mass spectrometry maldi imaging of colon cancer biomarkers: a new diagnostic paradigm

55ISSN 1752-036310.2217/17520363.3.1.55 © 2009 Future Medicine Ltd Biomarkers Med. (2009) 3(1), 55–69

Review

Mass spectrometry MALDI imaging of colon cancer biomarkers: a new diagnostic paradigm

Colorectal cancer (CRC), is the second-leading cause of cancer-related deaths in the USA, affecting both men and women. Current projections show little or no change since the publication of a morbidity and mortality study in 2005. The projected number of new cases for 2008 is 154,000, and the projected number of CRC cancer deaths for 2008 is 53,000. The standard diagnostic paradigm is based on histopathology of either biopsy or surgical specimens. This article suggests a new paradigm for colon cancer diagnosis and staging using matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI IMS or IMS). IMS may identify potential tumors in normal tissue of cancer patients and predict those cancer patients who are at risk for recurrent cancer.

keywords: biomarker, colorectal cancer, imaging, LCMs, liquid chromatography mass spectrometry, MALdI IMs, mass spectrometry, matrix-assisted laser desorption/ionization imaging mass spectrometry

Paul H Pevsner1†, Jonathan Melamed2, Tiffany Remsen1, Alexander Kogos1, Fritz Francois3, Paul Kessler1, Arnold Stern1 & Sury Anand4

†Author for correspondence: 1Department of Pharmacology, New York University School of Medicine, New York, NY, USA Tel.: +1 212 263 0233; [email protected] 2Department of Pathology, New York University School of Medicine, New York, NY, USA 3Department of Medicine, Division of Gastroenterology, New York University School of Medicine, New York, NY, USA 4Department of Medicine, Division of Gastroenterology, The Brooklyn Hospital Center, Brooklyn, NY, USA

Colorectal cancer (CRC) is the second-leading cause of cancer-related deaths in the USA, affect-ing both men and women. In an extensive study, Song demonstrated that fecal DNA testing every 5 years appeared both diagnostically effective and cost-effective compared with no screening, but inferior to other strategies such as fecal occult blood testing and colonoscopy. Fecal DNA test-ing could decrease the national burden if it could improve adherence with screening, particularly where the capacity to perform screening colonos-copy is limited [1]. In a follow-up study, the potential national impact of widespread screen-ing on clinical outcomes (with screening uptake of 75%) were as follows: the incidence of CRC could decrease by 17–54% to as few as 66,000 new cases per year, and deaths could decrease by 28–60% to as few as 23,000 per year, depend-ing on the strategy. The nonscreening total annual national CRC-related expenditures were estimated at US$8.4 billion. Screening would decrease expenditures for care by US$1.5–4.4 billion but total expenditures would increase to US$9.2–15.4 billion. Screening colonos-copy every 10 years would require 8.1 million colonoscopies per year including surveillance, with other strategies requiring 17–58% as many colonoscopies. The current national endoscopic capacity, as recently estimated, may be adequate to support widespread use of screening colonos-copy in the steady state. The impact of emerg-ing tests on colonoscopy demand will depend on the extent to which they replace screening

colonoscopy or increase screening uptake in the population. With improved screening uptake, total colonoscopy demand would increase, even assuming substantial use of virtual colonoscopy. The conclusion was that despite savings in care, widespread screening is unlikely to be cost sav-ing and may increase national expenditures by US$0.8–2.8 billion per year with conventional tests [2].

Current projections show little or no change since the publication of that study in 2005. The projected number of new cases for 2008 is 154,000, and the projected number of CRC cancer deaths for 2008 is 53,000 [101]. Even with surveillance, the prevalence and mortality of CRC cancer increases every year.

The standard diagnostic paradigm for CRC is based on histopathology of either biopsy or surgical specimens. This article suggests a new paradigm for CRC diagnosis and staging, a combination of histopathology and mass spec-trometry. CRC protein biomarkers have been identified by imaging mass spectrometry (IMS) in both tumors and histologically normal satellite tissue obtained from both endoscopic biopsies and surgical specimens. IMS may identify poten-tial tumors in normal tissue of cancer patients and predict those cancer patients who are at risk for recurrent cancer by their protein biomarkers. These protein biomarkers could become the basis for a low-cost, high-compliant, high-throughput mass spectrometry stool screening test that could have both high sensitivity and high specificity.

THEME: OncOlOgy

k.rowland

Text Box

For reprint orders, please contact: [email protected]

Biomarkers Med. (2009) 3(1)56 future science group

Review Pevsner, Melamed, Remsen et al.

Field defects or field cancerizationIn 1953, Slaughter and Southwick introduced the concept of field cancerization in head and neck tumors. They described this as an area of epi-thelium that has been preconditioned by an as yet unknown carcinogenic agent [3]. Braakhuis suggested a genetic explanation for ‘field cancer-ization’, the carcinogenesis model in which the

development of a field with genetically altered cells plays a central role. In the initial phase, a stem cell acquires genetic alterations and forms a ‘patch’ or clonal unit of altered daughter cells. These patches can be recognized based on muta-tions in tumor suppressor protein, TP53, and have been reported for head and neck, lung, skin and breast cancer. The conversion of a patch into

100

90

80

70

60

50

40

30

20

10

0

% In

ten

sity

1500 1550 1600 1650 1700 1750 1800

1[c].Mass/Charge

1480.80 {r3574}1517.73 {r6028}

1511.79 {r3030}

1529.75 {r5401}

1586.79 {r6213}

1598.77 {r6983} 1672.83 {r6399}

1653.00 {r3500}

1800.93 {r6...}

1790.90 {r6515}

1804.88 {r....}

1794.88 {r46..}

1769.00 {r3595}

Figure 1. Matrix-assisted laser desorption/ionization time-of-flight spectrum of colorectal adenocarcinoma and satellite normal tissue obtained from the on-tissue digest. Peptide mass fingerprinting from the imaging mass spectrometry image demonstrated the same two proteins, gi|119592539 hCG1787564 (Homo sapiens) Mass: 57590, and gi|119592490 hCG2040674 (Homo sapiens) Mass: 108178, in the colon adenocarcinoma and in the histopathologically normal satellite tissue. The Mascot search including peptides is shown in the Supplemental Material as an HTML.

