close

Вход

Забыли?

вход по аккаунту

?

686

код для вставкиСкачать
2353
The Levels of Expression of Galectin-1, Galectin-3,
and the Thomsen–Friedenreich Antigen and Their
Binding Sites Decrease as Clinical Aggressiveness
Increases in Head and Neck Cancers
Georges Choufani, M.D.1
Nathalie Nagy, M.D.2
Sven Saussez, M.D.1
Hadelin Marchant, M.D.3
Pierre Bisschop, M.D.3
Maria Burchert, Ph.D.4
André Danguy, Ph.D.5
Stéphane Louryan, M.D.6
Isabelle Salmon, M.D.2
Hans-Joachim Gabius, Ph.D.4
Robert Kiss, Ph.D.5
Sergio Hassid, M.D.1
1
Department of Otolaryngology and Head and
Neck Surgery, Cliniques Universitaires de Bruxelles, Hôpital Erasme; Brussels, Belgium.
2
Department of Pathology, Cliniques Universitaires
de Bruxelles, Hôpital Erasme, Brussels, Belgium.
3
Department of Otolaryngology and Head and Neck
Surgery, Hôpital Brugmann, Brussels, Belgium.
4
Institute of Physiological Chemistry, Faculty of
Veterinary Medicine, Ludwig-Maximilians University; Munich, Germany.
5
Laboratory of Histopathology, Université Libre de
Bruxelles, Brussels, Belgium.
6
Laboratory of Anatomy and Embryology, Faculty
of Medicine, Université Libre de Bruxelles, Brussels, Belgium.
BACKGROUND. The aim of this study was to investigate whether an increase in
malignancy level is accompanied by significant modifications of the expression of
galectin-1, galectin-3, and Thomsen–Friedenreich antigen (T antigen) as well as the
expression of binding sites for these three markers in head and neck squamous cell
carcinomas (HNSCCs).
METHODS. Immunohistochemical and glycohistochemical staining reactions were
carried out with antibodies, labeled lectins, and a custom-made neoglycoprotein
on the basis of histologic slides from a retrospective series of 40 normal and 75
HNSCC formalin fixed, paraffin embedded tissues, and were quantitatively described with the aid of computer-assisted microscopy.
RESULTS. Whatever the histologic type, the epithelial tissues in HNSCC exhibited
very significantly (P , 0.01 to P , 0.0001) lower amounts of galectin-1, galectin-3,
and T antigen and their respective binding sites than their corresponding normal
counterparts. The tumors of the larynx differed very significantly (P , 0.0001 to P ,
0.000001) from all the other tumor types. A loss of differentiation in the HNSCCs is
accompanied first by the loss of expression of galectin-3 and galectin-3-reactive
sites and then by that of the T antigen and its binding site(s). The opposite feature
was observed when the parameters associated with the TNM classification were
taken into account. The negative lymph node HNSCCs could be distinguished (P 5
0.02) from the positive lymph node HNSCCs on the basis of a loss of galectin-3
expression. The modifications occurring in the extent of expression of galectin-1
and galectin-1-reactive sites were relatively marginal in comparison with those
observed for galectin-3-dependent and T- antigen-dependent staining.
CONCLUSIONS. The decrease in the extent of expression of galectin-3 and galectin3-reactive sites, T antigen and T antigen-binding sites, and, to a lesser extent,
galectin-1 and galectin-1-reactive sites correlates significantly with an increasing
level of clinically detectable HNSCC aggressiveness. Cancer 1999;86:2353– 63.
© 1999 American Cancer Society.
Supported by grants from the Dr. M. Scheel-Stiftung
für Krebsforschung (Germany) and the Fonds de la
Recherche Scientifique Médicale (FRSM, Belgium).
KEYWORDS: galectins, lectins, head and neck cancers, prognosis, quantitative
measurements, computer-assisted microscopy.
The authors are especially indebted to Dr. J.L.
Wang for his kindness in providing the expression
vector for murine galectin-3.
C
R.K. is a Senior Research Associate with the Fonds
National de la Recherche Scientifique (FNRS, Belgium).
Address for reprints: Robert Kiss, Ph.D., Laboratory
of Histopathology, Faculty of Medicine, Free University of Brussels, 808 Route de Lennik, 1070
Brussels, Belgium.
Received June 24, 1999; accepted July 19, 1999.
© 1999 American Cancer Society
ancers of the upper aerodigestive tract are heterogeneous in their
neoplastic processes, each of which requires its own unique set of
epidemiologic, anatomic, pathologic, and therapeutic considerations,
as stated by Schantz et al.1 The predominant lesions in the case of
head and neck cancers are squamous cell carcinomas, which constitute a rather heterogeneous class and can be divided into welldifferentiated, moderately differentiated, and poorly differentiated
lesions (see Materials and Methods).
In accurately determining prognoses in head and neck squamous
cell carcinomas (HNSCCs), the monitoring of defined biochemical
characteristics can lead to new insights. Prognoses usually are esti-
2354
CANCER December 1, 1999 / Volume 86 / Number 11
mated by means of staging, but the role and actual
relevance of staging in the case of HNSCCs is undergoing continuous analysis.1 The common staging reported by the American Joint Committee on Cancer
Staging (AJCC)2 is based on the tumor-lymph nodemetastases (TNM) classification, with T representing
the extent of the primary disease, N representing the
extent of the regional lymph node metastases, and M
representing the measure of distant metastases.
