  , .

181: 323–329 (1997)

TUMOUR GROWTH FRACTION AND APOPTOSIS IN

SALIVARY GLAND ACINIC CELL CARCINOMAS.
PROGNOSTIC IMPLICATIONS OF Ki-67 AND bcl-2
EXPRESSION AND OF IN SITU END LABELLING
(TUNEL)
 . 1*,  2,   3,  2,  3 
 3
1
Department of Pathology II, University Hospital, Linköping, Sweden
2
Department of Otolaryngology, Head and Neck Surgery, University Hospital, Linköping, Sweden
3
Department of Pathology and Laboratory Medicine, European Institute of Oncology, Milan University School of Medicine, Milan,
Italy

SUMMARY
bcl-2 protein and Ki-67 (MIB-1) were studied in 32 acinic cell carcinomas (ACCs), all with a minimum of 5 years’ clinical follow-up.
Tumour apoptosis was evaluated by TdT dUTP nick end labelling (TUNEL) and by morphological criteria. Five patients died of their
disease. Patients with stage I tumours had significantly better survival compared with other stages (P<0·05). Patients with
MIB-1-negative tumours had significantly better survival than patients with MIB-1-positive tumours (P=0·05). This study confirms a
previous report that MIB-1 is an independent prognostic factor for survival in patients with ACC. Stage I tumours had high expression
of bcl-2 protein, but there was no difference when compared with other stages. TUNEL positivity was most prevalent in stage I tumours,
compared with stages II, III, and IV (P<0·05), probably indicating more apoptosis. This could imply a capacity of stage I tumours
(‘early tumours’) for early selection of tumour cells for elimination by apoptosis. There was no significant difference between expression
of bcl-2 and TUNEL, between these parameters and clinical outcome, or between any parameter and morphological subclassification.
We conclude that MIB-1 has prognostic value in ACC. Clinical staging, bcl-2, and TUNEL are also potentially useful as prognostic
markers. ? 1997 by John Wiley & Sons, Ltd.
KEY WORDS —apoptosis; bcl-2; TUNEL; end labelling; salivary neoplasm; Ki-67

INTRODUCTION been performed seeking for prognostic factors;5–11 cell

The new WHO classification of salivary gland
tumours1 is orientated to the routine work of the surgi-
cal pathologist. Useful commentaries on the classifica-
tion have also been published.2–4 Eighteen separate
types of salivary gland carcinoma can be distinguished
by precise histopathological features, as well as by
differences in treatment and prognosis.1 Nevertheless,
there are at least three of these carcinomas that can be
further subdivided. Acinic cell carcinoma (ACC), which
is one of these, may show four major histopathological
growth patterns. Acinic cell carcinomas are generally
low-grade malignancies, but some cases behave more
aggressively. Unfortunately, the subtypes are not always
readily recognised on histology and the clinical
behaviour is not always related to morphology nor is it
always predictable.
There is a need to increase our knowledge of the
net growth of tumours. Numerous studies have
*Correspondence to: Dr Henrik B. Hellquist, Department of
Pathology II, University Hospital, S-581 85 Linköping, Sweden.
E-mail: henrik.hellquist@pat.liu.se
Contract grant sponsor: Linköping University Hospital, Sweden.
CCC 0022–3417/97/030323–07 $17.50
? 1997 by John Wiley & Sons, Ltd.
proliferation, in particular, has been extensively studied,
although it does not correlate well with the overall
tumour growth rate.12 Two recent studies of cell pro-
liferation in ACC and mucoepidermoid carcinoma using
the MIB-1 antibody against the cell cycle-associated
Ki-67 antigen have revealed a correlation with progno-
sis,13,14 but estimation of the net tumour growth might
give additional useful information. The two main
mechanisms of cell loss from growing tumours are
apoptosis and necrosis.15 Apoptosis is characterized by
the successive fragmentation of DNA; in most cell types,
this eventually results in 180 bp sized fragments and
should have a major influence on net tumour growth.16
The bcl-2 viability-enhancing proto-oncogene encodes
for a 26 kD protein that protects cells from apoptosis,
favouring prolonged survival of normal as well as
neoplastic cells.17–21
We report a study of apoptosis in ACC in which
we have investigated the immunocytochemical expres-
sion of bcl-2 protein and DNA fragmentation by
TUNEL (TdT dUTP nick end labelling). The results
are correlated with histological subclassification,
MIB-1 labelling index, clinical staging, and biological
behaviour.
Received 15 May 1996
Accepted 27 August 1996
324 H. B. HELLQUIST ET AL.

