Download Glycoprotein Changes in Tumours: A Renaissance in Clinical

Clinical Science (1997) 93,287-293 (Printed in Great Britain)
287
Editorial Review
Glycoprotein changes in tumours: a renaissance in clinical
applications
Elizabeth F. HOUNSELL, Mia YOUNG' and Michael J. DAVIES2
MRC Glycoprotein StructurelFunction Group, Department of Biochemistry and Molecular Biology,
University College London, Gower Street, London W C l E 6BT. U.K.
1. Oligosaccharides linked to protein (glycoprotein)
or lipid (glycolipid) are the major components at
the outer surface of mammalian cells. Studies using
antibodies and lectins have shown in the past that
the oligosaccharides they recognize exhibit tumourassociated changes, i.e. they are carbohydrate
tumour-associated antigens.
2. The oligosaccharides have been further characterized in recent years by structural analysis using
high-resolution chromatographic techniques, MS
and NMR. NMR gives an oligosaccharide fingerprint that is characteristic of monosaccharide type
and linkage and which can be correlated with magnetic resonance spectroscopic data on fine-needle
tissue aspirates.
3. Also of relevance is the new understanding of the
molecular biology of MUC genes, which code for
mucin protein backbones, and of the glycosyltransferase genes, which determine oligosaccharide
structure and immunological recognition.
4. For these reasons, we believe that tumour-associated oligosaccharide changes should be revisited in
the context of what we now know about structure
and expression. This review synopsizes the past
data using the detection of carbohydrate tumourassociated antigens by binding of lectins and antibodies, and puts it into the context of NMR
fingerprints or signatures.
INTRODUCTION
Oligosaccharides linked to protein (glycoprotein)
or lipid (glycolipid) are the major components at the
outer surface of mammalian cells. They are often
characteristic of cell type and provide a protective
layer against chemical and enzymic attack. It has
become clear that they also function as highly
specific recognition molecules which play an import-
ant role in the physiological properties of the cell.
Many structural changes during malignant transformation into tumours have been described, as discussed herein, but the functional implications of the
changes are not completely understood. Alterations
in the structure of cell-surface glycoconjugates are
considered to be relevant to the abnormal properties of cancer cells, such as uncontrolled cell growth,
altered cell adhesion, avoidance of immunological
destruction, invasiveness and metastatic spread [l].
In medicine, a large amount of attention has been
given to these structural changes for possible utilization in diagnosis (to distinguish between benign and
malignant lumps) and in prognosis, which is
dependent on the maturity of the tumour and its
metastatic potential. In the absence of metastasis
most cancers would be cured by surgery, chemotherapy and radiation therapy [2]. The desire is to
measure the effects of medical treatment and
identify recurrence as soon as possible. In addition,
new therapeutic regimes are required.
The oligosaccharide sequences that have been
implicated in altered cell adhesion and metastasis
are chain-terminating monosaccharides which are
present in the correct spatial arrangement for recognition. Particularly important in this regard are the
monosaccharides fucose (Fuc) and sialic acid (SA).
They are held in distinct conformational space
specific for each carbohydrate-binding protein by
different linkage to backbones of galactose (Gal)
and N-acetylglucosamine (GlcNAc). Shown in Fig.
1, for example is the sequence called sialyl Lea
(SLea). The constituent monosaccharides are ring
structures (most easily seen for GlcNAc on the
right-hand side of Fig. 1) on which are appended
different functional groups such as the CH3 of Fuc
(shown top, middle of Fig. 1) and the NH-CO-CH3
(also abbreviated to NHAc) for GlcNAc and SA.
The CH3, CH2, CH and NH (the unshaded circles in
Key words: cancer, carbohydrate tumour antigens, magnetic resonance spectroscopy, monoclonal antibody, mucin, NMR spectroxopy.
Abbreviations: AP. adenomatous polyps; Fuc, fucose; GlcNAc, N-acetylglucosamine; HP. hyperplastic polyps; HPA, Helix pomotio agglutinin; mAb, monoclonal antibody; MRS,
magnetic resonance spectroscopy; PSA, prostate-specific antigen; SA, sialic acid; SLea,sialyl Lea;SLe", sialyl Le"; ST, sialyl T; ST,sialyl T"; TAA, tumour-associated antigen.
Correspondence: Dr E.F. Hounsell.
'Present address: Skyehusapoleket I Fredribtad, Postbokj 1026, I06 I Fredrikstad,Norway.
'Present address: Lonza Biologics PIC,228 Bath Road, Slough, Berkshire SL14DY, U.K.
E. F. Hounsell et al.
288
Fig. 1) give signals detected and characterized by
NMR spectroscopy. The position of each of the signals in the NMR spectrum is dependent on the
chemical environment of the atom within the molecule [3]. The CH3 of Fuc in SLe" can be seen to be
surrounded by oxygen atoms, for example (shaded
in Fig. 1, found within the ring and forming the linkages between the monosaccharides and unlinked
hydroxyl groups). Each different type of linkage of
Fuc to the backbone Gal and GlcNAc (Fig. 2) gives
a distinct signature in the NMR spectrum [4]. The
amount of each sequence present in tumours and
cell lines can now be explored by magnetic resonance spectroscopy (MRS) using easy to obtain twodimensional MRS images from fine-needle aspirates
of tumours or cancer cell lines. Studies so far have
shown different patterns of fucosylation dependent
on the stage of cancer [5, 61. This is an exciting new
development in tumour diagnosis which we discuss
here in the context of our knowledge about tumourassociated oligosaccharide changes.
