From T-antigen to plasmalogen-derived aldehydes: The identification of a marker of colorectal cancer in human rectal mucous

Article abstract of DOI:10.1139/v02-195

Recently, a simple noninvasive screening test for colorectal cancer was proposed, based on a hypothesis involving galactose-containing carbohydrate moieties such as the Thomsen-Friedenreich antigen. According to the hypothesis, such carbohydrate moieties, present in the human rectal mucous of patients with colorectal cancer, can be specifically oxidized with galactose oxidase to form substances that, upon reaction with Schiff reagent, yield purple (magenta) coloured compounds. While evaluating this proposed test, we discovered that the colour formation is not due to the proposed reaction between oxidized galactose moieties present in rectal mucous and Schiff reagent. We found instead that the mucous from colorectal cancer patients contains compounds that form purple (magenta) adducts with the Schiff reagent directly, i.e., they do not require oxidation by galactose oxidase. We have identified these compounds as long-chain aliphatic aldehydes, mainly palmitic aldehyde C₁₅H₃₁CH=O and stearic aldehyde C₁₇H₃₅CH=O. We have further found that the aldehydes originate from plasmalogens present in the phospholipid fraction of the mucous obtained from colorectal cancer patients. The aldehydes, present in plasmalogens as enol ethers, are released by the acidity of the Schiff reagent and in turn react with the Schiff reagent to form the coloured adducts. Correct identification of these markers could lead to the development of a more accurate colorectal cancer screening tool and to a deeper understanding of colorectal carcinogenesis.

Full text of DOI:10.1139/v02-195

1
09
From T-antigen to plasmalogen-derived aldehydes:
The identification of a marker of colorectal cancer
in human rectal mucous
Jiri J. Krepinsky, Gabor P. Kandel, Ka Sing Yeung, Jacek Chociej, Min Chen,
Gideon Cohen, Stephen P. Douglas, Rudolf Furrer, Vishal Kukreti,
Niculina Lupescu, Enoka Richens, and Keith L. Tanner
Abstract: Recently, a simple noninvasive screening test for colorectal cancer was proposed, based on a hypothesis in-
volving galactose-containing carbohydrate moieties such as the Thomsen–Friedenreich antigen. According to the hy-
pothesis, such carbohydrate moieties, present in the human rectal mucous of patients with colorectal cancer, can be
specifically oxidized with galactose oxidase to form substances that, upon reaction with Schiff reagent, yield purple
(
magenta) coloured compounds. While evaluating this proposed test, we discovered that the colour formation is not due
to the proposed reaction between oxidized galactose moieties present in rectal mucous and Schiff reagent. We found in-
stead that the mucous from colorectal cancer patients contains compounds that form purple (magenta) adducts with the
Schiff reagent directly, i.e., they do not require oxidation by galactose oxidase. We have identified these compounds as
long-chain aliphatic aldehydes, mainly palmitic aldehyde C H CH=O and stearic aldehyde C H CH=O. We have
1
5
31
17 35
further found that the aldehydes originate from plasmalogens present in the phospholipid fraction of the mucous ob-
tained from colorectal cancer patients. The aldehydes, present in plasmalogens as enol ethers, are released by the acid-
ity of the Schiff reagent and in turn react with the Schiff reagent to form the coloured adducts. Correct identification
of these markers could lead to the development of a more accurate colorectal cancer screening tool and to a deeper un-
derstanding of colorectal carcinogenesis.
Key words: T-antigen, plasmalogen-derived aldehydes, colorectal cancer marker.
Résumé : Récemment une nouvelle méthode simple et non invasive de dépistage du cancer colorectal a été proposée;
elle est fondée sur une hypothèse impliquant des portions de sucre contenant du galactose, tel l’antigène de Thomsen–
Friedenreich. Selon cette hypothèse, de telles portions de sucre présentes dans les muqueuses rectales de patients hu-
mains atteints d’un cancer colorectal peuvent être spécifiquement oxydées par l’oxydase du galactose pour former des
substances qui réagissent avec le réactif de Schiff et qui produisent des composés colorés en violet magenta. En éva-
luant la méthode proposée, on a découvert que la formation de la couleur n’est pas due à la réaction suggérée du réac-
tif de Schiff et des portions de galactose oxydé présentes dans la muqueuse rectale. On a plutôt trouvé que la muceuse
de patients atteints de cancer colorectal contiennent des composés qui, avec le réactif de Schiff, forment directement
des adduits colorés en violet magenta, c’est-à-dire qu’il n’est pas nécessaire de procéder par l’oxydation à l’aide
d’oxydase du galactose. On a identifié ces composés comme étant des aldéhydes aliphatiques à longues chaînes carbo-
nées, principalement l’aldéhyde palmitique, C H CH=O, et l’aldéhyde stéarique, C H CH=O. On a de plus trouvé
1
5
31
17 35
que ces aldéhydes trouvent leur origine dans les plasmalogènes présents dans la fraction phospholipide de la muqueuse
des patients atteints de cancer colorectal. Ces aldéhydes sont présents dans le plasmalogène sous la forme d’éthers éno-
liques, ils sont libérés par l’acidité du réactif de Schiff et ils réagissent alors avec le réactif de Schiff pour former les
adduits colorés. L’identification correcte de ces marqueurs pourrait conduire au développement d’un outil de dépistage
plus précis du cancer colorectal et à une meilleure compréhension du développement de ce type de cancer.
Mots clés : antigène T, aldéhydes dérivés du plasmalogène, marqueur du cancer colorectal.
[
Traduit par la Rédaction] Krepinsky et al. 117
Received 8 November 2001. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 30 January 2003.
Dedicated to the memory of Professor Raymond U. Lemieux.
1
J.J. Krepinsky, J. Chociej, M. Chen, G. Cohen, S.P. Douglas, R. Furrer, V. Kukreti, N. Lupescu, E. Richens, and
K.L. Tanner. Department of Medical Genetics and Microbiology, University of Toronto, Toronto, ON M5S 1A8, Canada.
G.P. Kandel. St. Michael’s Hospital, 30 Bond Street, Toronto, ON M5B 1W8, Canada, and Department of Medicine, University of
Toronto, Toronto, ON M5S 1A8, Canada.
K.S. Yeung. Department of Public Health Sciences, University of Toronto, Toronto, ON M5S 1A8, Canada.
1Corresponding author (e-mail: george.krepinsky@utoronto.ca).
Can. J. Chem. 81: 109–117 (2003)
doi: 10.1139/V02-195
© 2003 NRC Canada

1
10
Can. J. Chem. Vol. 81, 2003
Scheme 1. Chemical formulae of T-antigen before (I) and after
Introduction
(
hypothesized II) oxidation with galactose oxidase, palmitic (III)
and stearic (IV) aldehydes, and plasmalogens of ethanolamine
V) and choline (VI) types.
