Ambient Mass Spectrometry Imaging with ... - ACS Publications

Article pubs.acs.org/ac

Ambient Mass Spectrometry Imaging with Picosecond Infrared Laser Ablation Electrospray Ionization (PIR-LAESI) Jing Zou,† Francis Talbot,‡ Alessandra Tata,§,⊥ Leonardo Ermini,∥ Kresimir Franjic,† Manuela Ventura,⊥ Jinzi Zheng,⊥ Howard Ginsberg,#,∇ Martin Post,∥,◆ Demian R. Ifa,§ David Jaffray,⊥,○ R. J. Dwayne Miller,*,†,‡,¶ and Arash Zarrine-Afsar*,⊥,#,∇,○ †

Department of Physics, University of Toronto, 60 St George Street, Toronto, Ontario M5S 1A7, Canada Department of Chemistry, University of Toronto, 80 St George Street, Toronto, Ontario M5S 3H6, Canada § Department of Chemistry, York University, 4700 Keele Street, Toronto, Ontario M3J 1P3, Canada ∥ Program in Physiology and Experimental Medicine, Peter Gilgan Centre for Research and Learning, The Hospital for Sick Children, Toronto, Ontario M5G 0A4, Canada ⊥ Techna Institute for the Advancement of Technology for Health, University Health Network, Toronto, Ontario M5G 1P5, Canada # Department of Surgery, University of Toronto, 149 College Street, Toronto, Ontario M5T 1P5, Canada ∇ Keenan Research Centre for Biomedical Science, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8, Canada ○ Department of Medical Biophysics, University of Toronto, 101 College Street Suite 15-701, Toronto, Ontario M5G 1L7, Canada ◆ Institute of Medical Science, University of Toronto, Toronto, Ontario M5S 1A8, Canada ¶ Max Planck Institute for the Structure and Dynamics of Matter, Luruper Chaussee 149, 22761 Hamburg, Germany ‡

S Supporting Information *

ABSTRACT: A picosecond infrared laser (PIRL) is capable of cutting through biological tissues in the absence of significant thermal damage. As such, PIRL is a standalone surgical scalpel with the added bonus of minimal postoperative scar tissue formation. In this work, a tandem of PIRL ablation with electrospray ionization (PIR-LAESI) mass spectrometry is demonstrated and characterized for tissue molecular imaging, with a limit of detection in the range of 100 nM for reserpine or better than 5 nM for verapamil in aqueous solution. We characterized PIRL crater size using agar films containing Rhodamine. PIR-LAESI offers a 20−30 μm vertical resolution (∼3 μm removal per pulse) and a lateral resolution of ∼100 μm. We were able to detect 25 fmol of Rhodamine in agar ablation experiments. PIR-LAESI was used to map the distribution of endogenous methoxykaempferol glucoronide in zebra plant (Aphelandra squarrosa) leaves producing a localization map that is corroborated by the literature. PIR-LAESI was further used to image the distribution inside mouse kidneys of gadoteridol, an exogenous magnetic resonance contrast agent intravenously injected. Parallel mass spectrometry imaging (MSI) using desorption electrospray ionization (DESI) and matrix assisted laser desorption ionization (MALDI) were performed to corroborate PIR-LAESI images of the exogenous agent. We further show that PIR-LAESI is capable of desorption ionization of proteins as well as phospholipids. This comparative study illustrates that PIR-LAESI is an ion source for ambient mass spectrometry applications. As such, a future PIRL scalpel combined with secondary ionization such as ESI and mass spectrometry has the potential to provide molecular feedback to guide PIRL surgery. aser ablation using “nanosecond” infrared (IR) systems have been shown to offer desorption of biological tissues for the purpose of subsequent ionization and molecular imaging with ambient mass spectrometry.1,2 It has recently been shown that “picosecond” infrared lasers (PIRLs) offer efficient ablation of biological materials.3−7The origin of this efficiency lies within the rapid conversion into ablation of the impulsive energy deposited selectively into the OH vibration stretch mode of water within the tissue in the 3 μm region of the

L

© 2015 American Chemical Society

electromagnetic spectrum. In order to drive molecules into the gas phase, or to give rise to ablation, the absorbed laser energy must ultimately be transduced into translation motions. The PIRL ablation mechanism exploits the effectively direct Received: July 22, 2015 Accepted: November 11, 2015 Published: November 11, 2015 12071

DOI: 10.1021/acs.analchem.5b02756 Anal. Chem. 2015, 87, 12071−12079

Article

Analytical Chemistry

source for small molecule mass spectrometry imaging with the added bonus of minimal damage to ex vivo tissues being analyzed. A comparison between PIR-LAESI-MSI molecular maps and those of the well-established MSI techniques such as MALDI and DESI is also presented. A molecular imaging study comparing in detail the quality and the chemical sensitivity of PIR-LAESI to those of conventional LAESI is beyond the scope of our work here. However, we present preliminary characterization of limit of detection, the plume dynamics, and lateral and vertical resolution of PIR-LAESI using Zebra Plant (Aphelandra squarrosa) that has been studied in detail using conventional LAESI.2,18 We further show that PIR-LAESI is capable of ionizing proteins as well as phospholipids, opening up the possibility of performing in situ pathology of tumors.19,20