Similar to other tumor entities, the quest to define
markers with a demonstrable impact on clinical
courses is a potentially promising and therapeutically
relevant research topic. Several well-defined classes of
biomolecules have been monitored due to the assumed importance of cell-cell and cell-extracellular
matrix adhesion processes, cytoskeletal rearrangements for tumor cell invasiveness, and nuclear proteins for proliferative activity. For example, modifications in the expression levels of keratins,3–5
cadherins,6,7 cortactin,8 syndecan,9 nuclear matrix
proteins,10 and the organization of the basement
membrane11 have been reported as being correlated
to various extents with prognoses in patients with
HNSCC. In view of the potential of oligosaccharides to
serve as information-bearing molecules, attention
should be paid not only to protein-protein interactions. Obviously, the recognitive interplay between
these coding units and receptors, such as endogenous
lectins, likewise can regulate intercell interactions and
various cell activities.12 The monitoring of one class of
epitopes commonly has been performed by exogenous lectins. In this line of research, the detection of a
diagnostic and/or prognostic value for HNSCCs includes the peanut agglutinin (PNA), which is purified
from Arachis hypogaea and binds to the glycoproteins
that expose the galactose-b 1,3-N-acetylgalactosamine
sequence referred to as the Thomsen–Friedenreich
antigen (T antigen).13–15 Having explicitly implied the
functional significance of such epitopes, it is a logical
step to complement the application of an agglutinin
with the assessment of binding sites for the localized
sugar structure in the tissue. Chemical synthesis of the
ligand part and its immobilization on a suitable carrier
gains access to the required tool, i.e., a neoglycoconjugate.16,17 Such a probe already has been proven useful for localizing endogenous binding sites for the T
antigen.18 Thus, the concomitant analysis with an agglutinin and an appropriate neoglycoprotein is expected to advance our knowledge of distinct systems
of protein-carbohydrate interactions.
Due to the remarkable progress in animal lectinology, it is becoming possible to determine the expression of distinct endogenous lectins in tissues. If the
same concept is applied, then the concomitant monitoring of lectin expression by an antibody and of
lectin-reactive sites by the labeled tissue lectin constitutes a promising method for inferring physiologic
activities. Due to their involvement in various activities, including splicing, apoptosis, and cell-matrix interactions, members of the galectins group may influence aspects of tumor cell behavior.20 This reasoning
has prompted the monitoring of galectins in malignant human tumor specimens, such as sections from
breast, thyroid, and brain carcinoma, and also the
initial demonstration of the presence of galectin in
HNSCC cell lines and 35 primary HNSCCs.21–24 However, no study so far has applied biochemically purified galectins simultaneously to determine the extent
of lectin-reactive sites and galectin specific antibodies.
The objective of the current work was to answer
the pertinent question of whether increasing clinical
aggressiveness in HNSCCs, as evidenced by TNM staging, is correlated with modifications in the expression
of galectin-1, galectin-3, and galectin-reactive sites
and of T antigen/PNA-binding sites. The series of
specimens comprised 75 HNSCCs. Immunohistochemical and glycohistochemical staining reactions
were quantified by means of computer-assisted microscopy.
MATERIALS AND METHODS
Histopathologic and Clinical Data
The current series included 40 normal cases and 75
cancer cases. The 40 normal cases included 28 specimens from the oral cavity, 7 from the larynx, and 5
from the hypopharynx. All of these cases were fresh
material obtained either from autopsy or surgery performed for nontumor pathologies. There were no differences in immunohistochemistry patterns between
samples obtained directly from surgery and those obtained from cadavers (data not shown). The 28 oral
cavity cases included 6 samples from the tongue, 8
from the tonsils, 7 from between the tongue and the
tonsils, and 7 from the buccal mucosa. The patients
were age and gender matched with respect to the
cancer sample series.
The 75 HNSCC tumor specimens included 41
from the oral cavity, 23 from the larynx, and 11 from
the hypopharynx. The 41 oral cavity tumor specimens
included 10 from the tongue, 10 from the tonsils, 9
from between the tongue and the tonsils, and 12 from
the buccal mucosa. All of the patients (62 men and 13
women, ages 43– 83 years) had undergone partial or
radical surgery. The diagnoses were established on the
basis of the histologic criteria described by Hyams et
al.25 All of the 75 HNSCC tumor specimens used for
analysis were from primary tumors and not from metastases or recurrences.
The 75 specimens from patients with HNSCC
were categorized as well differentiated (n 5 35 speci-
Galectin Expression in Head and Neck Cancers/Choufani et al.
mens), moderately differentiated (n 5 30 specimens),
and poorly differentiated (n 5 10 specimens) on the
basis of the definitions given below. The well-differentiated lesions were characterized by the presence of
intercell bridges and numerous keratinization foci,
whereas the poorly differentiated lesions were characterized by the rare presence (or a notable absence) of
keratinization foci and the absence of intercell
bridges. Individual cell keratinization and the presence of intercell bridges are consistent with the moderately differentiated phenotype.
Clinical staging was carried out by using the TNM
classification system.2 There were 5 patients T1 tumors, 23 patients with T2 tumors, 19 patients with T3
tumors, and 28 patients with T4 tumors. Of these 75
HNSCC patients, 52 were lymph node negative, and 23
were lymph node positive. All 75 patients were M0. Of
the 75 HNSCC patients under study, 31 had undergone surgical treatment alone, and 38 had undergone
surgical treatment followed by radiotherapy. We
checked to determine whether these two therapeutic
regimens led to significantly distinct prognoses. According to what was observed during the relatively
short period of clinical follow-up, no such difference
existed (data not shown). We collected complete clinical follow-up data for 34 of the 75 HNSCC patients. Of
these 34 patients, 14 died ,12 months after their
initial diagnosis (the D , 12M group), whereas the
other 20 patients were still alive at the time of the
analysis and had survival periods exceeding 36
months from the initial diagnosis of their tumors (the
SURV . 36M group). The remaining 41 patients were
not included in the survival status-related analysis,
because their clinical follow-up periods ranged only
between 1 month and 21 months.