Colorectal carcinoma is the second most frequent cause of
cancer mortality in men and women in industrialized coun-
tries (1, 2). Primary prevention (i.e., averting the development
of the tumour by altering risk factors) is not yet feasible,
since the etiology of the disease is not adequately under-
stood. Therefore, there has recently been a great deal of in-
terest in secondary preventive measures (i.e., detection at an
asymptomatic, curable stage).
(
Recently, a promising screening test was proposed. It is
based on the observation that most colorectal neoplasms ex-
press a disaccharide D-Galp(β1-3)-D-GalpNAcα1 (also called
Thomsen–Friedenreich antigen or T-antigen) — bound as an
O-glycoside to (-)-Ser or (-)-Thr (I; Scheme 1) — among the
carbohydrate components of colonic mucin. This antigen is
usually detected by applying fluorescently labeled peanut
agglutinin (PNA) (3, 4) to histological specimens obtained
from the tumour. Obviously, obtaining biopsy specimens
from neoplasms inside the colon for screening purposes is
not possible. Therefore, a technique (5) was proposed
whereby this carbohydrate marker could be detected in colo-
rectal mucous collected during digital rectal examination.
The technique further assumes that the two galactose moi-
eties in T-antigen are oxidizable by galactose oxidase to C-6
aldehydes in dialdehyde II. These two C-6 aldehydes, on
subsequent reaction with Schiff reagent (6, 7), would yield
an adduct detectable by its purple (magenta) colour (5, 8).
Other galactose-containing oligosaccharides were subsequently
added to T-antigen as carbohydrate markers of colorectal
cancer (9). In addition to being simple and noninvasive, this
technique has the advantage that it samples mucous from the
whole colorectum because mucous from the total length of
the colon flows towards the rectum. (5)
Because of the potential advantage of such a screening
test, we initiated its clinical and laboratory evaluation in
1
989. In 1991 we discovered unexpectedly that colorectal
mucous from individuals with colorectal neoplasia devel-
oped a purple (magenta) colour after direct addition of
Schiff reagent to the mucous specimens, without the treat-
ment with galactose oxidase (10). The oxidation step was
somewhat puzzling, since the galactose oxidase treatment
could be shortened to 10 min (8) without any effect on the
test results, although the oxidation of a terminal galactose to
the C-6 aldehyde usually requires 15–30 min at 37°C, while
nonterminal galactose units are much less efficient substrates
and technician time. The first step toward this practical goal
was to identify the chemical structure(s) of the marker(s). In
this article we describe both the isolation and the identifica-
tion of the chemical structure of the markers. Although the
objective of this investigation has been the identification of
the chemical structure of the markers, patient data gathered
while collecting colorectal mucous in the process of the
identification of the markers are presented (cf. Fig. 1 for ex-
amples). These data suggest that further epidemiologic in-
vestigations of the usefulness of the markers in a screening
test for colorectal cancer are warranted.
(
10). Seeking an explanation to why colorectal mucus from
patients with colorectal cancer developed a magenta colour
directly upon addition of Schiff reagent, we first found that
the component(s) of the mucous responsible for the colour
development (referred to as a marker in this article) were
soluble in lipophilic solvents. This suggested that the marker
could not be a carbohydrate, which should be hydrophilic.
We also found that asialofetuin (a suitable model glyco-
protein because it contains T-antigen) does not undergo a
change in colour when reacted with Schiff reagent, whether
treated with galactose oxidase or not.
Since the marker reacted with Schiff reagent directly, we
speculated that such a marker could become the basis of a
simpler and more economical screening test (11) than the
test involving galactose oxidase. The elimination of the oxi-
dation step would reduce the costs of chemicals, equipment,
Materials and methods
Materials and general procedures
Thin layer chromatography (TLC) was performed on sil-
ica gel 60 plates (thickness 0.25 mm) or preparative plates
(thickness 1 mm). Analytical plates were visualized by
spraying with 50% aq H SO and heating to 200°C. Silica
2
4
gel (200–400 mesh, Toronto Research Chemicals) was used
for column chromatography. PE (brain extract) and PC
(heart extract) plasmalogens were purchased from Doosan
Serdary Research Laboratories (Toronto, Canada). Schiff re-
agent was prepared from p-rosaniline chloride (#1528, also
basic fuchsin or magenta from Sigma Chemical Co.,
©
2003 NRC Canada

Krepinsky et al.
111
Fig. 1. Proportion of positives on five groups of subjects. Mucus examples were (1) from colectomy specimens from colorectal cancer
patients obtained in several Toronto hospitals; (2) from individuals scheduled for colonoscopy at the Endoscopy suite, Wellesley Hospi-
tal, Toronto, subsequently diagnosed with cancer (mucus specimens were collected immediately before colonoscopy); (3) from individ-
uals scheduled for colonoscopy at the Endoscopy suite, Wellesley Hospital, Toronto, subsequently diagnosed with large (larger than
1
cm in diameter) polyps (mucus specimens were collected immediately before colonoscopy); (4) from individuals scheduled for col-
onoscopy at the Endoscopy suite, Wellesley Hospital, Toronto, subsequently diagnosed with small (1 cm or smaller in diameter) polyps
mucus specimens were collected immediately before colonoscopy); (5) symptoms-free young adults (without any lavage performed be-
fore mucus collection). Note that colorectal cancer often, but not always, develops in polyps.
(
St. Louis, Mo.) or from basic fuchsin (ICN, Aurora, Ohio).
Asialofetuin (A1908 and A4781) and galactose oxidase were
purchased from Sigma. HPLC was performed on a variety of
silica gel columns (5 µm, e.g., Hypersil) using a Beckman
System Gold with a Beckman UV 166P detector or a Varex
in Toronto, Ont., Canada. The colectomy specimens were
washed with water to remove blood and fecal matter 15–
20 min after surgery. Mucous was collected by gently scrap-
ing the surface with a small spatula. The scraped mucous
was placed into small plastic vials and frozen. One
colectomy specimen typically yielded 0.5–1.0 g of mucous;
some specimens, however, were much larger. To determine
whether the collected mucous gave a colour reaction with
Schiff reagent, the frozen vials were allowed to thaw at
room temperature for 60 min, and a small amount of mucous
on the tip of a spatula was smeared on the support and tested
as described in the following section.