coupling of the OH stretch mode through the hydrogen bond network of water into the O−O translation, the very motions needed to drive molecules into the gas phase. PIRL ablation occurs faster than any loss mechanism to this transduction such as thermal diffusion or acoustic propagation outside the excited zone. Most important, the whole process occurs faster than nucleation growth that may lead to cavitation and mechanical damage to surrounding tissues.7,8 Since the bulk of the absorbed energy above the ablation threshold goes directly into the ablation process, PIRL surgery does not lead to extensive thermal tissue damage outside the ablated zone or postoperative scar tissue formation.5−7 Furthermore, ablation of biological material by PIRL has been shown to result in no loss of biological activity, as evidenced by recent works suggesting that native enzymatic activity and viral infectivity are retained in the plume of PIRL ablated material.9,10 While the studies above provide convincing evidence to the suitability of PIRL as the next generation laser scalpel for scar free laser surgery, they do not evaluate the utility of PIRL as a desorption source for mass spectrometry. Recently, near real time capture and analysis of the plume of electrocautery by mass spectrometry has offered the tantalizing prospect of surgical guidance based on molecular information.11 Here, a “gentle” scalpel that offers, unlike electrocautery, removal of biological material without thermal damage, yet releases tissue fingerprints intact to provide guidance for surgery, is currently missing and must be developed. However, this will not materialize without coupling PIRL to an ionization source for direct mass spectrometry analysis of tissues. To this end, in the work described here we demonstrate that PIRL ablation coupled with electrospray ionization (ESI) using a laboratorybuilt interface similar to what is reported for conventional LAESI1,2is capable of delivering molecular maps of select metabolites in biological tissues. Here, we further showcase the utility of PIR-LAESI in mapping the distribution of a magnetic resonance imaging12 (MRI) contrast agent in biological tissues. The ability to verify the distribution of contrast agents within biological tissues is key to successful development of novel agents. To this end, elemental imaging of tissue-borne contrast agents by inductively coupled plasma mass spectrometry (ICPMS) has also been reported, suggesting a spatial distribution pattern that closely matches the MRI results.13,14Gadolinium-based contrast agents, routinely used in medical applications, have been subjected to matrix assisted laser desorption ionization mass spectrometry imaging (MALDI-MSI)15,16under vacuum, highlighting the utility of MALDI-MSI in the detection of MRI contrast agents within the tissue as a verification tool that complements in vivo MRI results. In the quest to facilitate the detection of contrast agents under ambient conditions while preserving the native molecular information such as metabolism, aggregation, and binding, etc., gentle laser-based mass spectrometry imaging technologies are particularly useful. In MALDI-MSI the analysis is done under vacuum and with the application of a matrix material that enhances the ionization. In addition to matrix deposition artifacts that could lead to false localization of tissue molecules,17 the requirement for vacuum further limits in vivo applications of MALDI for surgical guidance. Many ambient laser-based ion sources, on the other hand, do not require matrix compounds also alleviating the need for vacuum. While a complete comparison of the chemical sensitivity of PIR-LAESI to conventional LAESI using nanosecond mid-IR lasers is to be undertaken, our study here validates PIR-LAESI as an ion

■

EXPERIMENTAL SECTION Materials. Methanol (MeOH), ultrapure water (H2O) both HPLC-MS grade, agar (in the form of Molecular Biology grade agarose), and Rhodamine 640 were purchased from SigmaAldrich (Oakville, ON). L-α-Phosphatidylcholines (PC) from chicken eggs (average molecular weight ∼770) were purchased from Avanti Polar Lipids (Alabaster, AL). An amount of 10 μg of PC was diluted in chloroform, and 100 μL of the resulting mixture was then dissolved in 900 μL (1:1) MeOH/H2O solution with a approximate final concentration of 10−5 M. Cytochrome C was obtained through Sigma-Aldrich (Oakville, ON, Canada) and diluted in water to a final concentration of 10−4 M. Deionized water (18 MΩ) was used for all solvent mixtures in PIR-LAESI-MS experiments. Zebra plant was purchased from a local store. Limit of Detection Experiment. Aqueous solutions of Reserpine (Sigma-Aldrich (Oakville, ON, Canada)) in the concentration range 100 nM to 50 μM shown in Figure 2 were prepared and placed in a laboratory built reservoir connected to a liquid sample holder to constantly replenish the ablated liquid from the sample holder. Laser pulses at 100 Hz were sent to the liquid sample for a period of 1 min, and MS data were accumulated during this time resulting in an ion abundance value for reserpine m/z 609.2. At each reserpine concentration, 5 independent acquisitions of ion abundance were performed to form an average. Verapamil was purchased from SigmaAldrich as well. Aqueous solutions from 5 nM to 10 μM were prepared. Laser pulses at 100 Hz were sent to the liquid sample for a period of 1 min, and MS data were accumulated during this time resulting in an ion abundance value for verapamil m/z 455.6. Characterization of the PIRL Ablation Plume of Biological Tissues by Dark Field Imaging. To study the PIRL ablation plume dynamics we used a home-built timeresolved dark field imaging microscope (DFIM) as described.5 Here, the seed pulse from a passively mode-locked Nd:YLF laser with an energy of 1 nJ/pulse, and pulse duration of 100 ps (repetition rate ∼80 MHz), was split into two beams by a 50:50 beam splitter and the first branch amplified by a regenerative amplifier. The output of the amplifier (1.7 mJ/pulse) served as the pump beam for an optical parametric amplifier (OPA), seeded with a CW diode laser (DFB) operating at the wavelength of 1660 nm. The resulting OPA output in the midIR range, centered at 2880 nm, had an energy of 150 μJ/pulse. This mid-IR beam was focused on the sample by a CaF2 lens (f = 100 mm) (L4, in Figure 4) to a spot with a 1/e2 diameter of 100 μm, resulting in a maximum fluence of 1.9 mJ/cm2, that 12072