Immunohistochemistry and Ligandohistochemistry
The expression of galectin-1 and galectin-3 in the 40
normal cases and the 75 HNSCC patients under study
was demonstrated by means of specific polyclonal
antigalectin-1 and antigalectin-3 antibodies, respectively, and the expression of galectin-1-reactive and
galectin-3-reactive sites was assessed by biotinylated
galectin-1 and galectin-3 using routine procedures.26
The expression of the T antigen was detected by
means of the application of biotinylated PNA, whereas
that of the T antigen-binding sites was revealed by
means of a biotinylated carrier-immobilized T antigen
(i.e., by the use of a synthetic T antigen derived from
a chemical preparation as a p-aminophenyl derivative18) as a ligand part of a biotinylated neoglycoprotein. Methodologically, tissue specimens were fixed in
4% formaldehyde and embedded in paraffin, and
5-mm-thick sections were subjected to processing with
the various histochemical probes and kit reagents un-
2355
der study. The markers were prepared and biotinylated under activity-preserving conditions, as detailed
elsewhere.19,26,27 The galectins were purified, as described previously in detail.28,29 Incubation with labeled probes was carried out at room temperature for
60 minutes at a concentration of 10 mg/mL. The extent
of specifically bound markers (the biotinylated probes
or antibodies) was demonstrated by avidin-biotinperoxidase complex (ABC) kit reagents (Vector Laboratories, Burlingame, CA) with diaminobenzidineH2O2 as a chromogenic substrate, as detailed
elsewhere.17,19,30 –32 The control reactions included
competitive inhibitions to ascertain sugar specificity,
and the omission of the incubation step with a labeled
marker served to exclude any staining by the binding
of kit reagents, such as the mannose-rich glycoproteins horseradish peroxidase and avidin. Counterstaining with Toluidine blue concluded the processing.
Computer-Assisted Microscopy
The variables from the quantitative histochemical
stainings were computed by means of a SAMBA 2005
computer-assisted microscope system (Alcatel-TITN,
Grenoble, France) with a 320 magnification lens (aperture, 0.50). The way in which this system was used to
quantify the histochemical staining is detailed elsewhere.30 –32 The computer-assisted microscope and
related quantitative analyses were standardized as detailed elsewhere.33 Briefly, a negative histologic control slide (on which the primary antibody of the biotinylated glycoprobe was omitted) was analyzed for
each probe assayed for each of the 40 normal samples
and the 75 HNSCC samples under study. The software
included in the computer-assisted microscope automatically subtracted the labeling index (LI) and mean
optical density (MOD) values (see below) of the negative control sample from each corresponding positive
sample. Some CCD camera-based systems, like that
used in this study, exhibit inherent shading, i.e., equal
levels of incident illumination that do not produce
equal signal values from all of the image areas to be
analyzed. Shading, therefore, is one factor that limits
the performance of a computer-assisted microscope.
Specific software in our system was set up to evaluate
the level of magnitude of the shading phenomenon.
This software was used to measure optical densities
on a uniform sample (i.e., a test slide with different
neutral density Kodak Wratten filters; Eastman-Kodak,
Rochester, NY) in five different parts of the field of
vision. The test results showed that the computerassisted microscope was not affected significantly by
shading features (data not shown). Glare phenomenon also can affect computer-assisted microscoperelated measurements, because it concerns all types of
2356
CANCER December 1, 1999 / Volume 86 / Number 11
TABLE 1
Description of the 18 Quantitative Histochemical Variables Under Study
Quantitative parametersa
Tissue target (lectin or lectin-binding sites)
Probe used
Designation
Abbreviation
Galectin-1
Antigalectin-1 antibody
Galectin-1-binding sites
Biotinylated galectin-1
Galectin-3
Antigalectin-3 antibody
Galectin-3-binding sites
Biotinylated galectin-3
T antigen
Biotinylated PNA
T-antigen binding sites (endogenous PNA-like lectin)
Biotinylated T antigen-bearing neoglycoprotein
Labeling index
Mean optical density
Concentrational heterogeneity
Labeling index
Mean optical density
Concentrational heterogeneity
Labeling index
Mean optical density
Concentrational heterogeneity
Labeling index
Mean optical density
Concentrational heterogeneity
Labeling index
Mean optical density
Concentrational heterogeneity
Labeling index
Mean optical density
Concentrational heterogeneity
aGal1-LI
aGal1-MOD
aGal1-CH
Gal1-LI
Gal1-MOD
Gal1-CH
aGal3-LI
aGal3-MOD
aGal3-CH
Gal3-LI
Gal3-MOD
Gal3-CH
PNA-LI
PNA-MOD
PNA-CH
T-LI
T-MOD
T-CH
PNA: peanut agglutinin.
a
The labeling index (LI) relates to the percentage of tissue area specifically stained by a glycohistochemical probe. The mean optical density (MOD) relates to the staining intensity. The concentrational heterogeneity
(CH) represents the extent of variability calculated for the MOD.
phenomena due to extra light gathered in the image.
The quantity of glare is system dependent and is a
function of the relative size of the object in the illuminated area. Glare phenomenon is greater for highly
absorbent particles and may introduce errors into
some quantitative measurements. In the system used
in this study for the shading evaluation, as explained
below, specific software also was set up to evaluate the
level of magnitude of the glare phenomenon. This
phenomenon did not modify our results significantly
(data not shown). In cases in which an object to be
analyzed fills the whole field (like in the current study),
the glare phenomenon remains of very low magnitude.
Fifteen fields of between 60,000 mm2 and
120,000 mm2 epithelial cells were scanned per histologic slide. The LI refers to the percentage of epithelial tissue area specifically stained by a glycohistochemical probe, the MOD denotes staining
intensity, and the concentrational heterogeneity
(CH) feature expresses the concentration spread of
individual epithelial areas and is evaluated by the
coefficient of variation (CV 5 standard deviation
value/mean value) of the MOD values. Thus, 18
quantitative histochemical variables were available
for each of the 40 normal samples and each of the 75
HNSCC samples under study. These are listed and
explained in Table 1.
Data Analysis
Discriminant analysis, a multivariate statistical procedure that is described elsewhere,34 was used to evaluate the contribution of the measured quantities (i.e.,
LI, MOD, and CH) for each of the six probes to the
characterization of the various histopathologic groups
considered in the series of 75 HNSCC samples. In the
current study, the stepwise discriminant analysis program Statistica (Statsoft, Tulsa, OK) was used.
RESULTS
Expression of Galectin-1, Galectin-3, and T Antigen and
their Respective Binding Sites in Normal Samples versus
HNSCC Samples
The percentages (determined by the LI variable) of
epithelial cells that exhibited significant amounts of
acceptor sites for the biotinylated T antigen were very
significantly lower (P , 0.01 to P , 0.0001) in the
HNSCC samples compared with the normal samples
regardless of the histologic site considered (Fig. 1A). In
contrast, the amounts (determined by the MOD variable) of T antigen-binding sites expressed by the glycohistochemically T antigen positive epithelial cells
were relatively similar (P , 0.05 to P . 0.05) in the
normal samples and the HNSCC samples (Fig. 1B).
