direct tissue analysis by matrix-assisted laser desorption...

MaisOtyTcctim

*

†

S

A

0©

S

Direct Tissue Analysis byMatrix-Assisted Laser DesorptionIonization Mass Spectrometry:Application to Kidney Biology

Kristen D. Herring,* Stacey R. Oppenheimer, PhD,* and Richard M. Caprioli, PhD†

Summary: Direct tissue analysis using matrix-assisted laser desorption ionization mass spec-trometry (MALDI MS) provides in situ molecular analysis of a wide variety of biologicalmolecules including xenobiotics. This technology allows measurement of these species intheir native biological environment without the use of target-specific reagents such asantibodies. It can be used to profile discrete cellular regions and obtain region-specific images,providing information on the relative abundance and spatial distribution of proteins, peptides,lipids, and drugs. In this article, we report the sample preparation, MS data acquisition andanalysis, and protein identification methodologies used in our laboratory for profiling/imagingMS and how this has been applied to kidney disease and toxicity.Semin Nephrol 27:597-608 © 2007 Elsevier Inc. All rights reserved.Keywords: MALDI, direct-tissue analysis, imaging mass spectrometry, LCM, drug-toxicity,renal cancer

sm(emaac[aTu

icoaaa2mst

atrix-assisted laser desorption ionizationmass spectrometry (MALDI MS), intro-duced in the late 1980s,1,2 has become

n enabling analytic technology to study proteinsn biological systems because of its molecularpecificity and high-throughput capabilities.3,4

ne of the more recent developments to thisechnology is its application to direct tissue anal-sis for both molecular profiling and imaging.5

he technology allows for the analysis of mole-ules in tissues without the need for target-spe-ific reagents such as antibodies or the need forissue homogenization, thereby maintaining thentegrity of the tissue sample and allowing assess-

ent of molecular spatial distribution.

Department of Biochemistry, Vanderbilt University Medical Center, Nash-ville, TN.

Mass Spectrometry Research Center, Vanderbilt University Medical Center,Nashville, TN.

upported in part by National Institutes of Health/National Institute of GeneralMedical Sciences (NIGMS) GMS8008-08.

ddress reprint requests to Richard M. Caprioli, Vanderbilt University, 9160Medical Research Bldg III, 465 21st Ave S, Nashville, TN 37232-8575.E-mail: [email protected]

270-9295/07/$ - see front matter
c2007 Elsevier Inc. All rights reserved. doi:10.1016/j.semnephrol.2007.09.002
eminars in Nephrology, Vol 27, No 6, November 2007, pp 59

The principles of MALDI MS have been de-cribed.2,6-9 In brief, the analyte of interest isixed with an energy-absorbing compound

matrix) on a MALDI target plate. As the solventvaporates, the analyte co-crystallizes with theatrix. Inside the mass spectrometer, molecules

re desorbed from the sample by irradiation withn ultraviolet laser. In this process, analytes be-ome protonated and primarily give rise toM�H]� ions that subsequently are measuredccording to their mass-to-charge ratio (m/z).ypically, the time-of-flight (TOF) analyzers aresed to measure m/z values of ions (Fig. 1).8,9

MALDI MS can be used to profile moleculesn a sample (ie, analyze a small number of dis-rete spots scattered throughout the sample),r image a sample (ie, systematically rastercross a tissue section to produce an orderedrray of spots). In the latter, the laser performsraster over the tissue surface in a predefined-dimensional array or grid, generating a fullass spectrum at each grid coordinate. Mass

pectra acquired in this manner display ions inhe m/z range of 500 to more than 100,000,
orresponding to many hundreds of different
7-608 597

mnatcdm

ld

taft

FEdc

F(thn

598 K.D. Herring, S.R. Oppenheimer, and R.M. Caprioli

olecules present in the tissue. The coordi-ates of the irradiated spots are used to gener-te 2-dimensional ion density maps, or images,hat represent individual m/z values with theirorresponding intensities (Fig. 2). Likewise,rug analysis in tissue can be accomplished byonitoring the protonated drug or its metabo-

igure 1. MALDI-TOF. A sample is irradiated with a brnergy is absorbed by the matrix and transferred to theown a field-free drift tube. Ions collide with a detector aalibration converted to m/z.

Optical Image

Ion Image

A Optical ImageOptical Image

Ion Image

A

igure 2. Images generated by imaging MALDI MS. (A) Obottom) of a normal rat kidney section. Shown are 3 ionshe cortex, green: m/z 8,560 in the outer medulla, blue:ematoxylin-eosin stained (top) and ion images (bottom
ontumor tissue. Shown are 2 overlaid ions (red: m/z 11,090 loc
ites and their corresponding fragments at eachiscrete coordinate of such an array.