Article


Figure 1. Characterization of the PIRL ablation crater with optical imaging and PIR-LAESI mass spectrometry imaging (2% agar). (A) Crater from PIRL ablation on 2% agar visualized through dark field imaging (Figure 4B illustrates the dark field imaging setup used for this visualization). Here, three bursts of 10 PIRL pulses (fluence of 1.9 J/cm2 with the energy of 150 μJ/pulse) were delivered to a 2% agar gel. The crater created through the initial burst (i.e., first 10 pulses irradiating the sample) was not visible yet it produced MS signal (see panel B). (B) Abundance of Rhodamine 640 (m/z 491.3) in a 100 μm thick agar layer (see part D for the spectrum of the ionized Rhodamine m/z 491.3) as a function of PIRL pulse bursts (10 pulses per burst) using a home-built PIR-LAESI system depicted in part C. Panel C also defines LAESI source geometry used in our experiments. The lateral resolution (i.e., the laser focus spot) in the agar experiments is 100 μm. In part B the 2% agar substrate contained 100 μM Rhodamine 640 allowing the abundance of the post ionized Rhodamine m/z 491.3 to be used to confirm depth profiling shown in part A. The vertical resolution for MS analysis is 30 μm for agar-based samples (spectrum D).

could be tuned down and modulated by a half-wave plate/ polarizer combination on the pump of the OPA. The other branch of the seed pulse from the mode-locked Nd:YLF laser was sent to another regenerative amplifier, with the output energy of 1.5 mJ/pulse that was subsequently converted to its second harmonic (SHG) at 532 nm by a 10 mm long lithium triborate (LBO) crystal. For improvement of the dark field image quality, this beam (532 nm) was coupled to a 1.5 m long multimode fiber, eliminating speckles from the coherent illumination source. The output of the fiber was focused onto a diffuser (Thorlabs, DG10-1500) by a crown glass (BK7) lens ( f = 40 mm) serving as the illumination beam for DFIM. A metal rod with the diameter of 2 mm was placed in front of a lens group composed of two crown glass BK7 lenses (L2, L3, Figure 4) to block the central part of the illumination beam. The edges of the illumination beam were thus focused onto the plume to create scattered light illumination. The scattered light was collected by a microscope objective lens (Infinity PhotoOptical, Model K2, Objective CF4) mounted in front of a Nikon D3200 digital camera. The time delay between ablation and imaging events was controlled by a delay generator triggered by the leakage of the passively mode-locked Nd:YLF oscillator. Each picture was taken for a single shot ablation event by setting the delay generator to single pulse mode. Tissue Samples. All animal studies were conducted in accordance with institutional guidelines and approved by the Animal Use Committee. Severe combined immunodeficient (SCID) mice received iv injections of gadoteridol (∼20−100 μL/25 g). At 5 min post injection, the mice were sacrificed with an overdose of isoflurane and subjected to surgical removal of

the kidneys. Excised tissues were frozen using liquid N2 vapor and stored at −80 °C prior to sectioning (see below). PIR-LAESI-MSI Experiments. The PIRL laser (mid-IR pulses at λ = 2880 nm with an energy of 150 μJ/pulse, a pulse duration of ∼80 ps, and a repetition rate of 100 Hz) was used for all characterizations. The mass spectrometer used for this work was an orthogonal acceleration time-of-flight mass spectrometer (Q-TOF), a prototype of the QSTAR Elite Series (AB Sciex). The built-in nanospray ion source was replaced by a custom-made ion source combining electrospray ionization (ESI) and laser ablation under atmospheric pressure. The mid-IR beam was focused by a f = 100 mm CaF2 lens (Thorlabs) to 100 μm at focus in front of the MS on a sample plate holder, resulting in a fluence value of 1.9 J/cm2. The lateral resolution in the tissue mapping experiments is limited by this laser focus size. Note that the OPA beam has a beam quality parameter M2 ∼ 1.0, which allows us to easily focus its spot to 2 and cannot be focused to such small sizes. The distance between the laser focus and the entrance of the mass spectrometer (xs‑ms) was 5−8 mm (Figure 1C). The distance between the ESI needle and the MS (xesi‑ms) was 10− 13 mm (Figure 1C), and the distance between the sample holder and the ESI needle (ds‑esi) was 4 mm vertically. The emitter of the electrospray was made up of a stainless steel blunt tube with the inner diameter (i.d.) of 125 μm. The proximal (to the MS) end of the emitter tube was tapered down to be cone-shaped, having an outer diameter (o.d.) of 150−200 μm at the tip. 12073