1
Evaporative Light Scattering Detector IIA. H NMR spectra
were recorded at 500 MHz with a Varian Unity spectrometer
in the NMR spectroscopy laboratory of the Molecular Medi-
cine Research Centre, University of Toronto. Spectra were
obtained in CDCl , mixtures of CDCl –CD OD containing a
3
3
3
trace of TMS (δ = 0 ppm) as the internal standard, or D O
2
and are reported using a δ (ppm) scale. Mass spectra were
recorded with VG Analytical ZAB-SE, SCIEX API III, or
Voyager-Elite MALDI (Perseptive Biosystems) mass spec-
trometers in the mass spectrometry laboratory, Molecular
Medicine Research Centre, University of Toronto. O-
Separation of mucous components
Mucous collected as above was pooled (66 g) and lyo-
philized for 24 h to give a semi-solid residue (6.0 g). This
residue was consecutively extracted with hexane, ethylacetate,
chloroform–methanol (2:1), modified Folch extraction (12),
and water; the final residue after extractions was a gel-like
solid (colonic mucin). Ethylacetate extract gave a weak col-
our reaction with Schiff reagent, while the chloroform–
methanol extract gave a very strong colour reaction with
Schiff reagent. Both extracts giving colour with Schiff
reagent (i.e., positive) were combined and subjected to chro-
matography on a column of silica gel using solvents of
increasing polarity: hexane, ethylacetate, chloroform,
chloroform–methanol (with increasing contents of methanol),
and methanol. Chloroform–methanol (7:2.5) afforded posi-
tively reacting fractions, which were combined and evapo-
rated to dryness to give an amorphous residue containing
phospholipids (36.6 mg), which reacted strongly positive
with Schiff reagent (using conditions described in the
(
2,3,4,5,6-Pentafluorobenzyl)oximes of aldehydes were ana-
lyzed by GC–CI(–)-MS (reagent gas: ammonia) using a VG-
Trio 2A quadrupole mass spectrometer interfaced to a
Hewlett-Packard 5890 Series II gas chromatograph at the
Hospital for Sick Children, Toronto. Visible spectra were re-
corded with a Beckman DU 600 or a Varian CARY 300 Bio
UV–Vis spectrometer (in solution or on the plates in trans-
mittance mode). Reflectance scans were executed directly on
the plates with a Varian CARY 50 spectrometer with remote
reflectance probe.
Isolation of markers from surgical (colectomy) specimens
Mucous collection
Mucous was obtained from colectomy specimens from
surgeries performed in operating theatres of several hospitals
©
2003 NRC Canada

1
12
Can. J. Chem. Vol. 81, 2003
preceding paragraphs). This fraction was repeatedly
chromatographed on a series of silica gel columns using
chloroform with a gradient of methanol to give two highly
positively reacting fractions, PL-1 and PL-2, which were
was removed by filtration. The resulting colourless solution
was stored in a dark-glass bottle at 3–5°C.
Identification of compounds giving colour reaction with
Schiff reagent
1
further analyzed by H NMR spectroscopy.
2
,4-Dinitrophenylhydrazones of aldehydes from mucous
Mucous (4.27 g) obtained from a colectomy as described
Testing mucous obtained in digital rectal examination
Mucous collection
above and giving a colour reaction with Schiff reagent was
spread with a spatula on five sheets (10 × 4 cm) of polyester
fabric (polymacron) and allowed to dry for 1 h. The sheets
were soaked in phosphate buffer (pH 7) for 20 min, then
lifted from the buffer, washed with water until all the dark
coloured material was removed and the sheets were slightly
yellow, and air-dried. The sheets were then cut into smaller
pieces (1 × 2 cm) and immersed into a solution of 2,4-
dinitrophenylhydrazine hydrochloride (30 mL; prepared by
dissolving 2,4-dinitrophenylhydrazine hydrochloride (0.25 g)
in 6 N HCl (100 mL)) at room temperature for 30 min. The
polyester cuttings were then removed from the reagent solu-
tion and washed with water until the eluant remained colour-
less; the deep-yellow 2,4-dinitrophenyl hydrazones were
extracted from the cuttings by acetonitrile. Acetonitrile was
removed on a rotary evaporator; ethylacetate (10 mL) was
added and evaporated (to remove traces of water) to yield a
dry yellow residue. The residue was dissolved in dry ace-
tonitrile (1.5 mL) and applied on a silica gel 60 plate (20 ×
20 cm; thickness 1 mm) in a single line; the plate was devel-
oped with a hexane – ethyl acetate solution (4:1). The yel-
Mucous samples were collected from individual human
subjects by a physician during routine digital rectal examina-
tion using a gloved finger lubricated with MUKO or a simi-
lar lubricant that would not react with the Schiff reagent
under the testing conditions. Informed consent of the tested
individuals was obtained as per the protocol approved by the
Ethics Committee, University of Toronto. Mucous collected
on the gloved finger was smeared onto a support (e.g., glass
fibre (Whatman GF) or polyester (polymacron, Dupont,
Wilmington, Del.) fabric), showing through an aperture of
1
.0–1.3 cm in diameter and fastened between two glass
plates (7.5 × 7.5 cm) or sealed between two plastic plates (of
microscope slide size). The plates with smeared mucous
samples were sent to a laboratory for processing. If pro-
cessed at the place of collection, the work-up should have
started no earlier than 90 min after depositing the sample to
obtain sufficient specimen adherence to the fabric. The
plates carrying specimens could be stored for short periods
of time at room temperature (several days) or stored frozen
for prolonged periods of time (several months).
low band at R 0.37 containing 2,4-dinitrophenylhydrazones
f
Treatment of mucous specimens with Schiff reagent
Plates carrying mucous specimens were placed in the
holder part of a Wheaton glass staining dish (capacity 10
plates), immersed in phosphate buffer (0.1 M, pH 7.0), and
washed by gentle agitation for 10 min. The holder was then
transferred into a dish containing distilled water and gently
agitated for 3 min. The aqueous wash was repeated; any ex-
cess water was allowed to drip for 10–15 min to render the
plates and the holder visibly dry. The holder was then im-
mersed into a dish containing Schiff reagent for 2 min under
gentle agitation followed by three washes with distilled wa-
ter (3 min each). The plates were removed from the holder
and air dried. The pink-red-magenta colour (positive reac-
tion) appeared almost immediately in some cases (or within
a few minutes (20–25 min at the most)) after removal from
the Schiff reagent.
was cut out and extracted with 10% MeOH in CH Cl ; the
2
2
solvents were evaporated to give a deeply yellow solid,
which was subsequently analyzed by MS. Two major ions
were [M+H]+ (m/z = 421) and [M+H]+ (m/z = 449). H
1
NMR (CDCl ) confirmed the presence of 2,4-dinitrophenyl
3
derivatives of long chain aldehydes: 0.88 (t, 3H, CH , J =
3
6.84 and 7.08 Hz), 1.25–1.42 (m, CH ), 6.96 (t, CH=N),
2
7.53 (t, CH=N), 7.93 (d, H-6 (arom), J = 9.52 Hz), 7.96 (d,
H-6 (arom), J = 9.52 Hz), 8.29 (dd, H-5 (arom), J = 3.18
and 9.76 Hz), 8.33 (dd, H-5 (arom), J = 3.17 and 9.52 Hz),
9.12 (d, H-3 (arom), J = 2.45 Hz), 9.14 (d, H-3 (arom), J =
2.45 Hz), 11.00 (s, NH), 11.18 (s, NH).
Two controls were applied: (1) the above procedure up to
the chromatography was repeated with blank polyester
sheets (no mucous deposited) — no reaction with 2,4-di-
nitrophenylhydrazine was detectable; and (2) the above pro-
cedure was repeated using mucus from colectomies that did
not give a colour reaction with Schiff reagent — no reaction
with 2,4-dinitrophenylhydrazine was detectable.
