- Email: [email protected]
The assessment of proliferating cell nuclear antigen (PCNA) immunostaining in human benign and malignant epithelial lesions of the parotid gland
Pergamon
Copyright
PII:
The Assessment Immunostaining
Oral Oncolo~ Vol 33, No. I, pp. 29-35, 1997 < ,997 Elsewer Saence Ltd. All rights reserved Printed m Great Bntam 0964-,955:97 $17 oo+o.oo
SO964-1955(96)00038-3
of Proliferating Cell Nuclear Antigen (PCNA) in Human Benign and Malignant Epithelial Lesions of the Parotid Gland Q. Zhu,
Department
F.H.
of Anatomy, The
White
and G.L.
University
of Hong
Tipoe Kong,
Hong
Kong
Immunoreactivity of proliferating cell nuclear antigen (PCNA) was assessed in formalin-fixed, paraffin-embedded sections from human normal parotid gland (N; n = 12), chronic sialadenitis (CS; n = S), Warthin’s tumour (W; n = lo), benign pleomorphic adenoma (BPA; n = ll), mucoepidermoid carcinoma (MEC; n = 14), carcinoma in pleomorphic adenoma (CPA; n = 10) and adenoid cystic carcinoma (ACC; n = 12) of the parotid gland, using the monoclonal antibody PC 10. The morphometric parameters measured comprised PCNA labelling indices (PI = the numerical percentage of PCNA positive nuclei) and volume densities of PCNA positive nuclei ( Vv, PEP = the relative volume of positive nuclei per unit volume of reference epithelium). All parameters were expressed in relation to total positive, as well as to strongly- and weakly-positive nuclei. In general, the values of PCNA parameters increased progressively in benign lesions in comparison with the N group, and in malignant neoplasms in comparison with non-neoplastic groups and benign lesions. The stronglypositive parameters showed more statistically significant differences than weakly-positive ones, suggesting that weakly-stained nuclei may include some non-cycling cells and, therefore, that weakly-positive parameters may not be reliable proliferation markers. Values for all parameters in CPA were significantly higher than those in BPA, suggesting that these parameters may be used as diagnostic discriminators. Spearman rank correlation analysis showed a highly positive correlation between the morphometric parameters and the severity of the lesions. Furthermore, the mean values of PIsp were significantly higher in patients who died of the malignant tumours than in those patients who survived. Our results indicate that PCNA indices might be useful markers for discriminating between benign (BPA) and malignant tumours of the parotid gland and that the parameter PZsp may have prognostic applications. I’ 1997 Elsevier Science Ltd. All rights reserved. Key words: PCNA,
salivary
Oral Oncology, Vol. 33, No.
gland, proliferation,
1, pp. 29-35,
morphometry,
prognosis, diagnosis, neoplasia
1997
INTRODUCTION
Salivary gland neoplasms exhibit a wide spectrum of histopathological features and diverse biological behaviour which does not follow survival patterns of other tumours in the head and neck region [ 1, 21. It can be difficult to separate salivary gland neoplasms into benign and malignant categories because of their characteristically bland histomorphological appearances and incomplete diagnostic criteria. Furthermore, malignant transformation may occur in some benign lesions such as benign pleomorphic adenoma, especially when they are left untreated [3, 41. Correspondence to F. H. White. Received 10 May 1996; provisionally accepted 23 May 1996; revised manuscript received 14 June 1996.
Reliable quantitative indicators would enhance objectivity and improve cancer diagnostic and prognostic decision making. Proliferating cell nuclear antigen (PCNA) has been demonstrated to be a useful proliferation marker. In some neoplasms, PCNA immunoreactivity has been found to correlate well with the malignant grade of the lesions, with the outcome of the patients, and with other proliferation markers of the cell cycle, e.g. bromodeoxyuridine (BrdU) incorporKi-67 expression and ation, thymidine labelling, S + G*M phase fraction by flow cytometry [5, 61. The aim of this study was to evaluate cellular proliferation with PCNA, using quantitative methods in human parotid gland lesions, in order to determine whether these methods can provide useful information for both diagnosis and prognosis.