Article


The latter was used to process the mass spectral data and to generate 2D molecular maps. MALDI-MSI Experiments. The kidney tissue block was mounted onto the specimen disc of a cryostat (Leica Microsystems) using a small amount of optimal cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA) on the distal side of the sample as a support. As such, the front side of the sample, exposing the area to be sectioned, was completely free of OCT. The tissue sections were prepared at a thickness of 12 μm at −15 °C and mounted onto indium tin oxide (ITO)-coated glass slides. A thin matrix layer was applied to the tissue section using an automated MALDI plate matrix deposition system (TM-Sprayer, Leap Technologies, Carboro, NC). A total of 5 mL of 9-aminoacridine (9AA; 15 mg/mL in methanol) was sprayed per slide over 4 passes at 80 °C with a velocity of 400 mm/min and a line spacing of 3 mm. A time-of-flight tandem mass spectrometer (AB SCIEX TOF/TOFTM 5800 System, AB SCIEX, Ontario) was used to acquire the images. MALDI mass spectra were obtained using the third harmonic of a Nd:YAG laser (355 nm) with 3 ns pulse duration and 400 Hz repetition rate. The data were acquired in the negative- and positive-ion reflector modes using an external calibration method. The standard gadoteridol (Prohance, Bracco Imaging) was deposited on the ITO-coated slides to recalibrate and minimize the mass shift. In the imaging experiment, a total of 200 laser shots per pixel were delivered (1 s/pixel) with the step size (i.e., the interval between the data points) of 75 μm. The mass data were processed using a specialized script of the Analyst software (AB SCIEX) at a mass resolution of 0.1 amu, and the images were visualized using TissueView (AB SCIEX). Histology Analysis. For each tissue slice subjected to mass spectrometry imaging, a consecutive 5 μm slice was taken for the hematoxylin and eosin (H&E) staining. After MALDI imaging, the matrix was removed with cold methanol, and the tissue was fixed prior to standard H&E staining. We used standard protocols for the staining.

A mixture of H2O/MeOH (1:1, v/v) with 1% acetic acid served as the solvent for ESI. During the experiment, a flow rate of 1.5−2.0 μL/min was applied to the solvent with the potential on the emitter tube ranging from 2.5 to 4.0 kV depending on the flow condition to form a steady Taylor cone spray mode (verified optically through a mounted camera). A stainless steel sample plate was mounted on a 3D translation stage (with the X and Y axes motorized) equipped with a thermo electric cooler (TEC) from Thorlabs (T = −2 °C). To further assist ion collection by the mass spectrometer, a potential of 400 V was applied to the sample plate. The tissue slice 20 μm thin (from cryotome) was mounted directly on the metal plate. The generation of 2D maps was performed by synchronizing the translation of the sample with firing of the laser, through a home-built data acquisition software using LabVIEW 8.0. After the movement of the stage to expose a new spot on the sample surface (laser ablation pixel) was completed, the laser was fired by sending a trigger signal to the electronics controlling the amplifier. The acquisition software would then generate a scan file, recording the start time, and X and Y positions for each pixel. The scan ranges in X and Y, the laser repetition rate, the number of pulses per pixel, as well as the dwell time per pixel are all controlled by the acquisition software. In our experiments, 10−20 pulses at a repetition rate of 100 Hz were sent for ablation, followed by ∼3 s dwell time per pixel for the ESI process of the plume to be completed, and for the ions to be collected by the MS. The total recording time per pixel was 3.2 s. With a scan range of 5 × 5 mm2 and a step size of 200 μm between pixels, the total pixel count of an image was 676 corresponding to a total scan time of ∼36 min. These parameters were used to synchronize the scan time parameters with the chromatogram data from the mass spectrometer. After synchronization, each burst of ions observed in the Total Ion Count (TIC) and or extracted ion count (XIC) can uniquely be associated with the corresponding pixel, thereby allowing the construction of the molecular map. We further used the characteristic, laser dependent shape of the extracted ion chromatogram for the molecule of interest that illustrates areas of laser irradiation (see Figure S5, for gadoteridol adduct m/z 582.1) to verify the synchronization and to assess the degree of signal carry over between consecutive pixels. DESI-MSI Experiments. Tissues were sectioned using a cryotome (CM 1950, Leica) with a thickness of 20 μm. The tissue sections were mounted onto glass slides and placed on a laboratory-built prototype using tape and analyzed by DESIMS. All MS experiments were performed using a Thermo Fisher Scientific LTQ mass spectrometer (San Jose, CA) controlled by XCalibur 2.0 software (Thermo Fisher Scientific). The MS parameters and the DESI collection geometry were adjusted using the first slice, and DESI-MS imaging was performed using the second tissue slice without altering the collection geometry and collection parameters. An H2O/ MeOH (1:1) solution was used as the spray solvent at a flow rate of 1.5 μL/min. The tissues were scanned using a 2D moving stage with the spatial resolution of 150 μm (i.e., the step size between two consecutive DESI scan lines). The lines were scanned at a constant velocity in the range 248−414 μm/ s, and the scan time of the instrument was in the range 0.43− 0.56 s. Data were acquired in the positive ion mode over the mass range from m/z 500 to 900. MS spectra were processed using Qual Browser Xcalibur. ImageCreator version 3.0 was used to convert the Xcalibur 2.0 mass spectra files (.raw) into a format compatible with BioMap (http://www.maldi-msi.org/).