Preparation of suitable Schiff reagent
Many variations of the reagent are described (6). The
Schiff reagent used in this study was prepared as follows: p-
rosaniline (0.4 g; also basic fuchsin, magenta) was added to
distilled water (220 mL) at approximately 95–97°C. The
mixture was stirred well, brought to boil for 5 min, cooled to
2,4-Dinitrophenylhydrazone of palmitic aldehyde from a
plasmalogen
A mixture of PE plasmalogen (from brain extract; 10 mg
5
0°C, and filtered through doubly folded paper filter. Aque-
in 2 mL CHCl ), a solution of 2,4-dinitrophenylhydrazine
3
ous HCl (1 N, 34 mL) was added to the filtrate with stirring
and was allowed to cool to room temperature. Sodium bi-
sulfite (2.34 g) was added and the solution was stirred well —
the flask was sealed with parafilm — and stored in the dark
at room temperature for 4 days. To the resulting straw-
coloured solution was added charcoal (NORIT; 300 mg); the
mixture was vigorously stirred for 1.5 min, and the charcoal
hydrochloride (25 mL, prepared as above), and isooctane
(20 mL) was stirred for 1 h at room temperature. The aque-
ous phase was separated and washed with isooctane; organic
phases were combined and evaporated to dryness. The yel-
low residue was dissolved in dry acetonitrile (0.4 mL) and
applied on a silica gel 60 plate (20 × 20 cm; thickness
0.2 mm) in a single line, and the plate was developed with
©
2003 NRC Canada

Krepinsky et al.
113
over Na SO , and evaporated to dryness to give an amor-
2
4
phous solid containing both the aldehyde (30%) and starting
1
ester (70%). H NMR (CDCl ) (aldehyde): 9.77 (d, J =
3
1,2
tract by evaporation to give a yellow amorphous solid+,
1.95 Hz, CH CH=O); 2.42 (m, J = 1.95 Hz, CH CH=O);
(unreacted ester): 3.7 (s, CH OC[=O]-).
3
2 1,2 2
which was examined by MS. The major ion was [M+H]
1
(
m/z = 421). H NMR (CDCl ; major isomer is E (syn); mi-
3
nor isomer is Z (anti)): 0.88 (t, 3H, CH , J = 6.84 and
Preparation of 2,4-dinitrophenylhydrazone of stearic
aldehyde
3
7
.08 Hz), 1.25–1.42 (m, 24H, CH ), 1.61 (q, 2H, CH β to
2
2
CH, J = 7.33 Hz and J = 14.65 Hz, E isomer), 1.67 (q, 2H,
A solution of the crude aldehyde (50 mg) in CH2Cl2
(4 mL) with added solution of 2,4-dinitrophenylhydrazine
hydrochloride (26 mL, prepared as above) was stirred at
room temperature overnight. After work-up (analogous to
the work-up described for 2,4-dinitrophenylhydrazone of
palmitic aldehyde) and following thin layer chromatography
on silica gel, a yellow amorphous solid was obtained. MS:
CH β to CH, J = 7.06 Hz and J = 15.38 Hz, Z isomer), 2.38
2
(
ddd, 2H, CH α to CH, J = 5.44, 7.33, and 12.94 Hz, Z iso-
2
mer), 2.43 (ddd, 2H, CH α to CH, J = 5.70, 7.57 Hz and J =
2
1
5
5
2.94 Hz, E isomer), 6.96 (t, 1H, CH=N, J = 5.37 and
.62 Hz, Z isomer), 7.53 (t, 1H, CH=N, J = 5.37 and
.13 Hz, E isomer), 7.93 (d, 1H, H-6 (arom), J = 9.52 Hz, E
+
1
isomer), 7.96 (d, 1H, H-6 (arom), J = 9.52 Hz, Z isomer),
449 ([M+H] ). H NMR (CDCl3, major isomer: E, minor
isomer: Z): 0.88 (t, 3H, CH3, J = 6.84 and 7.08 Hz), 1.25–
1.43 (m, 28H, CH2), 1.61 (q, 2H, CH2 β to CH, J = 7.31 Hz
and J = 14.89 Hz, E isomer), 1.67 (q, 2H, CH2 β to CH, J =
7.08 Hz and J = 15.65 Hz, Z isomer), 2.38 (ddd, 2H, CH2 α
to CH, J = 5.44, 7.33, and 12.94 Hz, Z-isomer), 2.43 (ddd,
8
8
9
1
.29 (dd, 1H, H-5 (arom), J = 3.18 and 9.76 Hz, E isomer),
.33 (dd, 1H, H-5 (arom), J = 3.17 and 9.52 Hz, Z isomer),
.12 (d, 1H, H-3 (arom), J = 2.45 Hz, E isomer), 9.14 (d,
H, H-3 (arom), J = 2.46 Hz, Z isomer), 11.00 (s, 1H, NH, E
isomer), 11.18 (s, 1H, NH, Z isomer).
2
H, CH α to CH, J = 5.68, 7.57 Hz and J = 12.94 Hz, E
2
Preparation of palmitic aldehyde (13)
A solution of diisobutylaluminum hydride (DIBAL-H) in
toluene (1 M, 2.95 mL, 2 eq) was added dropwise over
isomer), 6.96 (t, 1H, CH=N, J = 5.38 and 5.37 Hz, Z iso-
mer), 7.53 (t, 1H, CH=N, J = 5.61 and 5.13 Hz, E isomer),
7
1
.93 (d, 1H, H-6 (arom), J = 9.52 Hz, E isomer), 7.96 (d,
H, H-6 (arom), J = 9.52 Hz, Z isomer), 8.29 (dd, 1H, H-5
5
min to a solution of methyl palmitate (0.4 g, 1.48 eq) in
(
(
(
arom), J = 3.17 and 9.52 Hz, E isomer), 8.33 (dd, 1H, H-5
arom), J = 3.17 and 9.52 Hz, Z isomer), 9.12 (d, 1H, H-3
arom), J = 2.44 Hz, E isomer), 9.14 (d, 1H, H-3 (arom), J =
dry toluene (10 mL) cooled to –70°C. The reaction mixture
was stirred for an additional 15 min, and excess reagent was
decomposed by 2 N HCl (10 mL); the mixture was allowed
to warm up to 0°C, and the aqueous layer was separated and
extracted with ethyl acetate (2 × 10 mL). The combined or-
ganic extracts were washed with water (3 × 5 mL), dried
over Na SO , and evaporated to dryness to give an amor-
2
1
.45 Hz, Z isomer), 11.00 (s, 1H, NH, E isomer), 11.18 (s,
H, NH, Z isomer).