The application of biotinylated PNA showed that
the percentages of cells exhibiting significant amounts
of T antigen also were significantly lower in the
Galectin Expression in Head and Neck Cancers/Choufani et al.
2357
Mean (bars) 6 standard
error of the mean values of the labeling
indices (A,C) and the mean optical densities (B,D) determined by using computerassisted microscopy for biotinylated T antigen (revealing glycohistochemically the
presence of T-antigen binding sites) and
biotinylated peanut agglutinin (PNA) (revealing glycohistochemically the presence
of T antigen) in normal cells and in epithelial cells from patients with head and
neck squamous cell carcinoma. Six histologic sites were analyzed: the buccal mucosa (B-M group), the tongue (TONG
group), the area between the tongue and
the tonsils (T-T group), the tonsils (TONS
group), the larynx (LAR group), and the
hypopharynx (HYPH group). Each histologic site included a minimum of 5 cases
and a maximum of 12 cases, either normal (open bars) or malignant (solid bars),
as detailed in Materials and Methods. The
number of asterisks indicates the level of
statistical significance: 1 asterisk, P ,
0.05; 2 asterisks, P , 0.01; three asterisks, P , 0.001; four asterisks, P ,
0.0001. a.u.: arbitrary unit.
FIGURE 1.
HNSCC samples compared with the normal samples
(Fig. 1C), but the differences were less pronounced
than what was observed for the biotinylated T antigenrelated glycohistochemical staining (Fig. 1A). In the
application of biotinylated T antigen (Fig. 1B), the
amounts of T antigen expressed by the glycohistochemically PNA positive epithelial cells were similar
(P . 0.05) in the normal samples and the HNSCC
samples (Fig. 1D).
The percentages of epithelial cells that were positive with respect to the galectin-1 binding sites (revealed by biotinylated galectin-1; Fig. 2A) and galectin-1 (revealed by antigalectin-1 antibody; Fig. 2C)
were significantly lower in the HNSCC samples compared with the normal samples. The amounts of galectin-1 binding sites in the biotinylated galectin-1
positive epithelial cells were similar (P . 0.05) in the
normal samples and the HNSCC samples except for
the buccal mucosa, in which the amounts of galectin-1
binding sites were significantly lower (P , 0.01) in the
epithelial cells from the HNSCC samples compared
with the normal samples (Fig. 2B). The same feature
was observed with respect to the amount of galectin-1
(Fig. 2D). In addition, the epithelial HNSCC cells from
the hypopharynx also exhibited significantly lower
(P , 0.05) amounts of galectin-1 compared with normal epithelial cells from the hypopharynx (Fig. 2D).
The percentages of epithelial cells that were positive with respect to the galectin-3 binding sites (revealed by biotinylated galectin-3; Fig. 3A) and galectin-3 (revealed by antigalectin-3 antibody; Fig. 3C)
were significantly lower in the HNSCC samples compared with the normal samples except for the larynx,
in which the percentages of immunohistochemically
positive galectin-3 epithelial cells were similar (P .
0.05) in the normal samples and the HNSCC samples
(Fig. 3C). The amounts of galectin-3 binding sites (Fig.
3B) and galectin-3 (Fig. 3D) were similar (P . 0.05) in
both HNSCC and normal epithelial cells.
Expression of Galectin-1, Galectin-3, and T Antigen and
their Respective Binding Sites in Relation to HNSCC
Histologic Types
Figure 4A–D illustrates the development of 4 of the
18 quantitative histochemical variables that were
available for each of the 75 HNSCC samples under
study. It is very difficult to formulate precise conclusions from such huge amounts of data. Therefore, a discriminant analysis was performed to select
the more informative variables. Together with the P
value of statistical significance, Table 2 reports the 3
most discriminatory quantitative histochemical
variables (ranked in decreasing order of discriminatory power) selected by discriminant analysis for
2358
CANCER December 1, 1999 / Volume 86 / Number 11
Mean (bars) 6 standard
error of the mean values of the labeling
indices (A,C) and the mean optical densities (B,D) determined by using computerassisted microscopy for biotinylated galectin-1 (revealing glycohistochemically the
presence of galectin-1 binding-sites) and
the antigalectin-1 antibody (revealing immunohistochemically the presence of galectin-1) in normal cells and in epithelial
cells from patients with head and neck
squamous cell carcinoma. For abbreviations and conventions, see the legend to
Figure 1.
FIGURE 2.
Mean (bars) 6 standard
error of the mean values of the labeling
indices (A,C) and the mean optical densities (B,D) determined by using computerassisted microscopy for biotinylated galectin-3 (revealing glycohistochemically the
presence of galectin-3 binding sites) and
for the antigalectin-3 antibody (revealing
immunohistochemically the presence of
galectin-3) in normal cells and in epithelial
cells from patients with head and neck
squamous cell carcinoma. For abbreviations and conventions, see the legend to
Figure 1.
FIGURE 3.
each of the paired comparisons. Each histologic
group harbored specific (P 5 0.023 to P , 0.000001)
glycohistochemical features except for the tumors
from the tongue-tonsil group when those tumors
were compared with those from either the tongue
group or the tonsil group (Table 2).
Expression of Glycohistochemical Markers in Relation to
the Embryologic Origins of HNSCC
Figure 4 shows that the origins of the tumors in this
series varied widely. Tumors from different sites have
different clinical courses that may be due to inherent
characteristics of the origin of the tissue. Some of the
tumors under study had an ectodermal origin,
whereas others had an endodermal origin. Thus, the
data were analyzed in terms of the embryologic origins of the HNSCC tumors. The endodermal group
includes tumors from the larynx, the base of the tonsils, the junction between the tongue and the tonsils,
the pharynx, and the hypopharynx. The ectodermal
group includes tumors from the mouth and those
corresponding to the anterior two-thirds of the
tongue. Despite the mixed embryologic origins of the
tonsil tumors, these tumors are considered to be
Galectin Expression in Head and Neck Cancers/Choufani et al.