■

RESULTS AND DISCUSSION Motivated by the success of conventional LAESI-MS using nanosecond lasers (referred to as LAESI-MS herein) and the ability of its picosecond counterpart, PIRL, to “gently” desorb biological materials without thermal damage, we sought to develop a new ionization source for the analysis of biological samples under ambient conditions. Figure 1 illustrates the ablative characteristics of PIRL using a 2% agar gel containing Rhodamine 640 as a reporter molecule for mass spectrometry. As in LAESI-MS,1,2 PIRL ablation is used to produce a plume directly from the target material. This plume is subsequently intercepted by electrospray to generate ions (Figure 1C). The depth of the laser crater produced in 2% agar visualized through optical dark field imaging grew approximately linearly with the number of pulses irradiating the sample, suggesting a ∼3−4 μm removal of material per pulse at laser fluence of 1.9 J/cm2 and the energy of 150 μJ/pulse (Figure 1A). The focus of the laser was 100 μm for this experiment. To ionize the molecules, the ablation plume was intercepted by the electrospray solvent, forming multiply charged ions captured by the analyzer (Figure 1C). PIR-LAESI requires a minimum of 10 pulses to produce a stable MS signal (Figure 1B). Therefore, from the abundance of the Rhodamine 640 molecule post ionized with ESI (m/z 491.3), PIR-LAESI has a vertical resolution of 30 μm. In other words, a single laser burst containing 10 pulses produces a 12074


Article


ionizing proteins as exemplified by successful ionization of cytochrome C, a ∼12 kDa protein which is multiply ionized (9+ to 17+ states) (Figure 3B). Note that in the absence of an ESI source for ionization we were not able to detect cytochrome C ions. This suggests that the ions produced during PIR-LAESI are largely formed during the ESI process. To further characterize PIR-LAESI ion source in the imaging mode we first used dark field imaging to study the effect of PIRL ablation on biological material. Dark field imaging of the ablation region from zebra plant leaf using single pulse irradiation gives rise to a highly directed plume of ablated material ejected from the surface of the sample (Figure 4A). The PIRL driven plume, imaged over various time points of 100 ns to 10 μs post ablation, shows the propagation and the rapid expansion of the plume material. Figure 4B illustrates the schematics of the dark field imaging setup used for visualizing the ablation plume from zebra plant leaves in Figure 4A and for the agar ablation plume shown in Figure 1A. Much like the conventional LAESI-MS imaging that uses a pulsed nanosecond midinfrared laser, PIR-LAESI was able to reproduce a molecular map for methoxykaempferol glucoronide (m/z 493.1) in zebra plant (Figure 4C). The distribution of the abundance of m/z 493.1 (heat map, Figure 4C) correlates well with the optical image of the plant leaf (gray scale, background, Figure 4C) indicating the location of the white stripes containing methoxykaempferol glucoronide. While a detailed comparative study between LAESI and PIR-LAESI in terms of ablation efficiency, dependence on tissue water content, and chemical sensitivity is being undertaken, we have been able to achieve a lateral resolution of 100 μm that is 3-fold smaller than what has been reported for LAESI.21 This improvement in lateral resolution is made possible by the better beam quality of picosecond sources that allow tighter focusing of the laser spot. Note that, with the application of fiber delivery of the IR source, the lateral resolution in LAESI down to 30 μm has been reported.25 The proposed PIR-LAESI ion source was further validated for animal tissue analysis using kidneys from mice injected with the contrast agent gadoteridol, as proxy for exogenous small molecules that could be tracked inside an organ and secreted intact via kidney filtration. This test enabled us to determine the localization of the contrast agent within the kidney with PIR-LAESI-MS imaging, and to correlate the resulting molecular maps to those from MALDI-MS and DESI-MS, the two leading mass spectrometry methods widely used in tissue imaging. Figure 5A shows the PIR-LAESI-MS spectrum of the medulla region of a kidney injected with gadoteridol. The ions of [gadoteridol + H]+ of m/z 560.1, [2gadoteridol + Na + H]2+ of m/z 571.1, [gadoteridol + Na]+ of m/z 582.1, [2gadoteridol + Na + K]2+ of m/z 590.1, and [gadoteridol + K]+ of m/z 598.1 can be seen. In the cortex region (Figure 5B), a much lower ion count for [gadoteridol + Na]+ of m/z 582.1 was observed (2-fold reduction in the ion count compared to medulla). The distribution of the contrast agent in the kidney was mapped using a home-built 2D translation stage synchronized to laser ablation as described in the Experimental Section. PIR-LAESI-MS images of [gadoteridol + H]+ of m/z 560.1 and [gadoteridol + Na]+of m/z 582.1 are reported in Figure 5C. Both adducts, in addition to having the highest accumulation in the medulla, are distributed in the whole tissue, delineating the entire shape of the kidney. The H&E image of a consecutive slice between the tissue sections used for PIR-

sufficiently large ion abundance that could be detected easily with the mass analyzer (Figure 1B). Three bursts containing 10 pulses each completely removed 100 μm of agar material. The ion abundance, however, decreased over the subsequent bursts, as also reported for conventional LAESI21 (Figure 1B). One laser burst (10 pulses) in this experiment contained 25 fmol of Rhodamine, clearly detectable by our mass analyzer (Figure 1D). To formally determine a limit of detection (LOD) for PIRLAESI-MS we used aqueous solutions of reserpine, routinely used in LOD experiments with mass spectrometry. Figure 2