Figure 2. Histologic section of colorectal cancer tumor stained with hematoxylin and eosin, 1 µ thick, micrograph 40×. Note the normal columnar epithelium of colon lumen (arrows). The arrowheads outline the edge of the colon adenocarcinoma.

www.futuremedicine.com 57future science group

Mass spectrometry MALDI imaging of colon cancer biomarkers Review

an expanding field is the next logical and criti-cal step in epithelial carcinogenesis. Additional genetic alterations are required for this step, and by virtue of its growth advantage, a proliferat-ing field gradually displaces the normal mucosa. In the mucosa of the head and neck, as well as the esophagus, such fields have been detected with dimensions of more than 7 cm in diam-eter. They are usually not detected by routine diagnostic techniques. Ultimately, clonal diver-gence leads to the development of one or more tumors within a contiguous field of preneoplastic cells. An important clinical implication is that fields often remain after surgery of the primary tumor and may lead to new cancers. These are currently designated by clinicians as a ‘second primary tumor’ or ‘local recurrence’, depending on the exact site and time interval. The develop-ment of an expanding preneoplastic field appears to be a critical step in epithelial carcinogenesis with important clinical consequences. The diag-nosis and treatment of epithelial cancers should not only be focused on the tumor but also on the field from which it developed [4]. Buckley described human chorionic gonadotropin (hCG) in 60 cases of adenocarcinoma of the colon and none in normal control tissue [5]. Subsequently, hCG was described by immunohistochemistry in infiltrating rectosigmoid carcinoma and in some of the histologically normal satellite tissue adjacent to carcinomas [6,7].

Fearon presented a genetic model for CRC tumor genesis. He proposed that CRC tumors appear to arise as a result of the mutational acti-vation of oncogenes coupled with the mutational inactivation of tumor suppressor genes (which predominate). Second, mutations in at least four to five genes are required for the formation of a malignant tumor. Fewer changes suffice

for benign tumor genesis. Third, although the genetic alterations often occur according to a preferred sequence, the total accumulation of changes, rather than their order with respect to one another, is responsible for determining the tumor’s biologic properties. Fourth, in some cases, mutant tumor suppressor genes appear to exert a phenotypic effect even when present in

Figure 3. Imaging mass spectrometry image of colorectal adenocarcinoma and satellite normal tissue obtained from the on-tissue digest described by the spectrum in Figure 1. The right outlined area corresponds to the tumor in the histology section, and the smaller left outlined region corresponds to the histologically normal satellite tissue. The identical human chorionic gonadotropin isoforms were identified in both loci. The smaller locus represents a satellite ‘field defect’ with potential for cancerization. The difference in grayscale between the tumor and normal satellite tissue is more obvious in the tumor than in the outlined area of satellite tissue with the same protein densities.

Box 1. sample A1 colon polyp liquid chromatography (electrospray) mass spectrometry.

GI � |30311, cytokeratin 18 (424 AA) (Homo sapiens).

GI � |28336, mutant β‑actin (β´‑actin) (Homo sapiens).

� *GI|113394, alcohol deydrogenase β subunit. Role in colon cancer.

GI � |4501881, α1 actin precursor (Homo sapiens).

GI � |4757756, annexin A2 isoform 2 (Homo sapiens).

� *GI|229751, chain A, α‑ferrous‑carbonmonoxy, β‑cobaltous‑deoxy hemoglobin (T state). Hyperinsulinemia > obesity > colorectal cancer.

GI � |178027, α‑actin.

GI � |63055057, hypothetical protein LOC345651 (Homo sapiens).

GI � |6650826, PRO2044 (Homo sapiens).

GI � |229752, chain B, α‑ferrous‑carbonmonoxy, β‑cobaltous‑deoxy hemoglobin (T state).

� *GI|40886941, haemoglobin β (Homo sapiens). Colon cancer proliferation.

GI � |29446, unnamed protein product (Homo sapiens).*Putative CRC proteins. The tandem mass spectrometry spectra of these proteins are shown in Figures 10–14.



the heterozygous state; thus, some tumor sup-pressor genes may not be ‘recessive’ at the cellular level. The general features of this model may be applicable to other common epithelial neo-plasms. He further proposed that an epigenetic

change, such as hypomethylation, could con-tribute to instability in the tumor cell genome and alter the rate at which genetic alterations, such as allelic losses, occur. He observed that the carcinomatous regions were derived from (and not simply adjacent to) the adenomatous regions. This was proven in several cases by the finding that the identical ras gene mutation was present in both regions. In all cases, however, the carcinomatous regions contained at least one alteration not found in the adenomatous region [8]. Scalmati described changes in the normal mucosa of patients with CRC carci-noma. The normal-appearing rectal mucosa of the affected patients showed a pattern of cell proliferation that clearly differs from that of normal people. The proliferative compartment of normal CRC mucosa of individuals at low risk for CRC cancer, and of normal rodents, is located in the lower part of the colonic crypts. The total labeling index (i.e., labeled cells ver-sus total cells) of low-risk controls and familial polyposis patients does not differ significantly. However, the patients show a shift of the pro-liferative compartment toward the top of the crypts. Thus, the main proliferative alteration noted in the colonic mucosa of these patients is not an increase in overall cell proliferation labeling index, but a lumenward displacement of the zone of active cell proliferation in the crypts [9].

Aberrant crypt foci (ACF) are putative pre-neoplastic lesions of colonic cancer. A diet con-taining 0.2% cholic acid, a reported colonic tumor promoter, has two prominent effects on the growth of ACF [10]:

Figure 4. Hematoxylin- and eosin-stained histologic section of histopathologically benign-appearing colon polyp. Note the central fibrotic contracted zone (opposing arrows). Note artery at lower growth margin of polyp infiltrating normal tissue (lower arrow). Tissue section micrograph 40×.


� *GI|229751, chain A, α‑ferrous‑carbonmonoxy, β‑cobaltous‑deoxy hemoglobin (T). Hyperinsulinemia>obesity>colon cancer.

� *GI|40886941, hemoglobin β (Homo sapiens). Colon cancer proliferation.

GI � |29446, unnamed protein product (Homo sapiens).

GI � |58177625, chain B, T‑to‑T (high) quaternary transitions in human hemoglobin: βy35f oxy (2mmIhp, 20% Peg) (1 test set).

GI � |61679764, chain B, T‑to‑T (high) quaternary transitions in human hemoglobin: βf42a deoxy low‑salt (1 test set).

GI � |179409, β‑globin.

GI � |66473265, β globin chain (Homo sapiens).

GI � |27574235, hemoglobin β, deoxy hemoglobin (A, C:v1m; B,D:v1m, V67w).

� *GI|5453712, galectin 4 (Homo sapiens). 50‑times lower in colon cancer than normal.

GI � |30908859, actin α1 skeletal muscle protein (Homo sapiens).

GI � |3114508, chain A, R state human hemoglobin (α V96w).

GI � |61679768, chain B, T‑to‑T (high) quaternary transitions in human hemoglobin: βf45a deoxy low‑salt (1 test set).

*Putative CRC proteins. The tandem mass spectrometry spectra of these proteins are shown in Figures 10–14.