Direct tissue analysis by MALDI has been usedo detect drugs and their metabolites10-13 as wells intact proteins and peptides14,15 directlyrom tissue. Profiling/imaging MS has been usedo study molecular aspects of cancer, providing

er pulse and molecules are desorbed from the surface.te, which becomes protonated and then is acceleratedend of this tube and their TOF is measured and through

Tumor

Non-Tumor[Cortex]

Tumor Margin

Area Imaged

Optical Image

Ion Image

Tumor

Non-Tumor[Cortex]

Tumor Margin

Area Imaged

Tumor

Non-Tumor[Cortex]

Tumor Margin

Area Imaged

Optical Image

Ion Image

l hematoxylin-eosin stained (top) and overlaid ion imagesented in color to indicate localization (pink: m/z 8,451 in965 localized primarily in the inner medulla). (B) Opticalhuman ccRCC tissue section with tumor and adjacent

ief lasanalyt the

BB

pticarepres

m/z 4,) of a

alized to tumor, green: m/z 4,563 in nontumor cortex).

tatmar

SP

DtioMooomhp

ncfefccak�2cMt

twbsptenmm1rs

pioimwe

fltcp�tciT3aTwpcTvm

tfignrtdhtsvjtdhtdpIa

MALDI MS and kidney biology 599

umor-specific markers as well as diagnostic-nd prognostic-specific markers.16,17 In this ar-icle, we describe the current profiling/imagingethodologies used in our laboratory and their

pplication to glomerulosclerosis, drug-inducedenal toxicity, and renal cancer.

AMPLE PREPARATION FORROFILING AND IMAGING TISSUES BY MS

irect MS analysis usually is performed on tissueshat have been excised and immediately frozenn liquid nitrogen to maintain tissue morphol-gy and to minimize molecular degradation.aintaining the integrity of the tissue through-ut this process maximizes the informationbtained in the analyses by ensuring that theriginal 3-dimensional structure is not compro-ised and the molecular species monitoredave not been altered as a result of the samplerocurement and preparation process.Frozen tissues are sectioned in a cryostat, if

ecessary, using a small amount of optimumutting temperature media on the cutting blockor support. Preferably, tissue should not bembedded in the polymer media because sur-ace contamination with polymer caused by theutting process tends to reduce ionization effi-iency. A cryostat temperature between �15°Cnd �25°C generally is optimal for sectioningidney tissue. Sections generally are cut at 12m thickness, but may range between 10 and0 �m depending on the application.4,10 Afterutting, sections are thaw-mounted onto aALDI target plate and washed and/or allowed

o dry in a vacuum desiccator.It is advantageous to wash tissue sections

hat are rich in salts and other contaminants asell as hemoglobin before matrix depositionecause these species can contribute to a highpectral baseline and generalized signal sup-ression. Because of the kidney’s role in filtra-ion of blood, washing kidney sections removesxcess hemoglobin and salts and enhances sig-al quality. The conventional tissue washingethod involves 2 separate 70% ethanol sub-ersions for 20 to 30 seconds followed by 10 to

5 seconds in 95% 200-proof ethanol (or 100%eagent grade ethanol). Ethanol is a known tis-
ue fixative in histology18 and delocalization of m
roteins appears not to be significant; however,t is noted that proteins soluble in these aque-us solutions may be removed during the wash-

ng step. Previous studies have indicated mini-al protein loss during this process, but theashing procedure should be confirmed for

ach tissue type.19,20

Selection of the matrix/solvent combinationor direct tissue analysis depends on the molecu-ar weight, hydrophobicity, and salt content ofhe analyte. Typically, 3,5-dimethoxy-4-hydroxy-innamic acid (sinapinic acid) is favorable forroteins (molecular weight � 2 kDa), and-cyano-4-hydroxycinnamic acid (CHCA) is op-

imal for peptides (500-2,000 Da). Small-mole-ule analysis, including drugs and lipids, usuallys performed with 2,5-dihydroxybenzoic acid.he optimal matrix concentration range is 10 to0 mg/mL sinapinic acid for protein analysisnd 10 to 20 mg/mL CHCA for peptide analysis.he usual matrix solvent is 50% acetonitrileith 0.1% trifluoroacetic acid (TFA). More non-olar solvents such as methanol or isopropanolan be used for more hydrophobic molecules.he optimal matrix/solvent combination mayary between tissue types and analytes andust be assessed for each study.Various methods of matrix deposition on

hin tissue sections have been used success-ully. For tissue profiling, matrix can be depos-ted manually either with a pipette or pulledlass capillary. Syringe pumps provide an alter-ative, more automated approach with highereproducibility and smaller matrix spot diame-ers. Robotic deposition offers superior repro-ucibility, smaller matrix spot diameters, and aigher throughput platform, and also allows forhe ability to perform histology-directed analy-es21 and imaging. Two types of robotic de-ices, an acoustic spotter19 and a chemical ink-et printer,22 have been used successfully forissue profiling and imaging. General matrixeposition guidelines for both robotic typesave been described,19,22 but for each tissueype, method optimization is recommended toetermine the number of matrix drops andasses required to obtain high-quality spectra.n some cases matrix seeding may be desired toid in the crystallization process.19 The current
inimum spot-to-spot spacing achievable with