Figure 2. Limit of detection (LOD) of PIR-LAESI-MS using aqueous solutions of reserpine and verapamil. A limit of detection of 100 nM with good linearity (R2 = 0.98) is obtained for reserpine. The error bars indicate one standard deviation around the average of 5 measurements of the ion abundance. At the limit of detection, we had a signal (m/z 609) to noise ratio of close to 5-fold (see inset). The LOD estimate for verapamil (m/z 455) is better than 5 nM (signal-tonoise ratio of close to 50).

suggests a limit of detection of 100 nM which constitutes a 5fold improvement over what is reported for LAESI-MS using the same compound.18 The LOD for verapamil with our PIRLAESI setup is better than 5 nM. This value significantly improves the previously reported value of 240 nM for conventional LAESI.18 With the addition of plume collimation a limit of detection of close to 1 nM is reported for nanosecond LAESI.22 PIR-LAESI, thus, allows a limit of detection in the low nanomolar range without the use of plume collimation. Many studies have suggested a utility for lipid profiling to classify different types of biological tissues including tumors with mass spectrometry.19,23 To this end, we show that PIRLAESI is capable of desorbing and ionizing phospholipids as shown in Figure 3A. Phospholipids extracted from chicken egg were subjected to PIR-LAESI-MS. We were able to resolve the dominant phospholipids of the egg yolk. Namely, sodiated, potassiated, and protonated adducts of PC (34:2) and PC (34:1) as well as protonated PC (38:5) and PC (38:4) and sodiated PC (36:1) were seen. These phospholipids are in agreement with a matrix assisted laser desorption ionization mass spectrometry (MALDI-MS) study of the phospholipids that constitute egg yolk.19,23,24 Further to lipid analysis, in the mass spectrometric study of biological materials peptide and protein fingerprinting is also considered to be a powerful tool.20 Here, we show that PIR-LAESI is capable of desorbing and 12075


Article


Figure 3. Characterization of phospholipids and proteins by PIR-LAESI mass spectrometry. (A) A PIR-LAESI-MS spectrum of the phospholipids extracted from a chicken egg. (B) A PIR-LAESI-MS spectrum of the cytochrome C protein indicating multiply charged species.

MS signal between PIRL pulses, further indicating that in the absence of laser ablation the ESI solvent spray does not produce any parasitic ions from the ablation plume carry over. For comparison, a slice of the same mouse kidney consecutive to the one analyzed by PIR-LAESI-MSI was subjected to MALDI-MSI, a well-established laser-based MS technique to corroborate PIR-LAESI images. MALDI-MSI experiments were run both in the positive and in the negative ion modes. Unfortunately, in the positive ion mode gadoteridol

LAESI-MS and DESI-MS imaging is also given in Figure 5C. The presence of gadoteridol in the entire kidney tissue from PIR-LAESI-MS maps is consistent with the results of in vivo MRI of the same kidney (Figure S1). Figure S5 shows the ion abundance of [gadoteridol + Na]+ of m/z 582.1 during the entire course of the imaging experiment. The inset in this figure shows the laser dependence of the [gadoteridol + Na]+ ion abundance. The ion intensity effectively diminishes in between laser pulses. This suggests that there is little to no carry over of 12076


Article


Figure 4. Characterization of the PIRL ablation plume with dark field imaging and with PIR-LAESI mass spectrometry imaging (zebra plant). (A) Dark field imaging of the ablated zebra plant leaf (single PIRL pulse) at various time points after the ablation pulse. (B) Schematics of the dark field imaging used. Here, BS is a 50% beam splitter for 1053 nm, M1 is a flat mirror with a high reflective coating at 1053 nm, HWP is a half wave plate for 1053 nm, M2 is a flat mirror with high reflective coating at 2910 nm, L1−3 are BK7 lenses with focal length values of L1 (f = 40 mm), L2 ( f = 25 mm), and L3 ( f = 40 mm), L4 is a CaF2 lens with f = 100 mm, OPA is an optical parametric amplifier, SHG is second harmonic generator, DFB is a distributed feedback diode laser (CW, fiber coupled). (C) The molecular map of methoxykaempferol glucoronide (m/z 493.1)21in a zebra plant leaf using PIR-LAESI (top), 100 μm laser spot size, with a step size (i.e., center to center distance between the two consecutive laser craters) of 200 μm. The molecular map of the same molecule with a replica LAESI setup using a nanosecond optical parametric oscillator (OPO) laser (bottom) at the fluence of 2 J/cm2 (300 μm laser spot size, with a step size of 400 μm). Each pixel received 20 laser pulses, with a repetition rate of 10 Hz. With PIRLAESI we were able to achieve a lateral resolution 3 times higher compared to LAESI with an OPO source at comparable laser fluence. Samples were ablated at −2.0 °C placed on a thermo electric cooler (TEC) as detailed in the experimental methods section.