A reduction in the number of ACF present in n

the colon due to either elimination or remodeling and;

Enhanced growth of remaining ACF to n

colonic cancer.

Increased bcl-2 expression was present in hyperplastic colonic polyps and in the majority of dysplastic lesions, from the earliest precursors through to large adenomas, high-grade flat dys-plasia and adenocarcinoma. Furthermore, bcl-2 expression was frequently abnormal in nondys-plastic epithelium surrounding dysplastic lesions, suggesting that altered expression occurred before the development of morphological dysplasia. Specifically, directly contiguous morphologically nondysplastic epithelium often showed abnor-mal bcl-2 expression throughout the full length of the crypt–villus axis. This expression pattern gradually diminished to involve only the crypt base (the normal pattern of expression), proceed-ing away from malignant or dysplastic lesions. Abnormal bcl-2 immuno reactivity in first, the earliest precursor dysplastic lesions and its per-sistence throughout neoplastic progression and second, contiguous morphologically unaltered nondysplastic epithelium, suggests that bcl-2 alterations occur early during the morphologi-cal and molecular sequence of events leading to gastrointestinal epithelial neoplasia [11].

Garewal proposed an assay for quantitation of bile acid-induced reduction in apoptotic abil-ity of CRC mucosa, which he implied increased

cancer risk. By applying a quantitative bile acid-induced apoptosis assay to CRC mucosal biop-sies, the percentage of apoptosis was found to be significantly reduced in CRC carcinoma patients.

Figure 5. Contiguous section to the H & e histolologic section in Figure 4 applied to MALdI conductive plate. The section was covered with Sinapic acid matrix by sublimation. Tissue section micrograph 40×.


GI � |4501881, α1 actin precursor (Homo sapiens).

� *GI|229751, chain a, α‑ferrous‑carbonmonoxy,k β‑cobaltous‑deoxy hemoglobin (T). Hyperinsulinemia>obesity>colorectal cancer.

� *GI|40886941, hemoglobin β (Homo sapiens). Colon cancer proliferation.


� *GI|5453712, galectin 4 (Homo sapiens). 50‑times lower in colorectal cancer than normal colon.

GI � |178045, gamma‑actin.

GI � |4501885, β‑actin (Homo sapiens).

GI � |88953571, PREDICTED: similar to prostate, ovary, testis expressed protein on chromosome 2 isoform 2 (Homo sapiens).

GI � |113413200, PREDICTED: similar to prostate, ovary, testis expressed protein on chromosome 2 (Homo sapiens).

GI � |89037243, PREDICTED: similar to actin‑like protein (Homo sapiens).

GI � |5901922, cell division cycle 37 proteins (Homo sapiens).

GI � |255317, nuclear autoantigen RA33 = A2 hnRNP homolog (human, peptide partial, 25 amino acid, segment 3 of 4).

GI � |106529, Igκ chain C region (allotype Inv [1, 2]) – human (fragment).


� *GI|37852, vimentin (Homo sapiens). Marker for colon cancer.*Putative CRC proteins. The tandem mass spectrometry spectra of these proteins are shown in Figures 10–14.



Furthermore, he proposed that cells with muta-tions and resistance to apoptosis in ‘normal’ appearing mucosa represented a field defect that could potentially become a carcinoma [12].

A novel antiapoptosis gene, designated ‘Survivin’ is an inhibitor of apoptosis protein – a group of proteins known to inhibit caspases, the proteolytic components of the apoptotic pathway implicated in the control of cell cycle progression. Survivin localizes to the intermito-chondrial membrane space in tumor cells. This localization accelerates tumor genesis in vivo. Survivin does not appear to be involved in the physiological regulation of apoptosis in adult colonic epithelium but is prominently expressed in CRC carcinoma. The mechanisms govern-ing expression of Survivin in malignant cells are presently unclear, but a complex response

to dedifferentiation of normal epithelium appears likely. Survivin mRNA expression was detected in a significantly greater pro-portion of CRC carcinoma than in normal mucosa samples (63.5 vs 29.1%, respectively; p <0.001). In no cases was Survivin mRNA detected in normal tissue when the associated cancer was Survivin-negative. Approximately half the number of Survivin-positive tumors, but none of the Survivin-negative tumors, was associated with normal mucosa that also expressed this gene. These data appear at vari-ance with previous studies that did not detect expression of either Survivin mRNA by in situ hybridization or Survivin protein by immuno-histochemistry in normal epithelium adjacent to tumors. Detection of mRNA transcripts by the more sensitive technique of RT-PCR20 sug-gests that Survivin expression may represent

9074.52 Da

m/z 9074.52SL 1/12D PolyP

Figure 8.9074.52 da. hCG2038976 present in infiltrating colon carcinomas.

8408.5 Da


Figure 7. 8408.5 da. Immunoglobulin‑κ light chain variable region: highly expressed in colon tumors.

6139.5 Da


Figure 6. 6139.5 da. Transmembrane tyrosine kinase receptor: upregulated in colon cancer.

20418.2 Da


Figure 9. 20418.2 da. Ras‑related protein Rap‑2c precursor: expressed in colon cancer.



an ‘intermediate’ biological change identifying histo logically normal mucosa at risk of neoplastic transformation.

Sarela also reported that histologically normal CRC epithelial cells from patients with a history of CRC carcinoma are subject to as yet unidenti-fied influences that result in significantly reduced apoptotic activity compared with that in similar cells from patients with no neoplasia [13].

Anti presented data that showed severe imbal-ance of cell proliferation and apoptosis in the left colon and in the rectosigmoid colon in subjects with a history of large adenomas. The normal segment-to-segment pattern of proliferation was basically preserved in patients with large adenomas, although the rates of proliferation themselves were markedly higher than those found in average-risk subjects. There was also a distinctive alteration in the distribution of pro-liferating cells within the colonic crypts, which was particularly marked in the left colon, and this change implied an additional risk for muta-tional events because highly vulnerable nondif-ferentiated cells at the mouth of the crypts were more exposed to intraluminal factors. Another interesting point was the relationship between changes in colonic epithelial cell kinetics and ACF, which are thought to be the microscopic precursors of both polyps and cancer. In CRC patients, hundreds and hundreds of ACFs have been observed in the normal-appearing mucosa of all segments of the colon. Only a few of these foci actually give rise to adenomas or carcino-mas. The majority of these ACFs never reach macroscopic dimensions and many even regress. Nonetheless, their widespread distribution illus-trates that in patients with CRC cancer, the nor-mal appearing mucosa of the entire colon pres-ents morphological changes with the potential to progress towards more advanced stages in the carcinogenetic pathway. The fact that an exten-sive tract of mucosa was found to harbor glands with a high proliferative propensity and reduced apoptotic activity, supports the ‘field canceriza-tion’ hypothesis, according to which longstand-ing exposure of the mucosa to carcinogens leads to diffuse biological abnormalities throughout large tracts of the colon, which can eventually produce cancers. It is also noteworthy that the severe imbalance between cell proliferation and apoptosis in the high-risk group was localized in the left colon where most colon malignancies develop [14].

