Laser

Oct 22, 2015 - With use of a model photoresist surface, known features as small as ∼1.5 μm were resolved. The capabilities of the system with real ...
29 downloads 12 Views 7MB Size
Download PDF
Article pubs.acs.org/ac

Characterization and Application of a Hybrid Optical Microscopy/ Laser Ablation Liquid Vortex Capture/Electrospray Ionization System for Mass Spectrometry Imaging with Sub-micrometer Spatial Resolution John F. Cahill, Vilmos Kertesz, and Gary J. Van Berkel* Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6131, United States S Supporting Information *

ABSTRACT: A commercial optical microscope, laser microdissection instrument was coupled with an electrospray ionization mass spectrometer via a low profile liquid vortex capture probe to yield a hybrid optical microscopy/mass spectrometry imaging system. The instrument has bright-field and fluorescence microscopy capabilities in addition to a highly focused UV laser beam that is utilized for laser ablation of samples. With this system, material laser ablated from a sample using the microscope was caught by a liquid vortex capture probe and transported in solution for analysis by electrospray ionization mass spectrometry. Both lane scanning and spot sampling mass spectral imaging modes were used. The smallest area the system was able to ablate was ∼0.544 μm × ∼0.544 μm, achieved by oversampling of the smallest laser ablation spot size that could be obtained (∼1.9 μm). With use of a model photoresist surface, known features as small as ∼1.5 μm were resolved. The capabilities of the system with real world samples were demonstrated first with a blended polymer thin film containing poly(2-vinylpyridine) and poly(N-vinylcarbazole). Using spot sampling imaging, sub-micrometer sized features (0.62, 0.86, and 0.98 μm) visible by optical microscopy were clearly distinguished in the mass spectral images. A second real world example showed the imaging of trace amounts of cocaine in mouse brain thin tissue sections. With use of a lane scanning mode with ∼6 μm × ∼6 μm data pixels, features in the tissue as small as 15 μm in size could be distinguished in both the mass spectral and optical images.

T

MSI, requiring little to no sample preparation.16−19 Desorption electrospray ionization (DESI), one of the more commonly used AP MSI techniques, routinely has spatial resolutions > 30 μm.19,20 In recent years a number of soft ionization AP MSI techniques have been described having spatial resolutions < 30 μm.21−29 To date some of the smallest reported data pixel sizes for soft AP MSI techniques include AP-MALDI,22−24 the direct liquid extraction-based technique termed single-probe,25,26 nano-DESI,27 and LA-APCI,28 which have achieved data pixel sizes of 3 × 3,24 2 × 5,25 1 × 5,27 and 4 μm × 8 μm,28 respectively. Recently, we introduced a laser ablation (LA)−liquid vortex capture (LVC)/ESI technique for high spatial resolution APMSI.29 The small pixel size achieved using the first prototype of LA-LVC/ESI-MS (2.5 μm × 2.5 μm229) indicated that with further improvement it had the potential to be among the

he analytical capabilities for mass spectrometry imaging (MSI) have continually improved over the past couple of decades and are now routinely applied to the characterization of complex systems.1−4 Advances in instrument design and sensitivity have enabled techniques such as secondary ionization mass spectrometry (SIMS),5−9 and matrix-assisted laser desorption ionization (MALDI)10−14 MSI to reach low micrometer or sub-micrometer pixel size and image resolution. SIMS and MALDI-MS, used to measure low molecular weight (e.g., inorganic compounds and small biomolecules) and high molecular weight molecules (e.g., proteins), respectively, while powerful techniques, require high vacuum, limiting the types of samples that can be imaged and often require extensive sample preparation (e.g., chemical matrix application).10,15 MSI at atmospheric pressure (AP) aims to remove the constraints of high vacuum, while achieving similar or improved ion sensitivities and MSI resolution to those achieved in vacuum. By utilizing AP ionization sources such as electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI), AP MSI has simplified procedures relative to MALDI© 2015 American Chemical Society

Received: August 27, 2015 Accepted: October 13, 2015 Published: October 22, 2015 11113

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

Figure 1. (a) Experimental design of the coupled LMD7000-LVC/ESI-MS system. (b) Schematic of the LVC probe design and positioning relative to the sample. (c) Photograph of the top of the LVC probe.

MSI. The design and operation of the LVC probe coupling is discussed, and the limits of the MSI capabilities determined using well-defined photoresist patterns are demonstrated. Application of the system for imaging chemical distributions in a polymer thin film and a thin tissue section is also presented.

highest spatial resolution AP MSI techniques available. The capabilities of our in-house-constructed system were limited by a number of factors including scanning stage, the stability of the optical platform, and the achievable diameter of the laser ablation spot size. In order to improve the technique’s capabilities beyond those currently demonstrated, a more robust platform for laser ablation was needed. Commercial laser microdissection (LMD) systems use a highly focused and precisely controlled laser beam to dissect regions of interest from tissue samples and collect them for offline analysis. LMD systems are capable of focusing laser beams to sub-micrometer spot sizes using sophisticated microscope lenses. For LA-MS techniques, LMD systems can be utilized as the apparatus for LA, taking advantage of the highly precise and focused laser beam for improved MSI resolution. For example, Becker and co-workers have used a commercially available LMD system coupled to inductively coupled plasma (ICP)-MS to reach spot sizes down to 3 μm.30−34 Similarly, Lorenz et al.28 utilized a laser LMD system, but due to limitations in material transport and ionization by the APCI source used, the smallest usable laser ablation spot size was ∼4 μm × 8 μm. In both cases the LMD system also functioned as a high resolution microscope with bright-field and fluorescence microscopy capabilities making their coupling with MS even more beneficial. Herein, we report on our coupling of a Leica LMD7000 laser microdissection instrument35,37 with an ESI quadrupole timeof-flight and a hybrid triple quadrupole ion trap mass spectrometer via a low profile LVC probe. The LMD7000 instrument could be used for optical and fluorescence microscopy as well as laser ablation for MSI. The data presented here demonstrate that this combination yields a multimodal, optical microscopy/mass spectrometry, imaging system capable of sub-micrometer mass spectral imaging resolution. No modification to the LMD7000 or mass spectrometer hardware was necessary to achieve this coupling other than changing the ESI emitter probe inner diameter to enable increased solvent self-aspiration rates. Software written in-house enabled fully automated operation of the system for

