Laser Capture Microdissection MALDI for Direct Analysis of Archival

After fixing, the slides were placed in deionized water, stained with hematoxylin stain (Richard Allan Scientific) for 30 s, immersed again in water f...
0 downloads 0 Views 288KB Size
Laser Capture Microdissection MALDI for Direct Analysis of Archival Tissue Sucharita H. Bhattacharya,† Anthony A. Gal,‡ and Kermit K. Murray*,† Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803 and Department of Pathology and Laboratory Medicine, Emory University, Atlanta, Georgia 30322 Received July 8, 2002

Abstract: MALDI mass spectra were obtained from cancer cells isolated by laser capture microdissection (LCM) of archived tissue. Frozen human lung tissue from adenocarcenoma and squamous cell carcenoma cases were cut into 5 to 15 µm thick sections, stained with hematoxylin and dehydrated. Cancer cells were isolated by LCM, mixed with matrix solution, and deposited on a MALDI target for mass spectrometric analysis. For comparison with LCM isolated cells, tissue sections were placed directly on the MALDI target without microdissection. Tissue sections frozen in optimal cutting temperature (OCT) solution and cut into 8 µm thick sections gave the best performance with direct MALDI analysis. Between 15 and 20 peaks were observed in the mass region between 1000 and 4000 Da, and roughly half of these peaks were common to either squamous cells or adenocarcenoma. Additional peaks were observed in the non-LCM mass spectra and these may result from biomolecules in the healthy tissue. When compared to fresh tissue, both LCM and non-LCM archived tissue produced fewer peaks, possibly due to degradation of the biomolecules in the archived tissue. Keywords: LCM • MALDI • cancer • tissue

Introduction MALDI is emerging as a valuable method for the direct analysis of biological analytes in tissue.1 Several attributes of MALDI make it an attractive choice for the analysis of biological samples including tolerance of salts and impurities and excellent mass resolution and accuracy. Direct tissue analysis by MALDI has been used to investigate peptides present in pond snail single nerve cells,2 neuroendocrine cells from the brain of the mollusk,3 peptides from sea slug atrial gland,4,5 insulin from rat pancreas,6 and hormone peptides in rat pituitary gland.6 In the tissue sections of rat pancreas and pituitary gland, biochemical imaging has been performed at a spatial resolution between 5 and 25 µm.1,6,7 MALDI tissue analysis is accomplished by cutting the tissue of interest into sections approximately 10-µm thick using a * To whom correspondence should be addressed. E-mail: kkmurray@ lsu.edu. † Department of Chemistry, Louisiana State University. ‡ Department of Pathology and Laboratory Medicine, Emory University. 10.1021/pr025547m CCC: $25.00

 2003 American Chemical Society

cryostat.1 After sectioning, the tissue can be analyzed either directly or after blotting onto a polymer membrane. For direct analysis,1-3 the tissue section is placed on the MALDI sample target and dried under vacuum or in a desiccator. A matrix solution is then deposited on the tissue and allowed to dry. A drawback of the direct analysis method is the difficulty in obtaining adequate analyte signal intensity, due to the high concentration of salts and other contaminants. Tissue blotting is an alternative to direct tissue analysis that is performed by holding the tissue section in contact with a polymer membrane for several minutes, which results in transfer of some of the biomolecules to the polymer.8 After the tissue is removed, the membrane is washed with cold water and a matrix solution is added and allowed to dry. The membrane is then attached to the MALDI target using conductive tape. Although the tissueblotting method requires extra sample preparation steps, the sample target can be washed to remove impurities while maintaining the spatial arrangement of molecules in the tissue.1 With both the direct and the blot tissue MALDI methods, biomolecules in the mass range between 2 and 100 kDa have been observed and in certain cases, peptide and proteins responsible for the peaks have been identified.1,8 In cases where the contributing biomolecule cannot be identified, it is often possible to distinguish between tissue types based on the mass spectral “fingerprint.”8 When tissue identification from molecular masses alone is difficult, structural mass spectrometry methods such as MALDI post-source decay6 and electrospray tandem mass spectrometry2 can be used. Laser capture microdissection (LCM) is a technique for separating and isolating specific areas from tissue sections with a spatial resolution down to the size of a single cell and is particularly useful for isolating homogeneous cell populations for proteomic and genomic study.9-12 The LCM device consists of an inverted light microscope in which tissue sections are placed between a thermoplastic film and a glass slide. Regions of tissue selected by the user are irradiated with an infrared laser, which melts the thermoplastic film and causes the cells to adhere. Multiple regions can be selected, after which the film and specific regions of interest are lifted free of the surrounding tissue. LCM has been coupled with direct MALDI analysis of cells isolated from fresh breast cancer tissue.13 Analyte peaks in the m/z range from 1 to 60 kDa were observed but contributing biomolecules could not be identified due to the limited mass resolution. However, a correlation was found between the mass spectral patterns and the cell type. MALDI analysis of LCM separated cells from fresh tissue shows much promise; however, LCM has the potential to be Journal of Proteome Research 2003, 2, 95-98