c1rmssdsscldtelt

H

RhMiorbmmahd

ml(rcttmepsifi

M

TdmsaTfmbmtasvw

Fcsi


ommercial robotic spotters is in the range of00 to 200 �m. For applications that require aesolution higher than 100 �m, spray coatingay be used. A spray coating device may be as

imple as a glass thin layer chromatographypray nebulizer that generates and sprays fineroplets of matrix under slight positive pres-ure, or automated spray devices. The matrixolution must wet the tissue surface to ensureo-crystal formation between matrix and ana-yte. For manual spraying methods, a 1-minuteelay between passes provides sufficient dryingime. Multiple passes are necessary to coat thentire tissue, but overcoating can suppress ana-yte signal. The matrix coverage can be moni-ored as necessary under a microscope.

istology-Directed MS Analysis

egion specificity is important for analysis ofeterogeneous tissue types such as kidney. OneALDI-compatible method to accomplish this

s to stain tissues using cresyl violet stain orther suitable stains (H&E stains give poor MSesults), allowing histology and MS analysis toe performed on the same tissue section withinimal interference in signal quality.23 Thisethod usually requires that tissue sections be

pplied to a conductive glass slide to obtainistopathologic detail. A new method, histology-irected MS analysis, uses digital imaging to

Optical Image of Tissue on MALDI plate

Optical Image of H&E Stained Tissue

Optical Image of H&E Stained Tissue Overlay

Optical Image of Tissue on MALDI plate

Optical Image of H&E Stained Tissue

Optical Image of H&E Stained Tissue Overlay

igure 3. Histology-directed MS sample preparation foellular areas (spots) of interest on a stained tissue sectiopotter for matrix deposition on a co-registered serial sec
n black for visualization) to give mass spectra.
erge traditional histopathology, or hematoxy-in-eosin tissue staining, with tissue profilingFig. 3). Details of this methodology have beeneported.21 In brief, a pathologist systematicallyan select areas on a stained tissue section withhe usual cellular specificity. By co-registration,he spot coordinates are transferred to a roboticatrix spotter and then to the mass spectrom-

ter. This allows the profiles and other data-rocessing algorithms such as statistical analy-is to be mapped directly onto the histologicmage and correlation between molecular pro-les and histopathology.

S Analysis

he mass spectral acquisition methods used inirect tissue analysis are performed in an auto-ated fashion. When profiling discrete matrix

pots, mass spectra are obtained by performingraster over the area of interest on the tissue.he number of spectra obtained and averaged

rom a single spot on tissue depends on howany shots can be acquired at one location

efore the signal is depleted and how big theatrix spot is. Most experiments average be-

ween 250 and 500 spectra on a single spotbout 150 �m in diameter. Mass spectral acqui-ition from spray-coated images does not in-olve averaging shots at different locationsithin a pixel. Instead, each ablated location

ical Image of Tissue after Spotting

25000

Inte

nsity

2000 m/z

Mass Spectrumical Image of Tissue

after Spotting

25000

Inte

nsity

2000 m/z 25000

Inte

nsity

2000 m/z 25000

Inte

nsity

2000 m/z

Mass Spectrum

e analysis. A pathologist systematically selects discretee coordinates of the spots are transferred to the roboticith subsequent laser irradiation of these spots (outlined

OptOpt

r tissun. Thtion w

rssc

aatcttrsTibttabnqb

D

Tetlsa(tatrpga

P

Itlpiotut

tstsfMuo

afptatptittltjbd

Oa

Rtrazpmastd(ppdimspp


epresents one pixel and the number of laserhots within that pixel depends on how manyhots can be obtained before the signal signifi-antly decreases.

Spectral preprocessing is performed on thecquired data to reduce interexperimentalnd intra-experimental variance. Such varia-ions may be produced from the ionization pro-ess, background noise, and calibration offsetshat result from biological or sample prepara-ion differences. Mass spectra are processed byemoving background noise, normalizing inten-ities, and performing a final mass calibration.he latter is used primarily for profiling exper-

ments in which multiple sample groups areeing compared, enhancing the efficacy of sta-istical algorithms to determine biological pat-erns and changes in these patterns. Imagingpplications typically do not require the recali-ration step, but background subtraction andormalization significantly enhance the imageuality. Illustrations of these processes haveeen presented.24

ata Analysis

he large, complex data sets produced by thesexperiments require robust bioinformatics toolso assess spectral patterns and to decipher mo-ecular species that differentiate one group ofamples from another. MS data from each groupre compared to determine spectral featurespeaks) that are significantly different betweenhe 2 subject groups. The experimental designnd biostatistical methods used are critical tohe success of such analyses. Many of the algo-ithms used to decipher the vast amount ofroteomic information were used initially forenomic microarray experiments and have beendapted for proteomic data set analyses.25-28