K]+, and [2gadoteridol + Na + K]2+ of m/z 582.1, 598.1, and 590.1, respectively, when subjected to DESI-MS. We used a kidney slice consecutive to the one analyzed in Figure 5 with PIR-LAESI. Figure S3A,B shows the averaged spectrum at an arbitrary point in the medulla and in the cortex, respectively. We further mapped gadoteridol distribution in the kidney by DESI-MSI (Figure S3C). Adduct [gadoteridol + Na]+ is accumulated in the medulla. Adduct [gadoteridol + K]+ is highly accumulated in the medulla and is ubiquitously distributed in the rest of the tissue, delineating the entire shape of the kidney slice. Both ambient mass spectrometry ionization techniques DESI and PIR-LAESI were found to have utility for in situ analysis of

was not efficiently ionized. As such, we used the negative ion mode results for our comparative molecular imaging. The MALDI-MS spectrum of gadoteridol obtained directly from the kidney shows the prominent ionic species [gadoteridol − H]− with the characteristic gadolinium isotopic pattern and the monoisotopic peak of m/z 558.1. Figure S2A,B shows, respectively, the averaged spectra in the medulla and at an arbitrary point in the kidney cortex sampled by MALDI imaging. In our case, MALDI-MSI did not show the presence of gadoteridol in the entire kidney tissue as revealed by PIRLAESI analysis in Figure 5. Gadoteridol, which accumulated in the kidney, appears as three cationized species [gadoteridol + Na]+, [gadoteridol + 12077


Article


Figure 5. PIR-LAESI-MS imaging of gadoteridol in mouse kidney in the positive ion mode. (A) PIR-LAESI-MS averaged spectrum of the medulla region showing [gadoteridol + H]+ of m/z 560.1, [2gadoteridol + Na + Na]2+ of m/z 571.1, [gadoteridol + Na]+ of m/z 582.1, [2gadoteridol + Na + K]2+ of m/z 590.1, and [gadoteridol + K]+ of m/z 598.1. (B) PIR-LAESI averaged spectrum of the cortex region with lower total gadoteridol ion count compared to the medulla region. In both spectra ion counts relative to [gadoteridol + Na]+ of m/z 582.1 are given. The total [gadoteridol + Na]+ of m/z 582.1 ion count in medulla and the cortex were 3.7 × 101 and 1.7 × 101, respectively. (C) PIR-LAESI-MS images of [gadoteridol + H]+ of m/z 560.1 and [gadoteridol + Na]+ of m/z 582.1. Here, the laser spot size is 100 μm, and the pixel step size (i.e., center-to-center distance between two adjacent laser spots) is 200 μm. There were 20 pulses sent to each pixel, and the sample thickness was 20 μm. As illustrated in the Supporting Information (Figure S4), the vertical resolution in this particular experiment was better than 20 μm as the sample was ablated frozen (at −2 °C), which avoided loss of water from the tissue. The dashed lines delineate the kidney and the medulla boundaries and are presented to guide the eye.

PIR-LAESI constitutes a 5-fold improvement over nanosecond LAESI. Furthermore, PIR-LAESI offers a 3-fold improvement in lateral resolution as a consequence of the improved beam quality for picosecond sources compared to nanosecond LAESI lasers.

the distribution of gadoteridol within biological tissues. Both DESI-MSI and PIR-LAESI-MSI showed a distribution of gadoteridol that was consistent with in vivo MRI results (Figure S1). While we demonstrated that DESI and PIR-LAESI can be used to track gadoteridol distribution in tissues, MALDI-MSI using the matrix material employed in this study only reported gadoteridol in the medulla where its concentration was the highest.

■

■

PROSPECTS AND OUTLOOK

Beyond the advantages in drug and metabolite distribution analysis, PIR-LAESI-MS may offer additional, yet to be fully tested, possibilities. During intraoperative applications PIRLAESI-MS could deliver molecular maps of the tissue without spraying solvent on open wounds as with DESI26,27 or burning the tissues as with electrocautery.11 The compatibility with contrast agent imaging could allow surgical navigation from MS analysis of the plume of ablated tumors targeted with exogenous imaging agents. Through tightly focusing the PIRL laser beam, intraoperative molecular maps could be available at a spatial resolution that exceeds that of current DESI-MS methods. Reducing the size of the open beam laser beam, without having to use a fiber system, can open up new

CONCLUSIONS PIR-LAESI is an ion source for mass spectrometry that is capable of endogenous and exogenous small molecule imaging. We have shown ionization of biological lipids and small molecules from animal and plant sources by PIR-LAESI. In addition, it is also possible to analyze large molecules such as proteins with this ion source. Lipid and protein profiling with MS is a powerful tool to characterize cancers and to also classify different types of human tumors.23 Our proof of principle demonstration of the utility of PIR-LAESI-MS in lipid analysis opens up a wide spectrum of applications in cancer characterization to this ion source. The limit of detection of 100 nM for 12078