Carbohydrate antigen structures
Historically, the first cellular carbohydrate (i.e.
oligosaccharide) antigens that were characterized
were the so called blood group antigens on human
erythrocytes [7], based on the presence of three
w
Fig. I. A minimum energy conformation of the oligosaccharide
sequence sialyl Leain the orientation shown
Fucal-3,
GlcNAc
,GalP1-4'
SAa2-3
SA is in this case N-acetylneuraminic acid (NeuSAc). The N-acetyl groups
(NH-CO-CH3) of NeuSAc and GlcNAc are indicated together with the CHI
of Fuc. The hydroxyl groups, 0 - 0 , and oxygens, 0 , are available for hydrogen binding to the protein-combining site of lectins or antibodies. CHI
groups and clustered ring C-Hs will stack onto aromatic side-chains of
amino acids.
Fucal-2Calpl-3GlcNAc
Type 1 H
Fucal-2Galpl-4GlcNAc
~ y p 2e n
Fucal-2Gal~l-3GalNAc
Typc3H
Fucal-3[Gnlpl-4~ClcNAc
Le'
Fucal-4[Gal~l-3~ClcNAc
Lea
Also Le' with Fucal-2 on the Gal is L d and similarly Le' with Fucal-2 on the Gal is Leb.
Gal may also have Galal-3 or CnlNAcal-3 attached which are blood group B or A.
respectively when the Fucal-2 is present eg ALd.
GalNAcal-3,
Galol-4,
Fucal-2'
FlcNAc
Fucal-3
ALd
Repeating Le' can occur with or without sialylation cg.
SAaZ-3Calol~
CjlcNAcpl-3Gnlpl-4
Fucal-3
(~ICNAC
Fucal-f
SLe'Le'
Fig. 2. Different types of fucose-containing sequence
structurally different fucosylated antigens, H, A, or
B (Fig. 2). The H structure, Fucal-2Gal/31-, may be
considered as a precursor giving rise to A or B by
the action of specific glycosyltransferases which add,
respectively, GalNAcal-3 or Galcrl-3 to the Gal in
the H structure. Later it was discovered that these
and related oligosaccharide sequences are found on
cells in most tissues [8-lo]. They may be presented
on glycolipids and different types of glycoproteins.
Two main classes of glycoprotein oligosaccharide
chains are defined, depending on their linkage to
either Asn (N-linked) or Ser/Thr (0-linked) amino
acids. The blood group and related antigens occur at
(distal) ends of both N- and 0-linked chains. In
mucins, they can occur in multivalent form on many
neighbouring 0-linked chains, which gives a cooperative effect for the interaction with antibodies or
naturally occurring carbohydrate-binding proteins
(lectins).
Much glycoproteins found in the lining of the
gastrointestinal and respiratory tracts are large, with
molecular weight from 1000-15000 kDa, and often
contain more than 50% carbohydrate, which gives
rise to structural diversity and antigenic polyvalency.
The protein backbones, coded by the MUC genes,
contain tandem repeats of around 20 amino acid
residues with a high content of Ser and Thr amino
acids. On average, one in every three Ser or Thr
residues may have GalNAc attached (the so-called
T" antigen), which can be further substituted by
other diverse monosaccharides to give eight different protein-to-oligosaccharide core regions [3, 81.
Each of these can have additional Fuc and SA residues or can be elongated by backbone sequences
which bear blood groups and other related antigens
with or without additional SA (Fig. 3). The oligosaccharide chains are built up sequentially by glycosyltransferase enzymes in the Golgi apparatus, for
example forming the colon cancer associated mucin
antigens T", T, sialyl T" (ST"), sialyl T (ST), Lex, sia-
289
Glycoprotein changes in turnours: a renaissance in clinical applications
GalNAc
GalPlJGalNAc
SAa2-CGalNAc
T
\
Glycoprotein (mucin)
ST"
SAa2-3Galpl-3GalNAc
sr
SAa2-6,
9aiNAc
GalP1-3
ST
SOr3GalP1-4,
GleNAcPl+
Fueal-f
FalNAc
SAa2-3GalPl-3
DNA (MUC genes)=, mRNA- protein backbone
T
SulphoLe'
Fig. 3. Common mucin core region and sialylated or sulphated oligosaccharide sequences
lyl Le" (SLe") and SLe" etc. (Figs. 2 and 3). Mucins
and mucin-type oligosaccharides are also found in
serum and on cell surfaces, where they have been
implicated in suppression of the immune response
to protein tumour antigens [ll]. Additionally,
mucins may specifically interact with antibodies
directed to the same oligosaccharides on the tumour
cell, thus inhibiting cell-mediated cytotoxicity at the
tissue. Sialylated mucin oligosaccharides are the
ligands for sialoadhesins, whereas in non-sialylated
form they may respond to hyperproliferative stimuli.
For example, the T antigen is unmasked in colon
carcinoma [12], and this can bind to peanut and
mushroom lectins. The former has been shown to
cause proliferation of human colonic epithelium and
colorectal cancer cell lines [12], whereas the latter
causes suppression of proliferation. These studies
highlight the fine differences in the specificity of
GalP1-3GalNAc-binding lectins and antibodies,
caused by the degree of clustering, tolerance of
other substituents, amino acid sequence in the peptide to which GalNAc is attached, etc.