Crowley-Weber demonstrated that nicotine, a component of cigarette smoke, and sodium deoxycholate (NaDOC), a cytotoxic bile salt

that increases in concentration in the gastroin-testinal tract after a high-fat meal, induce similar cellular stresses and that nicotine may enhance some of the NaDOC-induced stresses. Nicotine, at 0.8 mM, the very low submicromolar level occurring in the tissues of smokers:

Increases oxidative stress;n

Activates NF-n κB, a redox-sensitive transcription factor;

Activates the 78 kD glucose-regulated protein n

promoter, an indication of endoplasmic reticulum stress;

Induces apoptosis;n

Enhances the ability of NaDOC to activate n

the 153 kD growth arrest and DNA damage promoter, an indication of increased genotoxic stress;

Enhances the ability of NaDOC to activate n

the xenobiotic response element.

This leads to ‘field defects’ in the mucosa of patients with colon carcinoma [15].

Garawal described an early functional change characterizing field defects that seems to occur during progression to sporadic adenocarcinoma of the colon. This change was the loss of the capacity to undergo apoptosis in response to damage. If a cell acquires mutations or epimu-tations that cause apoptosis resistance, this can lead to increased clonogenic survival and con-sequent clonal expansion. In addition, suppres-sion of apoptosis leads to increased mutagen-esis. Other factors that may have similar effects include smoking, dietary alcohol, low intake of calcium or antioxidant vitamins, and the nondi-etary factors of obesity and low physical activity. It was determined that resistance to the induc-tion of apoptosis with a novel ex vivo bioassay, deoxycholate-induced apoptotic index was the most specific of the biomarkers and was present in 59% of the normal-appearing mucosal sam-ples from patients with colon cancer. Apoptosis resistance highly correlated with a low level of differentiation, assessed with Dolichos biflorus agglutinin (DBA) lectin staining. A high frac-tion (>35%) of tissue showing aberrant lectin staining (of nongoblet cells) was present only in the group of patients with colonic neoplasms, also making aberrant lectin reactivity a specific biomarker. The patchy nature of the field defects associated with apoptosis resistance and DBA staining underscores the necessity for multiple biopsies to assess colon cancer risk. The most important aspect of the study is the caveat of



‘patchiness’, since it will probably apply to other biomarkers as well. Although ACF and microadenomas are expected as biomarkers of colon cancer risk when an entire animal colon is examined with methylene blue staining, such lesions are not easily detected in biopsy-sized tissue samples taken from human colons. The plant lectin DBA binds preferentially to termi-nal N-acetyl-galactosamine of mucin and that terminal N-acetyl-galactosamine occurred in well-differentiated goblet cell mucin of the upper colonic crypts. Based on this, DBA can be used as a probe to assess the level of differentiation of goblet cells. Sparse or aberrant DBA reactivity is a sensitive biomarker for colon cancer risk and a high level of aberrant lectin staining in nongoblet cells is specific for high-risk patients. Overall, the results of the study suggest that the live cell bio-assay for apoptotic index and the more practical DBA staining assay on preserved tissue samples are promising biomarkers of colon cancer risk, but multiple samples must be obtained to give a valid indication of risk [16].

Chen compared gene expression levels in mor-phologically normal-appearing colon mucosa from cancer patients with those in mucosa from patients without cancer. One set of data con-sisted of samples from patients with cancer in the recto-sigmoid colon; the other set was from

patients with cancer in the ascending colon. In both studies, the values obtained from the can-cer patients were highly variable, much more so than the corresponding values from the controls. This finding parallels the observations made in APCmin and wild-type mice, although the variation in humans was even higher. Despite the great variability, expression levels for several genes were much higher in some samples from cancer patients than for any samples from con-trols. For example, four of the genes that were significantly upregulated in normal-appearing mucosa of APCmin mice – CXCR2, GRO-α, COX-2 and OPN – were upregulated in normal-appearing mucosa from some cancer patients to levels 50–200-times greater than those in controls. In addition, in some cancer patients, PPAR-α, -δ and -γ were downregulated 50–100-fold. A total of seven genes were significantly upregulated in morphologically normal mucosa from patients with recto-sigmoid cancer rela-tive to controls: MCSF-1, OPN, IL-8, COX-2, CXCR2, p21 and CD44. Two genes – PPAR-δ and -γ – were significantly downregulated. Similar results were obtained for the ascend-ing colon. Six of the seven genes significantly upregulated in recto-sigmoid mucosa were also upregulated in the ascending colon – MCSF-1, OPN, IL-8, COX-2, CXCR2 and CD44 – along

100

90

80

70

60

50

40

30

20

10

0203.0 344.2 485.4 626.6 767.8 909.0

0

867

794y6(+1)

b14(+2), y14(+2)

126.0y15(+2)

706b12(+2)

635633

y5(+1), y11(+2)

y10(+2)

340374

369

y7(+2)

y3(+1), y11(+3)

y8(+2)

487521

555b10(+2)

y9(+2)

m/z

% In

ten

sity

620.09(2) Spec no.1421 * [BP = 866.0, 126]

Alcohol dehydrogenase 1B (gi|113394)– Mascot score: 74– KFSLDALITHVLPFEK– Peptide expectation value: 1.9 e-005– From sample 7, spectrum 1421

y4(+1)

Figure 10. Alcohol dehydrogenase 1B (gi|113394).



with COX-1. Likewise, PPAR-δ and -γ were sig-nificantly downregulated in the ascending colon. The difference between cancer patients and con-trols was even more striking when the relative expression levels of three of the most upregulated genes – COX-2, OPN and MCSF-1 – were con-sidered together. In virtually every sample from a cancer patient, at least one of these three genes was significantly upregulated relative to its expression level in any sample from a control. This analy-sis suggested that expression levels of these three genes considered together may be sufficient to dis-tinguish normal colon mucosa in colon cancer patients from colon mucosa in controls [17].

Alterations in signaling pathways of FGFR, EGFR, ERBB2 and PI3K were detected in nearly two-thirds of breast and colorectal tumors that were comprehensively examined in a recent report from the Vogelstein laboratory. The com-plexity of genetic analysis of CRC was further discussed in the report in which the authors stated that: “these genetic analyses can only identify candidate genes that may play a role in cancer and do not definitively implicate any gene in the neoplastic process” [18].