■

EXPERIMENTAL SECTION Chemicals and Materials. LC-MS CHROMASOLV chloroform, methanol with 0.1% formic acid, and water with 0.1% formic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). The LMD-LVC/ESI-MS system was characterized using ink thin films on glass (blue and red Sharpie markers containing basic blue 7 and rhodamine B, respectively (Sanford, Oak Brook, IL, USA)) and photoresist test patterns on glass. Photoresist test patterns were fabricated by pipetting Shipley S1813 positive resist onto a clean microscope slide and then spinning at 3000 rpm for 30 s (EDC-650-15 spin coater; Laurell Technologies, North Wales, PA, USA). The slide was then baked at 120 °C for 60 s followed by exposure to UV light (365−440 nm) for 30 s through a glass mask with test patterns (HTA Photomask, San Jose, CA, USA). Lastly, the slide was developed using MF 319 developing solution for 30 s and rinsed with water. This process made photoresist films ∼1.8 μm thick. Poly(2-vinylpyridine) (P2VP; MW = 36 kDa) and poly(N-vinylcarbazole) (PVK; MW = 50 kDa) were purchased from Polymer Source, Inc. (Dorval, QC, Canada) and used without further purification. Polymer thin films were created by preparing ∼5 wt % solutions of each P2VP and PVK in chloroform. Aliquots of these solutions were then combined in a 1:2 P2VP:PVK (v/v) ratio for final P2VP and PVK concentrations of ∼1.7 and 3.3 wt %, respectively. A ∼1.1 μm thick polymer film (measured using a profilometer) was made by spin coating 50 μL of the mixed solution on a glass microscope slide using a custom spin coater. No annealing was needed to phase separate the polymers. Brain thin tissue sections obtained through the National Institute of Drug Abuse-Intramural Research Program, National Institutes of Health were prepared according to the following procedure. Adult male mice received an intraperitoneal (ip) injection of 11114

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

the probe. Additionally, the LMD7000 was covered in a plastic sheet to prevent ambient airflow in the sampling area. Data Collection and Image Creation. Control of the laser beam for LA, the trigger for mass spectral data acquisition, and creation of the mass spectral images were done as previously reported and are described in the Supporting Information.29 The LMD-LVC/ESI-MS system was capable of imaging in both lane scanning and spot sampling modes. Acquiring images in lane scanning mode required inputs of the length and spacing of lanes, both in pixels (px), and the scan speed (px/s). For spot sampling the number and spacing (px) of columns and rows of spots, laser duration, and mass spectral analysis time per spot were required. Values given in px or in px/s were transformed into actual distances using the ratio of one dimension of the optical field of view and the number of pixels representing that on the screen. The size of the viewable area (1024 × 755 px) was a constant value set by the LMD7000 operating software; as such, this value and the focusing objective selected governed minimum stepping size (μm/px) and actual size of the field of view (Table 1).

cocaine hydrochloride (30 mg/kg free base in normal saline) and were sacrificed within 30 min postdose. After sacrifice, the brains were immediately removed and frozen in dry ice chilled isopentane. Tissue sections were cut using a cryostat (Leica Microsystems CM3050S, Bannockburn, IL, USA) at −21 °C (cryochamber temperature) and −18 °C (specimen cooling temperature). The brain was attached to the cryostat specimen disk using ice slush made with distilled water, and 10 μm thick sections were cut and thaw mounted on 2 μm polyethylene naphthalate (PEN) membrane slides (Leica Microsystems, Wetzel, Germany). Instrumentation. The overall experimental setup is illustrated in Figure 1a. A Leica LMD7000 (Leica Microsystems, Wetzel, Germany) laser microdissection system35,36 having 5×, 10×, 20×, 40×, and 63× confocal dry objectives was used to acquire bright-field and fluorescent images of the samples. In addition, it served as the platform for LA by utilizing the included Nd:YLF laser operating at 349 nm with 5 kHz maximum repetition rate and 120 μJ maximum pulse energy. Laser beam positioning was accomplished using internal prism optics.35 The system used an inverted sample geometry (i.e., the sample is located on the lower side of a microscope glass slide) in combination with cutting laser-optics in transmission geometry for gravity-supported collection of ablated material. Schematic details of the low profile LVC probe are illustrated in Figure 1b. The probe used a coaxial tube design on the sampling end. The outer tube was stainless steel (∼1.28 mm i.d. × ∼1.61 mm o.d. × 1.1 cm length) connected to the mass spectrometer ion source electrical ground. A PEEK capillary (255 μm i.d. × 510 μm o.d. × ∼24 cm length; IDEX Health & Science LLC, Oak Harbor, WA, USA) was used as the inner tube. A 178 μm i.d. × 794 μm o.d. PEEK capillary was used for polymer and tissue samples which improved signal levels slightly. Each tube was secured in a commercially available fiveport PEEK manifold (IDEX) that was machined down from its original 31 mm circular base diameter to a 24 mm square shape (Supporting Information (SI) Figure S-1). Two of the four horizontal ports were plugged. Solvent was delivered through one of the remaining horizontal ports by an HPLC pump into the annulus region of the two coaxial tubes mounted in the vertical port of the manifold. When the solvent reached the top of the tubes, it was aspirated down the inner capillary, which exited the manifold through the second horizontal port, into the Turbo V ion source of a SCIEX TripleTOF 5600+ or QTRAP 5500 mass spectrometer (Sciex, Concord, Ontario, Canada). The nebulizing gas (nitrogen) in the ion source was used to adjust the solvent aspiration rate through the inner tube to slightly exceed the delivery rate of solvent to the probe from the HPLC. In doing so, a stable vortex, or whirlpool, drain was maintained at the sampling end of the probe to liquid capture laser ablated material from the sample mounted above the probe (see photograph in Figure 1c). To increase the accessible self-aspiration flow rate range of the system, the standard ESI emitter and nebulizer capillaries of the ESI probe were replaced by equivalent parts with larger internal diameters (530 and 150 μm, respectively). Approximate alignment of the LVC probe position into the center of the microscope field of view was done by altering the software “reference point” of the capture cap holder. The probe−substrate distance was kept at ∼1.2 mm. A sheath made of heat shrink tubing extending 1.1 mm from the probe (i.e., ∼0.1 mm from the substrate) was added to control airflow near