2359
FIGURE 4. Illustration of the development of the mean (open squares) and the related standard error (black rectangles) and deviation (bars) values in patients
with head and neck carcinoma of the buccal mucosa (BUC-MUC group; n 5 12), the tongue (n 5 10), the tonsils (n 5 10), the transitional area between the tongue
and the tonsils (TONG-TONS group, n 5 9), the larynx (n 5 23), and the hypopharynx (HYPOPHAR group, n 5 11). (A) The labeling index quantified with respect
to the T antigen. (B) The mean optical density for staining by the antigalectin-1 antibody. (C,D) The labeling index for staining by biotinylated galectin-3 and the
antigalectin-3 antibody, respectively.
endodermal because of their epithelial origins (squamous cell carcinoma).
Figure 5A, which should be compared with Figure
4A, shows that the HNSCC samples of ectodermal
origin (n 5 12 samples) exhibited a significantly higher
(P 5 0.008) mean LI value in relation to the T antigenbearing biotinylated glycoprotein compared with the
HNSCC samples of endodermal origin (n 5 63 samples). In contrast, Figure 5B, which should be compared with Figure 4D, shows that the reverse feature
was observed with respect to the anti-galectin-3 antibody (P 5 0.001).
Expression of Galectin-1, Galectin-3, and T Antigen and
their Respective Binding Sites in Relation to HNSCC
Differentiation Levels
Table 3 shows that the well-differentiated tumors (independent of histologic type; data not shown) present
significantly different glycohistochemical characteristics compared with the moderately differentiated (P 5
0.009) or the poorly differentiated (P 5 0.0004) tumors.
The moderately differentiated tumors differed weakly,
although significantly (P 5 0.02), from the poorly differentiated tumors (Table 3).
Expression of Galectin-1, Galectin-3, and T Antigen and
their Respective Binding Sites in Relation to the Clinical
Stage of HNSCC
Table 4 shows that tumors belonging to each of the
four clinical T stages exhibited significantly different
(P 5 0.02 to P 5 0.002) glycohistochemical characteristics except for stage T3, which did not differ significantly (P . 0.05) from stage T4. Lymph node negative
(N2) HNSCC tumors were weakly, although significantly, different (P 5 0.02) from lymph node positive
(N1) tumors. The three most discriminatory variables
in decreasing order of power were antigalectin-3
MOD, galactin-3 LI, and PNA MOD. All 75 HNSCC
samples under study included an M0 classification in
terms of TNM staging.
2360
CANCER December 1, 1999 / Volume 86 / Number 11
TABLE 2
Results Obtained by Means of Discriminant Analysis with the
Emphasis on Identifying the Most Informative Quantitative Variables
Describing the Expression of Galectin-1, Galectin-3, T Antigen, and
their Respective Binding Sites in Relation to Head and Neck
Squamous Cell Carcinoma Histologic Typea
Buc-Muc
Tongue
PNA-LI (1)
aGal1-LI (2)
T-LI (2)
P 5 0.006
Tong-Tons
PNA-MOD (2)
Gal3-MOD (1)
aGal3-MOD (2)
P 5 0.035
Tonsil
PNA-CH (2)
PNA-LI (1)
aGal1-MOD (2)
P 5 0.005
Larynx
aGal3-LI (2)
T-LI (1)
aGal3-MOD (2)
P , 0.000001
Hypophar
aGal3-LI (2)
T-LI (1)
aGal1-LI (2)
P 5 0.0006
Tongue
Tong-Tons
Tonsil
Larynx
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
ns
—
—
—
—
—
—
—
—
—
—
—
—
—
—
PNA-CH (2)
T-MOD (2)
T-CH (2)
P 5 0.023
—
ns
—
—
—
—
—
—
—
—
—
—
T-LI (1)
aGal1-LI (2)
aGal3-MOD (2)
P , 0.000001
aGal3-LI (2)
T-LI (1)
aGal1-LI (2)
P , 0.000001
aGal3-LI (2)
T-LI (1)
Gal3-CH (1)
P , 0.00001
—
—
—
—
aGal3-LI (2)
aGal1-LI (2)
T-LI (1)
P 5 0.00003
aGal3-LI (2)
aGal1-LI (2)
Gal1-MOD (1)
P 5 0.002
aGal3-LI (2)
Gal1-LI (2)
Gal1-MOD (1)
P 5 0.03
T-LI (2)
T-CH (2)
PNA-CH (1)
P 5 0.004
a
The head and neck squamous cell carcinoma histologic types include 12 tumors of the buccal mucosa
(the Buc-Muc group), 10 of the tongue (the Tongue group), 10 of the tonsils (the Tonsil group), 9
between the tongue and the tonsils (the Tong-Tons group), 23 of the larynx (the Larynx group), and 11
of the hypopharynx (the Hypophar group). The quantitative variables are described in Table 1. The
three most discriminatory quantitative histochemical variables in decreasing order of discriminatory
power are listed for each paired comparison along with the P level of statistical significance (ns, not
significant; i.e., P . 0.05). The (2) and (1) signs to the right of each quantitative parameter indicate
whether the mean parameter value in a given column group was lower (2) or higher (1) than the mean
value of this parameter in the corresponding line group. For example, the (1) value to the right of the
PNA-LI parameter under the Buc-Muc group column and in the Tongue group line indicates that the
mean PNA-LI value was higher in the Buc-Muc than in the Tongue group.
Expression of Galectin-1, Galectin-3, and T Antigen and
their Respective Binding Sites in Relation to Patient
Survival
The 14 patients with HNSCC in the D , 12M group
displayed significantly different (P 5 0.009) characteristics compared with the 20 patients in the SURV
. 36M group. The three most discriminatory variables in decreasing order were PNA LI, galectin-3
MOD, and T LI.
DISCUSSION
The current study characterizes the patterns of expression of 6 glycohistochemical markers in 40 normal
samples and in 75 patients with HNSCC. These 6
FIGURE 5.