Detection of alterations in antigen expression in tumours
A variety of carbohydrate markers have been used
to study the relevance of carbohydrate tumourassociated antigens (TAAs) in the clinical characterization of human carcinomas. The most frequent
alteration of antigen expression in tumours is
increased expression of naturally occurring carbohydrate TAA, with concomitant decreased expression of precursor sequence or alternate biosynthetic
product. Sometimes aberrant expression may also be
observed, which means that antigens not normally
found appear. Less often, deletion of detectable
expression occurs. Figure 4 illustrates how the
expression of the antigenic structures is controlled.
The protein backbone and the glycosyltransferases
are both encoded by their respective genes. The
transferases act by sequentially adding monosaccharides to the growing carbohydrate chain to
build up the mucin. In colon carcinomas, a 10-fold
increase in expression of the MUC protein back-
DNA (transferase)
- mRNA
/
protein enzymes
Fig. 4. Relationship of M U C and glycosyltransferase genes to the
expression of mucin carbohydrate TAAs
bone has been reported, along with increased
activity of the relevant transferases [13, 141,
although this is not always the case. It is therefore
suggested that several control mechanisms are
involved in the observed structural changes. Eight
MUC genes have now been characterized (Table 1).
The functional implications of the changes in carbohydrate structure are not completely understood at
present, but it is believed that the changes in cell
surface glycoconjugates are reflected in the altered
physiological properties and behaviour of the transformed cells.
The potential of the new molecular biological
methods for following alterations in mucin structure
(i.e. at the MUC and glycosyltranferase gene levels)
has not yet materialized. Traditionally, lectins were
used to distinguish carbohydrate epitopes. Lectins
are proteins or glycoproteins with specific affinity for
oligo- or mono-saccharides, or parts of larger structures, and as such, their specificities may not be
totally restricted to one antigen. The advent of polyclonal antisera increased the specificity of epitope
detection, although it was not until the arrival of
monoclonal antibodies (mAbs) that accurate epitope
detection became possible. mAbs have often been
found to be directed towards carbohydrate antigens
[3, 9, lo], specific examples of which will be given in
this review. Other mAbs have been shown to recognize conformational epitopes requiring both oligosaccharide and protein [ l l , 151. Today, improved
analytical methods, such as MS or HPLC, enable us
to determine carbohydrate structures from small
amounts of sample. It is therefore possible to confirm the presence of a suggested oligosaccharide
sequence on biopsies [12, 161. NMR spectroscopy
has been able to characterize oligosaccharides in
great detail from large amounts of mucin material,
but MRS has the potential to correlate this information with data obtained directly from biopsies as discussed below.
Table I. Characterized M U C genes
MUC I
MUC 2
MUC 3
MUC 4
MUC SAC
MUC 58
MUC 6
MUC 7
MUC 8
Breast and colon cell surface episialin
Colon and small intestine goblet cell secretion
Intestinal tissue
Trancheobronchialtract
Respiratorytract and goblet cell secretion
Submaxillary gland secretion
Gastric gland secretion
Salivary gland secretion
Respiratorytract
E. F. Hounsell et al.
290
Colon carcinoma
Breast tumours
Table 2 shows the carbohydrate TAA of special
importance in colon cancer [13, 14, 17-24], together
with examples of the mAbs (or for the T antigen the
peanut lectin PNA) used to detect each antigen. The
presence or absence of the antigen in normal tissue,
hyperplastic polyps (HP), adenomatous polyps (AP)
or carcinoma is shown. The ideal antigenic marker is
present in AP (which is harmless by itself, but has
the ability to turn into carcinoma) and carcinoma,
but not in normal tissue or AP. SLe" and LeX are
both expressed by normal tissue and HP as well as
AP and carcinomas [18]. That means that these antigens are of limited value as markers. The LeXLeX,
SLexLex and ST" antigens, which are all absent in
normal tissues and HP, are therefore the markers of
choice. A colonic carcinoma that had metastasized
to the liver expressed a higher level of the FH6 antigen (Table 2) than the primary tumours [21]. The
presence of ST" in mucinous colon carcinomas is
often associated with advanced disease and has a
poor prognosis [24,25].
Schwartz et
al.
[26] inhibited
mucin
0-glycosylation with benzyl-N-acetyl-galactosamine
(a competitive inhibitor of GalNAc-Sermhr transferase), giving reduced metastasis-related events
such as adhesion and invasiveness. These studies
suggested that, in addition to the increased expression of the unsialylated T antigen, there is an
increase in both sialylated and also highly fucosylated oligosaccharides (LexLex, SLexLex) in the
primary colon carcinoma and in metastatic sites. In
some ways these are contradictory results, showing
both an increase and decrease in sialylation levels,
but also illustrating the subtleties of possible variations in oligosaccharide structure. MRS analysis of
secreted glycoproteins from adenoma and carcinoma
cells of increasing tumorigenicity show an increase
in the number of different patterns of fucosylation
[5]. The most tumorigenic cells had so-called crosspeaks IIa and IIb, which fit best [3, 41 with the diagnostic profile for separate Fucal-2 (e.g. blood group
H) and Fucal-3 (Lex-type) linkages. Cross-peaks I11
and IV, also present in normal tissue, are most
likely to represent the two fucoses of the Leb-type
structure from their positions (chemical shifts) in
the MRS and NMR spectra.