rotein Identification

dentification of statistically relevant mass spec-ral peaks is necessary to gain insight into bio-ogical processes. There are 2 approaches torotein identification. The top-down approach

nvolves ionization and gas-phase fragmentationf the protein of interest inside the mass spec-rometer,29 whereas the bottom-up approachses MS to identify peptides obtained from pro-
ease digestion of that protein, often in a mix- t
ure of other proteolytic fragments.30 The re-ulting mass spectra are searched againstheoretical protein/peptide databases for corre-ponding sequence patterns. Searches are per-ormed using conventional algorithms such asASCOT and Sequest.31,32 When possible, man-al validation of reported identifications is rec-mmended to reduce false-positive results.

Commonly, the bottom-up strategy is usednd this involves homogenization of the tissueollowed by reverse-phase liquid chromatogra-hy (RP-LC) separation, where the eluate con-inuously is collected into fractions. Fractionsre vacuumed to dryness, dissolved in 40% ace-onitrile with 0.1% TFA, spotted onto a MALDIlate, and analyzed for the fractions containinghe peptides of interest. Based on the complex-ty of the fraction, one enzymatically can digesthe entire fraction or further separate the mix-ure by 1-dimensional gel electrophoresis fol-owed by excision and enzymatic digestion ofhe band of interest. Peptides then are sub-ected to further separation and fragmentationy RP-LC–tandem MS analysis and subsequentatabase searches.

n-tissue Digestionnd Protein Identification

ecent work has shown that digestion and iden-ification of proteins may be coupled with di-ect tissue analysis.22 The technique involvesutomatically depositing a spotted array of en-ymatic solution onto the tissue at room tem-erature. After hydrolysis, MALDI matrix (25g/mL 2,5-dihydroxybenzoic acid in 50% meth-

nol, 0.5% TFA) is deposited onto the array forubsequent MALDI MS analysis. Figure 4 illus-rates an image obtained from on-tissue trypticigestion of a clear cell renal cell carcinomaccRCC) section, where a unique distribution ofeptides is found in the tumor. This procedurerovides the ability to simultaneously assess theistribution of proteins and their correspond-

ng peptides by imaging MS (IMS) to obtain aore complete molecular signature of a tissue

ection. In addition, the in situ identification ofroteins requires less time than conventionalrotein identification strategies and the ability
o compare a given protein image to the image

otvmlo

A

GFsrtknTvggnatMrf

fnn

scapIMTsmnLatoptbua(

FtatiF


f its subsequent peptides increases identifica-ion confidence. This process is best used foralidation (ie, where the presence of one orore specific proteins is needed to be estab-

ished). For unknown proteins, the bottom-upr top-down procedure is recommended.

pplication to Kidney

lomerulosclerosisocal segmental glomerulosclerosis (FSGS), orcarring of the glomeruli, is found in scatteredegions of the kidney and impairs kidney func-ion. Its etiology and biological causes are un-nown, making treatment complicated and diag-osis difficult without a biopsy examination.33

he focal segmental characteristic of FSGS pro-okes the question of whether the nonscleroticlomeruli in sclerotic tissue already are pro-rammed to become sclerotic or whether theseonsclerotic glomeruli have less proscleroticctivation and therefore are more receptive toherapy. By using direct tissue analysis byALDI MS, Xu et al34 used a glomerulosclerotic

at model to determine if proteomic profiles

A

igure 4. On-tissue tryptic digestion. Kidney tissue conthickness and transferred to a MALDI target plate. The srray of 120 picoliter droplets of trypsin solution placed ao hydrolysis, matrix (CHCA) was deposited directly onmage of spotted trypsin and matrix array. (B) MALDI ioigure generated by M. Reid Groseclose.

rom sclerotic glomeruli could be differentiated a

rom nonsclerotic glomeruli, and, second, ifonsclerotic glomeruli have a prosclerotic phe-otype at the protein level.

To achieve the selective isolation and analy-is of glomeruli from tissue, laser capture mi-rodissection (LCM) was used. This techniquellows for the dissection of specific cells or cellopulations from a heterogeneous tissue section.n the FSGS study, LCM was combined withALDI MS for tissue protein profiling (Fig. 5).he LCM procedure requires that a thin tissueection (�8 �m thick) is mounted on a glassicroscope slide and dehydrated in serial etha-

ol dilutions followed by xylene rinses. TheCM procedure uses a narrow laser beam (di-meter, �7-30 �m) to irradiate a heat-sensitiveransparent polymer film, resulting in adhesionf the film to the section. The cells bound to theolymer are separated from the section whenhe polymer is removed. This polymer mem-rane is mounted onto a MALDI target platesing conductive double-sided tape. The cellsre spotted with very small volumes of matrixsinapinic acid) using a thin pulled capillary or

m/z 1988

Tumor

Margin

BNon-Tumor

ccRCC and adjacent nontumor was sectioned at 12 �mwas washed with ethanol and spotted with an orderedm lateral resolution using a robotic spotter. Subsequentspots and a peptide image was obtained. (A) Optical

ge of tryptic peptide m/z 1,988 localized to the tumor.

