Matrix-Assisted Laser Desorption/Ionization (MALDI) Mass Spectrometry Imaging of L-4-Phenylalanineboronic Acid (BPA) in a Brain Tumor Model Rat for Boron Neutron Capture Therapy (BNCT)

Boron neutron capture therapy (BNCT) is a cell-selective particle therapy for cancer using boron containing drugs. Boron compounds are accumulated in high concentration of tens ppm level of boron in target tumors to cause lethal damage to tumor tissue. The examination of boron distribution in target tumor and normal tissue is important to evaluate the efficiency of therapy. The matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) is a powerful tool to visualize the distribution of target analyte in biological samples. In this manuscript, we report a trial to visualize the distribution of a typical BNCT drug, L-4-phenylalanine boronic acid (BPA) in a brain tumor model rat using MALDI-MSI technique. We performed a MALDI-MSI with high mass resolution targeting to [BPA+H]+ at m/z 210 in a BPA-treated rat brain section using a spiral orbit-type time of flight (SpiralTOF) mass spectrometer. Several BPA ion species, [BPA+H]+, [BPA−H2O+Na]+, [BPA+DHB−2H2O+Na]+ and [BPA+DHB−2H2O+K]+ were detected separate from peaks originated from biomolecules or matrix reagent by achieving the mass resolving power of approximately 20,000 (full width at half maximum; FWHM) at m/z 210. The mass images with 60 μm spatial resolution obtained from these BPA ion species in a mass window of 0.02 Da revealed their localization in the tumor region. Additionally, the mass image obtained from [BPA+H]+ also likely showed the distribution of BPA inside the tumor. MALDI-MSI with high mass resolution targeting to [BPA+H]+ has a great potential to visualize the distribution of BPA in brain tissue with tumor.


INTRODUCTION
Boron neutron capture therapy (BNCT) is a cell-selective particle therapy for cancers. 1) L-4-phenylalanine boronic acid (BPA) is widely used as a boron-containing drug for BNCT. BPA is delivered to the cells by the L-type amino acid transporter (LAT1). Alpha (α) and lithium-7 particles emitted via boron ( 10 B) neutron capture reaction induced by irradiation of thermal (∼0.025 eV) or epithermal (0.5 eV ∼10 keV) neutrons, cause lethal e ect to cancer cells. Boron is required to be accumulate in concentration of approximately 20-40 µg/g boron in a target tumor to damage for cancer cells. 2) Furthermore, the abundance ratio of boron in tumor and normal tissue (T/N ratio) is one of the important parameters to determine the e ect of tumor treatment and damage for normal cells. erefore, it is important to investigate the concentration and distribution of boron compounds in tumor and normal tissue and the time pro le of uptake or excretion of boron compound in tumor tissue. Inductively coupled plasma atomic emission spectrometry (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS) are used in basic research to determine the boron concentration in tumor tissue. 3,4) In recent years, laser ablation-ICP-MS (LA-ICP-MS), 5) secondary ion mass spectrometry (SIMS) 6) and secondary neutral mass spectrometry (SNMS) 7) have been used for boron targeting imaging. However, these techniques have not been widely available yet and cannot obtain the image of biomolecules in tissue which can be interactive to boron compounds, to understand the mechanism of boron compound uptake.