Table 1. Screen Pixel (px) to Real Distance (μm) Conversion with the LMD7000 Systema objective

minimum stepping size (μm/px)

actual viewable area (μm × μm)

5× 10× 20× 40× 63×

1.702 0.851 0.426 0.213 0.136

1743 × 1285 871 × 643 436 × 321 218 × 161 140 × 103

a

Maximum viewable area was 1024 × 755 px.

Mass spectral data acquisition was started in the Analyst software (v 1.5 and 1.6) at the beginning of each lane (scanning mode) or single row of spots (spot sampling mode) and stopped at the end of each lane or row, creating a single mass spectral data file corresponding to one line of the mass spectral image. Mass spectra were acquired at a rate of 4 spectra/s in all experiments. Raw data files were converted into a TissueView (SCIEX) compatible format using another software developed in-house (TissueView Converter for WIFF Data). Chemical images of selected mass-to-charge (m/z) ratios were visualized by importing the resulting file into the TissueView software package. Since the whole field of view of the bright-field images were interrogated by the laser, the optical and mass spectral images were inherently co-registered allowing seamless postprocessing data overlay. Single reaction monitoring (SRM) scans were used for all images. The specific instrument settings used for optimum detection and imaging are provided in the Supporting Information.

■

RESULTS AND DISCUSSION LA Liquid Capture with Low Profile LVC Probe. In normal use, the LMD7000 transfers material by gravity-assisted dropping of a laser dissected sample into a collection device, such as the cap of a 0.6 mL centrifuge tube, placed directly below the dissection site.36 The LVC probe serves in place of these capture caps, functioning to capture laser ablated material directly into a solvent flow stream which immediately transports the material to the ESI source of the mass spectrometer for ionization and mass spectral analysis. Given the desire to modify the LMD7000 as little as possible and 11115

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

variation in this region due to laser-probe alignment. When laser ablating material in the field of view near the interior edge of the outer probe and beyond, the signal dropped off for both normal and reversed transects. For y-axis transects the edges of the probe were outside of the viewable area and thus the signal was relatively constant across the whole transect. The viewable ranges for all other focusing objectives were within the center region of the probe and exhibited no significant variations in signal. Thus, only when using the 5× objective was it necessary to avoid LA near the edges of the viewable sample area. Minimum LA Spot Size Generating Detectable Mass Spectral Signal. Single laser spots from a thin film of ink on glass containing rhodamine B were used to judge the smallest laser ablation spot size that could generate detectable mass spectral signal. The thin ink film (∼300 nm thick) was not solely rhodamine B, but this dye is in high concentration. Signals from single spots (100 shots/spot) of rhodamine B with ∼1.9 μm diameter spot size, the smallest spot size resulting in complete ablation of the sample, are shown in SI Figure S-5a. The time for appearance of signal after the laser firing was ∼6 s with signal persisting ∼1.8 s (baseline to baseline). This delay time for signal appearance and the signal peak width were caused by the time to transport material from the sampling end of the LVC probe to the ionization source and by dispersion of the captured sample in the solvent during transport, respectively. The average signal-to-noise (S/N) of the signal peak was very high (∼703) for a 1.9 μm laser spot indicating that the signal from even smaller ablation regions would be observable. Oversampling can be used to achieve lower “effective” spot sizes by stepping the laser position a smaller distance than the laser spot diameter. Since laser movement is currently controlled by the positioning of the mouse cursor on the location of interest in the microscope field of view, the smallest movable distance should be 1 pixel. This corresponds to a different minimum stepping distance for each microscope objective (Table 1). To verify this experimentally, mass spectral signals from a basic blue 7 containing inked film were acquired using different stepping distances (1, 2, 4, and 8 px) and objectives (5×, 10×, 20×, 40×, and 63×) (SI Figure S-6). The laser spot size differed for each objective used but in every case was oversampled. Average peak areas from three replicate measurements were normalized to values acquired at 8 px stepping distance for each objective (SI Figure S-6a). The average RSD within each measurement is plotted in SI Figure S-6b. Error bars in SI Figure S-6 represent the standard deviation across the three replicates. On average signal intensities followed the expected decrease in the ablated area with reduced stepping distance (i.e., slope = ∼1); however, % RSD increased substantially with stepping distances < 4 px. Note that this was independent of the objective used (i.e., the actual distance stepped). As a result, the smallest accurate stepping distance for the laser beam was 4 px in either lane scanning or spot sampling imaging modes, corresponding to a minimum oversampling ablation size of 0.544 μm when using the 63× objective. An oversampling ablation size of 0.544 μm resulted in a S/N of ∼71 for rhodamine B (SI Figure S-7), indicating that smaller oversampling ablation sizes could be used if not limited by the stepping resolution of the system. Imaging of Photoresist Substrates. Photoresist substrates, which can be fabricated with customized, reproducible features down to a few micrometers, were used to initially evaluate the imaging resolution of the system. Signal intensities