These studies demonstrate a wide spectrum of genetic and protein analyses of CRC and histo-logically normal tissue surrounding CRC tumors.

However, there is little duplication of genes and proteins between the various authors. Preliminary work in this laboratory suggested that mass spec-trometry of the CRC and normal satellite tissue could identify specific reproducible protein bio-markers that would characterize tissue at risk for CRC. Specifically, matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI IMS) of tissue and liquid chromatography (electro-spray) mass spectrometry (LCMS) of tissue protein extracts. The goal would be to use these putative proteins as the basis for a screening test of stool.

Imaging mass spectrometryA brief discussion of IMS will put the technique in context for CRC diagnosis. The history of IMS began with single-cell studies of Aplysi californica neurons by matrix-assisted laser desorption/ioniza-tion time-of flight mass spectrometry. This was probably the first direct tissue MALDI identifi-cation of peptides and tissue profiling based on ion density [19]. A refinement of this technique, imaging MALDI was described in human buc-cal mucosa, and rat pituitary and pancreas glands using two different approaches: direct targeting of the tissue itself and by analysis of blotted tar-gets previously exposed to the tissue [20]. An ‘MS Image Tool’, using the peptide neurotensin (peak

Chain A, α-ferrous-carbonmonoxy, β-cobaltous-deoxy hemoglobin (T-state) (gi|229751)– Mascot score: 102– VGAHAGEYGAEALER– Peptide expectation value: 4.0 e-005– From sample 9, spectrum 894

100

90

80

70

60

50

40

30

20

10

0

510.64(2) Spec no.894 * [BP = 417.3, 421]

% In

ten

sity

250.0 398.6 547.2 695.8 844.4 993.0

0

b10(+1)b9(+1)814.4

b8(+1)y7(+1)y6(+1)

y5(+1)

648.4

b7(+1)614.9

558.3

543.3534.8

522.8

489.3

472.3458.3

448.2

443.3

b9(+2)

418.3407.7b8(+2)

y7(+2)

b4(+1)

y6(+2)

y5(+2)300.2

305.2

y2(+1)

b10(+2)y4(+1)

y3(+1)b11(+2)

b12(+2)

422.7

421.0

m/z

Figure 11. Chain A, α‑ferrous-carbonmonoxy, β-cobaltous-deoxy hemoglobin (T state) (gi|229751).



at mass/charge [m/z] 1674) significantly improved the data acquisition, speed of IMS and utilization of the technique [21]. In a laser capture microdissec-tion study of both tumor and normal human breast tissue fixed in ethyl alcohol, and stained with hema-toxylin and eosin, normal tissue, carcinoma in situ, invasive carcinoma and metastatic carcinoma could be distinguished by their different MALDI spectra. The introduction of ‘BioMap’ software, by apply-ing a baseline correction to the spectra and integrat-ing over the peak of interest, demonstrated mouse brain images of amyloid-β, and Aβ peptides. That report was proof-of-principle that MALDI images of tissue could be obtained based upon the mass spectrometry m/z peak of interest [22]. Subsequent profiling and IMS of normal mouse epididymis identified different protein activity (ion densities) throughout the sections [23].

The negative effects on IMS resolution from destructive tissue freezing artifacts, excessive dehy-dration due to ethanol fixation, paraformaldeyde cationization, embedding artifacts from OCT polymer and agar, and coarse matrix crystal size were first described in a report on spatial profiling of invertebrate ganglia [24]. In a follow-up report, they suggested a solution to the problem of freez-ing artifacts using glycerol and the related com-pounds ethane-1, 2-diol and propane-1, 2-diol to stabilize cellular membranes [25].

A subsequent report suggested direct liquid nitrogen immersion of tissue in aluminum wrap-ping as a means of rapid fixation, but ignored the known consequences of freezer artifact [26]. Adjunctive histologic staining with methylene blue-stained tissues on standard metal plates or indium-tin coated glass slides were shown to be compatible with IMS, but cresyl violet stain decreased IMS signal intensity [27].

Tissue blotting with trypsin digestion for tandem mass spectrometry database analysis was shown to be useful in analyte localization, but was destructive to tissue morphology [28]. Stoeckli et al. described a less destructive trypsin digest step to their prior tissue blotting technique for IMS [29].

Another method, matrix-enhanced secondary ion mass spectrometry (SIMS), was described and used for direct molecular imaging of the ganglia of the freshwater snail Lymnaea stagnalis [30]. However, this technique presented signifi-cant limitations and was proved unsuitable for direct tissue imaging.

A refinement of IMS, oversampling with com-plete sample ablation at each sample position on the target plate, provided significant resolution enhancement with a translation stage raster step size of 25 µM. A 40 µM object could now be resolved with a 100 µM laser [31].

301.1

319.2

y5(+2)

329.1347.2

y3(+1)y6(+2)

380.2

421.3

421.9

y7(+2)

415.9

411.3

429.3

428.3

440.3

458.3

y8(+2)480.3

476.3474.3

y4(+1)

525.9

525.4

b11(+2), y9(+2)

y10(+2)

553.9554.3 588.4

y11(+2)602.4y5(+1)

650.9

660.9

660.4

659.9

363.2

100

90

80

70

60

50

40

30

20

10

0

% In

ten

sity

254.0 338.8 423.6 508.4 593.2 678.0

m/z

0

392.0

Galectin 4 (gi|5453712)– Mascot score: 79– VGSSGDIALHINPR– Peptide expectation value: 5.8 e-005– From sample 9, spectrum 934

479.31(2) Spec no.934 * [BP = 421.3, 392]

Figure 12. Galectin 4 (gi|5453712).



Tissue treatments with organic solvents such as chloroform, acetone, hexane, tolu-ene or xylene were shown to be an effective and rapid method for signal enhancement in MALDI direct tissue profiling [32]. These stud-ies demonstrated that solvent treatments par-tially removed lipids from the tissue surface. Compared with previous studies with ethanol, chloroform/xylene solvent rinsing is more spe-cific for lipid removal and does not generate delocalization or extraction of most soluble pep-tides/proteins as tested by immunohistochemis-try experiments. Among all the tested solvents, chloroform and xylene produced the greatest increase in MALDI signal intensity and num-ber of detected peptides/proteins. However, this treatment does not reduce salt adducts as does alcohol treatment. The results suggest that it is possible to detect, after organic rinsing treat-ments, compounds, such as peptides/proteins present in the cytoplasm, that were masked by lipids in the tissue.

Clench et al. reported development and appli-cation of a method using 9 aminoacridine as a matrix for negatively charged ions in MALDI imaging [33].

Crossman demonstrated the need for thin sec-tions to avoid differential extraction efficiency of matrix solvent in different tissues [34].