O-(2,3,4,5,6-Pentafluorobenzyl) oximes of aldehydes (14)
The above compounds were prepared from O-(2,3,4,5,6-
pentafluorobenzyl)hydroxylamine (250 µL of a 0.05 M solu-
tion in sodium acetate buffer, pH 5), added, in separate ex-
periments, to the phospholipid mixtures PL-1 and PL-2 (cf.
section “Separation of mucous components”) as well as to
authentic specimens of palmitic and stearic aldehydes (1 mg
in 100 µL of water), vortexed for 1 min, and allowed to react
for 30 min. Then 1 N HCl (10 µL) was added, and the reac-
tion mixture was extracted three times with hexane (1 mL).
The combined hexane extracts were dried over sodium sul-
fate and evaporated to dryness under a stream of nitrogen;
the residue was redissolved in hexane (50 µL). This solution
2
4
phous solid containing both the aldehyde (60%) and starting
1
ester (40%). H NMR (CDCl ) (aldehyde): 9.77 (d, J
=
3
1,2
1
.95 Hz, CH CH=O), 2.42 (m, J 1,2 = 1.95 Hz, CH CH=O);
2
2
(
unreacted ester): 3.7 (s, CH OC[=O]-).
3
Preparation of 2,4-dinitrophenylhydrazone of palmitic
aldehyde
The crude aldehyde (50 mg) dissolved in isooctane
25 mL) and a solution of 2,4-dinitrophenylhydrazine hydro-
(
chloride (25 mL, prepared as above) was stirred at room
temperature for 1 h. After work-up (analogous to the work-
up described for 2,4-dinitrophenylhydrazones of aldehydes
from plasmalogens) and ensuing chromatography, a yellow
(
1 µL injections) was used for GLC–MS identification of the
+
aldehydes. Negative ion CI-MS (reagent gas: ammonia) was
executed on a quadrupole mass spectrometer VG-Trio 2A in-
terfaced with a Hewlett-Packard 6890 Series II GC using a
DB-5 capillary column (length: 30 m, diameter: 0.25 µm),
employing helium as the carrier gas at a flow rate of
amorphous solid was obtained. MS: [M+H] m/z = 421. The
1
chemical shifts in the H NMR spectrum were identical with
1
the shifts in the H NMR spectrum of 2,4-dinitrophenyl
hydrazone of stearic aldehyde (cf. below).
–
1
1
3
2
mL min , and with the temperature increasing from 60 to
10°C while maintaining the injection port temperature at
50°C (15).
Preparation of stearic aldehyde
A 1 M solution of diisobutylaluminum hydride (DIBAL-H)
in toluene (2.95 mL, 2 eq) was added dropwise over 5 min
to a solution of methyl stearate (0.44 g, 1.48 eq) in dry tolu-
ene (10 mL) cooled to –70°C. The reaction mixture was
stirred for an additional 15 min, and excess reagent was de-
composed by 2 N HCl (10 mL); the mixture was allowed to
warm up to 0°C, and the aqueous layer was separated and
extracted with ethyl acetate (2 × 10 mL). The combined or-
ganic extracts were washed with water (3 × 5 mL), dried
Results
Aldehyde identification
Analytical chemists have been using Schiff reagent for de-
cades to demonstrate that the aldehydic group is present in
the examined samples. However, although the reagent is
©
2003 NRC Canada

1
14
Can. J. Chem. Vol. 81, 2003
very sensitive, it is extremely difficult to isolate the products
of its reaction with aldehydes (6, 7). 2,4-Dinitrophenylhy-
drazine in a strongly acid solution, on the other hand, pro-
duces chemically stable 2,4-dinitrophenylhydrazones, which
can be isolated and purified; often they can be obtained in a
crystalline form. Employing modern spectroscopic methods,
the chemical structure of such pure compounds can be deter-
mined with relative ease. As no aldehydic hydrogen H-C=O
could be identified in the NMR spectra of the extracts from
which 2,4-dinitrophenylhydrazones were prepared, the alde-
hydes had to be released from precursors during the treat-
ment with Schiff reagent. Since both Schiff reagent and 2,4-
dinitrophenylhydrazine operate in strongly acidic environ-
ments, it is reasonable to presume that the acidity of the re-
agents allowed for the release of the aldehydes from their
precursors. Therefore, 2,4-dinitrophenylhydrazones (2,4-
DNPH) are valuable derivatives for the identification of the
structure of markers. Indeed, palmitic and stearic aldehydes
were identified via their 2,4-dinitrophenylhydrazones as the
aldehydes originating in colorectal mucous from cancer pa-
aliphatic chains of fatty acid ester groups. The less polar
fraction consisted of a mixture of phospholipids containing
ethanolamine (PL-1), while the more polar fraction con-
tained a mixture of phospholipids with choline (PL-2); nei-
ther of the two fractions was further separable by
chromatography on silica gel. In the fraction PL-1, hydrox-
yls in the 2-position and partially in the 1-position of sn-
glycerol were esterified with fatty acids. The ethanolamine
1
(-OCH CH NH ) moiety was identified by H resonances at
2
2
2
δ = 3.13 ppm (triplet) and at δ = 4.03 ppm (overlapping with
H-3 of sn-glycerol). The presence of a signal of an α,β-un-
saturated ether was confirmed through a doublet (J1
=
,2
6.1 Hz) at δ = 5.90 ppm, which was assigned to the vinylic
proton O-CH=CH-. This proton was further coupled to a
quartet at δ = 4.38 ppm and a multiplet at δ = 2.02 ppm in a
COSY spectrum confirming the presence of the O-CH=CH-
CH moiety in the molecule. Other allylic CH groups giv-
2
2
ing rise to signals at δ = 2.82–2.88 ppm were probably pres-
1
ent in the olefinic chains of aliphatic acids. Two
H
multiplets assigned to H-2 of the glycerol moiety centered at
δ = 5.18 ppm and δ = 5.25 ppm suggested that this fraction
contained two derivatives of ethanolamine-containing phos-
pholipids. The integrations of the vinylic signal and the H-2
glycerol signal at δ = 5.18 ppm were of the same intensity
(1:1). The estimate of the ratio (7:3) between the 1-alk-1′-
enyl-2-acyl-sn-glycerophospholipids (= plasmalogens) and
1,2-diacyl-sn-glycerophospholipids was made on the basis of
a comparison of the integration of signals for CH-2 of the
glycerol moiety (cf. above).
1
tients by the combination of MS and H NMR spectroscopy.