Table I—Clinical data of 32 patients with salivary acinic cell haematoxylin and eosin and other sections were placed
carcinoma on pretreated slides (poly--lysine) for heating later in a
microwave oven before immunostaining and TUNEL.
Localization, All cases were reviewed and histologically subclassified
sex Therapy according to the predominant growth pattern, i.e. micro-
cystic, solid, papillary-cystic, or follicular type.1 The
T1N0M0 P: 5M, 6F SU: 11 slides were all coded and the examiners had no access to
Cases 1–12 S: 1M SU and RT: 1 clinical data.
(All stage I) Sections for immunocytochemistry were treated with
3 per cent hydrogen peroxide (H2O2) in distilled water
T1N1M0 P: 2F SU: 2 to inhibit endogenous peroxidase activity and then
Cases 13 and 14 immersed in boiling citrate buffer, pH 6, in a microwave
(All stage III)
oven,22 two changes of 5 min each. After washing in tap
T2N0M0 P: 3M, 9F SU: 9 and distilled water, the sections were subsequently incu-
Cases 15–27 S: 1F SU and RT: 4 bated with (1) a 1:20 dilution of normal rabbit serum for
(All stage II) 20 min at room temperature; (2) 1:50 and 1:100 dilutions
of the monoclonal antibodies to bcl-2 protein and Ki-67
T2N1M0 P: 1F SU: 1 (MIB-1, Immunotech, France), respectively, overnight
Case 28 at +4)C; (3) a 1:200 dilution of a biotinylated rabbit
(Stage III) antiserum to mouse immunoglobulins for 30 min at
room temperature; and (4) a 1:100 dilution of the
T3N0M0 P: 1F SU and RT: 1 streptavidin-biotinylated peroxidase complex for 30
Case 29
(Stage III)
min, also at room temperature. Peroxidase activity was
developed in diaminobenzidine chromogen substrate.
T3N1M0 P: 1F SU: 1 All reagents were bought from DAKO (Copenhagen,
Case 30 Denmark).
(Stage III) A known positive control for bcl-2 [a formalin-fixed,
paraffin-embedded follicular lymphoma carrying the
T3N2M0 P: 1F SU: 1 t(14;18) chromosomal translocation] was immuno-
Case 31 stained with the test samples. Negative control sections,
(Stage IV) where the specific monoclonal antibody was replaced
with the immunoglobulin fraction of non-immune
T4N1M0 P: 1M SU and RT: 1
Case 32
mouse sera, remained unstained. The immunoreactivity
(Stage IV) for bcl-2 was evaluated with regard to the staining
intensity, and to the percentage of immunoreactive
M=male; F=female; P=parotid gland; S=submandibular gland;
tumour cells (extent). Neoplastic cells showing definite
SU=surgery; RT=radiotherapy. cytoplasmic staining of the same intensity as shown by
The figures in each column represent the number of patients. the lymphoma cells or higher were graded as 3+.
Tumour cells with a moderately intense staining reaction
were graded 2+ and weakly positive cells 1+. Negative
MATERIALS AND METHODS tumour cells were recorded as negative. Furthermore, in
every tumour there were scattered lymphocytes, aggre-
Materials gates of lymphocytes, and/or lymphoid tissue, which
The series comprised 32 cases of ACC, all with a all served as internal positive controls. In cases where
minimum follow-up of 5 years, but many with consider- the intensity of immunoreactivity varied within the
ably longer. Twelve patients were lost to follow-up, tumour, the dominant areas were selected. The extent of
although they were followed for a minimum of 5 years. immunoreactivity was evaluated as 1+ if less than 25 per
Sixteen of the 32 cases were kindly donated by the cent of tumour cells were positive, 2+ if 25–75 per cent
Salivary Gland Register, University of Hamburg, were positive, and 3+ if more than 75 per cent of cells
Germany. Twelve patients had tumours of stage I, 13 were positive. The MIB-1 immunoreactivity in tumour
stage II, five stage III, and two stage IV (Table I). There cells (strongly but also moderately stained cells) was
were ten males and 22 females (mean age 50·2 years). evaluated either as less than 10 per cent of cells positive
Thirty tumours were in the parotid and two in the (=negative) or as more than 10 per cent positive (=posi-
submandibular gland. Surgery was performed in all tive). The entire tumour (all sections available) was
cases and in seven patients, surgery was also combined screened in each case, often involving several thousands
with radiotherapy (Table I). of cells.
After deparaffinization, sections for the TUNEL reac-
tion were digested by 20 ìg/ml proteinase K for 15 min.
Methods
After four washes in distilled water and quenching in 2·0
From each case, 4 to 5 ìm thick sections were cut per cent H2O2, the Apoptag> kit was applied according
from representative blocks of the original surgical to the manufacturer’s instructions (Oncor, Gaithersburg,
specimen. One section was routinely stained with MD, U.S.A.). Briefly, the TUNEL method is a tailing
 APOPTOSIS IN ACINIC CELL CARCINOMA 325