Pevsner demonstrated direct cellular MALDI identification of proteins in fixed cells and tis-sues without freezer artifact, tissue corrosion by matrix solvents or the use of tissue blotting [35,36]. This was confirmed in a later report [37].

Metal-assisted SIMS, a variation on SIMS, as well as MALDI IMS can provide images from tissue, but the duration of these protocols were highly dependent on sample size and technique parameters. The duration of these studies averaged approximately 5 h [38].

Agar et al. studied multiple solvent/matrix combinations. However, the tissue sections at the electron microscopic level demonstrated both freezing artifact and structural distortion, indicat-ing that their method disrupts normal subcellular structures such as mitochondria [39].

Baluya et al. reported a variation on inkjet-printed matrix application to tissue specimens [40], previously described by Sloane et al. [41]. The matrix application was of better quality and more reproducible than from specimens pre-pared by the electrospray and airbrush methods, but was still not completely uniform. A uniform method of tissue matrix application is sublima-tion. Sublimation is solvent free, rapid and was successfully used to identify lipids in brain tis-sue [42], and more recently described for pro-teins or peptides [43–46]. The challenge of IMS in

191.4 375.2

y3(+1)y4(+1)

b4(+1)

635.3

b5(+1) 690.9748.4

y12(+2)

689.9

868.5

869.5

100

90

80

70

60

50

40

30

20

10

0140.0 343.2 546.4 749.6 952.0 1156.0

0

108.0

m/z

% In

ten

sity

867.99(2) Spec no.967 * [BP = 689.9, 108]

Vimentin (gi|37852)– Mascot score: 61– LQDEIQNMKEEMAR– Peptide expectation value: 3.5 e-005– From sample 9, spectrum 967

Figure 13. Vimentin (gi|37852).



formalin-fixed, paraffin-embedded tissue was first addressed by Fournier [47] and then by Pevsner [45], and later by Stauber [48].

The ‘field cancerization’ and IMS literature combined with our own experience led to a series of experiments that resulted in direct identification of proteins in CRC tumors and their satellite tissue.

The biopsy tissue was cryoprotected by imme-diate immersion in a mixture of dimethyl sul-foxide (2%), glycerol (20%) and ethanol (78%). Cryosections, 1 µ thick, could be obtained with-out freezer artifact that would otherwise destroy the tissue architecture [43]. Contiguous 1 µ sec-tions were obtained for histology, IMS and pro-tein extraction with high pressure (Barocycler®, Pressure BioSciences Inc.). The complete pro-tocols are detailed in the references [43–45,49–51]. Trypsin 1 µg/µl in 100 mM ammonium bicar-bonate, pH 8, was applied to the tissue for IMS examination. Tissue corrosion was avoided by sublimation of MALDI matrix onto the tissue. No solvent was used. The Shimadzu Axima, TOF2 was operated in reflectron mode (acceler-ating voltage 20 kV, reflectron voltage 25 kV). Each spectrum represented the cumulative aver-age of 50 profiles (spectra) per spot with a 70 µ spot interval in a rectangle serpentine raster of each tissue section. The calibrated mass accuracy

in reflectron mode was ±0.3 Da, or 50 ppm for the parent ions, 75% lower than a recent IMS report [52].

A spectrum obtained from the tumor and sat-ellite tissue (Figure 1) was analyzed with Mascot software. The search criteria included only one missed cleavage. Peptide mass fingerprinting (PMF) from the IMS image demonstrated the same two proteins, gi|119592539 hCG1787564 (Homo sapiens) Mass: 57590, and gi|119592490 hCG2040674 (Homo sapiens) Mass: 108178 in the colon adenocarcinoma and in the histo-pathologically normal satellite tissue (supplementary

Figure 1) (see online www.futuremedicine.com/toc/bmm/3/1).

The hematoxylin and eosin stained histological section with the tumor is seen in Figure 2.

The reconstructed IMS image is shown in Figure 3. The right outlined area corresponds to the tumor in the histology section, and the smaller left outlined region corresponds to the histologi-cally normal satellite tissue. The identical hCG isoforms were identified in both loci. The smaller locus represents a satellite ‘field defect’ that is his-tologically normal, but biochemically similar to the tumor.

A prevailing view of progession from normal tissue to CRC is conversion of abnormal crypt foci with or without intermediate polyps to

100

90

80

70

60

50

40

30

20

10

0259.0 431.2 603.4 775.6 947.8 1120.0

0

5555.0

m/z

% In

ten

sity

657.91(2) Spec no.1016 * [BP = 659.4, 5555]

Hemoglobin β (gi|40886941)– Mascot score: 212– VNVDEVGGEALGR– Peptide expectation value: 2.6 e-006– From sample 8, spectrum 1016

y10(+1)y9(+1)y8(+1)

661.4

662.4642.4

y6(+1)603.4

b5(+1)485.3458.2

417.3

410.3383.2

y3(+1)344.2315.2

y4(+1)

660.4

y7(+1)

Figure 14. Hemoglobin β (gi|40886941).



CRC. We examined several polyps with IMS, and LCMS to identify putative CRC biomark-ers. Figures 4 & 5 are contiguous sections of colon tissue. The hematoxylin- and eosin-stained sec-tion is shown in Figure 4, and the sinapic acid matrix sublimated section is shown in Figure 5.

The presence of these two proteins in the his-tologically normal tissue represents a potential marker for field cancerization or field defect, since the exact same proteins were found in the CRC.

The transition from normal colon tissue to CRC is ACF to polyp to CRC. Therefore, we chose to examine multiple polyps for the pres-ence of putative CRC proteins, especially at their periphery, the point of greatest neovascularity and growth. This is the transition zone between the polyp and normal peripheral tissue.

A hematoxylin- and eosin-stained histologic section of a histopathologically benign-appear-ing colon polyp is shown in Figure 4. There is a central fibrotic zone and a peripheral zone of increased vascularity at the transition from polyp to normal tissue.

A contiguous section to the H & E histo-logic section in Figure 4 was applied to a MALDI conductive plate. The section was covered with Sinapic acid matrix by sublimation (Figure 5).

IMS of these undigested polyp tissues all revealed tentative protein identifications (obtained by query of the NCBInr database with the mass of the proteins) that were vis-ible at the transition zone between the polyp and the histologically normal peripheral tissue (Figures 6–9).

Def initive protein identif ications were obtained from trypsin-digested tissue protein extracts from the transition zone of these polyps

by LCMS (Box 1–3). The underlined proteins are putative CRC proteins. The tandem mass spec-trometry spectra of these proteins are shown in Figures 10–14.