The development of the mean (open squares) and the related
standard error (black rectangles) and deviation (bars) values in patients with
head and neck carcinoma according to their embryologic origins, i.e., endodermal (ENDO) (n 5 62) versus ectodermal (ECTO) (n 5 13). For conventions, see
Figure 4.
markers include the binding sites for the T antigen, as
revealed by lectinhistochemistry with biotinylated
PNA; the endogenous binding sites for this disaccharide, as revealed by glycohistochemistry with a biotinylated T antigen-carrying neoglyoprotein; the binding
sites for galectin-1, as revealed by lectinhistochemistry
with biotinylated galectin-1; the binding sites for galectin-3, as revealed by lectinhistochemistry with biotinylated galectin-3; and galectin-1 and galectin-3, as
detected by immunohistochemistry with specific
polyclonal antigalectin-1 and antigalectin-3 antibodies, respectively. When monitoring tissue lectin expression, it is essential to determine the presence of
lectin and lectin-reactive sites in serial sections. The
percentage of tissue area specifically stained by the
labeled probe, the histochemical concentration of the
probe, and the heterogeneity-homogeneity distribution level of the probe’s histochemical staining were
Galectin Expression in Head and Neck Cancers/Choufani et al.
TABLE 3
Results Obtained by Means of Discriminant Analysis with the
Emphasis on Identifying the Most Informative Quantitative Variables
Describing the Expression of Galectin-1, Galectin-3, T Antigen, and
their Respective Binding Sites in Relation to Head and Neck
Squamous Cell Carcinoma Differentiation Levela
Well
Mod
Mod
Gal3-MOD (1)
aGal3-MOD (1)
Gal3-LI (1)
P 5 0.009
Poor
PNA-LI (1)
T-LI (2)
Gal1-CH (1)
P 5 0.0004
—
—
—
—
PNA-LI (1)
Gal3-MOD (1)
Gal1-CH (1)
P 5 0.02
The data presentation is identical to that in Table 2. The variables are listed in Table 1. The Well (n 5
35), Mod (n 5 30), and Poor (n 5 10) groups cover well, moderately, and poorly differentiated head and
neck squamous cell carcinoma, respectively.
a
TABLE 4
Results Obtained by Means of Discriminant Analysis with the
Emphasis on Identifying the Most Informative Quantitative Variables
Describing the Expression of Galectin-1, Galectin-3, T Antigen, and
their Respective Binding Sites in Relation to the T Variable of the
TNM Classification Systema
T1
T2
T-MOD (2)
PNA-LI (2)
Gal3-MOD (1)
P 5 0.03
T3
aGal3-MOD (1)
Gal3-MOD (1)
P 5 0.002
T4
aGal3-MOD (1)
Gal3-MOD (1)
P 5 0.002
a
T2
T3
—
—
—
—
—
—
—
—
T-MOD (2)
aGal1-MOD (2)
aGal3-CH (1)
P 5 0.04
—
—
—
—
T-MOD (2)
T-CH (2)
PNA-LI (1)
P 5 0.009
—
ns
—
—
For conventions and abbreviations, see Table 3.
quantified by using computer-assisted microscopy, as
detailed elsewhere for other marker types.30 –32 Thus,
18 quantitative histochemical variables described
each of the 40 normal samples and the 75 HNSCC
tumors under study. Data treatment was carried out
by means of discriminant analysis for the 75 HNSCC
tumors.
A comparison of the normal samples and the
HNSCC samples revealed that the percentage of cells
exhibiting 1 of the 6 biological markers under study
was systematically lower in HNSCC samples than in
2361
normal samples with respect to a given histologic tissue type. One exception appears in the larynx, in
which the percentage of cells exhibiting galectin-3 was
similar in HNSCC and normal samples. In contrast,
the amounts of each marker exhibited by cells that
remained glycohistochemically positive or immunohistochemically positive with respect to a given maker
were relatively similar in normal and HNSCC samples.
The data obtained for the expression of the T
antigen/T antigen-binding site are in agreement with
reports previously published in the literature by other
authors. Gussack14 reported that the T antigen belongs
to the family of oncofetal antigens in HNSCC. A particularity of this class of tumors is that the expression
of the oncofetal antigen decreases as HNSCC aggressiveness increases at both the biologic level and the
clinical level.13–15 A similar correlation was observed in
the current study. Having introduced the corresponding neoglycoprotein, we extended these observations
to the T antigen-binding site level.
With respect to the galectins, we confirm the reported presence of these endogenous lectins in head
and neck tumors.22 From this basis, we have been able
to extend the field of the knowledge of galectin expression and, especially, of the galectin ligand sites. Our
data indicate that there are quantitative parameters to
describe modifications in galectin-3 and galectin-3reactive sites that mainly distinguish well-differentiated HNSCC from moderately differentiated HNSCC,
whereas there are mainly quantitative variables to describe modifications in the T-PNA expression pattern
that distinguish well-differentiated from poorly differentiated HNSCC (see Table 3). In fact, the mean value
(6 standard error of the mean) for each of these quantitative variables decreased significantly (P , 0.05 to
P , 0.01) from well-differentiated HNSCC through
moderately to poorly differentiated HNSCC (data not
shown). This suggests that a concomitant decrease in
galectin-3 and galectin-3-binding site expression is an
event that occurs early in the loss of HNSCC differentiation, whereas a concomitant decrease in PNA-T
expression occurs later in the process of transformation of well-differentiated HNSCC into moderately to
poorly differentiated HNSCC.
With respect to the clinical T stage, an inverse
feature was observed compared with what appeared
to occur with respect to the levels of differentiation.
Indeed, the T1 HNSCC tumors differed mainly from
the T2 tumors on the basis of a decrease in T antigen
and T antigen-reactive site expression (see Table 4),
whereas the T1 HNSCC tumors differed mainly from
the T3 and T4 tumors on the basis of a decrease in
galectin-3 and galectin-3-reactive site expression (see
Table 4). The grading and staging of the tumors from
the different sites are not presented in Materials and
2362
CANCER December 1, 1999 / Volume 86 / Number 11
Methods, because the grades and stages covered all of
the sites. For example, not all 5 T1 tumors were from
a specific site: One tumor was from the oral cavity, 1
tumor was from the tonsils, 1 tumor was from the
pharynx, and 2 tumors were from the larynx. In addition, the 10 poorly differentiated tumors also were
from different sites. Thus, the data interpretation presented here from the statistical analyses was not
skewed by any bias.