reaction where terminal deoxynucleotidyl transferase Table II—Follow-up of 32 patients with salivary acinic cell
(TdT) catalyses a template-independent addition of carcinomas
deoxyribonucleotide triphosphate to the 3*-OH ends of
single- or double (blunt ends)-stranded DNA. The incor- Recurrence Follow-up
porated nucleotides form a random heteropolymer of Case No. Stage (months) (months)
digoxigenin-11-dUTP and dATP, in a ratio that has
been optimized for anti-digoxigenin antibody binding. 1 I No DUD (174)
The anti-digoxigenin antibody fragment carries a conju- 2 I No NED (127)
gated reporter enzyme (peroxidase) to the reaction site. 3 I No NED (138)
TUNEL positivity was evaluated as 1+ when less than 4 I No NED (150)
5 per cent of tumour cells showed distinctive nuclear 5 I No NED (64)
staining, with an intensity equalling or almost equalling 6 I No NED (60)
7 I * *
that of the control; 2+ if more than 5 per cent but less 8 I * *
than 25 per cent were positive; and 3+ when more than 9 I * *
25 per cent of tumour cells were positive. As with 10 I * *
MIB-1, the entire tumour was screened (thousands of 11 I * *
cells). Tumours with no positive cells, or the occasional 12 I No NED (60)
positive cell only, were recorded as negative. TdT was 13 III * *
replaced with water serving as a negative control, and as 14 III * *
a positive control, DNase I was added (20 min at 37)C) 15 II No NED (65)
after quenching in H2O2, thus producing DNA breaks in 16 II No NED (121)
virtually all cells.23 17 II Reg+Dist (7) DOD (12)
18 II Loc+Reg (14) DOD (120)
Due to shortage and loss of representative material, it 19 II No NED (101)
was possible to perform the TUNEL investigation and 20 II No NED (66)
MIB-1 immunostaining in only 27 and 16 cases, respect- 21 II Reg (84) NED (130)
ively, whilst bcl-2 staining was performed in 31 of the 32 22 II No NED (65)
cases. 23 II Dist (12) DOD (24)
24 II No NED (56)
25 II * *
Statistics 26 II * *
Differences between MIB-1 positivity and tumour 27 II * *
28 III * *
stage were analysed by Fisher’s exact test. Differences 29 III Reg (43) DOD (142)
between TUNEL positivity and tumour stage were 30 III * *
analysed by the Spearman rank correlation test. Simi- 31 IV Dist (48) AWD (72)
larly, Spearman rank correlation was used in analysing 32 IV Dist (7) DOD (7)
differences between expression of bcl-2 and staging, and
also for investigating any correlation between expression Loc=local recurrence; Reg=regional metastasis; Dist=distant
of bcl-2 and TUNEL. The Kruskal–Wallis rank test was metastasis; NED=no evidence of disease; AWD=alive with disease;
used to analyse any correlation between bcl-2 expression DUD=dead of unrelated disease; DOD=dead of disease.
and morphology and between TUNEL positivity and *Follow-up not available.
morphology. The Mann–Whitney U-test was used to
compare expression of bcl-2 and rate of survival, and
to compare TUNEL and rate of survival. recurrence 43 months after the initial treatment and then
died of her disease 99 months later, a total of 142
months after diagnosis (Table II). Both stage IV patients
RESULTS developed distant metastases, one of them after 7
months, who then died of his disease (case 32; Table II).
Clinical follow-up and histopathology The other patient (case 31), 21 years old at the time of
In five of 12 patients with stage I, complete clinical diagnosis, developed distant metastasis after 48 months,
follow-up was not available. All seven of the 12 stage I but is still alive 72 months after diagnosis (Table II).
patients with adequate follow-up showed no recurrence Statistical analysis revealed that patients with stage I
or other evidence of disease. One patient died of tumours had a significantly better survival rate com-
unrelated disease after more than 14 years (case 1) pared with the group of patients with stage I, II, and III
(Table II). tumours (P<0·05), but no difference was observed
In ten of the 13 patients with stage II disease, between stage I tumours and stage II, stage III, or stage
follow-up was available. Four patients developed recur- IV tumours if the latter were analysed separately and not
rences and/or metastatic disease, after 7, 14, 84, and 12 as a group.
months, respectively. Three of these four patients died of The tumours were reclassified histologically before the
their disease 12, 120, and 24 months after diagnosis start of the study and most cases showed a mixture of
(cases 17, 18, and 23, respectively) (Table II). Complete growth patterns.1 Fifteen cases showed a predominantly
clinical data were available in one of the five patients microcystic pattern (cases 4, 6, 7, 10, 11, 13, 19–22,
with stage III disease. She first presented with a local 24–27, and 31); nine cases were of the solid type (cases
326 H. B. HELLQUIST ET AL.