maintain all automated functions, the LVC probe used here was designed to fit into the unmodified capture cap holder supplied with the system (see Figure 1b and SI Figure S-1). The LVC probe to surface distance was chosen based on several factors including achieved signal intensity, signal reproducibility, and ease of movement of the probe within the system. A probe−substrate distance of ∼0.5 mm yielded the highest signal level and reproducibility. However, with this small probe to surface separation the top of the probe would sometimes clip the sample holder during movement of the probe or sample. In addition, this close spacing often led to alteration of the inked and photoresist features by condensation of solvent vapor on the sample slide making imaging impractical. As a compromise, the probe−substrate distance was kept at ∼1.2 mm which allowed free movement within the system and eliminated solvent vapor related issues. To compensate for poorer signal reproducibility caused by this larger than optimum spacing, a sheath made of heat shrink tubing was placed on the PEEK base of the probe extending to within ∼0.1 mm from the sample surface. The flexibility of the tubing allowed close probe−substrate distance without limiting movement within the system. The use of the sheath improved signal reproducibility dramatically as illustrated by a comparison of the signals recorded from five single LA spots of basic blue 7 containing ink film with and without the sheath (SI Figure S-2). Without the sheath, the signal was very variable, having a relative standard deviation (%RSD) of 76%, with occasionally a complete lack of signal from an ablation event (e.g., shot 4 in SI Figure S-2b). In contrast, signal levels were very reproducible (6% RSD) with the sheath in place. This sheath stopped stray air currents from entering the ablation region, but also may have helped to direct the LA plume toward the center of the capture probe. In the overaspiration mode in which the capture probe operates, some air is entrained into the liquid as the solvent is aspirated to the inner capillary of the probe. By limiting the air entrance into the ablation region to only a small gap near the sample surface, the linear velocity of gas flow into the space above the probe was increased and likely directed gas flow along the sample surface and then down into the center of the probe (SI Figure S-3). This gas flow may enhance the reliability and efficacy of LA plume capture and warrants further study. LA-Probe Alignment. With the LMD-LVC/ESI-MS combination, mass spectral images are generated by LA of material while moving the laser beam across the substrate rather than by moving the sample stage. As such the LA-probe alignment depends on the magnification of the objective and where in the field of view ablation is taking place. The laser position in the field of view changes as mass spectral images are acquired. For all but the 5× objective, the field of view was smaller than the inner diameter of the probe outer capillary (∼1.3 mm), so this changing alignment was not expected to be a factor. To verify this, signal levels from x- and y-axis transects of a basic blue 7 containing inked film acquired by spot sampling across the viewable area using a 5× objective were collected (SI Figure S-4). Three transects were acquired to give an estimation of average signal and variance at each point and then were normalized. To mitigate potential error due to variation in basic blue 7 concentration across the substrate, the substrate was rotated 180° and transects reacquired in the same nominal area (gray traces in SI Figure S-4). As can be seen in SI Figure S-4, values toward the center of the probe lie within one standard deviation of each other, indicating no significant signal 11116

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

Figure 3. Bright-field images of three photoresist patterns with smallest features (a) 1.5, (f) 2.3, and (k) 3.1 μm in length. MSI of the same photoresist features using pixel sizes of (b, g, l) 0.544, (c, h, m) 0.816, and (d, i, n) 1.088 μm in scanning mode and (e, j, o) 0.816 μm in spot sampling mode. Scale bars are 5 μm. Instrument parameters are given in the Supporting Information.

Figure 2. (a) Bright-field image of patterned photoresist substrate. (b) LMD-LVC/ESI-MS image (198 × 77 px) of photoresist acquired using 0.544 μm square pixels. Scale bars are 10 μm. Instrument parameters are given in the Supporting Information.