Q. Zhu et al.
30
MATERIALS Tissue
AND
METHODS
determined by replacing a negative control.
antibody
with PBS as
specimens
All tissues used in the present study were formalin-fixed, paraffin-embedded blocks obtained from 57 patients who underwent primary surgery between 1977 and 1987 in the Stomatological Hospital, West China University of Medical Sciences, People’s Republic of China. They comprised histologically-normal tissues (N; n = 12), chronic sialadenitis (CS; n = 8), Warthin’s tumour (W; n = lo), benign pleomorphic adenoma (BPA; IZ = 1 l), carcinoma in pleomorphic adenoma (CPA; n = lo), mucoepidermoid carcinoma (MEC; n = 1 l), carcinoma in pleomorphic adenoma (CPA; n = 12). All cases were from the parotid gland. CPAs were characterised by coexistence of both benign and malignant elements, with their histopathological subtypes comprising mucoepidermoid carcinoma (n = l), adenoid cystic carcinoma (n = 2), squamous cell carcinoma (n = 2), myoepithelial carcinoma (n = 2), adenocarcinoma (n = 1) and undifferentiated carcinoma (n = 2). All the lesions were classified according to the WHO classification [7]. Four micrometre haematoxylin and eosin stained sections were reviewed to confirm the original diagnosis by the authors. Only lesions in which diagnoses were unanimous were included in the study. The normal tissues, which showed apparently normal histological features on qualitative examination, were obtained from BPAs and were at least 1 cm distant from the capsule of the tumour. Immunohistochemisty
Immunohistochemical studies were performed with the avidin-biotin-peroxidase complex (ABC) method. Four micrometre sections were dewaxed, rehydrated and treated with 0.75% hydrogen peroxide in phosphate saline solution (PBS) for 20 min at room temperature to block endogenous peroxidase activity. The following procedure was performed with ABC kits (Dakopatts, Carpinteria, California, U.S.A.). The sections were incubated with normal horse serum for 20 min at 37°C to block non-specific antibody binding. This was followed by incubation overnight at 4’C with mouse antihuman monoclonal antibody PC 10 (Dako) diluted to 1: 100 in PBS. The sections were then incubated with biotinylated antimouse immunoglobulin at 37’C for 30 min and subsequently for 30 min with the avidin-biotin complex at 37°C. Antigen localisation was visualised with a peroxidase substrate solution containing 0.05% 3,3’-diaminobenzidine tetrahydrochloride and 0.03% hydrogen peroxide in Tris buffer solution. Th.e sections were lightly counterstained with Harris’s haematoxylin and mounted in Permount (Fisher Scientific, Fairlawn, New Jersey, U.S.A.). The specificity of the immunohistochemical reactions was
Table No. patients
No. blocks/ patient
77 *Samples
the primary
comprised
1. Sampling No.
sections/ block
1 nuclei
of all normal
procedure
1 and pathological
epithelial
Assessment
and quantification
PI = Y?
staining
x 100.
NEN
(ii) PZsr was determined by the ratio of strongly-positive nuclei and the total number of nuclei in the reference epithelium multiplied by 100. NspN is the number of stronglypositive nuclei in the reference epithelium P&P
NSPN = _I NEN
x 100.
by the ratio of weakly-positive (iii) PZT was determined nuclei and the total number of nuclei in the reference epithelium multiplied by 100; NW,, is the number of weaklypositive nuclei in the reference epithelium NWPN
PIWP = 7
x 100
1vEN
Volume densities. (i) The volume density of positive nuclei in the reference epithelium (V,, pEp) was defined as the ratio of the number of uoints falling on both stronglv-_ and
for PCNA
quantification
No. fields/ section 10
cell types.