After the colon, the breast is the most prevalent
organ where mucin carbohydrate structural changes
have been characterized as TAAs. Table 3 lists different antigens with the lectin (Table 4) or mAb
used for identification, and an indication of the usefulness of the antigen as a marker in breast cancer.
Each is described further below. In contrast to colon
cancer, the structures T, ST, T" and ST" are all considered to be of little diagnostic value [27]. Lex has
also been found to be an unsuitable marker in
breast cancer, but a potential role in invasiveness
has been suggested [28].
Lectin reactivity to breast cancer antigens have
been extensively studied. Walker [29] included in his
regime a selection of lectins and studied their value
as markers of short-term prognosis in breast cancer.
None could give prognostic information over and
above that provided by histological evaluation. However, the lectin from the garden snail Helix pomutiu
(HPA) that recognizes terminal GalNAc residues is
more promising and has been investigated by several
groups, but the results obtained are conflicting [30].
HPA-binding was observed to normal breast epithelium and to the majority of carcinomas, and it was
later found that HPA recognizes a glycoprotein that
is associated with metastasis and poor prognosis in
breast cancer [31]. However, the International (Ludwig) Breast Cancer Study Group [32] reported no
clinical predictive value of HPA; the work continues
to characterize breast cancer glycans with terminal
immunodominant GalNAc, which could be implicated in aggressive biological behaviour [33].
In general, histological studies are very dependent
on the patients that are included and the methodology used. Discrepancies in results are also due to
the tumour cells not staining evenly. Some cells are
stained and some not. How many cells should be
positive in order to conclude that the tumour is
stained? These problems may be overcome by moving away from histological examination (with lectins
or antibodies) and identifying oligosaccharides from
previous lectin and antibody recognition studies as
markers that can be targeted for MRS analysis.
Fine-needle aspiration biopsy specimens have been
used for diagnosis of breast and other carcinomas
using mAb B72.3, but the antigenic determinant was
Table 2. Glycoprotein markers in colonic tissues. N = normal, HP = hyperplastic polyps, AP = adenomatous polyps,
C = carcinoma.
Antigen
SLea
Lex
Le'Le"
SLe'Le"
T"
T
ST"
MAb/lectin
N
HP
AP
C
Ref.
MSWl 13
t
t
-
t
t
t
t
-
tt
tt
tt
tt
tt
tt
tt
ttt
ttt
ttt
ttt
ttt
ttt
ttt
[I81
[191
[201
SSEA-I
FH4
FH6
MLS 128
PNA
JTlk
PI1
[221
~ 3 1
~ 4 1
29 I
Glycoprotein changes in turnours: a renaissance in clinical applications
Table 3. Glycoprotein markers in breast cancer
Structure
mAb/lectin
Use
T,ST,T",ST"
Le"
Carbohydrate
GalNAc
Epithelial mucin
Epithelial mucin
Non-sialylatedmucin
Sialylatedglycoprotein
Acidic glycoprotein
Mucin, NeuSGc
Structure
t
?
t It
t
t
IgM
Lectins
HPA
872.3
(315-3, MCA
83D4
BTAA
a 4 9
3E1.2
Ref.
I
text)
.!-
t
t
not characterized. The mAbs bl2/MCA and
DF3/CA15-3, also uncharacterized, appear to be
good prognostic indicators utilized in commercial
tests [35].
Pancino et al. [36] characterized a 300-1000 kDa
antigen of human breast carcinoma which reacted
strongly with wheat germ agglutinin and peanut
agglutinin, but gave no response to lentil or concanavalin A lectins (Table 4). The activity was strongly
reduced by periodate oxidation, which is a method
to destroy the carbohydrate, and also by trypsin
digestion, which cleaves the protein chain. The
activity was retained after neuraminidase treatment
which results in the cleavage of sialic acids. These
results indicate that a carbohydrate epitope is
involved, but not SA.
Pal et al. [37] purified a breast TAA (BTAA) with
a molecular weight of 85 kDa that may be useful in
diagnosis and prognosis. It was neuraminidase-,
trypsin- and papain-sensitive, thermostable at high
temperatures and stained by periodate/Schiff
reagent for detecting carbohydrates. The authors
concluded that breast TAA is a carbohydrate epitope involving SA. Another useful marker for stage
IV breast cancer (CA549) was also found to recognize an acidic glycoprotein [38]. This study also
tested mAbs raised against human milk fat globule
membranes, which are known to be a rich source of
fucosylated carbohydrate antigens that have also
been characterized by NMR [31, 391. Also easily
identified by NMR [4] is the difference between two
members of the SA family, i.e. N-acetylneuraminic
acid (NeuSAc), which we have been discussing so far
as SA, and N-glycolylneuraminic acid (NeuSGc),
which until recently was thought not to occur in
humans. However, Devine et al. [40] reported the
Table 4. The specificity of several of the lectins used in diagnosis
Lectin
Abbrevation
Specificity
Peanut agglutinin
Wheat germ agglutinin
ConcanavalinA
Lentil
Helix pornatia agglutinin
PNA
WGA
Con A
Lentil
H PA
/l-~-Gal( I-3)~-GalNAc
D-GIcNAc,NeuSAc,o-GalNAc
u-D-M~,a-D-Glc
N-D-M~UI
u-D-G~~NAc
presence of a mucin carbohydrate epitope defined
by antibody 3E1.2, which recognized an 0-linked
glycan containing NeuSGc. This finding has been
corroborated by Hanisch et al. [41] who have identified specific changes in glycosylation (glycoforms) of
the MUC-1 protein from a breast tumour surgical
resection and from T47D cells. These showed a lack
of fucosylated Le blood-group-related antigens and
the appearance of novel core-type ST antigen,
Gal~l-3[Neu5Gca2-61GalNAc.