Table 2. RPF Metric Calculated for Each Image Shown in Figure 3

from single 1.9 μm spots of photoresist (m/z −227.2 → −107.2) having a calculated S/N = 254 are shown in SI Figure S-5b. Figure 2 shows an exemplary image using the appropriate laser scan rate, lane stepping size, and mass spectrometer acquisition rates to achieve the smallest possible square pixel size for the mass spectral images with the present system (0.544 μm × 0.544 μm). Each letter on the photoresist has a thickness of ∼5 μm and spacing between letters of ∼3 μm. The highlighted point of interest in Figure 2 is the smallest feature in the image, ∼2 μm, which is resolved in the mass spectral image. To provide a more quantitative measure of the imaging resolution capabilities of the present system using differing measurement conditions, photoresist patterns having 2.3, 3.1, and 4.6 μm features with 1.5, 2.3, and 3.1 μm spacing between them, respectively, were imaged using various degrees of oversampling in scanning and spot sampling modes (Figure 3). For each mass spectral image acquired there were features resembling those known from bright-field images (e.g., Figure 3l−n); however, by simple visual inspection the quality of the images was obviously different. An unbiased approach to image resolution quality was developed and applied here. To determine to what degree a known feature is resolved, a resolving power factor (RPF) can be calculated: RPF =

RPF square pixel size (μm) 0.544 0.816 1.088 0.816

2.3 μm feature, 1.5 μm spacing

3.1 μm feature, 2.3 μm spacing

Scanning Mode 0.75 ± 0.07 0.89 ± 0.05 0.29 ± 0.10 0.60 ± 0.03 0.59 ± 0.26 Spot Sampling Mode 0.97 ± 0.05 0.99 ± 0.01

4.6 μm feature, 3.1 μm spacing 0.97 ± 0.03 0.81 ± 0.08 0.60 ± 0.05 1.01 ± 0.02

occurred continuously as the laser was moved. Therefore, when imaging in scanning mode, the signal for a given pixel will be a mixture of signals from a larger area dependent on the washout profile. Spot sampling was the most sensitive method of acquiring images and generated images of the highest quality, but suffered from significantly longer acquisition times (∼10− 25 times) than lane scanning mode as every pixel took a minimum of 1.8 s to acquire. When in spot sampling mode using a 0.816 μm square pixel, all features were completely resolved, with RPFs of 1.01 ± 0.02, 0.99 ± 0.01, and 0.97 ± 0.05 for features with 3.1, 2.3, and 1.5 μm spacing, respectively (Figure 3e,j,o). These improved metrics came at the cost of analysis time. For example, the acquisition time for the image in Figure 3b acquired using lane scanning mode was ∼30 min while that in Figure 3e acquired using spot sampling mode required ∼3 h. As can be seen from the images in Figure 3 and calculated RPF values in Table 2, the determination of whether a feature is resolved becomes somewhat arbitrarily defined by the user. That is, at what RPF does one judge an image acceptable for reporting a resolved feature? The features in the images in Figure 3h,n lie on the threshold of being resolved (RPF ∼ 0.60), while the features in the images in Figure 3g,l,m are resolved more confidently (RPF > 0.80). To provide context to a stated imaging resolution, we believe it is important to include an image quality metric in conjunction with the stated value. Thus, for the photoresist substrates used here, we report that

dMV dMB

where dMV is the is the difference in intensity between peak max and peak valley and dMB is the intensity difference between measured peak max and baseline (SI Figure S-8). A RPF of 1.0 means the feature is completely resolved. RPF values for each image in Figure 3 were calculated (Table 2) as described in the Supporting Information. RPF increased with decreasing pixel size, with a maximum RPFs of 0.97 ± 0.03, 0.89 ± 0.05, and 0.75 ± 0.07 achieved for features with 3.1, 2.3, and 1.5 μm spacing, respectively, using the 0.544 μm × 0.544 μm pixel size. In spot sampling mode, the signal for a given pixel was the integral of the entire signal washout profile (∼1.8 s, SI Figure S5b) for one ablation event, whereas in scanning mode ablation 11117

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

overlapping layers, and agreed well with the bright-field and fluorescence images. As expected the P2VP mass spectral image (Figure 4c) was not as crisp as the mass spectral image of PVK (Figure 4d) due to the longer washout time of P2VP signal. Despite this, features as small as ∼1.5 μm were still resolved. The yellow highlighted regions in Figure 4 indicate features, each measured to be ∼1.5 μm, that are resolved in all four images. Another series of images was taken from a different area of the polymer thin film, this time using spot sampling imaging mode. Parts a and b of Figure 5 show bright-field and

this LMD-LVC/ESI-MS combination can confidently resolve a 1.5 μm feature with a RPF = 0.75 ± 0.07 (Figure 3b) and RPF = 0.97 ± 0.05 (Figure 3e) when using lane scanning and spot sampling modes, respectively. LMD-LVC/ESI-MS Imaging of a Phase Separated Polymer Thin Film. To demonstrate the capabilities of the LMD-LVC/ESI-MS on a real world sample, a phase separated polymer thin film comprised of two commonly studied polymers, poly(2-vinylpyridine) (P2VP) and poly(9-vinylcarbazole) (PVK), was investigated. PVK is a hole-transporting polymer often used for light-emitting diodes while P2VP is commonly used to make block co-polymers.37 The chemical structures as well as the mass spectrum of each obtained in a LA spot sampling experiment for P2VP and PVK are shown in SI Figure S-9a,b and d,e, respectively. The base peak in the spectrum from P2VP was consistent with the protonated monomer (m/z 106, C7H8N+; SI Figure S-9b). The base peak in the spectrum from PVK was m/z 180 (C13H10N+) corresponding in mass to the loss of CH2 from the protonated monomer of this polymer (m/z 194; SI Figure S-9e). For imaging, SRM detection of each polymer was optimized using the product ions from each of these precursor ions (P2VP, m/z 106 → 78; PVK, m/z 180 → 152) (SI Figure S-9c,f, respectively). A notable difference observed between P2VP and PVK samples was the washout time of the signal. PVK had a washout time similar to that observed with most other analytes (∼1.8 s, baseline−baseline); however, the P2VP signal profile was broader (∼3.8 s, baseline−baseline) apparently due to limited solubility (data not shown). Parts a and b of Figure 4 show bright-field and fluorescence images, respectively, of the polymer blend using the 63×

