spectra acquired from mouse brain tissues and corresponding blanks using TiO 2 sub-micron particle-, TiO2-DA sub-micron particle-, and TiO2-DA monolith-assisted LDI MS.    Table S5. ANOVA results for data presented in Figure 5C and Figure S12B.