Glycoprotein antigens in other neoplasms and general
clinical implications
Other studies have addressed changing carbohydrate
structure in a range of tumours, including prostate,
ovary, pancreas and lung. Table 5 shows some of
those for the prostate and pancreas where the
importance of mucin-type carbohydrate has been
characterized [42-SO]. TURP-27 is related to the
neural cell adhesion molecules recognized by the
mAb HNK-1, the epitope for which is known to
involve sulphated and polysialylated oligosaccharides
such as those on neural cell adhesion molecules.
Polysialic acid has also been shown to occur on
0-linked chains in breast cancer and leukaemia cell
lines [51], whereas that on neural cell adhesion
molecules is on N-linked chains. TAAs, which are
glycoproteins where the carbohydrate chains are of
the N-linked type, are found, such as the prostatespecific antigen (PSA) which is a glycosylated serine
protease of the kallikrein family [52, 531. For the
antigen 5T4 [15], a marker of several tumours, the
carbohydrate and protein are involved together in
antigenicity. 5T4 is involved in extravasation of the
placenta, which can be considered as a model for
metastasis and an oncodevelopmental paradigm.
Similarly, oligosaccharides were first discovered as
differentiation/developmental markers, particularly
Lex and L e y [3, 9, 10, 441 and SLe" (CA19-9 [48]). It
was shown that they had specific roles in morphogenesis by interacting with the carbohydrate-binding
proteins, called selectins. Results have been
obtained which indicate that SLex and SLea antigens
Table 5. Glycoprotein markers in the prostate and pancreas. PAC,
prostate carcinoma-associatedglycoprotein complex.
Antigen
Prostate
Sialoglycoproteincomplex PAC
Ld, H-type 2
Ley, di-Le" tri-Lex
TAG-72 mucin
Mucin sialylated oligosaccharides
Pancreas
Sialylated mucin
SLea
Sialylated mucin
Lea/type I mucin
mAb
TUW-2 7
BI
83
142,431
B72.3, CC83, CC89
PR92
1451
~461
DUPAN-2
CA19.9
Span-I, la3, Nd2
BW494
[47l
~481
MI
WI
1501
292
E. F. Hounsell et al.
on the surface of colon cancer cells function as
receptors for E and P selectin expressed on vascular
endothelium [54, 551. This may be the initial event
before extravasation of cancer cells from the capillaries [56].
That many of the markers reoccur for different
tumours points to a similar mechanism if they are
involved in pathogenesis. For example, PSA also
appears in lung and breast tumours [57]. Devine et
al. [58] compared PSA with other more highly glycosylated antigens as a TAA in colon, breast, prostate,
pancreas, lung, bladder and ovary. Changes in the
lung have also been correlated with mRNA levels of
several MUC genes [59], and the levels of mRNA of
a number of other MUC genes have been found
increased in different cancer tissues (recently
reviewed in [60]). Elevated serum levels of CA15-3
[35] and other mucins have been reported in
patients with lung, pancreatic, colon, breast and
ovary cancer [61]. MRS studies have looked at ovary
biopsies where the Fuc IIa and IIb cross-peaks, discussed above for colon studies, correlate with
increased tumorigenesis [6]. Compared with NMR
data [4, 8, 621, this pattern may be due to an
increase in the CA19.9 antigen (also implicated in
the pancreas [48]). The presence of SLe" and SLe"
may have a reciprocal presentation with respect to
T" and ST" [56], which is important because ST" can
mediate inhibition to natural killer cell cytotoxicity
[631.
These clues as to mechanism are being
approached with respect to new therapeutic regimes.
For example, O'Boyle et al. [64] vaccinated 20
cancer patients with mucin containing T" and ST"
determinants. A significant rise in the corresponding
antibodies was observed, which has prompted further clinical trials. The use of synthetic peptides
corresponding to the protein core of the polymorphic glycosylated epithelial mucin as a vaccine in
cancer patients is being explored [65], and the high
density of carbohydrate moieties peculiar to
tumours has prompted the re-exploration of mAb
directed towards them for use in stimulation of antibody-dependant cellular cytotoxicity [66]. Conjugates with mAb have been extensively investigated
as possible drug-targeting systems. In a recent
example using a radio-immunoconjugate against
mucin antigens, Peterson et al. [67] cured six of
seven mice breast tumours by treatment with tungsten-labelled mAb. The antigens were the major
components of the human milk fat globule membrane, which expresses high levels of fucosylated and
sialylated oligosaccharides [3, 9, 391.
In summary, epithelial cells of the colon, breast,
lung, ovary, prostate and pancreas express cell-surface much antigens that change during malignant
transformation. There is a link between changes in
the structure of cell carbohydrate and the abnormal
properties of cancer cells. The structural changes in
colon cancer are recognized by mAbs and confirmed
by analytical methods. A large amount of attention
has been given to these structural changes for possible utilization in the diagnosis of benign and
malignant tumours, measurement of the effectiveness of treatment, and prognosis based on tumour
maturity and metastatic potential. MRS offers a new
technique with which to access tumour type and
stage. By cross-matching to known oligosaccharide
antigen NMR chemical-shift signatures, this will
provide further knowledge with which to interpret
pathogenic mechanisms in the light of our increasing
understanding of endogenous carbohydrate-binding
proteins and much biosynthesis. The additional
potential outcome is in the design of new therapeutic regimes, which are highly desirable, e.g. drug
targeting with mAbs, vaccination with mucin antigens and manipulation of MUC or glycosyltransferase genes.