First, strong molecular ions were found in FAB-MS, which
corresponded to the 2,4-dinitrophenylhydrazones of palmitic
+
+
(
[M+H] : m/z = 421) and stearic ([M+H] : m/z = 449) alde-
1
hydes. This finding was confirmed by H NMR spectros-
copy, by comparison of the NMR spectra of the 2,4-DNPH
of aldehydes obtained from mucous fractions with synthetic
authentic specimens, and by a further comparison with the
1
previously published H NMR data. As expected, two geo-
metric isomers were present in all specimens of 2,4-DNPH
of aldehydes and assigned as described in reference 16. The
In the more polar fraction PL-2, OH-2 and, partially, OH-1
in sn-glycerol were again esterified with fatty acids. The sig-
nals at δ = 3.24 ppm and δ = 3.40 ppm were identified as be-
longing to two different phospholipids containing a choline
moiety (CH CH -N(CH ) ). As in the PL-1 fraction, this
1
H NMR chemical shifts of the 2,4-dinitrophenylhydrazones
of palmitic and stearic aldehydes were found to be practi-
cally identical. These results indicated that the above alde-
hydes are the markers detected by Schiff reagent in
colorectal mucous. The next question to be answered was:
From what precursors were the aldehydes released?
2
2
3 3
fraction contained an α,β-unsatured ether, which was con-
firmed through a doublet at δ = 5.90 ppm assigned to the
vinylic proton O-CH=CH-. Two proton multiplets, centered
at δ = 5.16 ppm and δ = 5.25 ppm and assigned to H-2 of the
glycerol moiety, suggested that this fraction contained two
derivatives of choline-containing phospholipids. The integra-
tion of the vinylic signal and the H-2 sn-glycerol signal at
δ = 5.16 were again of the same intensity (1:1). The ratio be-
tween the 1-alk-1′-enyl-2-acyl-sn-glycerophospholipids (=
plasmalogens) and 1,2-diacyl-sn-glycerophospholipids in
PL-2 was estimated to be 4:1 by comparison of the integra-
tion of signals for CH-2 of the glycerol moiety (cf. above).
During the chromatography of colorectal mucous, we ob-
served that the fractions containing phospholipids produced
the strongest colour with Schiff reagent. Since the only
phospholipids that could release the long chain aliphatic al-
dehydes by treatment with acids are the plasmalogens, we
therefore attempted to detect plasmalogens in the phospho-
lipid fraction. To purify the plasmalogens, the phospholipid
fraction was further rechromatographed on silica gel col-
umns. Two well-separated groups of phospholipids (PL-1
and PL-2) were obtained. The first group contained
ethanolamine and the other contained choline classes of
phospholipids; both groups gave strong colouration on treat-
ment with Schiff reagent. As these groups could not be fur-
Stereochemistry of the glycerol moiety was not deter-
mined, although it was assumed that it does not differ from
stereochemistry commonly found in natural phospholipids
1
based on comparisons of the H NMR spectra. Similarly, the
1
ther fractionated, we employed H NMR spectroscopy to
geometry of the vinylic ether double bond in plasmalogens
isolated from the colorectal mucus of cancer patients and
plasmalogens from other natural sources was found to be
identical (based on the identity of the coupling constant of
confirm the presence of the plasmalogens in these fractions
(
and consequently, to confirm the hypothesis that the alde-
hydes were released from the plasmalogens).
1
the vinylic signal for H at δ = 5.90 ppm).
Plasmalogen identification
The aldehydes released by hydrolysis of PL-1 and PL-2
were identified as stearaldehyde and palmitaldehyde by
comparison with authentic specimens of O-(2,3,4,5,6-penta-
fluoro)benzyl oximes of the aldehydes using MS combined
with GLC. These derivatives exhibited positive ions at m/z
1H NMR analysis (17) showed that both fractions were
phospholipids with long fatty chain acids (sn-glycero-3-
1
phosphoryl derivatives) based on H resonances characteris-
tic for glycerol and CH -groups of polar head groups and for
2
©
2003 NRC Canada

Krepinsky et al.
115
4
[
35 ([M – H]), 415 ([M – 20]), and 350 ([M – 50]) — with
M – 20] predominating owing to the loss of HF (palmit-
aldehyde) — and at m/z 463 ([M – H]), 443 ([M – 20]), and
13 ([M – 50]), again with [M – 20] predominating, for
formed, and the choice of a variant of Schiff reagent.
Through these investigations we could confirm that nearly
all mucus specimens from patients with colorectal cancer
were positive. For instance, the aldehydes were detected in
close to 94% of the mucus samples removed from colectomy
specimens obtained in several Toronto hospitals from 101
patients diagnosed with colorectal cancer by colonoscopy
4
stearaldehyde.
Structure of the marker
(
cf. Fig. 1). In rectal mucous specimens collected from colo-
Thus, we have identified the markers as a mixture of long
chain aliphatic aldehydes released from plasmalogens, pre-
dominantly palmitic (III) and stearic (IV) aldehydes (cf.
Scheme 1). Plasmalogens are phospholipids of both the
ethanolamine V and the choline VI types (cf. Scheme 1) in
which the aldehydes are cryptic in the form of unsaturated
ethers (18). A minor contribution from other aldehydes can-
not be excluded; similarly, acetal phospholipids in which the
aldehydes are bound as acetals to the C1–C2 of glycerol
may be minor precursors. Because of the acid lability of un-
saturated ether bonds, the aldehydes are released from plas-
malogens under very mild acid conditions (19); therefore, it
is not surprising that the aldehydes are released upon expo-
sure to the strongly acidic Schiff reagent. The free alde-
hydes, on reaction with the leuko (colorless) form of the dye
in Schiff reagent, provide shades of magenta colour consid-
ered characteristic of aldehydes (7, 8).
rectal cancer patients attending the Endoscopy suite of the
Wellesley Hospital (after the lavage of the colon but before
the colonoscopy) and in the rectal mucous of patients with
polyps a high proportion of positive results was obtained (cf.
Fig. 1). Similar results were obtained with mucous from pa-
tients before the lavage (data not shown). In symptom-free
individuals (47 students of medicine, ages 20–29), the alde-
hydes were not detectable in mucus from 43 individuals.
Mucous specimens were treated with Schiff reagent as de-
scribed in the Materials and methods section.