Table III—Histological subclassification, staging, bcl-2 immunoreactivity, MIB-1, TUNEL positivity, and
follow-up in 32 cases of acinic cell carcinoma

bcl-2
Case Sub- Follow-up
No. type Stage Extent Intensity MIB-1 TUNEL months

1 S I " " + " DUD (174)

2 S I +++ ++ " ++ NED (127)
3 S I +++ +++ " " NED (138)
4 M I ++ ++ " +++ NED (150)
5 S I ++ ++ " +++ NED (64)
6 M I +++ + ND +++ NED (60)
7 M I +++ ++ ND +++ *
8 M I +++ ++ ND ++ *
9 P I " " ND ND *
10 M I " " ND ++ *
11 M I +++ ++ ND +++ *
12 S I ++ + ND +++ NED (60)
13 M III ND ND ND + *
14 S III +++ +++ ND " *
15 S II ++ + " " NED (65)
16 P II +++ ++ + ++ NED (121)
17 F II +++ +++ " + DOD (12)
18 P II +++ ++ " +++ DOD (120)
19 M II +++ +++ " ++ NED (101)
20 M II +++ +++ " + NED (66)
21 M II +++ ++ " " NED (130)
22 M II +++ +++ " " NED (65)
23 S II +++ + + + DOD (24)
24 M II +++ + ND ND NED (60)
25 M II +++ ++ ND ND *
26 M II + + ND + *
27 M II " " ND +++ *
28 M III +++ + ND ND *
29 F III +++ +++ " " DOD (142)
30 P III +++ ++ ND +++ *
31 M IV + ++ ND ND AWD (72)
32 F IV +++ +++ + +++ DOD (7)

S=solid; M=microcystic; P=papillary-cystic; F=follicular; " =negative; ND=not done.

*Patient alive 5 years after diagnosis; no more information available.
DUD=dead of unrelated disease; DOD= dead of disease; NED=no evidence of disease; AWD=alive with disease.
+, + +, etc., see Materials and Methods section.

