Direct Acquisition of Matrix-assisted Laser Desorption/Ionization Time-of-Flight Mass Spectra from Laser Capture Microdissected Tissues

Recent studies have demonstrated the usefulness of laser capture microdissection (LCM) in gene expression studies of pure cell populations derived from healthy and diseased tissues (1)(2)(3). However, mRNA expression does not necessarily correlate with protein abundance or predict posttranslational modifications (4). Currently, a few groups have reported details of protein expression profiling of cells captured by LCM, using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) followed by peptide mass fingerprinting of selected proteins (5)(6). Although such an approach represents the gold standard for global analysis of protein expression, 2D-PAGE often is impractical for diagnostic applications. We describe a rapid and sensitive method to obtain an abridged protein expression profile from microdissected cells by the direct acquisition of matrix-assisted laser desorption/ionization time-of-flight (MALDI-ToF) mass spectra from LCM transfer films. LCM uses an infrared laser to transiently and focally melt an ethylene vinyl acetate (EVA) thermoplastic film applied to a tissue section. During the infrared laser pulse, the EVA transfer film fuses with the underlying cells, which remain adherent when the film is separated from the tissue section. In most applications, the transfer film is subsequently immersed in a lysis buffer to extract soluble molecules from the captured cells and their microenvironment, and the lysate is used in further analysis. Our approach is to obtain MALDI-ToF mass spectra directly from microdissected cells of interest. MALDI-ToF allows the simultaneous molecular mass characterization of a broad variety of biological molecules up to 100 kDa or higher. Mass measurements of 1 pmol of an analyte present in a 1-μL specimen are routine for pure analytes and simple mixtures. MALDI-ToF matrix solution is applied directly to the cell-adherent surface of LCM films. The matrix solution extracts soluble molecules from the captured cells and their microenvironment, which co-crystallize with the matrix as the solvent evaporates. Excitation with an ultraviolet laser pulse disrupts specific chemical bonds within matrix molecules, rapidly vaporizing the crystals, ionizing and liberating the incorporated biomolecules into the gas phase for mass analysis. Mass spectra are readily collected from the crystal-coated film using a standard MALDI-ToF mass spectrometer. Notably, the MALDI-ToF spectra do not require elaborate sample preparation, e.g., chromatography, affinity capture, or other purification step. Furthermore, no instrumentation beyond the LCM device and a MALDI-ToF mass spectrometer is required. The EVA substrate is not ionized, nor does it suppress ionization. Matrix co-crystallized LCM samples are stable for at least 2 weeks under reduced pressure. All tissue utilized for this study was obtained from a modified radical mastectomy specimen from a single patient. Tissue in excess of what was necessary for diagnostic purposes was obtained <15 min after removal from the patient, embedded in cryostat mounting medium (TISSUE-Tek O.C.T.; Sakura Finetek U.S.A.), and frozen in liquid nitrogen. The tissues were sectioned at 8 μm in a cryostat, mounted on uncoated glass slides, and stored immediately at −80 °C. At the time of microdissection, slides containing frozen sections were immediately fixed in 700 mL/L ethanol for 30 s and stained with hematoxylin and eosin, followed by 5-s dehydration steps in 700 mL/L, 950 mL/L, and undiluted ethanol and a final 5-min dehydration step in xylene. Once air-dried, the sections were laser microdissected with a PixCell I LCM system (Arcturus Engineering). Approximately 2500 morphologically normal breast stroma cells, normal breast epithelial cells, malignant invasive breast carcinoma cells, and malignant metastatic (to an axillary lymph node) breast carcinoma cells were laser captured, according to the standard protocol of Emmert-Buck et al. (1). Each population was estimated to be >98% homogeneous as determined by microscopic visualization of the captured cells. After LCM, the EVA transfer film was gently peeled from the CapSureTM (Arcturus Engineering) and bisected; one-half of the film was attached to a standard Bruker mass spectrometer stainless steel target (Bruker Daltonics), with the cell-adherent surface up. Common rubber cement worked