Discussion
A number of clinical (8) and population studies on the
mucous test employing the sequence galactose oxidase –
Schiff reagent have shown (22) that the test has the potential
The lack of colour of the Schiff reagent depends critically
(
23) to become an alternative to the fecal occult blood test
on the presence, in sufficient concentration, of SO and HCl
2
(FOBT) (24) as an efficaceous, simple, and economical
screening tool for colorectal cancer (25). Some of the short-
comings of the galactose oxidase – Schiff reagent test appear
to be due to the targeting of the incorrect carbohydrate
marker, one which required repeated, extensive washings and
treatments with enzymes (galactose oxidase and catalase),
leading to the loss of the marker from test plates. Our find-
ing — that the marker of the mucous test for colorectal can-
cer screening is of a lipid rather than of a carbohydrate
character (including the disaccharide T-antigen) — makes it
possible to correlate more accurately the presence of the
marker with the disease and to develop a more reliable
screening test for the disease. Significantly increased quanti-
ties of the plasmalogens in neoplasms of the large bowel
in the reagent. Should this concentration decrease, the Schiff
reagent may be partially restored to the original red colour
of p-rosaniline. Therefore, depending on the aldehyde con-
centration, the shade of colour on the solid support may be
observed as a mixture of red (20) and magenta (purple). The
adduct of p-rosaniline–aldehyde (resulting from the reaction
of an aldehyde with Schiff reagent) has a typical absorption
>
560 nm; therefore, the presence of such an absorption band
in a reflectance spectrum should resolve the occasional col-
our identification problems (21). Schiff reagent further re-
sorts to the original colour of p-rosaniline on treatment with
an excess of basic substances, for instance, basic amino ac-
ids such as lysine, histidine, and arginine; nucleosides such
as cytidine; or simple bases such as sodium carbonate. Thor-
ough washing with water and a buffer solution ensures that
all such result-altering components of mucous are removed
from the specimens before treatment with the Schiff reagent.
(
26, 27) and some other tissues (28, 29) were independently
31
found by P NMR spectroscopy. Since the amounts of both
the ethanolamine and choline families of plasmalogens in
neoplastic tissues appear to be significantly increased com-
pared with disease-free tissues, it appears likely that it is a
biochemical pathway leading to the unsaturated ether forma-
tion that has a connection with the neoplastic transformation.
Plasmalogens are a class of ubiquitous membrane phos-
pholipids. Although the highest levels are normally found in
tissues such as brain and heart, small concentrations have
been reported in most other normal tissues, and small in-
creases in plasmalogen levels in cancer cells have been pre-
viously observed (30, 31). Increased concentrations in the
colon were ascribed to a combination of an increased level
of the synthetic enzyme CTP – phosphocholine cytidyltrans-
ferase and a decreased concentration of hydrolytic enzymes,
phospholipases C and D (32). However, biosynthesis of the
“aldehydic moiety” of plasmalogens is varied (33, 34).
Presently, a connection between colorectal cancer and
plasmalogens can only be hypothesized. For instance, plas-
malogens, because of their effect on membrane fluidity, may
Potential of the detection of marker aldehydes in a
screening test for colorectal cancer
In the course of the marker identification studies it was
necessary to use mucous from individuals whose disease sta-
tus was known (usually determined by colonoscopy). It was
also necessary to determine whether the treatment of the col-
lected mucous with Schiff reagent resulted in colour forma-
tion (positive reaction). Therefore, rectal mucous was
obtained from patients with colorectal cancer, from patients
with various other pathological findings, and from individu-
als who had no complaints and thus were presumed to be
disease free. It was also necessary to establish suitable con-
ditions (as described in the Material and methods section) of
the treatment of specimens, such as length of time for the re-
action with the Schiff reagent, washings with buffers and
water, different supports on which the treatment was per-
©
2003 NRC Canada

1
16
Can. J. Chem. Vol. 81, 2003
facilitate the cellular uptake of carcinogenic substances (35),
e.g., of dietary (36) or microbial origin (37, 38). It is of in-
terest to note that the majority of strict anaerobes contain
detectable plasmalogens, while aerobic and facultative anaero-
bic bacteria lack plasmalogens (39). However, plasmalogen-
like fecal mutagens (fecapentaenes, produced in the human
colon by several anaerobes) seem to play no role in the etiol-
ogy of human colon cancer (19).
Cancer Institute of Canada (NCIC), the Natural Sciences and
Engineering Research Council of Canada (NSERC), and
ProcyonBiopharma, Inc. (Montréal, Que., Canada) is grate-
fully appreciated.
References
An aldehyde-based test for colorectal cancer screening
would be expected to have an advantage over fecal occult
blood test (FOBT) and colonoscopy (for screening purposes
only!). It could, potentially, uncover earlier stages of the
neoplastic progression than both FOBT and colonoscopy be-
cause the latter two techniques target more developed can-
cer. Therefore, the aldehyde test could detect cases not
sufficiently developed to be detected by colonoscopy. For in-
stance, we have found several patients with strongly positive
aldehyde mucus tests who were given a clean bill of health
by colonoscopy (therefore they were rated as “false”
positives). The same patients returned to the clinic within no
later than 1–2 years with frank cancer. Given the lengthy
progression of colon carcinogenesis, it is not unlikely that
these individuals already harbored the disease at the time of
the first aldehyde test. Since it is believed that the progres-
sion from early neoplastic changes to frank cancer may span
decades (40), an earlier detection of the disease may lead to
the realization of the hope that drugs could control and per-
haps reverse the disease progression. Early results with non-
steroid antiinflammatory drugs (e.g., aspirin) (41), inhibitors
of cyclooxygenase-2 (COX-2) (42), and identification of
molecular targets of these drugs (43) suggest that such hopes
are not unrealistic. Equally encouraging is the prospect that
the early neoplastic changes could be reverted, for example,
through dietary or other lifestyle measures.
1. H.E. Mulcahy, M.J.G. Farthing, and D.P. O’Donoghue. Br.
Med. J. 314, 285 (1997).
2. R.F. Thoeni. In Neoplasms of the digestive tract: imaging,
staging, and management. Edited by M.A Meyers. Lippincott-
Raven Publishers, Philadelphia, PA. 1998. p. 237.
3
. S.H. Itzkowitz, M.Yuan, C.A. Montgomery, T. Kjeldsen, H.K.
Takahashi, W.L. Bigbee, and Y.S. Kim. Cancer Res. 49, 197
(
1989).
. G.D. MacLean and B.M. Longenecker. Semin. Cancer Biol. 2,
33 (1991).
. H. Xu, K. Sakamoto, and A.M. Shamsuddin. Arch. Pathol.
Lab. Med. 116, 1234 (1992).
4
5
4
6
7
. F.H. Kasten. Int. Rev. Cytol. 10, 1 (1960).
. (a) J.H. Robins, G.D. Abrams, and J.A. Pincock. Can. J.
Chem. 58, 339 (1980); (b) G.S. Nettleton and A.-M. Carpen-
ter. Stain Technol. 52, 63 (1977).
8
9
. A.M. Shamsuddin. Anticancer Res. 16, 2193 (1996).
. A.M. Shamsuddin. U.S. Patent 5 348 860, September 20, 1994.
1
0. C.G. Hamberg and M. Tolvanen. In Guide to techniques in
glycobiology. Methods in Enzymology. Vol. 230. Edited by
W.J. Lennarz and G.W. Hart. Academic Press, San Diego, CA.
1994. p. 32.
11. K.S. Yeung, J.Chociej, G.P. Kandel, and J.J. Krepinsky. Proc.
Annu. Meet. Am. Assoc. Cancer Res. 36, A1251 (1995).
12. J. Folch, M. Lees, and G.H. Sloan-Stanley. J. Biol. Chem. 226,
497 (1957).