ainingectiont 250 �these

n ima

robotic matrix deposition device.

wrcrngprtksi

nem

DTdrsrmt

Fwa and


The analysis of LCM-captured cells combinedith MALDI MS profiling and statistical analysis

esulted in the proteomic differentiation andlassification of normal, nonsclerotic, and scle-otic glomeruli. Statistical analysis revealed thatonsclerotic glomeruli in the presence of pro-ressive renal scarring also have an alteredroteomic profile that is more similar to scle-otic than normal glomeruli. For example,hymosin �4 was identified as one of severaley differentiators between nonsclerotic andclerotic glomeruli. Its expression also was

igure 5. LCM preparation for MALDI MS analysis. A thith a laser beam through a thermoplastic polymer film cttached to the cap. The cap is affixed to a MALDI plate

ncreased in these samples as compared with t

ormal glomeruli, suggesting its role as anarly activator and indicator of proscleroticechanisms.

rug Nephrotoxicityhe kidney is particularly susceptible to toxicamage because of its large surface area and itsole in blood filtration. Current methods of as-essing nephrotoxicity, such as monitoring se-um creatinine levels and changes in cellularorphology, often fail to detect changes until

he onset of toxic nephropathy. Characteriza-

e section is placed on a glass slide where it is irradiatedhen the cap is removed, cells that were irradiated remainmatrix is applied for subsequent MS analysis.

in tissuap. W

ion of toxicity markers with profiling and im-

ahe

pSpmsftlmsit3aauwest(feTwitr

ia

RAgwoelscatitTtat

litcpsoftal

Ftfia n.35


ging MALDI MS may provide a robust andigh-throughput method to detect toxicity inarly lead compounds.

A pilot study was performed to examine theroteomic changes induced by gentamicin (Sigma,t. Louis, MO), an antibiotic known to causeroximal tubule damage in kidneys. Experi-ental details for this study have been de-

cribed.35 Briefly, male Wistar rats were treatedor 7 consecutive days with 100 mg/kg/d gen-amicin or vehicle and killed 24 hours after theast treatment. Kidney sections either were

anually spotted for profiling or automaticallypotted with an acoustic robotic spotter formaging. Mass spectra from each region con-ained hundreds of peaks in the m/z range of,500 to 25,000. Some signals were confined tonatomic regions, such as the cortex, medulla,nd papilla, giving each region its own molec-lar signature. Notably, a peak at m/z 12,959as found to be statistically significant in differ-

ntiating control and treated animals. IMS re-ults indicated that this feature was localized tohe cortex, the site of damage in these animalsFig. 6). This protein was extracted directlyrom the cortex of 3 tissue sections and thextracts were pooled and separated by RP-LC.he LC fraction containing the peak of interestas subjected to RP-LC–tandem MS analysis and

dentified as transthyretin, a marker of nutri-ional status36-38 and a ligand of the megalin

igure 6. Imaging results from gentamicin study. Ratsheir kidneys excised. Differential protein expression wasgure shows the image analysis of m/z 12,959 differentiaverage peaks in the spectrum. Adapted with permissio

eceptor,39 which binds gentamicin.40,41 The t

dentification was confirmed by Western blotnd immunohistochemistry.

enal Tumor Marginsnother study currently ongoing is the investi-ation of the molecular distributions in ccRCCith respect to molecular tumor margins. Onef the major concerns in clinical oncology isnsuring complete tumor removal to minimizeocal recurrence and ensure long-term patienturvival.42-44 Depending on the tumor type, re-urrence can occur a few months to a few yearsfter removal of the primary tumor, suggestinghat there are underlying molecular processesn the remaining normal tissue that go undetec-ed using current histopathologic techniques.he discovery of new molecular markers for

hese processes has the potential to providedditional complementary approaches to iden-ify abnormal tissue environments.