1–3, 5, 8, 12, 14, 15, and 23); five cases showed a
predominantly papillary-cystic pattern (cases 9, 16, 18,
28, and 30); and three were classified as follicular ACC
(cases 17, 29, and 32) (Table III and Figs 1A–1D).
Comparison was made between morphology and clinical
outcome in those cases where complete follow-up data
were available (in total 20 cases). None of the six cases
with microcystic ACC developed any recurrence
(cases 4, 6, 19, 20, 22, and 24), whereas one of the eight
cases with a solid growth pattern developed recurrence/
metastatic disease after 12 months; this patient died of
the disease after another 12 months (case 23). Of the two
cases with a papillary growth pattern, one developed
local recurrence and distant metastasis and later died of
the disease (case 18). The other patient with papillary-
cystic ACC and follow-up (case 16) has no evidence of
disease 121 months after diagnosis. The three patients
with the follicular type of ACC have all developed
metastatic disease and died of their disease after 120, 142
and 7 months, respectively (cases 17, 29, and 32) (Tables
II and III).
bcl-2
bcl-2 immunostaining was performed in all cases but
one (case 13), all with adequate positive controls (inter-
nal lymphocytes included). In 21 tumours, more than 75
per cent of the cells were positive, most of which (14
cases) had an intensity equalling the control (3+) (Fig.
1E). In four cases, bcl-2 positivity was found in 25–75
per cent of tumour cells (all but one stage I tumours)
(Table III). Two cases (cases 26 and 31; stage II and IV,
respectively) had less than 25 per cent bcl-2-positive
cells, and four were devoid of bcl-2 immunoreactivity
(cases 1, 9, 10, and 27; stage I and II, respectively)
(Tables II and III). Analysing expression of bcl-2 and
 APOPTOSIS IN ACINIC CELL CARCINOMA 327

Fig. 1 —(A) Photomicrograph of a microcystic acinic cell carcinoma (ACC) with numerous microcysts. (B) The solid variant
of ACC. Note the aggregate of lymphocytes in the right upper corner—lymphocytes are common in ACC. (C)
Photomicrograph showing a papillary-cystic variant of ACC. (D) The follicular type of ACC, with thyroid-like follicles filled
with fluid. (E) A solid ACC stained for bcl-2 protein. Note the positive lymphocytes in the right upper corner. (F) A solid
ACC stained for MIB-1 with several (>10 per cent) positive cells (case 23). (G) TUNEL performed in a case of microcystic
ACC. This particular field of the tumour shows 4–5 positive cells only, although screening of the entire tumour showed
positivity (>25 per cent positive cells; case 4). (H) A positive control of an ACC treated with DNase I. Virtually all cells are
TUNEL-positive