95

Published on Web 11/22/2002

technical notes

LCM of Archival Tissue

even more widely applicable if it can be performed on archival specimens. The tissue obtained from pathology workups is typically frozen and stored for months to years and increasingly utilized in many clinical and pathological investigations. The tissue is typically frozen in formalin, ethanol or a mixture of poly(ethylene glycol), and poly(vinyl alcohol). The latter solution is available under the trade name Optical Cutting Temperature Medium (OCT). Although all of the above solutions help to preserve proteins, nucleic acids, and other biomolecules, OCT medium is particularly useful in limiting biomolecule cross-linking and degradation.14 The goal of this work was to investigate direct MALDI-MS of LCM isolated cells from archived frozen tissue. This is a methodological study of frozen samples of lung cancer adenocarcinoma or squamous cell carcinoma without any specific clinical or pathological modifiers. Tissue specimens containing two histologically dissimilar types of lung cancer, squamous cell carcinoma, and adenocarcinoma were subjected to LCM and the resulting cancer cells were analyzed in a commercial MALDI mass spectrometer. Details of the optimized procedure are given below.

Experimental Section Archived lung tissue specimens were obtained from the Department of Pathology at Emory University (approval IRB74799). The frozen tissue samples were anonymously obtained without respect to a specific identifiable patient. The tissue used for the experiments was preserved as follows: fresh lung tissue containing lung cancer was subdivided into 8 to 15 mm pieces and completely covered with OCT (Tissue-Tek, Sakura, Finetek, Torrence, CA) medium. The dissected tissue was immediately placed in dry ice for 3 to 5 min, then covered with aluminum foil and stored indefinitely at -80 °C. Formalin (Fisher Scientific, Houston, TX) and ethanol (Richard Allan Scientific, Kalamazoo, MI) were also used as freezing medium for the storage of tissues. When ready for sectioning and staining, the tissue was removed from the foil and placed on an aluminum stub for cutting in the cryostat. Repeated thawing of the tissue causes ice-crystal formation, hence tissue was slowly thawed in the cryostat at -20 °C. The tissue was cut into 5-15 µm thick sections, placed on a glass slide, and fixed by immersion in 70% ethanol for 10 s to preserve tissue morphology for staining. After fixing, the slides were placed in deionized water, stained with hematoxylin stain (Richard Allan Scientific) for 30 s, immersed again in water for 30 s, then washed in a pH 8.5 ammonia “bluing solution” (Richard Allan Scientific) for 30 s to increase the intensity of the stain. The tissue was then washed in 70% ethanol and stained with eosin (Richard Allan Scientific) for 90 s. Dehydration was accomplished with two 10 s washes in 95% ethanol and two 10 s washes with 100% ethanol. Finally, the slide was placed in xylene (Richard Allan Scientific) for 30 s and allowed to dry. An analogous tissue preparation without LCM was performed after staining and dehydration for comparison with the LCM preparation. Stained frozen section slides were subjected to LCM using a Pixcell 100 instrument (Arcturus Engineering, Mountain View, CA). The tissue was mounted on a glass microscope slide that was then placed on the LCM microscope stage. A LCM cap with thermoplastic film was mounted above the tissue section. The laser beam was adjusted to 7.5 µm diameter to allow microdissection of groups of identified cancer cells. After microdissection, the LCM cap with the microdissected cells was placed 96

Journal of Proteome Research • Vol. 2, No. 1, 2003

Figure 1. Optical microscope photographs of stained and dehydrated lung tissue sections containing squamous cell carcinoma (a) prior to LCM and (b) after LCM. Arrows indicate the regions containing the cancer cells before and after LCM.