Figure 5. (a) Bright-field, (b) fluorescence, and (d) merged P2VP/ PVK mass spectral image of 2:1 PVK/P2VP polymer blend acquired in spot sampling imaging mode (0.544 μm/px). (c) Intensity profile along the highlighted trace shown in panel a. Measured distances baseline−baseline were 0.98, 0.86, and 0.62 μm for features no. 1, no. 2, and no. 3, respectively, and were measured in the mass spectral image (d). Purple color in panel d indicates a layered P2VP PVK structure. Scale bars are 6 μm. Instrument and image parameters are given in the Supporting Information.

fluorescence images, respectively, along with the merged P2VP/ PVK mass spectral image (Figure 5d) that was acquired using a 0.544 μm × 0.544 μm data pixel size. This sampling mode substantially increased the imaging time, but completely eliminated the slight blurring effect that was observed in the P2VP image in Figure 4c. For the most part, the mass spectral data show that P2VP and PVK were completely spatially separated in the film, except in a few instances of layered P2VP/PVK (purple color in the top regions of Figure 5d). The spatial resolution in this image was examined by following the P2VP and PVK signal trace along the yellow line drawn in the bright-field image in Figure 5a. The spacing in the transition across these three features (labeled no. 1, no. 2, and no. 3 in Figure 5a) in the bright-field image were measured to be 0.98, 0.86, and 0.62 μm, respectively (Figure 5c). All of these features were clearly resolved in the P2VP/PVK mass spectral image (Figure 5d; RPF > 0.9) indicating sub-micrometer mass spectral imaging resolution. LMD-LVC/ESI-MS Imaging of Cocaine Dosed Tissue Sections. Drug dosed tissue sections provide a sample to test this system for imaging of a trace component in a complicated matrix. As an exemplary sample, brain thin tissue sections from mice dosed with cocaine (30 mg/kg) were imaged. Parts b and c of Figure 6 shows bright-field and fluorescence images,

Figure 4. (a) Bright-field, (b) fluorescence, (c) P2VP MSI, and (d) PVK MSI of 2:1 PVK:P2VP polymer blend acquired in lane scanning imaging mode (0.544 μm/pixel). Scale bars are 15 μm. Highlighted regions in panel a indicate ∼1.5 μm features also measured in images b−d. Instrument and image parameters are given in the Supporting Information.

objective. PVK domains could be differentiated from P2VP domains by the autofluorescence of the PVK polymer. The image could then be used to corroborate the mass spectral chemical image of the polymer blend. Mass spectral images of P2VP (m/z 106 → 78, green) and PVK (m/z 180 → 152, red) acquired in scanning imaging mode with a 0.544 μm × 0.544 μm data pixel size, the smallest data pixel size possible with the current LMD-LVC/ESI-MS hybrid system, are shown in Figure 4c,d, respectively. P2VP and PVK mass spectral images had predominantly inverted spatial distributions indicating that the two polymers were mostly separated, rather than having 11118

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Analytical Chemistry

■

Article

CONCLUSIONS

A hybrid optical microscopy/mass spectrometry imaging system has been achieved using a low profile LVC probe to combine a commercial laser microdissection instrument with an ESI mass spectrometer. This hybrid system is capable of multimodal imaging, having bright-field and fluorescence microscopy capabilities, as well as high spatial resolution AP MSI. The coupling was accomplished using a simple, robust, and inexpensive coaxial capillary, continuous flow, LVC probe to catch and transport laser ablated material to the mass spectrometer for ionization and detection. No LMD7000 or mass spectrometer hardware modifications were necessary to achieve this coupling other than changing the ESI emitter probe inner diameter. Software written in-house enabled fully automated operation of the system for MSI. It should be noted that though ESI was used in the work presented here, other AP ionization sources (e.g., APCI) can be implemented easily with this system. The use of the LMD7000 for LA-LVC/ESI-MS has enabled improved image quality and resolution over our previous report using an in-house-constructed LA system and LVC/ESI-MS. In our previous work a minimum square pixel size of 2.5 μm was used due to limits in signal resulting in an imaging resolution of 6 μm estimated using stamped ink grids.29 In comparison, a ∼0.544 μm square pixel was able to generate enough signal from rhodamine B containing ink, photoresist surfaces, and polymer thin films for use in mass spectral imaging. Using model photoresist, surfaces features as small as ∼1.5 μm were resolved. This improved imaging resolution is due in part to the small laser spot size provided by the LMD7000 system (∼1.9 μm versus ∼50 μm) which reduces the degree of oversampling needed to reach smaller pixel sizes, the precise laser control of the system, and increased signal reproducibility through the use of a sheath around the LVC probe. The hybrid system was applied toward two real world samples by imaging the chemical distribution of two polymers, P2VP and PVK, in a phase separated polymer thin film and by imaging trace levels of cocaine in a dosed mouse brain tissue section with data pixels sizes of ∼0.544 μm × 0.544 μm and ∼6 μm x 6 μm, respectively. Fluorescence images acquired using the same platform identified PVK in the polymer blend and improved contrast in the tissue images. Mass spectral images of the polymer film acquired using spot sampling mode distinguished features as small as ∼0.62 μm (measured using optical microscopy). This is the smallest feature measured to date by an AP-MSI technique, the next best being our own atomic force microscopy thermal desorption/secondary ionization/mass spectrometry imaging technique.39 This imaging resolution rivals the best possible with vacuum MSI techniques such as MALDI and SIMS. It should be noted that although the LMD-LVC/ESI-MS system uses an ultraviolet (UV) laser, little fragmentation has typically been observed over that caused by ESI alone. This is because the signal observed using this technique is predominantly due to the neutral particulates ejected during the laser ablation event and captured in the LVC probe, rather than from direct formation and fragmentation of gas phase ions and their fragments.40 The use of special membrane sample slides, such as that used here with the thin tissue section, (e.g., PEN membrane slides) designed for high absorption of the UV laser may assist in this molecular preservation. Although the LMDLVC/ESI-MS measures most small and large species intact, in