of PCNA
PCNA counting was performed with a light microscope (Olympus) equipped with a 36 point eyepiece graticule. The total side length of the lattice was 0.25 mm. The total area of the lattice was 0.0625 mm2 and each individual square of the graticule was 0.0025 mm2 at the final magnification of 400 x Microscopical fields were systematically sampled from each section. The sampling procedure is summarised in Table 1. The minimal sample size for each parameter was determined by progressive mean plot analysis [8] from one representative section for each patient which yielded a percentage error of the final mean of <5%. All epithelial cells which contained reaction product in their nuclei were considered to be positive. Positive nuclei were further subclassified as strongly positive, which showed intense brown staining with either a granular or diffuse pattern and weakly positive, which possessed light diffuse staining. No attempt was made to distinguish between the different types of the epithelial cells such as duct-like cells, myoepithelial cells, fibroblast-like cells, plasma-like cells, and clear cells in either control or neoplastic lesions. All counting was performed blind, without prior knowledge of clinical and follow-up data. The following parameters were determined. PCNA labelling indices. (i) PI was defined as the percentage of both strongly- and weakly-positive nuclei in the reference epithelium. NPN is the number of positive nuclei in the reference epithelium and NEN is the total number of nuclei in the reference epithelium
No. nuclei*/ field 200
Total no. nuclei/patient 2000
PCNA weakly-positive nuclei (Z’PN) and the number falling on the reference epithelium (Pan) h'.PEP
=-.
Staining
in Salivary Tumours
of points
(ii) The volume density of strongly-positive nuclei (Vv, sFEP) was defined as the ratio of the number of points falling on strongly-positive nuclei Z’s,,, and the number of points falling on the reference epithelium PSPN = ~ PEP
(iii) The volume density of weakly-positive nuclei (Vv, wPEF) was defined as the ratio of the number of points falling on weakly-positive nuclei (PwPN) and the number of points falling on the reference epithelium (Pa,,) h.X'PEP
PWPIG
= ~.
PEP
Statistical
analyses
The mean for each parameter was calculated per individual patient and the standard error of the mean was estimated for each group. The Chi-square test for goodness of fit was used to determine the normal distribution of the data. Log transformation was performed if the data were found to be not normally distributed. One-way analysis of variance and Duncan’s multiple range test were performed to detect the differences between the groups. The Spearman rank test was used to detect the correlation between the PCNA parameters and the severity of the disease. Statistical analyses were performed using Statgraphics Plus software. RESULTS Qualitative
were
detected
scattered
throughout
the neoplastic
PPN PEP
h.SPEP
nuclei cells.
31
observations
All PCNA staining was confined to nuclei. In normal tissues, PCNA-positive nuclei were occasionally found in acinar, ductal and all myoepithelial cells associated with both acinar and ductal systems (Fig. 1). The frequency of positive nuclei in the CS group was slightly increased when compared with the N group (Fig. 2). In Warthin’s turnours, positive nuclei were observed in both inner and outer epithelial layers as well as in the germinal centres of the lymphoid follicles (Fig. 3). In the BPA group, positive nuclei were infrequent and were located in myoepithelial sheets and both the luminal and outer layer cells of the duct-like structures, whereas the nuclei in myxoid and chondroid areas were negative (Fig. 4). The positive reaction was apparently increased in the three malignant groups. In the MEC group, positive reactions were present in epidermoid and intermediate cells (Fig. 5), but mucous cells were all negative. In the ACC group, PCNA-positive nuclei were scattered throughout the neoplastic epithelial components in both solid and cribriform areas (Fig. 6). The distribution of PCNA-positive nuclei in the CPA group varied according to the histological subtypes (Fig. 7). The distribution of positive nuclei in the subtypes of ACC and MEC were similar to those mentioned above. In the squamous cell carcinoma subtype, positive nuclei were localised mainly in basal cells and occasional parabasal cells of the epithelial islands. In the myoepithelial carcinoma subtype, positivity was shown in all neoplastic myoepithelial nuclei. In adenocarcinoma, positive