on a centrifuge tube containing the MALDI matrix solution and the tube was inverted for 5 min. A 3 µL aliquot of the solution containing matrix and cells was deposited on the MALDI target and allowed to dry. A saturated matrix solution of 2,5dihydroxybenzoic acid (Sigma, St. Louis, MO) was used for the results shown below, but a sinapinic acid matrix gave comparable results. Tissue that was not subjected to LCM was transferred to the MALDI target and the matrix solution was added and the solvent allowed to dry. Mass analysis was done using a Bruker Reflex III MALDI-TOF mass spectrometer in reflectron mode.

Results and Disussion Initial experiments were performed to identify the optimum tissue thickness for LCM analysis. Tissue sections between 5 and 15 µm thick in steps of 1 µm were prepared and placed on microscope glass slides, stained, and dehydrated. Archived tissue sections 8 µm thick were found to be a good compromise between speed of dehydration, which favors the thin sections, and ease of LCM, which favors thick sections. Mass spectra were obtained from mixtures of matrix and OCT, hematoxylin, and eosin stain and the spectra indicate that these compounds contribute to interference only in the m/z range below 1000. Mass spectra obtained from intact tissue frozen in OCT were compared to intact tissue frozen in formalin or ethanol. The

technical notes

Figure 2. Direct MALDI mass spectra of LCM isolated (a) squamous cell carcinoma and (b) adenocarcinoma cells.

tissue frozen in formalin or ethanol resulted in mass spectra with more low mass peaks and fewer peaks in the high mass region compared to mass spectra from tissue frozen in OCT. Archived frozen lung tissue from five cases of squamous cell carcinoma and four cases of adenocarcinoma were studied. The tissue had been frozen in OCT and stored an average of 17 months, with the oldest tissue being 36 months old and the newest tissue, 2 months old. Two LCM and one non-LCM tissue preparations were performed for each tissue sample: 15 sections for squamous cell carcinoma tissue and 8 sections for adenocarcinoma tissue. Approximately 1000 cells were collected for each section. Optical microscope photographs of 12 µm thick tissue sections are shown in Figure 1. Figure 1a shows a tissue section of intact squamous cell carcinoma that has been stained and dehydrated prior to LCM. The arrows indicate the cancerous cells, which have large and nonuniform nuclei in comparison to the normal nuclei. The surrounding connective tissue and cytoplasm is stained pink due to the hematoxylin and eosin stain and can be easily distinguished from the cancerous cells. Figure 1b shows a second tissue section of squamous cell carcinoma after LCM has been performed. The arrows indicate the areas where the cancerous cells have been removed. Direct LCM MALDI mass spectra obtained from 8 µm sections of tissue frozen in OCT are shown in Figure 2. Figure 2a shows a representative mass spectrum from LCM isolated squamous cell carcinoma cells, and Figure 2b is a representative mass spectrum from adenocarcinoma cells. In all of the mass spectra obtained, between 15 and 20 peaks were observed in the m/z range from 1000 to 4000. No peaks were observed above m/z 4000 for any of the tissue samples. Approximately

Bhattacharya et al.

Figure 3. Direct MALDI mass spectra of intact (non-LCM) tissue sections for (a) squamous cell carcenoma and (b) adenocarcinoma.

half of the peaks observed were common to mass spectra for the particular tissue type. For the squamous cell carcinoma, the m/z values for the common peaks were 1329, 1479, 2112, 2867, and 3400 and for adenocarcinoma, the m/z values for the common peaks were 1738, 2083, 2511, 2650, 3376, 3447, and 3809. MALDI mass spectra were obtained from non-LCM squamous cell and adenocarcinoma tissue that was sectioned and stained for comparison with the LCM mass spectra. Mass spectra from 8 µm thick non-LCM sections are shown in Figure 3. Figure 3a shows a mass spectrum from squamous cell carcenoma tissue and Figure 3b shows a mass spectrum from adenocarcinoma tissue. The non-LCM tissue mass spectra are in general similar to the LCM mass spectra: peaks that were common to the LCM mass spectra were also found in the corresponding non-LCM mass spectra. In addition, there were several peaks found in mass spectra from both types of nonLCM tissue at m/z 4161, 7615, 7940 and 10 030. These mass spectral features were not seen in LCM MALDI mass spectra and possibly result from MALDI ionization of nonneoplastic cellular constituents.