Article


(20) Fournier, I.; Wisztorski, M.; Salzet, M. Expert Rev. Proteomics 2008, 5, 413−24. (21) Nemes, P.; Barton, A. A.; Li, Y.; Vertes, A. Anal. Chem. 2008, 80, 4575−82. (22) Stolee, J. A.; Vertes, A. Anal. Chem. 2013, 85, 3592−8. (23) Eberlin, L. S.; Norton, I.; Dill, A. L.; Golby, A. J.; Ligon, K. L.; Santagata, S.; Cooks, R. G.; Agar, N. Y. Cancer Res. 2012, 72, 645−54. (24) Fuchs, B.; Schiller, J.; Süss, R.; Schürenberg, M.; Suckau, D. Anal. Bioanal. Chem. 2007, 389, 827−34. (25) Shrestha, B.; Patt, J. M.; Vertes, A. Anal. Chem. 2011, 83, 2947− 55. (26) Santagata, S.; Eberlin, L. S.; Norton, I.; Calligaris, D.; Feldman, D. R.; Ide, J. L.; Liu, X.; Wiley, J. S.; Vestal, M. L.; Ramkissoon, S. H.; Orringer, D. A.; Gill, K. K.; Dunn, I. F.; Dias-Santagata, D.; Ligon, K. L.; Jolesz, F. A.; Golby, A. J.; Cooks, R. G.; Agar, N. Y. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 11121−6. (27) Calligaris, D.; Caragacianu, D.; Liu, X.; Norton, I.; Thompson, C. J.; Richardson, A. L.; Golshan, M.; Easterling, M. L.; Santagata, S.; Dillon, D. A.; Jolesz, F. A.; Agar, N. Y. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 15184−9. Eberlin, L. S.; Norton, I.; Orringer, D.; Dunn, I. F.; Liu, X.; Ide, J. L.; Jarmusch, A. K.; Ligon, K. L.; Jolesz, F. A.; Golby, A. J.; Santagata, S.; Agar, N. Y.; Cooks, R. G. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 1611−6.

possibilities in more accurately delineating disease margins during surgery compared to conventional LAESI.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b02756. Additional MRI details and images as well as additional MS data (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Nemes, P.; Vertes, A. J. Visualized Exp. 2010, DOI: 10.3791/ 2097. (2) Nemes, P.; Barton, A. A.; Vertes, A. Anal. Chem. 2009, 81, 6668− 75. (3) Franjic, K.; Cowan, M. L.; Kraemer, D.; Miller, R. J. Opt. Express 2009, 17, 22937−59. (4) Cowan, M. L.; Bruner, B. D.; Huse, N.; Dwyer, J. R.; Chugh, B.; Nibbering, E. T.; Elsaesser, T.; Miller, R. J. Nature 2005, 434, 199− 202. (5) Franjic, K.; Miller, D. Phys. Chem. Chem. Phys. 2010, 12, 5225− 39. (6) Jowett, N.; Wöllmer, W.; Mlynarek, A. M.; Wiseman, P.; Segal, B.; Franjic, K.; Krötz, P.; Böttcher, A.; Knecht, R.; Miller, R. J. JAMA Otolaryngol Head Neck Surg 2013, 139, 828−33. (7) Amini-Nik, S.; Kraemer, D.; Cowan, M. L.; Gunaratne, K.; Nadesan, P.; Alman, B. A.; Miller, R. J. PLoS One 2010, 5, 10.1371/ journal.pone.0013053. (8) Kraemer, D.; Cowan, M. L.; Paarmann, A.; Huse, N.; Nibbering, E. T.; Elsaesser, T.; Miller, R. J. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 437−42. (9) Kwiatkowski, M.; Wurlitzer, M.; Omidi, M.; Ren, L.; Kruber, S.; Nimer, R.; Robertson, W. D.; Horst, A.; Miller, R. J.; Schlüter, H. Angew. Chem., Int. Ed. 2015, 54, 285−8. (10) Ren, L.; Robertson, W. D.; Reimer, R.; Heinze, C.; Schneider, C.; Eggert, D.; Truschow, P.; Hansen, N. O.; Kroetz, P.; Zou, J.; Miller, R. J. Nanotechnology 2015, 26, 284001. (11) Balog, J.; Sasi-Szabó, L.; Kinross, J.; Lewis, M. R.; Muirhead, L. J.; Veselkov, K.; Mirnezami, R.; Dezső , B.; Damjanovich, L.; Darzi, A.; Nicholson, J. K.; Takáts, Z. Sci. Transl. Med. 2013, 5, 194ra93. (12) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem. Rev. 1999, 99, 2293−352. (13) Sussulini, A.; Wiener, E.; Marnitz, T.; Wu, B.; Müller, B.; Hamm, B.; Sabine Becker, J. Contrast Media Mol. Imaging 2013, 8, 204−9. (14) Pugh, J. A.; Cox, A. G.; McLeod, C. W.; Bunch, J.; Writer, M. J.; Hart, S. L.; Bienemann, A.; White, E.; Bell, J. Anal. Bioanal. Chem. 2012, 403, 1641−9. (15) Aichler, M.; Huber, K.; Schilling, F.; Lohöfer, F.; Kosanke, K.; Meier, R.; Rummeny, E. J.; Walch, A.; Wildgruber, M. Angew. Chem., Int. Ed. 2015, 54, 4279−83. (16) Acquadro, E.; Cabella, C.; Ghiani, S.; Miragoli, L.; Bucci, E. M.; Corpillo, D. Anal. Chem. 2009, 81, 2779−84. (17) Hankin, J. A.; Barkley, R. M.; Murphy, R. C. J. Am. Soc. Mass Spectrom. 2007, 18, 1646−52. (18) Nemes, P.; Vertes, A. Anal. Chem. 2007, 79, 8098−106. (19) Eberlin, L. S.; Dill, A. L.; Golby, A. J.; Ligon, K. L.; Wiseman, J. M.; Cooks, R. G.; Agar, N. Y. Angew. Chem., Int. Ed. 2010, 49, 5953−6. 12079