Quantitative
observations
The results of the quantitative data and statistical analyses are summarised in Tables 2 and 3. Generally, the values for the majority of parameters increased in malignant lesions when compared to non-neoplastic groups and benign lesions. Z’CNA labelling indices. I&, was significantly higher in CS and W groups than in the N group and was significantly higher in malignant lesions (CPA, MEC and ACC) than in N and BPA groups. The value of PZsr in BPA was significantly higher than in N, but was lower than in all the other groups. The CPA group showed significantly the highest value when compared with N, CS, W and BPA, but no significant differences were detected in comparisons between other malignant groups. The mean values of P&, in MEC and ACC groups were higher than in CS and W groups, but a statistically significant difference was only detected between CS and ACC. Although the values of PZw, increased in malignant groups in comparison with N, CS and benign lesions, the only significant differences detected for this parameter were between BPA and ACC groups, which had significantly lowest and highest values, respectively. The trend for PI was similar to that for PZwp, except that the value for BPA was significantly different compared with all other groups. Volume densities. The mean values of Vv,,,E, increased significantly in benign lesions when compared with N, and in malignant groups when compared with N, CS and benign lesions. The values in CPA and ACC were significantly higher than in the other groups. The MEC value was significantly higher than those for N, W and BPA. V,,,,,, failed to detect differences between each group except for ACC, where the value was consistently higher than all the other groups. The trend for V,,,,, was similar to that for V,,,,,, and statistically significant differences were detected between ACC and all other groups, between CPA and N, between CPA and CS, between CPA and benign groups and between CPA and ACC. Speaman
rank
correlation
analyszs
Highly positive correlations were observed between all parameters and all groups, indicating that the increasing values of each parameter strongly correlated with increasing severity of the lesions. Prognostic
value
of the PCNA
quantitative
methods
Follow-up data were collected from all 36 patients with malignant tumours (Table 4). Amongst them, 1 patient with CPA died on a date unknown to us, 1 patient with CPA was followed-up for only 12 months, 1 patient with ACC survived with a coexisting carcinoma of the urinary bladder and another patient with MEC died of diabetes. These patients were excluded from the analysis. Of the remaining patients, 12 died of the tumour between 5 and 132 months after the operation and 20 patients survived to the end of the follow-up date. Of these 20 patients, the survival time ranged from 60 to 156 months and amongst them, 7 patients had one or more recurrent tumours and 13 patients were without any recurrence between the operation
32
Q. Zhu
Fig.
1. One
PCNA-positive Fig.
Fig. Fig. Fig.
Fig. Fig.
3. Nuclei
from
basal
nucleus
nuclei
6. Numerous 7. Positive
is localised
2. PCNA-positive
4. A positive 5. Positive
nucleus
nuclei epithelial
is located
are present PCNA-positive
nuclei
are present
et al.
in an acinar cell in normal parotid gland (x400). t = strongly-positive fl= weakly-positive nucleus. are distributed in the intercalated duct cells in chronic sialadenitis (x400).
cells and
Corn
in a myoepithelial
in epidermoid nuclei
lymphocytes
of Iymphoid
cell in the myoepithelial
follicles nest
and
stain
intermediate cells in mucoepidermoid are negative (x400). are scattered in myoepithelial cells in cribriform
in the neoplastic
epithelial
cells
in an adenocarcinoma
intensely
in benign
with
pleomorphic
carcinoma, adenoid
adenoma but
cystic
in pleomorphic
PCNA
mucous
nucleus;
(x400). (x400). cell nuclei
carcinoma
(x400).
adenoma
(x200).