staging (all stages together and separately) failed to four different morphological subtypes of ACC, nor
identify any significant differences (Spearman rank, survival rate, and expression of bcl-2 (Kruskal–Wallis,
P=0·27). Nor were there any differences between the P=0·56 and Mann–Whitney, P=0·12, respectively).
328 H. B. HELLQUIST ET AL.
There was no correlation between expression of bcl-2
and TUNEL positivity (Spearman rank, P=0·08).
MIB-1
Four tumours were MIB-1-positive, with more than
10 per cent of the tumour cells positively stained (cases
1, 16, and 32) (Fig. 1F). One of the patients died of her
disease (case 32; stage IV; Tables II and III). Compari-
son between MIB-1 positivity and clinical outcome
(death) revealed a significantly better survival in those
patients with MIB-1-negative tumours (Fisher’s exact
test, P=0·05) (Tables II and III).
TUNEL
TUNEL was performed in 27 cases and positivity was
most prevalent in stage I tumours, which were positive
in all cases but two (Fig. 1G). Four of the 11 stage II
tumours investigated with TUNEL were positive, and
one of three stage III tumours. The only stage IV
tumour investigated was also positive (Table III). No
significance or particular pattern could be observed
concerning TUNEL positivity and morphological sub-
classification. Positivity was seen in seven of 13 investi-
gated microcystic acinic cell carcinomas: three of eight
of the solid type, three of three papillary-cystic, and
one of three follicular carcinomas (Table III). Statisti-
cal analysis of correlation between different staging
and TUNEL positivity failed to reach significance
(Spearman rank, P=0·28), whilst patients with stage I
tumours against all other stages had a better survival
(Spearman rank; P=0·05). There was no significant
difference between different morphological subtypes
and TUNEL positivity (Kruskal–Wallis, P=0·29), nor
between significantly better survival rate and positive
TUNEL (Mann–Whitney, P=0·068). The DNase-
treated specimens showed almost 100 per cent positivity
(Fig. 1H).
DISCUSSION
The principal aim of the present investigation was to
study parameters which can be related to net tumour
growth, namely the expression of bcl-2 protein, MIB-1
positivity, and TUNEL positivity. The results were also
analysed with regard to the clinical behaviour of ACC.
Five out of 32 patients (and one patient with distant
metastases but still alive: case 31) have died of their
disease (17 per cent). We believe that this percentage is
somewhat high for a tumour so generally accepted to
be of low-grade malignancy. Although the expression
of MIB-1 in ACC correlates with prognosis,13 clinical
staging and histopathological grading have hitherto
been the most reliable prognostic indicators for ACC
and for salivary gland tumours in general. We found the
MIB-1 counting method13,14 to be accurate and repro-
ducible, but it is cumbersome and does not easily lend
itself to routine diagnostic work. We estimated MIB-1
positivity by using the cut-off point of 10 per cent found
by Skalova et al.13,14 We thus only counted less or more
than 10 per cent, which is easier and quicker and also
enabled us to screen the entire tumour. In spite of this
less precise method, we were able to confirm the findings
of Skalova et al.13,14 MIB-1 thus clearly seems to be an
independent prognostic factor for survival rate in ACC.
The bcl-2 gene was originally discovered by Tsujimoto
et al. to be involved in the majority of non-Hodgkin’s
B-cell lymphomas [t(14;18) chromosomal trans-
locations].24–26 High levels of bcl-2 protein, aberrant
patterns of bcl-2 protein production, or both have also
been observed in a variety of solid tumours.19,27,28 The
bcl-2 gene (3·5 kb 5* exon and 5 kb 3* exon separated by
a >65 kb intron) encodes a 26 kD protein. The bcl-2
protein family consists of bcl-2 and bcl-XL (blockers)
and bcl-XS and bax (inducers), which tend to form
dimers and/or heterodimers. The bax gene protein prod-
uct (221 percent homology to the bcl-2 protein) can
form heterodimers with bcl-2 and thus abrogate its
ability to suppress apoptosis.29 Acinic cell carcinoma,
regarded as a tumour of low-grade malignancy, could
thus be expected to express the bcl-2 protein at a high
level. This assumption is also based on previously
reported observations that bcl-2 expression in certain
solid human tumours appears to be associated with less
aggressive behaviour.19,20 In the present study, 25
tumours had a high level of bcl-2 protein (+ + or + + +)
and only four of 31 tumours totally lacked immuno-
reactivity for the protein, thus supporting previous
observations. However, we also detected high levels of
bcl-2 in ACC in all five patients that we know have died
of their cancer, which indicates that as a prognostic
marker, the expression of bcl-2 has to be interpreted
with great caution. Furthermore, we were unable to