Mass spectrometry imaging (MSI) enables multiple molecule-speci c compound mapping and identifying molecular species in tissue section. 8,9) is technique has been used to visualize the localization of endogenous biomolecules such as protein, peptides, 10,11) lipids, 12) amino acids, 13) fatty acids, 14) neurotransmitters 15,16) and steroid hormones 17) as well as exogenous compounds such as pharmaceutical drugs in various tissues. 18,19) ere are several scanning types of MSI techniques, such as matrix-assisted laser desorption/ ionization (MALDI), desorption electrospray ionization (DESI), 20) tapping mode scanning probe electrospray ionization (t-SPESI), 21) SIMS, 22) and visualization of analyte abundance in micro-dissected tissues determined by liquid chromatography mass spectrometry (LC/MS) measurement. 23) MALDI-MSI has the advantages of high sensitivity and simultaneous detection of multiple organic compounds with 20-200 µm spatial resolution and enables to visualize the distribution of analytes in multiple tissues, for example tumor and normal cell region in a single image. However, biological tissues are complex mixtures of biomolecules and a great number of multiple peaks are detected. In addition, multiple peaks originated from matrix compounds are also detected in the mass range less than m/z 500 in MALDI measurements. Sugiura et al. reported an example of several overlapping peaks within approximately 0.1 Da around m/z 394, which is the m/z value of protonated molecule of a drug, in a liver sample in a MALDI-MSI measurement. 24) erefore, peak separation is necessary to distinguish the target analyte from biomolecules or matrix compounds in biological sample. Tandem mass spectrometry (MS/MS), and use of mass spectrometer with high mass resolving power such as Fourier transform ion cyclotron resonance mass spectrometry (FT-ICRMS) are representative methods to avoid overlapping signals from isobaric compounds in tissues. e amount of administered drug distributed in tissue is usually much smaller than that of biomolecules or matrix compounds. e peaks of biomolecules or matrix compounds adjacent to target analytes not only cause the problem of overlapping, but also make it di cult to detect target analytes due to ionization suppression. MALDI-MSI technique coupled with on tissue derivatization 25,26) is often used to enhance the signals of analyte, which have low ionization e ciencies and also low abundance in tissue. e combination of peak separation by MS/MS or high mass resolution and derivatization technique is applied to detect the administered drug in tissue, however it is sometimes hard to perform due to low peak intensity of product ions in MS/MS measurement or low e ciency of derivatization reaction on tissue section and so on.
In contrast, the e ective concentration level of boron drug in BNCT is comparable to that of lipids in brain sample to which MALDI-MSI is applied without derivatization so that it is likely to be detectable without derivatization. However, the peak separation of target analyte from biomolecules or matrix compound with high mass resolution is needed to detect BNCT drugs because most of BNCT agents are small molecules with a high probability of overlap with isobaric compounds in tissues, for example amino acids, neurotransmitters or matrix compounds. A spiral orbittype time of ight (SpiralTOF) mass spectrometer has a 17 m ight path and the mass resolving power is more than 30,000 (full width at half maximum; FWHM) in a range of m/z 1000-2500. Satoh et al. reported the high mass resolution of SpiralTOF enabled to separate two peaks of phospholipid at m/z 848.556 and m/z 848.645 (0.09 Da di erences) and the mass images obtained from each peak in mass window of 0.1 Da illustrated a di erent localization in tissue, although the mass images obtained from each peak in mass window of 0.2 Da revealed overlapped image. 27) Using SpiralTOF mass spectrometer, we demonstrated an e ective approach of MALDI-MSI to detect BNCT drug, BPA, in a rat brain tissue with a kind of tumor, melanoma treated with BPA without derivatization or MS/MS measurement, aiming to achieve mass resolving power of approximately 20,000 (FWHM) around protonated BPA molecule at m/z 210.
Brain tissue e protocol of the experiments in this study was approved by the Animal Care and Use Committee of Osaka University (approval number: 2019-1-3). An 8 week old, 305 g, male jcl:SD rat transplanted with melanoma (B16F10) in the brain was purchased from CLEA Japan, Inc. (Tokyo, Japan). A BPA solution (10.6 mL, 20 mg/mL) was administered to a rat via the cervical vein for 2 h aiming to 20 µg/g or more of boron to be accumulated in melanoma described in previous study. 3) e brain sample of rat was immersed in isopentane chilled in liquid nitrogen for tissue frozen and stored at −80°C until use.
Tissue sectioning e frozen rat brain sample was sliced into 16 µm thickness using a cryostat microtome (CM1100, Leica Biosystems, Nusslosh, Germany) and mounted on a stainless steel sheet (50 µm thickness, 20 mm ×20 mm, Iwata industry, Tokyo, Japan). e brain section samples were stored at −80°C until use. e sections were brought to 25°C in a desiccator for 30 min before matrix application. Before matrix application, 0.5 µL of TFANa solution (2 mg/ mL) dissolved in ACN and 1 µL of PEG200 and PEG600 mixture solution dissolved in ACN (5 mg/mL) were spotted outside the brain section on the stainless steel sheet for mass calibration, and also 100 pmol and 20 pmol of BPA standard solution was spotted the other place on the stainless steel sheet for the con rmation of detection.