1
1
3. E. Winterfeldt. Synthesis, 617 (1975).
4. F.J.G.M. van Kuijk, D.W. Thomas, R.J. Stephens, and
E.A. Dratz. Biochem. Biophys. Res. Commun. 139, 144
In summary, we believe that the identification of alde-
hydes from plasmalogens — rather than galactose-containing
oligosaccharides — as the substances in colonic mucous that
seem to correlate with colorectal cancer and that react with
Schiff reagent, will make it possible to design meaningful
epidemiologic studies that will confirm whether the alde-
hydes are indeed markers of colorectal cancer suitable for
population screening.
(
1986).
1
1
1
5. X.P. Luo, M. Yazdanpanah, N. Bhooi, and D.C. Lehotay. Anal.
Biochem. 228, 294 (1995).
6. S.F. Tayyari, J.L. Speakman, M.B. Arnold, W. Cai, and
M.A. Behforouz. J. Chem. Soc. Perkin Trans. 2, 2195 (1998).
7. M.L. Sparling, R. Zidovetzki, L. Muller, and S.I. Chan. Anal.
Biochem. 178, 67 (1989).
1
1
2
8. F. Paltauf. Chem. Phys. Lipids, 74, 101 (1994).
9. J.J. Krepinsky. Prog. Biochem. Pharmacol. 22, 35 (1988).
0. H. Zollinger. Colour chemistry. Syntheses, properties and ap-
plications of organic dyes and pigments. 2nd ed. VCH Pub-
lishers, Weinheim. 1981. p. 398
1. J.J. Krepinsky, R. Furrer, M. Chen, S.P. Douglas, and J. Camp-
bell. J. Chem. Soc. Chem. Commun. 2003. In preparation.
2. D.Y. Zhou, F.C. Fang, Y.L. Zhang, Z.S. Lai, W.D. Zhang, L.B.
Li, G.L. Xu, T.M. Wan, D.S. Pan, D. Zhou, Y.C. Zhang, and
S.B. Li. Chin. Med. J. 106, 739 (1993).
Acknowledgments
We are grateful to our clinicical colaborators without
whom the study would not be possible: Claude Burul and
Thomas Harmantas (St. Joseph’s Health Centre, Toronto,
Ont., Canada); Jean Couture, Robin S. McLeod, Carol J.
Swallow, Helen M. McRae, Zane Cohen, R. Warren, J. Hartwick,
Alan Wolff, and Joanne Boone (Mt. Sinai Hospital, Toronto,
Ont., Canada); Gregory Haber, Paul Kortan, Joseph
Mamazza, and Norman E. Marcon (St. Michael’s Hospital,
Toronto, Ont., Canada); and Michael A. P. Smith (Rudd
Clinic, Toronto, Ont., Canada). We thank the staff of the
Molecular Medicine Research Centre, University of Toronto,
for NMR and MS analyses and Dr. D.C. Lehotay, Hospital
for Sick Children, Toronto, for GC–MS analyses of O-
2
2
2
3. M.A. Shamsuddin, G.T. Tyner, and G.Y. Yang. Cancer Res. 55,
1
49 (1995).
2
2
4. S. Otto and S. Eckhardt. J. Surg. Oncol. 75, 220 (2000).
5. I.T. Said, A.M. Shamsuddin, M.A. Sherief, S.G. Taleb,
W.F. Aref, and D. Kumar. Histol. Histopathol. 14, 351 (1999).
26. T.E. Merchant, J.N. Kasimos, P.W. de Graaf, B.D.
Minsky, L.W. Glerke, and T. Glonek. Int. J. Colorectal Dis. 6,
121 (1991).
(
2,3,4,5,6-pentafluorobenzyl) oximes of aldehydes. The fi-
nancial support of this research program from the National
©
2003 NRC Canada

Krepinsky et al.
117
2
7. T.E. Merchant, P.M. Diamantis, G.Y. Lauwers, T. Haida,
J.N. Kasimos, J. Guillem, T. Glonek, and B.D. Minsky. Can-
cer, 76, 1715 (1995).
36. B.N. Ames and L.S. Gold. Biotherapy, 11, 205 (1998).
37. J.B. Brooks, W.B. Cherry, L. Thacker, and C.C. Alley. J. In-
fect. Dis. 126, 143 (1972).
2
2
8. T.E. Merchant, P. Meneses, L.W. Gierke, W. Den Otter, and T.
38. K. Kanazawa, F. Konishi, T. Mitsuoka, A. Terada, K. Itoh,
S. Narushima, M. Kumemura, and H. Kimura. Cancer, 77
(Suppl. 8), 1701 (1996).
39. H. Goldfine and P.-O. Hagen. In Ether lipids. Edited by
F. Snyder. Academic Press, New York, NY. 1972. p. 329.
40. J.-L. Tsao, Y. Yatabe, R. Salovaara, H.J. Järvinnen, J.-P.
Mecklin, L.A. Aaltonen, S. Tavaré, and D. Shibata. Proc. Natl.
Acad. Sci. U.S.A. 97, 1236 (2000).
Glonek. Br. J. Cancer, 63, 693 (1991).
9. T.E. Merchant, B.D. Minsky, G.Y. Lauwers, P.M. Diamantis,
T. Haida, and T. Glonek. NMR Biomed. 12, 184 (1999).
0. F. Snyder and R. Wood. Cancer Res. 28, 972 (1968).
1. F. Snyder and C. Snyder. Prog. Biochem. Pharmacol. 10, 1
3
3
(
1975).
3
3
2. D.-A. Dueck, M. Chan, K. Tran, J.T. Wong, F.T. Jay, C.
Littman, R. Stimpson, and C.P. Choy. Mol. Cell. Biochem.
41. E. Giovannucci, K.M. Egan, D.J. Hunter, M.J. Stampfer,
G.A. Colditz, W.C. Willett, and F.E. Speizer. N. Engl. J. Med.
333, 609 (1995).
1
62, 97 (1996).
3. F. Paltauf. In Ether lipids. Edited by H.K. Mangold and
F. Paltauf. Academic Press, New York, N.Y. 1983. p. 107.
4. H. Hayashi and M. Hara. J. Biochem. 121, 978 (1997).
5. S. Christen, T.M. Hagen, M.K. Shigenaga, and B.N. Ames. In
Microbes and malignancy (Infection as a cause of human can-
cer). Edited by J. Parsonnet. Oxford University Press, London.
42. M. Tsuji, S. Kawano, and R.N. DuBois. Proc. Natl. Acad. Sci.
3
3
U.S.A. 94, 3336 (1997).
43. I. Shureiqui, D. Chen, J.J. Lee, P. Yang, R.A. Newman,
D.E. Brenner, R. Lotan, S.M. Fischer, and S.M. Lippman. J.
Natl. Cancer Inst. 92, 1135 (2000).
1
999. p. 35.
©
2003 NRC Canada