ACKNOWLEDGMENT
The authors are grateful to Mrs Gail Evans for
excellent secretarial assistance.
REFERENCES
I. Bhavanandan VP. Cancer-associatedmucin and mucin-type glycoprotein.
Glycobiology 1991; 1: 493-503.
2. Muramatsu T. Carbohydrate signals in metastasis and prognosis of human
carcinomas. Glycobiology 1993; 3: 294-6.
3. Hounsell EF. Physicochemical analysis of oligosaccharide determinants of
giycoproteins. Adv Carbohydr Chem Biochem 1994; 50: 3 I 1-50.
4. Hounsell EF, Davies MJ, Renouf DV. Studies of oligosaccharide and glycoprotein
conformation. Biochem SOC Trans 1992; 2 0 259-63.
5. Mackinnon WB, Huschtscha L, Dent K, Hancock R ParaskevaC, Mountford CE.
Correlation of cellular differentiation in human colorectal carcinoma and
adenoma cell lines with metabolite profiles determined by ' H magnetic
resonance spectroscopy. Int J Cancer 1994; 5 9 248-6 I.
6. Mackinnon WB, Russel P, May GL, Mountford CE. Characterizationof human
ovarian epithelial tumors (ex vivo) by proton magnetic resonance spectroscopy.
Int J Gynecol Cancer 1995; 5: 21 1-2 I.
7. Watkins WM. Biochemistry and genetics of the ABO Lewis and P blood group
systems. Adv Hum Genet 1980; 1 0 1-136,379-85.
8. Hounsell EF, Davies MJ, Renouf DV. 0-linked protein glycosylation structure and
function. GlycoconjugateJ 1996; 13: 19-26.
9. Feizi T. Demonstration by monoclonal antibodies that carbohydrate structures of
glycoproteins and glycolipids are onco-developmental antigens. Nature (London)
1985: 3 1 4 53-7.
10. Hakomori S. Glycosphingolipids. Sci Am, 1986; 2 5 4 32-41.
I I. Ccdington IF, Haavik S. Epiglycanin, a carcinoma-specific mucin-type glycoprotein
of the mouse TA3 tumour. Glycobiol 1992; 2 173-80.
12. Campbell JB, Finnie IA, Hounsell EF, Rhcdes ]A Direct demonstration of
increased expression of Thomsen-Friedenreich (TF) antigen in colonic
adenocarcinoma and ulcerative colitis mucin and its concealment in normal
mucin. J. Clin Invest 1995; 95: 571-6.
13. Hanski C, Hanisch FG, Recken RO. Alteration of mucin-boundcarbohydrate
moieties in malignant transformation of colonic mucosa. Cancer J 1992; 5
332-42.
14. Dahiya R, Kwak K-S, Byrd JC, Ho S,Yoon W-H, Kim YS. Mucin synthesis and
secretion in various human epithelial cancer cell lines that express the MUC-I
mucin gene. Cancer Res 1993; 53: 1437-43.
15. Statzynska T, Marsh PJ. Schofield PF, Roberts SA. Myers KA. Stern-PL prognostic
significance of 5T4 oncofetal antigen expression in colorectal carcinoma. Br
J Cancer 1994; 6 9 899-902.
16. Davies MJ, Hounsell EF. Comparison of separation modes for high-performance
liquid chromatography of glycoprotein- and proteoglycan-derived
oligosaccharides. J Chromatogr 1996; 720 227-34.
17. Yang J-M, Byrd JC, Siddiki BB, Chung Y-S, Okuno M, Sowa M, Kim YS. Matta KL,
Brockhausen I. Alteration of 0-glycan biosynthesis in human colon cancer tissues.
Glycobiology 1994; 4 873-84.
Glycoprotein changes in tumoun: a renaissance in clinical applications
18. Kitagawa H, Nakada H, NumataY. Kurosaka A, Fukui S. Funakoshi I, Kawasaki
T, Yamashina I.A monoclonal antibody that recognises sialyl-Leaoligosaccharide.
but is distinct from NS 19-9 as to epitope recognition. J Biochem (Tokyo) 1988
104 817-21.
19. Solter D, Knowles BB. Monoclonal antibody defining a stage-specific mouse
embryonic antigen (SSEA-I). Proc Natl Acad Sci USA 1978 7 5 5565-9.
20. ltzkowitz S. Carbohydrate changes in colon carcinoma. APMIS Suppl 1992; 27:
173-80.
21. Cleary KR Ota DM, Hoff SD, JrimuraT. Monoclonalantibody defining a stage
specific mouse embryonic antigen (SSAA-I). Matsushita Lab Invest 1990 63:
780-9 I.
22. Numata Y, Nakada H, Fukui S, Kitagawa H, Ozaki K, lnoue M, Kawasak T,
Funakoshi I, Yamashina I. A monoclonal antibody directed to T” antigen. Biochem
Biophys Res Commun 1990; 170 981-5.