The progressive loss of differentiation and the
progression from T1 clinical stage to T4 clinical stage
apparently was not reflected in the same way at the
level of the modification of the glycohistochemical
characteristics monitored. Table 4 also shows that
modifications to the levels of expression of galectin-1
and galectin-1-binding sites correlate only marginally
with the progression from T1 through T2, T3, and T4
stages. Galectin-1 and galectin-3 along with their ligands evidently behave disparately. In this connection, we observed that all 5 T1 HNSCC tumors under
study were either well differentiated or moderately
differentiated. In contrast, 5 of 23 T2 tumors under
study were poorly differentiated, whereas 14 of 28 T4
HNSCC tumors were well differentiated (data not
shown). This feature means that the differentiation
level and clinical T stage clearly reflected the distinct
biologic status in HNSCC.
The clinical progression of patients with HNSCC
from N2 to N1 was paralleled biologically by a decrease in the amount of galectin-3 expression (as evidenced by the antigalectin-3 MOD variable, which
quantifies the MOD of the antigalectin-3 antibodyrelated immunohistochemical staining), a decrease in
the percentage of the cells exhibiting galectin-3-reactive sites (as evidenced by the galectin-3 LI variable),
and a decrease in the amount of T antigen expression
(as evidenced by the PNA MOD variable; data not
shown). Once again, the expression of galectin-1 and
galectin-1-binding sites did not appear to play a major
role in the distinction that could be made between N2
and N1 HNSCC (data not shown).
It should be emphasized that fewer than half of
the cases that we studied had any long term clinical
follow-up that enabled prognostic conclusions to be
made. These cases covered all of the anatomic sites
and tumor grades and stages. We are now enlarging
our series significantly, i.e., up to 300 cases, so that
independent prognostic analyses can be performed
for each anatomic site.
With respect to the 34 patients for whom we were
able to collect sufficient clinical data regarding survival status (those who died in ,12 months [the D ,
12M group] and those who survived for .36 months
[the SURV . 36M group] after the initial diagnosis of
HNSCC), it appeared that the HNSCC tumors from the
D , 12M group displayed weak but nevertheless significant differences in terms of glycohistochemical
characteristics. The weak level of statistical significance is due to the rather small number of patients
included in the analysis (14 patients vs. 20 patients), a
factor that precludes making any definitive conclusions. However, the current results raise initial evidence for the suggestion of an independent prognostic
value for the quantitative determination of lectin-related characteristics compared with conventional clinical prognosticators like the TNM variables. There
were 10 T3 or T4 tumors for the 14 D , 12M patients,
as expected; however, there were 4 T2 tumors and 7
N1 tumors compared with 7 N2 tumors. This means
that these T2 and/or N2 tumors were not accounted
for by the prognostic value of the TNM classification
system, which, in the current study, underestimated
the unfavorable prognosis of the patients analyzed. In
the same way, there were 5 T3 tumors, 2 T4 tumors,
and 2 N1 tumors in the SURV . 36M group. Thus, the
TNM classification system overestimated the potentially unfavorable prognosis of the patients analyzed.
Having indicated by this study the usefulness of monitoring endogenous lectin expression, other lectin-related characteristics should be taken into consideration to contribute independent prognostic values to
the TNM classification system. One example may be
sarcolectin, which is a sialic acid-binding and lymphokine-binding protein35 that acts as a human interferon
a/b antagonist and a growth regulator,36 for which we
recently demonstrated a prognostic value in oral cavity and lung tumors.32,37
In conclusion, the characterization of the expression of galectin-3 and the T antigen along with their
respective binding sites and, to a lesser extent, of
galectin-1 and galectin-1-binding sites appears to contribute additional prognostic value to conventional
clinical staging of patients with HNSCC. Nevertheless,
the fact remains that the current study, which included 75 patients with HNSCC, is preliminary. Indeed, there are 6 different anatomic sites (which
would be complemented by further sites), there are 3
different tumor grades, and there are potentially 8
different tumor stages, representing a potential tumor
diversity of at least 144 tumor groups. We are now
collecting HNSCC samples from various hospitals to
set up a series of cases that include several hundred
specimens so that significant numbers of cases can be
reviewed from each tumor group.
REFERENCES
1.
Schantz SP, Harrison LB, Forastiere AA. Tumors of the nasal
cavity and paranasal sinuses, nasopharynx, oral cavity, and
oropharynx. In: DeVita VT Jr., Hellman S, Rosenberg SA,
editors. Cancer: principles and practice of oncology. vol 1,
5th ed. Philadelphia: JB Lippincott Co., 1997:741– 847.
Galectin Expression in Head and Neck Cancers/Choufani et al.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
American Joint Committee on Cancer. Manual for staging of
cancer. 3rd ed. Philadelphia: JB Lippincott Co., 1988.
Vaidya MM, Borges AM, Pradhan SA, Rajpal RM, Bhisey AN.
Altered keratin expression in buccal mucosal squamous cell
carcinoma. J Oral Pathol Med 1989;18:282– 6.
Klijanienko J, El-Naggar A, De Braud F, Micheau C, Janot F,
Luboinski B, et al. Keratins 6, 13 and 19. Differential expression in squamous cell carcinoma of the head and neck. Anal
Quant Cytol Histol 1993;15:335– 40.
Balm AJ, Hageman PC, Van Doornewaard MH, Groeneveld
EM, Ivanyi D. Cytokeratin 18 expression in squamous cell
carcinoma of the head and neck. Eur Arch Otorhinolaryngol
1996;253:227–33.
Mattijssen V, Peters HM, Schalkwijk L, Manni JJ, van’t HohGrrotenboer B, de Mulder PH, et al. E-cadherin expression
in head and neck squamous cell carcinoma is associated
with clinical outcome. Int J Cancer 1993;55:580 –5.
Schipper JH, Unger A, Jahnke K. E-cadherin as a functional
marker of the differentiation and invasiveness of squamous
cell carcinoma of the head and neck. Clin Otolaryngol 1994;
19:381– 4.
Patel AM, Incognito LS, Schechter GL, Wasilenko WJ, Somers KD. Amplification and expression of EMS-1 (cortactin) in
head and neck squamous cell carcinoma cell lines. Oncogene 1996;12:31–5.