demonstrate statistical significance between the expres-
sion of bcl-2 and survival rate, nor any differences
between bcl-2 and morphological subtypes, nor any
relationship between bcl-2 and TUNEL positivity.
The tumour suppressor gene p53 plays an important
role as a regulator of bcl-2 and bax gene expression,
both in vivo and in vitro.30 There is mounting evidence
that the bcl-2 gene and bcl-2-binding proteins like
BAG-1 (bcl-2-associated anthanogene 1) are involved in
apoptosis, with anti-cell death activities.31 The four
bcl-2-negative cases (two microcystic, one solid, and
one papillary type) in the present study may represent
tumours with alternative apoptotic pathways, or
tumours devoid of that particular protein but having,
for example, BAG-1. From our present study we con-
clude that bcl-2 is of little use in ACC as a single marker,
but together with other parameters may add valuable
information and increase our understanding of the basic
mechanisms in different apoptotic pathways. Increased
understanding may also have practical implications,
since apoptosis that is induced by essentially all cyto-
toxic anti-cancer drugs and radiation can be blocked by
bcl-2.32 We were not able to perform immunostaining
for bax. We expected to find an inverse relationship
between bcl-2 and TUNEL but we did not. The rela-
tively small number of cases could be one explanation
for the lack of statistical significance; another, probably
more plausible explanation is that a variety of apoptotic
pathways exist and that bcl-2 is not involved in all.
 APOPTOSIS IN ACINIC CELL CARCINOMA
The TUNEL reaction has rapidly become a widely
accepted and sensitive indicator of apoptosis at the
single cell level. Its full specificity remains to be clarified,
as TUNEL is a tailing reaction and theoretically could
incorporate dUTP to any DNA nick break and not only
those occurring during apoptosis. It would seem logical,
in terms of low net tumour growth, if TUNEL is highly
expressed. Eight of 12 patients with no evidence of
disease expressed TUNEL (+ + and + + +) (and 13 out
of 26 patients that were free of disease after 5 years but
for whom no further information was available), whilst
three out of five that died lacked TUNEL (negative or
+) (Table III). Although these data failed to reach
statistical significance (P=0·068), they may indicate that
there is a tendency for a positive correlation between
TUNEL positivity and better survival. In our view,
TUNEL is a method well worth using in studies of other
tumours, since it is easy to perform and adequate
positive and negative controls are readily obtained.
Proliferation-associated antigens such as Ki-67 have
been extensively investigated in human tumours. The
simplicity of the technique used in the present study (i.e.,
less or more than 10 per cent positive nuclei, and all
tumour sections screened) revealed similar results to
those previously reported.13 We therefore conclude that
in ACC, Ki-67 positivity (MIB-1, Immunotech, France)
is an independent prognostic factor for survival. To the
best of our knowledge, this is one of the very few
tumours where a single marker has repeatedly been
shown to be reliable; it could be used prospectively in
individual cases.
We conclude by noting the usefulness of MIB-1; the
significantly better prognosis for stage I tumours com-
pared with stages I, II, and IV together; and the minor
importance of morphological subclassification of ACC
in relation to prognosis. We also emphasize the overall
high expression of bcl-2 protein and TUNEL positivity
in stage I tumours. We were, however, unable to detect
any inverse relationship between bcl-2 positivity and
TUNEL positivity.
ACKNOWLEDGEMENTS
Alessandra Di Bacco is the recipient of a fellowship
from A.I.R.C., Milan, Italy. The fellowship (HBH) from
the Swedish Medical Research Council (Stockholm)–
Consiglio Nazionale Della Ricerche (Rome), and sup-
port from the Linköping University Hospital Funds
(HBH) are gratefully acknowledged.
REFERENCES
1. Seifert G, Batsakis JG, Brocheriou C, et al. Histological typing of salivary
gland tumours. WHO International Histological Classification of Tumours.
2nd edn., Berlin: Springer-Verlag, 1991.
2. Seifert G. Histopathology of malignant salivary gland tumours. Eur J
Cancer, Oral Oncol 1992; 28B: 49–56.
3. Seifert G, Sobin LH. The World Health Organization’s histologic classifi-
cation of salivary gland tumours. A commentary on the second edition.
Cancer 1992; 70: 379–385.
4. Simpson RHW. Classification of tumours of the salivary glands. Histo-
pathology 1994; 24: 187–191.
5. Hickman RE, Cawson RA, Duffy SW. The prognosis of specific types of
salivary gland tumors. Cancer 1984; 54: 1620–1624.
329
6. Jensen OJ, Poulsen T, Schiødt. Mucoepidermoid tumors of salivary glands.
A long term follow-up study. APMIS 1988; 96: 421–427.
7. Hamper K, Schimmelpenning H, Caselitz J, et al. Mucoepidermoid tumors