Figure 6. (a) Whole tissue image of a 10 μm thick mouse brain section from mouse dosed with cocaine. (b) Bright-field and (c) fluorescence images of tissue and (d) mass spectral image of cocaine (m/z 304 → 182) of the highlighted region in panel a. Arrows point to features ∼15 μm in size. Scale bars are 200 μm. The mass spectral image was 143 × 107 pixels (6 μm/px) acquired using the 10× objective. Instrument parameters are given in the Supporting Information.

respectively, in addition to a mass spectral image for cocaine (Figure 6d) from the highlighted region of the mouse brain shown in Figure 6a. The mass spectral image was acquired in scanning mode (m/z 304 → 182) with a data pixel size of 6 μm × 6 μm using the 10× objective. Smaller pixel sizes resulted in mass spectral signal levels too low to get high contrast images. On the basis of other literature reports, the average concentration of cocaine in this tissue was in the low parts per million range.38 The size of the area imaged was 858 μm × 642 μm taking ∼1.5 h to complete. The yellow arrows in Figure 6b−d point to features as small as 15 μm that could be distinguished in both the optical and mass spectral images. 11119

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry

Institutes of Health) are thanked for supplying the mouse brain tissue. The tissue imaging work at Oak Ridge National Laboratory (ORNL) was supported by, and the QTRAP 5500 and TripleTOF 5600+ mass spectrometers used in this work were provided on loan by, SCIEX through a Cooperative Research and Development Agreement (CRADA NFE-1002966). Julian Burke (Leica Microsystems) is thanked for the loan of the LMD7000 instrument, Leslie L. Wilson (ORNL) is thanked for preparing the photoresist coated substrates, and Vera Bocharova is thanked for help with polymer film preparation. The instrument advancement, fundamental metric studies, and polymer imaging work was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division.

the case of the synthetic polymers studied here some breakage of the polymer backbone did occur as a result of the laser ablation process. This resulted in the formation of distinctive low mass n-mers that were used to selectively detect the respective polymer species. Since the various proteins we have tested (bradykinin, angiotensin III, and cytochrome C (data not shown)) have not exhibited any additional fragmentation the intact ablation or breakage of polymer backbones, whether synthetic or natural, using LMD-LVC/ESI-MS deserves some additional study to fully understand the process. All indications are that the present system can be improved to yield even higher spatial resolution mass spectral images. Currently the minimum pixel size for our hybrid system, ∼0.544 μm, seems to be limited by ablation area spot to spot variability. Use of a higher focusing objective (150×) should enable LA spots sizes as small as ∼0.8 μm providing the possibility of data pixels sizes significantly smaller than the current ∼0.544 μm × 0.544 μm achieved with a 1.9 μm LA spot size and oversampling. Some ablated material can be seen to redeposit on the sample surface indicating that sample transfer efficiency can be improved upon, which will result in higher signal levels for a given ablation spot size. Modifications to the LVC probe design, dimensions, and gas and liquid flows will likely improve collection efficiency, mass transport, ionization, and ultimately mass spectral detection. Since signal reproducibility was greatly improved through the addition of a sheath to the LVC probe, additional research into the effects of airflows to guide material capture into the LVC probe may be a worthwhile endeavor.