Inki P, Joensuu H, Grenman R, Klemi P, Jalkanen M. Association between syndecan-1 expression and clinical outcome in squamous cell carcinoma of the head and neck. Br J
Cancer 1994;70:319 –23.
McGaffery JD, Gapany M, Faust RA, Davis AT, Adams GL,
Ahmed K. Nuclear matrix proteins as malignant markers in
squamous cell carcinoma of the head and neck. Arch Otolaryngol Head Neck Surg 1997;123:283– 8.
Sakr WA, Zarbo RJ, Jacobs JR, Crissman JD. Distribution of
basement membrane in squamous cell carcinoma of the
head and neck. Hum Pathol 1987;18:1043–50.
Gabius HJ. Concepts of tumor lectinology. Cancer Invest
1997;15:454 – 64.
Wolf GT, Carey TE, Schmaltz SP, McClatchey KD, Poore J,
Glaser L, et al. Altered antigen expression predicts outcome
in squamous cell carcinoma of the head and neck. J Natl
Cancer Inst 1990;82:1566 –72.
Gussack GS. Oncofetal antigen expression in head and neck
carcinoma. Laryngoscope 1992;102:168 –76.
Bryne M, Gravdahl C, Koppang HS, Kjaerheim A, Dabelsteen
E. Is the carbohydrate sialosyl-Tn a marker for altered, nonmalignant activity in squamous epithelium in the head and
neck region? J Pathol 1995;175:237– 42.
Gabius HJ. Tumor lectinology: at the intersection of carbohydrate chemistry, biochemistry, cell biology and oncology.
Angew Chem Int Edit 1988;27:1267–76.
Danguy A, Kayser K, Bovin NV, Gabius HJ. The relevance of
neoglycoconjugates for histology and pathology. Trends Glycosci Glycotechnol 1997;7:261–75.
Gabius HJ, Schröter C, Gabius S, Brinck U, Tietze LF. Binding of T-antigen-bearing neoglycoprotein and peanut agglutinin to cultured tumor cells and breast carcinomas. J Histochem Cytochem 1990;38:1625–31.
Gabius HJ, Gabius S, editors. Lectins and glycobiology. Heidelberg: Springer Verlag, 1993.
Gabius HJ. Animal lectins. Eur J Biochem 1997;243:543–76.
Gabius HJ, Brehler R, Schauer A, Cramer F. Localization of
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
2363
endogenous lectins in normal human breast, benign breast
lesions and mammary carcinomas. Virch Arch B Cell Pathol
Mol Pathol 1986;52:107–15.
Gillenwater A, Xu XC, El-Naggar AK, Clayman GL, Lotan R.
Expression of galectins in head and neck squamous cell
carcinoma. Head Neck 1996;18:422–32.
Xu XC, El-Naggar AK, Lotan R. Differential expression of
galectin-1 and galectin-3 in thyroid tumors. Potential diagnostic implications. Am J Pathol 1995;147:815–22.
Bresalier RS, Yan PS, Byrd JC, Lotan R, Raz A. Expression of
the endogenous galactose-binding protein galectin-3 correlates with the malignant potential of tumors in the central
nervous system. Cancer 1997;80:776 – 87.
Hyams VJ, Batsakis JG, Michaels L, editors. Tumors of the
upper respiratory tract and ear. Atlas of tumor pathology.
Washington, DC: Armed Forces Institute of Pathology, 1988.
Gabius HJ, Bardosi A. Neoglycoproteins as tools in glycohistochemistry. Progr Histochem Cytochem 1991;22:1– 66.
Gabius HJ, Gabius S, editors. Glycosciences: status and perspectives. London: Chapman & Hall, 1997.
Gabius HJ. Influence of type of linkage and spacer on the
interaction of b-galactoside-binding proteins with immobilized affinity ligands. Anal Biochem 1990;189:91– 4.
Siebert HC, Adar R, Arango R, Burchert M, Kaltner H, Kayser
G, et al. Involvement of laser photo CIDNP (chemicallyinduced dynamic nuclear polarization)-reactive amino-acid
side chains in ligand binding by galactoside-specific lectins
in solution. Eur J Biochem 1997;249:27–38.
Remmelink M, Salmon I, Goldschmidt D, Decaestecker C,
Nemec E, Berthe JV, et al. Quantitative measurements of
desmin and vimentin immunostains and cell density in
leiomyomas and leiomyosarcomas. Anal Cell Pathol 1996;
12:25– 44.
Goldschmidt D, Decaestecker C, Berthe JV, Gordower L,
Remmelink M, Danguy A, et al. The contribution of computer-assisted methods for histopathological classification
of adipose tumors. Lab Invest 1996;75:295–306.
Saussez S, Marchant H, Nagy N, Decaestecker C, Hassid S,
Jortay A, et al. Quantitative glycohistochemistry defines new
prognostic markers in cancers of the oral cavity. Cancer
1998;82:252– 60.
Camby I, Nagy N, Lopes MB, Schäfer BW, Maurage CA,
Ruchoux MM, et al. Supratentorial pilocytic astrocytomas,
astrocytomas, anaplastic astrocytomas and glioblastomas
are characterized by a differential expression of S100 proteins. Brain Pathol 1999;9:825– 43.
Johnson RA, Wicheren DW. Applied multivariate statistical
analysis. 3rd ed. New York: Prentice-Hall International,
1992.
Zeng FY, Gerke V, Gabius HJ. Characterization of the macrophage migration inhibitory factor-binding site of sarcolectin and its relationship to human serum albumin. Biochem
Biophys Res Commun 1994;200:89 –94.
Chany-Fournier F, Jiang PH, Chany C. Sarcolectin and interferon in the regulation of cell growth. J Cell Physiol 1990;
145:173– 80.
Kayser K, Bovin NV, Korchagina EY, Zeilinger C, Zeng FY,
Gabius HJ. Correlation of expression of binding sites for
synthetic blood group A-, B- and H-trisaccharides and for
sarcolectin with survival of patients with bronchial carcinoma. Eur J Cancer 1994;30A:653–7.
Документ
Категория
Без категории
Просмотров
2
Размер файла
184 Кб
Теги
686
1/--страниц
Пожаловаться на содержимое документа