of the salivary glands. Correlation of cytophotometrical data and prognosis.
Cancer 1989; 63: 708–717.
8. Hamper K, Caselitz J, Arps H, Askensten U, Auer G, Seifert G. The
relationship between nuclear DNA content in salivary gland tumors and
prognosis: comparison of mucoepidermoid tumors and acinic cell tumors.
Arch Otolaryngol 1989; 246: 328–332.
9. Greiner TC, Robinson RA, Maves MD. Adenoid cystic carcinoma. A
clinicopathologic study with flow cytometric analysis. Am J Clin Pathol
1989; 92: 711–720.
10. Clode AL, Fonseca I, Santos JR, Soares J. Mucoepidermoid carcinoma of
the salivary glands: a reappraisal of the influence of tumor differentiation on
prognosis. J Surg Oncol 1991; 46: 100–106.
11. Fonseca I, Clode AL, Soares J. Mucoepidermoid carcinoma of major and
minor salivary glands. A survey of 43 cases with study of prognostic
indicators. Int J Surg Pathol 1993; 1: 3–12.
12. Quinn CM, Wright NA. The clinical assessment of proliferation and growth
in human tumours: evaluation of methods and applications as prognostic
variables. J Pathol 1990; 160: 93–102.
13. Skalova A, Leivo I, von Boguslawsky K, Saksela E. Cell proliferation
correlates with prognosis in acinic cell carcinomas of salivary gland origin.
Immunohistochemical study of 30 cases using the MIB 1 antibody in
formalin-fixed paraffin sections. J Pathol 1994; 173: 13–21.
14. Skalova A, Lehtonen H, von Boguslawsky K, Leivo I. Prognostic signifi-
cance of cell proliferation in mucoepidermoid carcinomas of the salivary
gland: clinicopathological study using MIB 1 antibody in paraffin sections.
Hum Pathol 1994; 25: 929–935.
15. Wyllie AH. The biology of cell death in tumors. Anticancer Res 1985; 5:
131–136.
16. Arends MJ, McGregor AH, Wyllie AH. Apoptosis is inversely related to
necrosis and determines net growth in tumors bearing constitutively
expressed myc, ras and HPV oncogenes. Am J Pathol 1994; 144: 1045–1057.
17. Vaux DL, Cory S, Adams JM. Bcl-2 gene promotes haematopoietic cell
survival and cooperates with c-myc to immortalize pre-B cells. Nature 1988;
335: 440–445.
18. Nunez G, London L, Hockenbery D, Alexander M, McKearn J, Korsmeyer
SJ. Deregulated Bcl-2 gene expression selectively prolongs survival of
growth factor-deprived hemopoietic cell lines. J Immunol 1990; 144: 3602–
3610.
19. Pezzella F, Turley H, Kuzu I, et al. bcl-2 protein in non-small-cell lung
carcinoma. N Engl J Med 1993; 329: 690–694.
20. Pilotti S, Collini P, Rilke F, Cattoretti G, Del Bo R, Pierotti MA. BCL-2
protein expression in carcinomas originating from the follicular epithelium
of the thyroid gland. J Pathol 1994; 172: 337–342.
21. Doglioni C, Dei Tos AP, Laurino L, Chiarelli C, Barbareshi M, Viale G.
The prevalence of BCL-2 immunoreactivity in breast carcinomas and its
clinicopathological correlates, with particular reference to oestrogen recep-
tor status. Virchows Arch 1994; 424: 47–51.
22. Shi S-R, Key ME, Kalra KL. Antigen retrieval in formalin-fixed, paraffin-
embedded tissues: an enhancement method for immunohistochemical stain-
ing based on microwave oven heating of tissue sections. J Histochem
Cytochem 1991; 39: 741–748.
23. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell
death in situ via specific labeling of nuclear fragmentation. J Cell Biol 1992;
119: 493–501.
24. Tsujimoto Y, Cossman J, Croce CM. Involvement of the bcl-2 gene in
human follicular lymphoma. Science 1985; 228: 1440–1443.
25. Bakhshi A, Jensen JP, Goldman P, et al. Cloning the chromosomal
breakpoint of t(14,18) human lymphomas: clustering around JH on chro-
mosome 14 and near a transcriptional unit on 18. Cell 1985; 41: 899–906.
26. Cleary ML, Sklar J. Nucleotide sequence of a t(14;18) chromosomal
breakpoint in follicular lymphoma and demonstration of a breakpoint-
cluster region near a transcriptionally active locus on chromosome 18. Proc
Natl Acad Sci USA 1985; 82: 7439–7443.
27. McDonnell TJ, Troncoso P, Brisbay SM, et al. Expression of the proto-
oncogene bcl-2 in the prostate and its association with emergence of
androgen-independent prostate cancer. Cancer Res 1992; 52: 6940–6944.
28. Lu Q-L, Elia G, Lucas S, Alero Thomas J. Bcl-2 proto-oncogene expression
in Epstein–Barr-virus-associated nasopharyngeal carcinoma. Int J Cancer
1993; 53: 29–35.
29. Oltvai ZN, Milliman CL, Korsmeyer SJ. Bcl-2 heterodimerizes in vivo with
a conserved homlog, Bax, that accelerates programmed cell death. Cell
1993; 74: 609–619.
30. Miyashita T, Krajewski S, Krajewski M, et al. Tumor suppressor p53 is a
regulator of bcl-2 and bax gene expression in vitro and in vivo. Oncogene
1994; 9: 1799–1805.
31. Takayama S, Sato T, Krajewski S, et al. Cloning and functional analysis of
BAG-1: a novel bcl-2-binding protein with anti-cell death activity. Cell
1995; 80: 279–284.
32. Reed JC. BCL-2 protein function: insights from protein–protein inter-
actions. Proc 2nd Ann Apoptosis Res, Cambridge, MA, 1995.