Conclusion & future perspectiveAll of the polyps had putative CRC proteins in the transition zone between polyp and periph-eral tissue. This finding suggests that these polyps have already mutated to CRC, and these patients may be at risk for CRC from other patches of ACF that could develop into polyps and CRC. A pathologist reporting no tumor in the normal tissue would be correct. However, histology alone would significantly affect therapy and postoperative surveillance by underestimating the extent of metaplastic or malignant disease. We suggest that IMS and LCMS, and histopathology will provide a more comprehensive diagnosis for CRC tumors. This diagnostic paradigm should provide the patient and his/her physicians with a more complete diagnosis and may lead to better out-come of CRC. With future refinement, LCMS of extracted stool proteins could become the basis for an acceptable, high-throughput mass spectrometry screening test.

Financial & competing interests disclosureThe authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employ‑ment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

executive summary

Colorectal cancerColorectal cancer (CRC) is the second‑leading cause of cancer‑related deaths in the USA, and affects �both men and women.

Diagnostic paradigmThis report suggests a new paradigm for CRC diagnosis and staging, a combination of �histopathology and mass spectrometry.

Field defects or field cancerizationThe diagnosis and treatment of epithelial cancers should not only be focused on the tumor but also �on the field from which it developed.

Imaging matrix-assisted laser desorption/ionization imaging mass spectrometryMatrix‑assisted laser desorption/ionization images of tissue can be obtained based upon the mass �spectrometry mass/charge [m/z] peak of interest.

New diagnostic CRC paradigmImaging mass spectrometry, liquid chromatography mass spectrometry and histopathology will �provide a more comprehensive diagnosis for CRC tumors, and demonstrate the full extent of the tumor.



Bibliography Song K, Fendrick AM, Ladabaum U: 1

Fecal DNA testing compared with conventional colorectal cancer screening methods: a decision ana lysis. Gastroenterology 126, 1270–1279 (2004).

Ladabaum U, Song K: Projected national 2

impact of colorectal cancer screening on clinical and economic outcomes and health services demand. Gastroenterology 129, 1151–1162 (2005).

Slaughter DP, Southwick HW, Smejkal W: 3

Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer 6, 963–968 (1953).

Braakhuis BJ, Tabor MP, Kummer JA, 4

Leemans CR, Brakenhoff RH: A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res. 63, 1727–1730 (2003).

Buckley CH, Fox H: An 5

immunohistochemical study of the significance of HCG secretion by large bowel adenocarcinomata. J. Clin. Pathol. 32, 368–372 (1979).

Fukayama M, Hayashi Y, Koike M: Human 6

chorionic gonadotropin in the rectosigmoid colon. Immunohistochemical study on unbalanced distribution of subunits. Am. J. Pathol. 127, 83–89 (1987).

Yamaguchi A, Ishida T, Nishimura G 7 et al.: Human chorionic gonadotropin in colorectal cancer and its relationship to prognosis. Br. J. Cancer 60, 382–384 (1989).

Fearon ER, Vogelstein B: A genetic model for 8

colorectal tumorigenesis. Cell 61, 759–767 (1990).

Scalmati A, Lipkin M: Proliferation and 9

differentiation biomarkers in colorectal mucosa and their application to chemoprevention studies. Environ. Health Perspect. 99, 169–173 (1993).

Magnuson BA, Shirtliff N, Bird RP: 10

Resistance of aberrant crypt foci to apoptosis induced by azoxymethane in rats chronically fed cholic acid. Carcinogenesis 15, 1459–1462 (1994).

Bronner MP, Culin C, Reed JC, Furth EE: 11

The bcl‑2 proto-oncogene and the gastrointestinal epithelial tumor progression model. Am. J. Pathol. 146, 20–26 (1995).

Garewal H, Bernstein H, Bernstein C, 12

Sampliner R, Payne C: Reduced bile acid-induced apoptosis in ‘normal’ colorectal mucosa: a potential biological marker for cancer risk. Cancer Res. 56, 1480–1483 (1996).

Sarela AI, Macadam RC, Farmery SM, 13

Markham AF, Guillou PJ: Expression of the antiapoptosis gene, Survivin, predicts death from recurrent colorectal carcinoma. Gut 46, 645–650 (2000).

Anti M, Armuzzi A, Morini S 14 et al.: Severe imbalance of cell proliferation and apoptosis in the left colon and in the rectosigmoid tract in subjects with a history of large adenomas. Gut 48, 238–246 (2001).

Crowley-Weber CL, Dvorakova K, Crowley C 15

et al.: Nicotine increases oxidative stress, activates NF-κB and GRP78, induces apoptosis and sensitizes cells to genotoxic/xenobiotic stresses by a multiple stress inducer, deoxycholate: relevance to colon carcinogenesis. Chem. Biol. Interact. 145, 53–66 (2003).

Bernstein H, Holubec H, Warneke JA 16 et al.: Patchy field defects of apoptosis resistance and dedifferentiation in flat mucosa of colon resections from colon cancer patients. Ann. Surg. Oncol. 9, 505–517 (2002).

Chen LC, Hao CY, Chiu YS17 et al.: Alteration of gene expression in normal-appearing colon mucosa of APC(min) mice and human cancer patients. Cancer Res. 64, 3694–3700 (2004).

Leary RJ, Lin JC, Cummins J18 et al.: Integrated ana lysis of homozygous deletions, focal amplifications, and sequence alterations in breast and colorectal cancers. Proc. Natl Acad. Sci. USA 105, 16224–16229 (2008).

Garden RW, Moroz LL, Moroz TP, Shippy 19

SA, Sweedler JV: Excess salt removal with matrix rinsing: direct peptide profiling of neurons from marine invertebrates using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. J. Mass Spectrom. 31, 1126–1130 (1996).

Caprioli RM, Farmer TB, Gile J: Molecular 20

imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal. Chem. 69, 4751–4760 (1997).

Stoeckli M, Farmer TB, Caprioli RM: 21

Automated mass spectrometry imaging with a matrix-assisted laser desorption ionization time-of-flight instrument. J. Am. Soc. Mass Spectrom. 10, 67–71 (1999).

Stoeckli M, Chaurand P, Hallahan DE, 22

Caprioli RM: Imaging mass spectrometry: a new technology for the ana lysis of protein expression in mammalian tissues. Nat. Med. 7, 493–496 (2001).

Chaurand P, Caprioli RM: Direct profiling 23

and imaging of peptides and proteins from mammalian cells and tissue sections by mass spectrometry. Electrophoresis 23, 3125–3135 (2002).

Kruse R, Sweedler JV: Spatial profiling 24

invertebrate ganglia using MALDI MS. J. Am. Soc. Mass Spectrom. 14, 752–759 (2003).

Rubakhin SS, Greenough WT, Sweedler JV: 25

Spatial profiling with MALDI MS: distribution of neuropeptides within single neurons. Anal. Chem. 75, 5374–5380 (2003).