Conclusions A method has been developed for MALDI analysis of frozen archival tissue in which the tissue is cut into 8 µm thick sections, stained, dehydrated, and the cells of interest isolated by LCM. After addition of matrix, the cells can be analyzed by MALDI. It was found that tissue frozen and stored in OCT medium resulted in mass spectra with more mass spectral peaks and fewer interferences compared to tissue treated with Journal of Proteome Research • Vol. 2, No. 1, 2003 97

technical notes

LCM of Archival Tissue

formalin or ethanol. For the analysis of archived human lung tissue containing adeoncarcinoma and squamous cell carcinoma, it was found that approximately half of the observed peaks in the mass range between 1000 and 4000 Da were indicative of either squamous cell carcinoma or adeoncarcinoma and could be used as a fingerprint of those cancers. In general, direct MALDI mass spectra of archived tissue, both LCM and non-LCM (intact), have fewer peaks than spectra from fresh tissue, either direct,1 blotted,8 or when compared to LCM MALDI of fresh tissue.13 The observation of fewer peaks may indicate a partial degradation of the biomolecules in the archived tissue.

Acknowledgment. The authors thank Dr. Jan Pohl and the Emory University Microchemical Facility for help in obtaining the MALDI mass spectra. This work is supported by NSF Grant Nos. CHE0096457 and CHE0196568. References (1) Todd, P. J.; Schaaff, T. G.; Chaurand, P.; Caprioli, R. M. J. Mass Spectrom. 2001, 36, 355-369. (2) Dreisewerd, K.; Kingston, R.; Geraerts, W. P. M.; Li, K. W.; Int. J. Mass Spectrom. Ion Processes 1997, 169/170, 291-299.

98

Journal of Proteome Research • Vol. 2, No. 1, 2003

(3) Jimenez, C. R.; Li, K. W.; Dreisewerd, K.; Spijker, S.; Kingston, R.; Bateman, R. H.; Burlingame, A. L.; Smit, A. B.; van Minnen, J.; Geraerts, W. P. M. Biochemistry 1998, 37, 2070-2076. (4) Li, L.; Garden, R. W.; Romanova, E. V.; Sweedler, J. V. Anal. Chem. 1999, 71, 5451-5458. (5) Garden, R. W.; Moroz, L. L.; Moroz, T. P.; Shippy, S. A.; Sweedler, J. V. J. Mass Spectrom. 1996, 31, 1126-1130. (6) Caprioli, R. M.; Farmer, T. B.; Gile, J. Anal. Chem. 1997, 69, 47514760. (7) Stoeckli, M.; Chaurand, P.; Hallahan, D. E.; Caprioli, R. M. Nat. Med. 2001, 7 (4), 493-496. (8) Chaurand, P.; Stoeckli, M.; Caprioli, R. M. Anal. Chem. 1999, 71, 5263-5270. (9) Emmert-Buck, M. R.; Bonner, R. F.; Smith, P. D.; Chuaqui, R. F.; Zhuang, Z.; Goldstein, S. R.; Weiss, R. A.; Liotta, L. A. Science 1996, 274 (5289), 998-1001. (10) Bonner, R. F.; Emmert-Buck, M. R.; Cole, K.; Pohida, T.; Chuaqui, R.; Goldstein, S.; Liotta, L. A. Science 1997, 278 (5342), 14811483. (11) Simone, N. L.; Bonner, R. F.; Gillespie, J. W.; Emmert-Buck, M. R.; Liotta, L. A. Trends Genet. 1998, 14 (7), 272-276. (12) Wulfkuhle, J. D.; McLean, K. C.; Paweletz, C. P.; Sgroi, D. C.; Trock, B. J.; Steeg, P. S.; Petricoin, E. F. Proteomics 2001, 1 (10), 12051215. (13) Palmer-Toy, D. E.; Sarracino, D. A.; Sgroi, D.; Le Vangie, R.; Leopold, P. E. Clin. Chemistry 2000, 46 (9), 1513-1516. (14) Anderson, L.; Seilhamer, J. Electrophoresis 1997, 18, 533-537.

PR025547M