23. Ryder SD, Smith ]A, Rhodes JM. Peanut lectin: a mitogen for normal human
colonic epithelium and human HT29 colorectal cancer cells. J Natl Cancer lnst
1992; 8 4 1410-6.
24. Siddiki B, Ho JJL, HuangJ, Bryd JC, Lau E, Yuan M, Kim YS. Monoclonalantibody
directed against colon cancer mucin has high specificity for malignancy. Int
J Cancer 1993; 5 4 467-74.
25. Schumacher U. Higgs D, Loizidou M, PickeringR, Leathem A, Taylor 1. Helix
pomatia agglutinin binding is a useful prognostic indicator in colorectal carcinoma.
Cancer 1994; 7 4 3 104-7.
26. Schwartz B, Bresalier RS, Kim YS. The role of mucin in colon cancer metastasis.
Int J Cancer 1992; 5 0 60-5.
27. Schmitt FC, Figueirdo P. Lacerda M. Simple mucin-type carbohydrate antigens
(T, sialosyl-T. T” and sialosyl-T”) in breast carcinogenesis. Virchows Arch 1995;
427: 25 1-8.
28. Brooks SA, Leatham, AJC. Expressionof the CD15 antigen (Lewis x) in breast
cancer. Histochemical J 1995; 27: 689-93.
29. Walker RA. Assessment of milk fat globule membrane antibodies and lectins as
markers of short-term prognosis in breast cancer. Br J Cancer 1990; 6 2 462-6.
30. Walker RA. Helix pomotio and prognosis of breast cancer. Br J Cancer 1993; 6 8
453-4.
3 I. Brooks SA, Leathem AJC. Prediction of lymph node involvement in breast cancer
by detection of altered glycosylation in primary tumour. The Lancet 1991; 338
7 1-4.
32. Gustenon BA et al. Prognostic value of Helix pomotio in breast cancer. Br
J Cancer, 1993; 6 8 146-50.
33. Brooks SA, Leathem AJC. Expression of alpha-GalNAc glycoproteins by breast
cancers. BrJ Cancer 1995; 71: 1033-8.
34. JohnstonWW, Szpak CA, Lottich SC, Thor A, Schlom J.Use of a monoclonal
antibody (872.3) as a novel immunohistochemical adjunct for the diagnosis of
carcinomas in fine needle aspiration biopsy specimens. Human Pathol 1986; 17:
501-13.
35. Cohen AD,Gopas J, Karplus G, Cohen Y. CA15-3 mucin-like carcinomaassociated antigen and tissue polypeptide-specific antigen: correlation to disease
state and prognosis in breast cancer patients. Isr J Med Sci 1995; 3 I: 155-9.
36. Pancino G, Osinaga E, Charpin C, Mistro D. Barque JPh, Roseto A. Purification
and characterisation of a breast-cancer-associatedglycoprotein not expressed in
normal breast and identified by monoclonal antibody 83D4. Br J Cancer I99 I;
63: 390-8.
37. Pal S,Sanyal U, Chattopadhyay U. Purification and characterizationof a new
85-kDa glycoprotein antigen from human breast tumor. Int J Cancer 1995; 5 0
759-65.
38. Chan DW, Beveridge RA, Bhqava, A+ Wilcox PM, Kennedy J, Schwartz MK.
Breast cancer marker CA549. Am J Clin Pathol 1994; 101: 465-70.
39. Gooi HC, Jones NJ,Hounsell EF, Scudder P, HilkensJ. Hilgers J, Feizi T.Novel
antigenic specificity involving the blood group antigen, Lea, in combination with
onco-developmental antigen, SSEA- I, recognised by two monoclonal antibodies
to human milk-fat globule membranes. Biochem Biophys Res Commun 1985;
131: 543-50.
40. Devine PL, Clark BA, Birrell, GW,Layton GT, Ward BG, Alewood PF, McKenzie
IFC. The breast tumour-associated epitope defined by monoclonal antibody
3EI .2 is an 0-linked mucin carbohydrate containing N-glycolylneuraminicacid.
Cancer Res 1991; 51: 5826-36.
4 I. Hanisch F-G, Stadie TRE, Peter-Katalinic J. MUC I glycoforms in breast cancer
cell line T47D as a model for carcinoma-associatedalterations of 0-glycosylation.
Eur J Biochem 1996; 236 3 18-27.
42. Wright GL, Beckett ML, Lipford GB, Haley CL, Schellhammer PF. A novel
prostate carcinoma-associatedglycoprotein complex (PAC) recognized by
monoclonal antibody TURP-27. Int J Cancer 1991; 47: 717-25.
293
43. Lipford GB, Wright GL Jr.Comparative study of monoclonal antibodies TURP-27
and HNK-I: their relationship to neural cell adhesion molecules and prostate
tumor-associated antigens. Cancer Res 1991; 51: 2296-301.
44. Pastan I, Lovelace ET, Gallo MG. Rutherford AV, MagnaniJL. Willingham MC.
Characterizationof monoclonal antibodies BI and B3 that react with mucinous
adenocarcinomas. Cancer Res 1991; 51: 3781-7.
45. Myers RB, Meredith RF. Schlom J, LoBuglio AF,Bueschen AJ, Wheeler RH,
Stockard CR, Grizzle WE. Tumour associated glycoprotein-72 is highly expressed
in prostatic adenocarcinomas. J Urol 1994; I 5 2 243-6.