Schwartz SA, Reyzer ML, Caprioli RM: 26

Direct tissue ana lysis using matrix-assisted laser desorption/ionization mass spectrometry: practical aspects of sample preparation. J. Mass Spectrom. 38, 699–708 (2003).

Chaurand P, Schwartz SA, Billheimer D, 27

Xu BJ, Crecelius A, Caprioli RM: Integrating histology and imaging mass spectrometry. Anal. Chem. 76, 1145–1155 (2004).

Bunch J, Clench MR, Richards DS: 28

Determination of pharmaceutical compounds in skin by imaging matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun. Mass Spectrom. 18, 3051–3060 (2004).

Rohner TC, Staab D, Stoeckli M: MALDI 29

mass spectrometric imaging of biological tissue sections. Mech. Ageing Dev. 126, 177–185 (2005).

Altelaar AF, van Minnen J, Jimenez CR, 30

Heeren RM, Piersma SR: Direct molecular imaging of Lymnaea stagnalis nervous tissue at subcellular spatial resolution by mass spectrometry. Anal. Chem. 77, 735–741 (2005).

Jurchen JC, Rubakhin SS, Sweedler JV: 31

MALDI-MS imaging of features smaller than the size of the laser beam. J. Am. Soc. Mass Spectrom. 16, 1654–1659 (2005).

Lemaire R, Wisztorski M, Desmons A32 et al.: MALDI-MS direct tissue ana lysis of proteins: improving signal sensitivity using organic treatments. Anal. Chem. 78, 7145–7153 (2006).

Burrell M, Earnshaw C, Clench M: Imaging 33

matrix-assisted laser desorption/ionization mass spectrometry: a technique to map plant metabolites within tissues at high spatial resolution. J. Exp. Bot. 58, 757–763 (2007).

Crossman L, McHugh NA, Hsieh YS, 34

Korfmacher WA, Chen JW: Investigation of the profiling depth in matrix-assisted laser desorption/ionization imaging mass spectrometry. Rapid Commun. Mass Spectrom. 20, 284–290 (2006).

Pevsner PH, Naftolin F, Hillman DE35 et al.: Direct identification of proteins from T47D cells and murine brain tissue by matrix-assisted laser desorption/ionization post-source decay/collision-induced dissociation. Rapid Commun. Mass Spectrom. 21(3), 429–436 (2007).

Pevsner PH, Naftolin F, Miller DC36 et al.: Direct matrix assisted laser desorption ionization-mass spectrometry identification of tubulin ß-2 chain and other proteins in murine brain and spinal cord tissue. J. Soc. Gynecol. Investig. 13, A1–B10 (2006).



Groseclose MR, Andersson M, Hardesty 37

WM, Caprioli RM: Identification of proteins directly from tissue: in situ tryptic digestions coupled with imaging mass spectrometry. J. Mass Spectrom. 42, 254–262 (2007).

Altelaar AF, Luxembourg SL, McDonnell LA, 38

Piersma SR, Heeren RM: Imaging mass spectrometry at cellular length scales. Nat. Protoc. 2, 1185–1196 (2007).

Agar NY, Yang HW, Carroll RS, Black PM, 39

Agar JN: Matrix solution fixation: histology-compatible tissue preparation for MALDI mass spectrometry imaging. Anal. Chem. 79(19), 7416–7423 (2007).

Baluya DL, Garrett TJ, Yost RA: Automated 40

MALDI matrix deposition method with inkjet printing for imaging mass spectrometry. Anal. Chem. 79, 6862–6867 (2007).

Sloane AJ, Duff JL, Wilson NL41 et al.: High-throughput peptide mass fingerprinting and protein macroarray ana lysis using chemical printing strategies. Mol. Cell Proteomics 1, 490–499 (2002).

Hankin JA, Barkley RM, Murphy RC: 42

Sublimation as a method of matrix application for mass spectrometric imaging. J. Am. Soc. Mass. Spectrom 18, 1646–1652 (2007).

Pevsner P, Naftolin F, Vecchione D43 et al.: Microtubule-associated proteins (MAP) and motor molecules: direct tissue MALDI

identification and imaging. British Mass Spectrometry Society, Edinburgh, Scotland, September, 2007.

Pevsner P, Vecchione D, Stall B, Remsen T, 44

Anand S, Stern A: Colon cancer: protein biomarkers in tissue and body. British Mass Spectrometry Society, Edinburgh, Scotland, September 2007.

Pevsner P, Vecchione D, Remsen T45 et al.: Colorectal carcinoma – field defects in satellite tissue. British Mass Spectrometry Society, Edinburgh, Scotland; Robinson College, Cambridge, UK, December 2007.

Puolitaival SM, Burnum KE, Cornett DS, 46

Caprioli RM: Solvent-free matrix dry-coating for MALDI imaging of phospholipids. J. Am. Soc. Mass Spectrom. 19, 882–886 (2008).

Lemaire R, Desmons A, Tabet JC, Day R, 47

Salzet M, Fournier I: Direct ana lysis and MALDI imaging of formalin-fixed, paraffin-embedded tissue sections. J. Proteome Res. 6, 1295–1305 (2007).

Stauber J, Lemaire R, Franck J48 et al.: MALDI imaging of formalin-fixed paraffin-embedded tissues: application to model animals of Parkinson disease for biomarker hunting. J. Proteome Res. 7, 969–978 (2008).

Vecchione D, Naftolin F, Remsen T, Kessler P, 49

Stern A, Pevsner P: Stroke: prophylactic estrogen (estradiol) therapy. British Mass

Spectrometry Society, Edinburgh, Scotland; Robinson College, Cambridge, UK, December 2007.

Remsen T, Kessler P, Francois F, Stern A, 50

Anand S, Pevsner P: Imaging MALDI of colorectal carcinoma: field defects in satellite tissue. American Chemical Society, Washington, DC; Queensbourough College, Queens, NY, USA, May 2008.

Pevsner P, Formenti S, Remsen T51 et al.: Mass spectrometry of buccal mucosa: biomarkers for biodosimetry in radiation incidents. Armed Forces Radiobiology Research Institute, Bethesda, MD, USA, May 2008.

Groseclose MR, Massion PP, Chaurand P, 52

Caprioli RM: High-throughput proteomic ana lysis of formalin-fixed paraffin-embedded tissue microarrays using MALDI imaging mass spectrometry. Proteomics 8, 3715–3724 (2008).

WebsitenColon and rectum cancer (invasive). 101

http://seer.cancer.gov/csr/1975_2005/results_merged/sect_06_colon_rectum.pdf

mass spectrometry maldi imaging of colon cancer biomarkers: a new diagnostic paradigm

Documents