Matrix application
One microliter of 10 mg/mL of matrix DHB solution (85% ACN aqueous solution) was used for MALDI-MS measurement of BPA standard spot on a stainless steel sample plate. A matrix DHB solution dissolved in 85% ACN aqueous solution containing 0.1% TFA (v : v) was used for MSI measurement to detect the protonated BPA molecule in preference to other adduct molecules. e matrix DHB solution (1 mL, 40 mg/mL) was sprayed onto the brain section using an airbrush sprayer with 0.2 mm i.d. nozzle (Tamiya, Tokyo, Japan) and the section was dried for 10 min at 25°C. e brain section on stainless steel sheet was placed on a stainless steel sample plate using a carbon tape.

MALDI-MS data acquisition
MALDI-MS measurement were performed using a MALDI-SpiralTOF/TOF mass spectrometer (JMS-S3000, JEOL Ltd., Akishima, Japan) with a 349 nm solid-phase laser operating at a frequency of 250 Hz. e mass spectrometer was operated with positive polarity in spiral mode and the spectra were acquired in the range of m/z 205-550 to detect the protonated BPA molecule at m/z 210 and the other BPA ion species. Aiming to secure the mass resolving power of 20,000 (FWHM) at m/z 210, the delayed extraction time (200 ns) and the other operating parameters of the mass spectrometer were optimized. An external mass calibration with PEG was performed prior to the sample measurements.
MS/MS experiments by high-energy collision-induced dissociation were performed to con rm the production of BPA-DHB complexes. Isolation of the precursor ion with a width of 1 Da was carried out using a timed ion gate and the product ions in the collision cell were further accelerated with 20 kV and detected through an o set parabolic ion mirror. e collision cell was operated in the presence of a helium gas. e MSI scanning for the brain section was carried out at 60 µm spatial resolution (pixel size). Five spectra were accumulated in each pixel with 625 laser shots. e laser diameter was approximately 20 µm. e laser power and detector voltage were set to 42% and 58% respectively, to detect peaks of BPA ion species with detectable intensity while avoid overlapping with peaks originated from biomolecules or matrix. e regions of interest (ROI) were manually designated in tumor region or cerebral cortex region in the brain section, and the averaged spectra by a pixel were extracted from ROI using msMicroImager Extract so ware (ver. 3.0.0.1, JEOL Ltd., Akishima, Japan). Internal mass calibration using [DHB−H+ 2K] + at m/z 230.946 were performed for each averaged spectrum obtained from ROI because the thickness of the brain section (16 µm) corresponds to the mass di erence of 0.2 mDa in spiral mode. Another MSI measurement was also performed in the mass range m/z 20-215 to investigate the localization of matrix DHB, potassium and sodium in a brain tissue section.
[DHB+ K] + at m/z 192.990 was used for the internal mass calibration. e mass images were extracted from averaged spectra in a mass window of 0.02 Da and visualized using msMicroImager View so ware (ver. 3.0.0.1, JEOL Ltd., Akishima, Japan).