It is important to be able to measure the mo-ecular characteristics of the tumor and surround-ng tumor margin of ccRCC. The radical nephrec-omy procedure is the standard treatment forcRCC, but it is excessive for small tumors oratients in whom the tumor is confined to aolitary functioning kidney, has arisen bilaterally,r when the remaining kidney could be underuture threat from renal deficiencies. The par-ial nephrectomy procedure maximizes themount of kidney spared but presents a chal-enge for clinicians because of concerns over

osed with gentamicin once daily for 7 days, killed, andared between control and dosed kidneys using IMS. Thisressed in the cortex of dosed tissue with corresponding

were dcomplly exp

umor involvement of the surgical margin.45-47

FhatmTum


igure 7. Tumor margin analysis by MALDI MS. (A) Optical image of section with matrix spots with regions of interestighlighted. (B) Optical image of the corresponding hematoxylin-eosin–stained section. (C) The top figure shows themplitude of 6 m/z values plotted as a function of distance (of corresponding matrix spot analyzed) from the histologicumor margin in one representative tissue sample. Each line represents a different m/z value, each identified as aember of the mitochondrial electron transport chain. These are smoothed trend lines of the original scatterplot data.

he bottom figure is a plot of the first derivative of the trend line. Both plots illustrate that the features represented arenderexpressed in the tumor and in the margin normal tissue. These features remain underexpressed until approxi-
ately 4,000 �m past the histologic tumor border.

itthshftMeAemfs(tttmonewpnBiapc

C

DtmlTiadtmnuai

R

1

1

1

1

1

1

1

1


Proteomic technologies such as profiling/maging MALDI MS may allow for the elucida-ion of tumor margin characteristics to fur-her understand tumor spread. Early resultsave identified molecular signatures thathow atypical tissue significantly beyond theistologic margin. MALDI-TOF MS was per-ormed on 34 ccRCC biopsy specimens con-aining tumor and adjacent nontumor tissue.atrix arrays were deposited on the tissue, cov-

ring the tumor and nontumor tissue (Fig. 7A).utomated MS acquisition was performed onach of the matrix spots. After processingass spectra, statistical analysis was per-

ormed on the regions as follows: tumor ver-us normal and margin normal versus normalFig. 7B). Many of the differences betweenissue on the normal side of the margin andissue distant normal also were observed be-ween tumor and nearby histologically nor-al tissue. Protein identification revealed that

ne of the families contributing to these ab-ormal characteristics was mitochondriallectron transport proteins. These proteinsere underexpressed in the tumor as com-ared with normal and underexpressed in theear-normal compared with normal (Fig. 6C).ased on the results as well as other stud-

es,48-56 it is probable that increased glycolysist the expense of mitochondrial oxidativehosphorylation plays a significant role incRCC tumor spread into the normal tissue.

ONCLUSIONS

irect tissue analysis by MALDI MS is an impor-ant technology for assessing the localization ofolecular species and for revealing the under-

ying molecular signatures indicative of disease.he molecular species identified in these exper-

ments can provide insight into mechanismsnd etiology of disease. Applications to kidneyiseases have just begun, but the potential ofhis technology to help unravel the complexolecular processes involved are extraordi-

ary. With increasing advances in the field, thetility of this technology will continue to evolvend will play a fundamental role in understand-
ng kidney biology.
EFERENCES1. Hillenkamp F, Karas M, Beavis RC, Chait BT. Matrix-

assisted laser desorption/ionization mass spectrome-try of biopolymers. Anal Chem. 1991;63:1193A-203A.

2. Karas M, Hillenkamp F. Laser desorption ionization ofproteins with molecular masses exceeding 10,000daltons. Anal Chem. 1988;60:2299-301.

3. Chaurand P, Caprioli RM. Direct profiling and imagingof peptides and proteins from mammalian cells andtissue sections by mass spectrometry. Electrophoresis.2002;23:3125-35.

4. Schwartz SA, Reyzer ML, Caprioli RM. Direct tissueanalysis using matrix-assisted laser desorption/ioniza-tion mass spectrometry: practical aspects of samplepreparation. J Mass Spectrom. 2003;38:699-708.

5. Caprioli RM, Farmer TB, Gile J. Molecular imaging ofbiological samples: localization of peptides and proteinsusing MALDI-TOF MS. Anal Chem. 1997;69:4751-60.

6. Karas M, Gluckmann M, Schafer J. Ionization in matrix-assisted laser desorption/ionization: singly charged mo-lecular ions are the lucky survivors. J Mass Spectrom.2000;35:1-12.

7. Zenobi R, Knochenmuss R. Ion formation in MALDImass spectrometry. Mass Spectrom Rev. 1998;17:337-6.

8. Wiley WC, McLaren IH. Time-of-flight mass spectrom-eter with improved resolution. Rev Sci Instruments.1955;26:1150-1157.

9. Cotter RJ. The new time-of-flight mass spectrometry.Anal Chem. 1999;71:445A-51A.

0. Khatib-Shahidi S, Andersson M, Herman JL, GillespieTA, Caprioli RM. Direct molecular analysis of whole-body animal tissue sections by imaging MALDI massspectrometry. Anal Chem. 2006;78:6448-56.

1. Reyzer ML, Caprioli RM. MALDI-MS-based imaging ofsmall molecules and proteins in tissues. Curr OpinChem Biol. 2007;11:29-35.