■

■

(1) Addie, R. D.; Balluff, B.; Bovée, J. V. M. G.; Morreau, H.; McDonnell, L. A. Anal. Chem. 2015, 87, 6426−6433. (2) Crecelius, A. C.; Vitz, J.; Schubert, U. S. Anal. Chim. Acta 2014, 808, 10−17. (3) Spengler, B. Anal. Chem. 2015, 87, 64−82. (4) Kriegsmann, J.; Kriegsmann, M.; Casadonte, R. Int. J. Oncol. 2015, 46, 893−906. (5) Brown, A.; Vickerman, J. C. Surf. Interface Anal. 1984, 6, 1−14. (6) Aoyagi, S.; Fletcher, J. S.; Sheraz, S.; Kawashima, T.; Razo, I. B.; Henderson, A.; Lockyer, N. P.; Vickerman, J. C. Anal. Bioanal. Chem. 2013, 405, 6621−6628. (7) Piehowski, P. D.; Kurczy, M. E.; Willingham, D.; Parry, S.; Heien, M. L.; Winograd, N.; Ewing, A. G. Langmuir 2008, 24, 7906−7911. (8) Benninghoven, A. Surf. Sci. 1994, 299-300, 246−260. (9) Levisetti, R.; Chabala, J. M.; Li, J.; Gavrilov, K. L.; Mogilevsky, R.; Soni, K. K. Scanning Microsc. 1993, 7, 1161−1172. (10) Norris, J. L.; Caprioli, R. M. Chem. Rev. 2013, 113, 2309−2342. (11) Zavalin, A.; Todd, E. M.; Rawhouser, P. D.; Yang, J.; Norris, J. L.; Caprioli, R. M. J. Mass Spectrom. 2012, 47, 1473−1481. (12) Gessel, M. M.; Norris, J. L.; Caprioli, R. M. J. Proteomics 2014, 107, 71−82. (13) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T.; Matsuo, T. Rapid Commun. Mass Spectrom. 1988, 2, 151−153. (14) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299−2301. (15) Wiangnon, K.; Cramer, R. Anal. Chem. 2015, 87, 1485−1488. (16) Venter, A. R.; Douglass, K. A.; Shelley, J. T.; Hasman, G.; Honarvar, E. Anal. Chem. 2014, 86, 233−249. (17) Venter, A.; Nefliu, M.; Cooks, R. G. TrAC, Trends Anal. Chem. 2008, 27, 284−290. (18) Huang, M. Z.; Yuan, C. H.; Cheng, S. C.; Cho, Y. T.; Shiea, J. Annu. Rev. Anal. Chem. 2010, 3, 43−65. (19) Wu, C. P.; Dill, A. L.; Eberlin, L. S.; Cooks, R. G.; Ifa, D. R. Mass Spectrom. Rev. 2013, 32, 218−243. (20) Campbell, D. I.; Ferreira, C. R.; Eberlin, L. S.; Cooks, R. G. Anal. Bioanal. Chem. 2012, 404, 389−398. (21) Yanes, O.; Woo, H.; Northen, T. R.; Oppenheimer, S. R.; Shriver, L.; Apon, J.; Estrada, M. N.; Potchoiba, M. J.; Steenwyk, R.; Manchester, M.; Siuzdak, G. Anal. Chem. 2009, 81, 2969−2975. (22) Schober, Y.; Guenther, S.; Spengler, B.; Rompp, A. Anal. Chem. 2012, 84, 6293−6297. (23) Guenther, S.; Rompp, A.; Kummer, W.; Spengler, B. Int. J. Mass Spectrom. 2011, 305, 228−237. (24) Römpp, A.; Spengler, B. Histochem. Cell Biol. 2013, 139, 759− 783. (25) Rao, W.; Pan, N.; Yang, Z. J. Am. Soc. Mass Spectrom. 2015, 26, 986−993. (26) Pan, N.; Rao, W.; Kothapalli, N. R.; Liu, R. M.; Burgett, A. W. G.; Yang, Z. B. Anal. Chem. 2014, 86, 9376−9380. (27) Laskin, J.; Heath, B. S.; Roach, P. J.; Cazares, L.; Semmes, O. J. Anal. Chem. 2012, 84, 141−148.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b03293. Software description, calculation of RPF from experimental data, Figure 3 instrument settings, and images of PEEK manifold, extracted ion chromatograms, hypothesized airflow schematic, normalized signal intensities, relative signal intensities, exemplary calculation, and P2VP and PVK structures and mass spectra (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: 865-574-1922. Notes

This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a nonexclusive, paidup, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Amina S. Woods, Shelley Jackson, and Aurelie Roux (National Institute of Drug Abuse-Intramural Research Program, National 11120

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121

Article

Analytical Chemistry (28) Lorenz, M.; Ovchinnikova, O. S.; Kertesz, V.; Van Berkel, G. J. Rapid Commun. Mass Spectrom. 2013, 27, 1429−1436. (29) Ovchinnikova, O. S.; Bhandari, D.; Lorenz, M.; Van Berkel, G. J. Rapid Commun. Mass Spectrom. 2014, 28, 1665−1673. (30) Wu, B.; Niehren, S.; Becker, J. S. J. Anal. At. Spectrom. 2011, 26, 1653−1659. (31) Wu, B.; Becker, J. S. Int. J. Mass Spectrom. 2012, 323-324, 34− 40. (32) Wu, B.; Becker, J. S. Int. J. Mass Spectrom. 2011, 307, 112−122. (33) Becker, J. S.; Niehren, S.; Matusch, A.; Wu, B.; Hsieh, H. F.; Kumtabtim, U.; Hamester, M.; Plaschke-Schlutter, A.; Salber, D. Int. J. Mass Spectrom. 2010, 294, 1−6. (34) Sussulini, A.; Becker, J. S. Talanta 2015, 132, 579−582. (35) Weiss, A. U.S. Patent US 7035004 B2, 2006. (36) Leica laser microdissection system, http://www.leicamicrosystems.com/products/light-microscopes/life-science-research/ laser-microdissection/details/product/leica-lmd7000/ (accessed Oct. 1, 2015). (37) Matyjaszewski, K., Mö ller, M., Eds. Polymer Science: A Comprehensive Reference; Elsevier B.V.: Amsterdam, 2012. (38) Azar, M. R.; Acar, N.; Erwin, V. G.; Barbato, G. F.; Morse, A. C.; Heist, C. L.; Jones, B. C. Pharmacol., Biochem. Behav. 1998, 59, 637− 640. (39) Ovchinnikova, O. S.; Tai, T.; Bocharova, V.; Okatan, M. B.; Belianinov, A.; Kertesz, V.; Jesse, S.; Van Berkel, G. J. ACS Nano 2015, 9, 4260−4269. (40) Cahill, J. F.; Kertesz, V.; Ovchinnikova, O. S.; Van Berkel, G. J. J. Am. Soc. Mass Spectrom. 2015, 26, 1462−1468.

11121

DOI: 10.1021/acs.analchem.5b03293 Anal. Chem. 2015, 87, 11113−11121