RESULTS AND DISCUSSION
e observed BPA ion species by BPA standard measurements e chemical structures of BPA and matrix DHB are shown in Fig. 1. Boron has a unique isotope pattern, and the calculated protonated BPA molecule at m/z 210.093 ([BPA+ H] + ) and its isotope ions are also shown in Fig.  1. Before MSI measurement, we performed MALDI-MS measurement of BPA standard spot with matrix DHB to con rm the detection of BPA ion species. Janda et al. reported that several analyte-DHB complexes, phosphocholine, leucine, folic acid, etc. were found in the mass spectra obtained from multiple MSI measurements. 28) Table 2. Supplementary Fig. 1  We also examined the detection of BPA standard with matrix CHCA. e intensity of [BPA+ H] + with CHCA was approximately 1/5 of that with DHB, although there was no multiple ion formation, such as BPA-CHCA complexes. erefore, matrix DHB was adopted for BPA mass imaging in brain section. Figure 3 shows the brain tissue section with melanoma administered with BPA and the location of ROI designated as a tumor region. MSI data acquisitions were performed for the half area of brain section with melanoma. Figure  4 shows the averaged mass spectrum by a pixel extracted from the region of tumor. Also, the enlarged mass spectrum around m/z 210.093 ([BPA+ H] + ) and the mass images showed the distribution to the whole area of the brain section and it could be assigned to an endogenous biomolecule. To con rm the peak assignment of [BPA+ H] + in the sample, the BPA standard was spotted on the tumor or cerebral cortex a er MSI measurement and then additional MSI data acquisitions were carried out around the BPA spot area in the tumor or cerebral cortex region. Figure 5 shows the mass spectra before/a er the BPA standard spotting on the tumor region (10 pmol) or cerebral cortex region (50 pmol). e relative intensity of m/z 210.085 to m/z 209.915 increased in the averaged mass spectrum obtained from the BPA-spotted tumor region (Fig. 5B). e increase in the relative peak intensity at m/z 210.085 to m/z 209.914 was also observed in the averaged mass spectrum obtained from the BPA-spotted cerebral cortex region (Fig. 5D). ese results suggested that the peak at m/z 210.084 was assigned to [BPA+ H] + in the sample. e mass di erence (−9 mDa o ) from the calculated value was due to the slight overlap with adjacent peaks, which were originated from biomolecules or matrix DHB. e mass resolution of 20,000 (FWHM) around m/z 210 was e ective to separate peaks around [BPA+ H] + in the sample and to identify the peak of [BPA+ H] + . erefore, it is highly possible to obtain mass image less overlapping with biomolecules or matrix DHB. However, the di erence in color between tumor region and cerebral cortex region in the mass image obtained from [BPA+ H] + should be carefully considered because the color contrast in mass image is based on the peak intensity in each pixel which is probably a ected by the ionization e ciency of BPA in tumor or cerebral cortex. Actually, the BPA standard spotting   were also detected. e assignments of these peaks were con rmed by BPA-spotting experiment. e supplementary Fig. 2 shows the averaged mass spectra by a pixel around [BPA+ DHB−2H 2 O+ K] + in the BPA-spotting experiment. e observed m/z values of BPA species detected in the brain tissue administered with BPA and the observed m/z values of BPA-spotting experiments were less than 9 mDa (mass error <43 ppm) as shown in  Fig. 6(A). And the mass images obtained from dehydrated DHB ion species detected in the same mass range were also shown in Fig. 6(A). e mass image of [2DHB− H 2 O+ K] + seemed like similar localization in the tumor re-  We also carried out another MSI measurement in the low mass range at m/z 20-215 using another brain section administered with BPA to investigate the distribution of matrix DHB, sodium ion and potassium ion in the brain section. e mass images obtained from [BPA+ H] + , [DHB+ H] + , sodium ion and potassium ion were shown in Fig. 6

CONCLUSION
We developed an e ective method of MALDI-MSI with high mass resolution targeting to [BPA+ H] + for BNCT drug, BPA in a brain tumor model rat.
e peak of [BPA+ H] + at m/z 210.084 was detected separate from other peaks originated from biomolecules or matrix DHB in the brain sample administered with BPA by achieving mass resolution of approximately 20,000 (FWHM) at m/z 210. e mass image obtained from [BPA+ H] + revealed the distribution in the tumor region in the brain tissue section and also likely showed the localization inside the tumor. e mass image obtained from BPA ion species including [BPA−H 2 O+ Na] + , [BPA+ DHB−2H 2 O+ Na] + and [BPA+ DHB−2H 2 O+ K] + illustrated the distribution in tumor region in the brain tissue section and they could be used to validate the mass image obtained from [BPA+ H] + or estimatie the time pro le of BPA uptake or excretion. is MSI method has a great potential to evaluate the e ect of therapy and damage to normal tissue with near-cellular spatial resolution in BNCT. e color contrast in tumor against cerebral cortex in mass image was probably a ected by the di erence of BPA ionization e ciency in each tissue. Takeo and Shimma 29) reported quantitative mass spectrometry imaging (qMSI) combined with the quantitation using LC-MS/MS. In our next investigation, a method to correct the contrast in mass image including qMSI technique will be developed in order to estimate the T/N ratio in a brain tissue.