2. Troendle FJ, Reddick CD, Yost RA. Detection of phar-maceutical compounds in tissue by matrix-assistedlaser desorption/ionization and laser desorption/chemical ionization tandem mass spectrometry witha quadrapole ion trap. J Am Soc Mass Spectrom.1999;10:1315-21.

3. Reyzer ML, Hsieh Y, Ng K, Korfmacher WA, CaprioliRM. Direct analysis of drug candidates in tissue bymatrix-assisted laser desorption/ionization mass spec-trometry. J Mass Spectrom. 2003;38:1081-92.

4. Stoeckli M, Chaurand P, Hallahan DE, Caprioli RM.Imaging mass spectrometry: a new technology for theanalysis of protein expression in mammalian tissues.Nat Med. 2001;7:493-6.

5. Chaurand P, Cornett DS, Caprioli RM. Molecular im-aging of thin mammalian tissue sections by mass spec-trometry. Curr Opin Biotechnol. 2006;17:431-6.

6. Schwartz SA, Weil RJ, Johnson MD, Toms SA, CaprioliRM. Protein profiling in brain tumors using mass spec-trometry: feasibility of a new technique for the analysisof protein expression. Clin Cancer Res. 2004;10:981-7.

7. Schwartz SA, Weil RJ, Thompson RC, Shyr Y, Moore
JH, Toms SA, et al. Proteomic-based prognosis of brain

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4


tumor patients using direct-tissue matrix-assisted laserdesorption ionization mass spectrometry. Cancer Res.2005;65:7674-81.

8. Srinivasan M, Sedmak D, Jewell S. Effect of fixativesand tissue processing on the content and integrity ofnucleic acids. Am J Pathol. 2002;161:1961-71.

9. Aerni HR, Cornett DS, Caprioli RM. Automated acous-tic matrix deposition for MALDI sample preparation.Anal Chem. 2006;78:827-34.

0. Xu BJ, Caprioli RM, Sanders ME, Jensen RA. Directanalysis of laser capture microdissected cells byMALDI mass spectrometry. J Am Soc Mass Spectrom.2002;13:1292-7.

1. Cornett DS, Mobley JA, Dias EC, Andersson M, Ar-teaga CL, Sanders ME, et al. A novel histology-directedstrategy for MALDI-MS tissue profiling that improvesthroughput and cellular specificity in human breastcancer. Mol Cell Proteomics. 2006;5:1975-83.

2. Groseclose MR, Andersson M, Hardesty WM, CaprioliRM. Identification of proteins directly from tissue: insitu tryptic digestions coupled with imaging massspectrometry. J Mass Spectrom. 2007;42:254-62.

3. Chaurand P, Schwartz SA, Billheimer D, Xu BJ, Cre-celius A, Caprioli RM. Integrating histology and imag-ing mass spectrometry. Anal Chem. 2004;76:1145-55.

4. Norris JL, Cornett DS, Mobley JA, Andersson M, SeeleyEH, Chaurand P, et al. Processing MALDI mass spectrato aid biomarker discovery and improve mass spectralimage quality. Int J Mass Spectrom. 2007;260:212-21.

5. Tusher VG, Tibshirani R, Chu G. Significance analysisof microarrays applied to the ionizing radiation re-sponse (vol 98, pg 5116, 2001). Proc Natl Acad SciU S A. 2001;98:10515.

6. Naiman DQ. Random data set generation to supportmicroarray analysis. Methods Enzymol. 2006;411:312-25.

7. Hedenfalk I, Duggan D, Chen Y, Radmacher M, BittnerM, Simon R, et al. Gene-expression profiles in hereditarybreast cancer. N Engl J Med. 2001;344:539-48.

8. Storey JD, Tibshirani R. Statistical significance forgenomewide studies. Proc Natl Acad Sci U S A. 2003;100:9440-5.

9. Reid GE, McLuckey SA. ‘Top down’ protein charac-terization via tandem mass spectrometry. J Mass Spec-trom. 2002;37:663-75.

0. Wysocki VH, Resing KA, Zhang Q, Cheng G. Massspectrometry of peptides and proteins. Methods.2005;35:211-22.

1. Hirosawa M, Hoshida M, Ishikawa M, Toya T. MASCOT:multiple alignment system for protein sequences basedon three-way dynamic programming. Comput Appl Bio-sci. 1993;9:161-7.

2. Link AJ, Eng J, Schieltz DM, Carmack E, Mize GJ, MorrisDR, et al. Direct analysis of protein complexes usingmass spectrometry. Nat Biotechnol. 1999;17:676-82.

3. Ponticelli C, Villa M, Banfi G, Cesana B, Pozzi C, PaniA, et al. Can prolonged treatment improve the prog-nosis in adults with focal segmental glomerulosclero-
sis? Am J Kidney Dis. 1999;34:618-25.
4. Xu BJ, Shyr Y, Liang X, Ma LJ, Donnert EM, RobertsJD, et al. Proteomic patterns and prediction of glo-merulosclerosis and its mechanisms. J Am Soc Neph-rol. 2005;16:2967-75.