well and produced no detectable peaks above 200 Da. Approximately 1 μL of MALDI-ToF matrix solution, consisting of saturated sinapinic acid in 300 mL/L acetonitrile containing 1 mL/L trifluoroacetic acid, was uniformly distributed over the transfer film. Acetonitrile concentrations as high as 500 mL/L yielded similar spectra. An alternative matrix that has worked well is 10 g/L 2,5-dihydroxybenzoic acid in 300–500 mL/L acetonitrile containing 1 mL/L trifluoroacetic acid. MALDI-ToF spectra were collected in positive-ion mode, using a Bruker Reflex III or Biflex III MALDI-ToF mass spectrometer (Bruker Daltonics) in linear mode. A deflecting voltage was applied to filter out peaks <2000 m/z. A spectrum consisted of the sum of 10 subspectra from different locations on the film, each collected with 50 laser shots. Raw spectra were smoothed using the nine-point Savitzky-Golay algorithm and baseline-subtracted using Bruker data reduction software. As a challenge in selecting and analyzing pure cell populations, normal breast stroma, normal ductal epithelium, ductal carcinoma in situ, and invasive ductal carcinoma were microdissected from a single frozen section of human breast (see Fig. 1 ).Distinct spectra were obtained from 1250 cells from each of the four cell types. The stromal cells revealed several prominent peaks in the 4500–7000 Da range, which were attenuated or absent in the spectra from cells of epithelial derivation. A series of high-mass peaks from 45 to 60 kDa distinguished the invasive carcinoma spectrum from that of normal epithelium and from the stromal spectrum as well. The carcinoma in situ spectrum had features intermediate between the normal epithelium and the invasive tumor. Several common features were also evident in each spectrum. A frozen section of ductal carcinoma metastatic to a lymph node was also available from the same patient. As evidence for the reproducibility of the method, the spectra derived from the invasive and metastatic ductal carcinoma were virtually identical (see Fig. 1 ). Are these putative markers artifacts of LCM? The infrared-induced fusion causes local heating to 90 °C (7), but subsequent analysis with 2D-PAGE has revealed no noticeable impact on the concentrations of expressed proteins or their integrity (5). Extended case series are required to demonstrate whether these findings are consistent among different patients. Capturing cells from one frozen section requires <30 min, and proceeding from transfer film to 20–50 spectral peaks is equally fast. This compares favorably with the many hours of laser-capture time and days of electrophoresis needed to visualize several hundred spots on 2D-PAGE from LCM (5)(6). Moreover, the mass resolution of the 5- to 10-kDa MALDI-ToF peaks obtained from the LCM cells was 1–5%, which is better than the than the typical 5–10% resolution obtained with a 2D-PAGE system. Refinements in matrix deposition, biomolecule extraction, and desalting techniques may improve signal quality. Although the mass/charge ratio (m/z) of a mass speak is insufficient to unambiguously establish the identity of any given marker, we show that spectral patterns are reproducible to the cell type and provide a credible indicator of lineage. Specific marker identification through PAGE and peptide mass fingerprinting remains an option but is not a diagnostic prerequisite. These spectral markers are not necessarily etiologically related to the disease process but may reflect generalized changes in cell physiology. Four cell populations microdissected from a single frozen section of human breast: normal stromal cells, normal epithelial cells, ductal carcinoma in situ, and invasive ductal carcinoma (rows 1–4, respectively). Photomicrographs illustrate the tissue before LCM as well as the captured cells adherent to the transfer films (columns 1 and 2, respectively). In addition, a population of ductal carcinoma cells metastatic to a lymph node from the same patient were microdissected from a separate frozen section (row 5). Here column 2 illustrates the lymph node cells remaining in the frozen section after LCM. The mass spectra acquired directly from these transfer films are shown for mass/charge (m/z) of 3500–18 000 and 20 000–70 000 (columns 3 and 4, respectively). This work was funded in part by NIDDK SBIR Grant 1R43 DK54118-01. D.E.P-T. was funded in part by NIH:NCI Training Grant T32-CA09216. We thank Paul Kowalski of Bruker Daltonics for access to the mass spectrometers.