46. Kim YD, Robinson DY, Manderino GL, Tribby IIE. Tomita JT. Molecular
characterizationof the epitope in prostate and breast tumour-associated PR92
antigen. Cancer Res 1989; 49: 2379-82.
47. Sawabu N, Toya D, Takemori Y, Hattori N, Fukui M. Measurement of a
pancreatic cancer-associated antigen (DUPAN-2) detected by a monoclonal
antibody in sera of patients with digestive cancers. Int J Cancer 1986 37: 693-6.
48. Wu IT, Chang J. ChromatographiccharacterizationofCAI 9-9 molecules from
cystic fibrosis and pancreatic carcinoma. J Clin Lab Anal 1992; 6 209- 15.
49. HoJJL, Bi N, Yan P, Yuan M, Norton KA, Kim YS. Characterizationof new
pancreatic cancer-relative monoclonal antibodies directed against purified mucin.
Cancer Res I99 I; 5 I: 372-80.
50. Hanisch F-G, Auerbach B, Bosslet K, Kolbe K, Karsten U, Nakahara N, Ogawa T,
Uhlenbruck G. Monoclonal antibody BW494 defines a blood group Lewis ‘/type I
chain-related antigen on carcinoma-associatedmucins. Biol Chem 1993; 374
1083-9 I.
51. Martersteck CM, Kedenha NL, Drapp DA, Tsui TG, Colley KJ. Unique alpha28-sialylated glycoproteins in breast cancer and leukemia cells. Glycobiology 1996;
6 289-301
52. Schellhammer PF, Wright GL. Biomolecular and clinical characteristics of PSA and
other candidate prostate tumour markers. Urol Clin North Am 1993; 20:
597-606.
53. Armbuster DA. Prostate specific antigen: biochemistry, analytical methods and
clinical applications. Clin Chem 1993; 3 9 181-95
54. Hanski C, Hanski M-L, Zimmer T, Ogorek D, Devine P. Riecken E-0.
Characterization of the major sialyl-Lex-positivemucins present in colon, colon
carcinoma, and sera of patients with c o l o r e d cancer. Cancer Res 1995; 5 5
928-33.
55. Mannori G, Crottet P, Cecconi 0, Hanasaki K, Aruffo Nelson RM, Varki A,
Bevilacqua MP. Differentialcolon cancer cell adhesion to E-, P-, and I-selectin:
role of mucin-type glycoproteins. Cancer Res 1995; 5 5 4425-3 I.
56. Fukuda, M. Possible roles of tumour-associated carbohydrate antigens. Cancer
Res 1996; 5 6 2237-44.
57. Levesque M, DCosta Yu H, Tadross Diamondis EP. lmmunoreactiveprostatespecific antigen in lung tumors. J Clin Lab Anal 1995; 9: 375-9
58. Devine PL, McGuckin MA, Ramm LE, Ward BG, Pee D, Long S. Serum mucin
antigens CASA and MSA in tumoun of the breast, ovary, lung, pancreas, bladder,
colon and prostate. Cancer 1993; 7 2 2007- 15.
59. Yu C-J, Yang P-C, Shew J-Y, Hong T-M, Yang S-C, Lee Y-C, Lee, L-N, Luh K-T,
Wu C-W. Mucin mRNA expression in lung adenocarcinomacell lines and tissues.
Oncology 1996; 53: I 18-26.
dycoproteins
in neoplasia. Glycoconjugate
J.
60. Kim YS, Gum I, Brockhausen I. Mucin -.
.
1996; 13:693-707.
61. Yadema KA. Kenemans P. Wobbes T. van K a m ~GI, de Bruiin HW, Thomas CM,
Massuger LF, Schijf CP, Eon GG, Vermorken JB, Viorhorst F, Hilgers10.
Carcinoma-associatedmucin serum markers CAM26 and CAM29 efficacy in
detecting and monitoring patients with cancer of the breast, colon, ovary,
endometrium and cervix. Int J Cancer 1991; 47: 170-9.
62. Hounsell EF. ‘ H NMR in the structural and conformational analysis of
oligosaccharidesand glycoconjugates. Progr NMR Spectrosc. 1995; 27: 445-74
63. Ogata S, Maimonis PJ. lQOWiQ SH. Mucins bearing the cancer-associated sialosylT” antigen mediate inhibition of natural killer cell cytoxicity. Cancer Res 1996;
5 2 474 1-6.
64. OBoyle KP, Zmore R, Adluri S, Cohen A, Kemeny N. Welt S. Lloyd KO,
Oettgen HF. Old LJ, Livingston PO. Immunizationof colorectal cancer patient
with modified ovine submaxillary gland mucin and adjuvants induces IgM and IgG
antibodies to sialylated T”. Cancer Res 1992; 52 5663-7.
65. Rughetti A, Turchi V, Appollonj C, Scambia G, Panici PB, Roncucci G, Mancuso S,
Frati L, Nuti M. Human B-cell immune response to the polymorphic epithelial
mucin. Cancer Res 1993; 5 3 2457-9.
66. Vitetta ES, Uhr JW. Monoclonalantibodies as agonists: an expanded role for their
use in cancer therapy. Cancer Res 1994; 5 4 5301-9.
67. Peterson JA,Couto JR, Taylor MR, Ceriani RL. Selection of tumour-specfic
epitopes on target antigens for radioimmunotherapyof breast cancer. Cancer
Res 1995; 5 5 5847-5 I.