5. Meistermann H, Norris JL, Aerni HR, Cornett DS,Friedlein A, Erskine AR, et al. Biomarker discovery byimaging mass spectrometry: transthyretin is a biomar-ker for gentamicin-induced nephrotoxicity in rat. MolCell Proteomics. 2006;5:1876-86.

6. Bernstein LH, Ingenbleek Y. Transthyretin: its re-sponse to malnutrition and stress injury. Clinical use-fulness and economic implications. Clin Chem LabMed. 2002;40:1344-8.

7. Ingenbleek Y, Young V. Transthyretin (prealbumin)in health and disease: nutritional implications. AnnuRev Nutr. 1994;14:495-533.

8. Lasztity N, Biro L, Nemeth E, Pap A, Antal M. Proteinstatus in pancreatitis—transthyretin is a sensitive bi-omarker of malnutrition in acute and chronic pancre-atitis. Clin Chem Lab Med. 2002;40:1320-4.

9. Sousa MM, Norden AG, Jacobsen C, Willnow TE,Christensen EI, Thakker RV, et al. Evidence for therole of megalin in renal uptake of transthyretin. J BiolChem. 2000;275:38176-81.

0. Verroust PJ, Birn H, Nielsen R, Kozyraki R, Chris-tensen EI. The tandem endocytic receptors megalinand cubilin are important proteins in renal pathology.Kidney Int. 2002;62:745-56.

1. Moestrup SK, Cui S, Vorum H, Bregengard C, BjornSE, Norris K, et al. Evidence that epithelial glycopro-tein 330/megalin mediates uptake of polybasic drugs.J Clin Invest. 1995;96:1404-13.

2. Balch GC, Mithani SK, Simpson JF, Kelley MC. Accu-racy of intraoperative gross examination of surgicalmargin status in women undergoing partial mastec-tomy for breast malignancy. Am Surg. 2005;71:22-8.

3. Looser KG, Shah JP, Strong EW. The significance of“positive” margins in surgically resected epidermoidcarcinomas. Head Neck Surg. 1978;1:107-11.

4. Nathan CO, Amirghahri N, Rice C, Abreo FW, Shi R,Stucker FJ. Molecular analysis of surgical margins inhead and neck squamous cell carcinoma patients.Laryngoscope. 2002;112:2129-40.

5. Dechet CB, Zincke H, Sebo TJ, King BF, LeRoy AJ,Farrow GM, et al. Prospective analysis of computer-ized tomography and needle biopsy with permanentsectioning to determine the nature of solid renalmasses in adults. J Urol. 2003;169:71-4.

6. Nguyen TT, Parkinson JP, Kuehn DM, Winfield HN.Technique for ensuring negative surgical margins dur-ing laparoscopic partial nephrectomy. J Endourol.2005;19:410-5.

7. Sengupta S, Zincke H. Lessons learned in the surgicalmanagement of renal cell carcinoma. Urology. 2005;66:36-42.

8. Acker T, Fandrey J, Acker H. The good, the bad and theugly in oxygen-sensing: ROS, cytochromes and prolyl-
hydroxylases. Cardiovasc Res. 2006;71:195-207.

4

5

5

5

5

5

5

5


9. Gatenby RA, Gawlinski ET. The glycolytic phenotype incarcinogenesis and tumor invasion: insights throughmathematical models. Cancer Res. 2003;63:3847-54.

0. Giaccia AJ, Simon MC, Johnson R. The biology ofhypoxia: the role of oxygen sensing in development,normal function, and disease. Genes Dev. 2004;18:2183-94.

1. Guzy RD, Schumacker PT. Oxygen sensing by mito-chondria at complex III: the paradox of increasedreactive oxygen species during hypoxia. Exp Physiol.2006;91:807-19.

2. Haase VH. The VHL/HIF oxygen-sensing pathway andits relevance to kidney disease. Kidney Int. 2006;69:
1302-7.
3. Hervouet E, Demont J, Pecina P, Vojtiskova A,Houstek J, Simonnet H, et al. A new role for the vonHippel-Lindau tumor suppressor protein: stimulationof mitochondrial oxidative phosphorylation complexbiogenesis. Carcinogenesis. 2005;26:531-9.

4. Raghunand N, Gatenby RA, Gillies RJ. Microenviron-mental and cellular consequences of altered bloodflow in tumours. Br J Radiol. 2003;76:S11-22.

5. Simonnet H, Alazard N, Pfeiffer K, Gallou C, Beroud C,Demont J, et al. Low mitochondrial respiratory chaincontent correlates with tumor aggressiveness in renalcell carcinoma. Carcinogenesis. 2002;23:759-68.

6. Warburg O. The metabolism of tumors. London: Con-
stable Press; 1930.

direct tissue analysis by matrix-assisted laser desorption...

Documents