PCNA Table
Groups N(n
W(n
0.29 (f0.05)
= 8)
BPA(n = MEC(n = CI’A(n = ACC(n =
14)
9.93 (*2.60) 5.96
12)
(+1.16) f = standard
error
3.83 (~0.6 1 J 5.75 (iO.52, 7.09 (&0.71)
4.25 (+0.59)
4.69 (iO.80)
10)
labelling
PI
3.54
(f0.34) 0.82 (+o. 17)
11)
PCNA
(f0.60) 3.36 (k0.32)
2.40 (kO.24) 2.83
= 10)
results:
1.51 (t-0.5.5) 4.82
2.32 (f0.67) 9.51
(iO.57) 5.03 (+I .20)
(fl.26) 14.95 (f3.55)
14.38 (k3.44)
.20.33 (k3.98)
DISCUSSION The proliferative activity of a tumour is defined as the growth fraction, the number of cycling cells relative to the total number of cells [9]. Estimation of growth fraction has been used to determine the aggressiveness and progression of various tumour types [lo]. Currently, the techniques for analysis of proliferative activity include mitotic counting, AgNOR DNA flow cytometry, counting, BrdU (bromodeoxyuridine) incorporation and thymidine labelling. Each method measures a different parameter of proliferative activity and has its limitations [lo]. Using the technique of AgNOR counting, highly significant differences have been detected between the number of AgNORs in benign and malignant salivary gland tumours [ll-141. Although the technique is easy to carry out, the nature and significance of the measurement remains uncertain [5]. More recently, Ki67, a proliferation-associated antigen has attracted much attention. Murakami et al. [14] reported a significant differences in the frequency of Ki-67-positive cells between carciTable
3. Results
of statistical
mm
PI
Z’Zsr
analysis
vV,SPEP
vV,WPEP
h,PEP
*
*
*
*
cs
*
*
**
*I
*
*
W BPA MEC CPA ACC
**
*
* *
* * * *
* * ** *
N
indices
and volume
vV.WPEP
VV,SPEP
0.0002 (*o.ooo 0.0082 (*0.0015) 0.0032 (+o.ooos, 0.0020 (~0.0008, 0.0142 (*0.002x) 0.0299 (+0.0058) 0.0274 (~0.0054,
densities
1)
VV,PEP
0.0034 (~o.ooox,
0.0035 (+0.000x,
0.0089 (+0.0015, 0.0058
0.0171 (*0.0027) 0.0090
(~0.0018, 0.0070 (*0.0028)
(f0.0024) 0.0090 (iO.0032)
0.0137 (+0.003x)
0.0279 (+0.0057)
0.0166 @0.0043) 0.0538
0.0465 (kO.0097) 0.0813
(+0.0127)
(kO.0164)
of mean.
and the end of the follow-up date. These 20 patients were classified under the surviving group. The values for PZsp and PI in the group that died of the tumour were significantly higher than those in the group that survived. There was a significant difference in the parameter PZsP between the group that died of the tumour and the group that survived, with values for the former being higher than those of the latter (Table 4). PZwP and volume density parameters showed no statistically significant differences between the two groups (Table 4).
Group
33
in Salivary Tumours
m-w
PkP
= 12)
CS(n
2. Quantitative
Staining
*
*
*
* *** * **
One-way ANOVA followed down the column indicates nificant at 95% confidence
* * * * by Duncan’s that group interval.
* ** * multiple comparisons
** * *
*
*
range test; Overlap are not statistically
* of (*) sig-
noma and benign pleomorphic adenoma and between benign pleomorphic adenoma and normal tissues. However, Ki-67 requires fresh frozen tissues. A new monoclonal antibody, MIB 1, has been raised against the recombinant Ki67 which can be used on conventional paraffin-embedded tissues [ 151. PCNA is a highly conserved 36 kd non-histone nuclear protein that functions as an accessory protein to DNA polymerase delta in DNA synthesis. It plays a role in the initiation of cell proliferation, necessary for DNA replication. It is expressed in late Gi and increases through S and Gs, being lowest during mitosis and, hence, is considered as a cellular proliferation marker [ 161. PC 10 monoclonal antibody against PCNA is designed to recognise PCNA in conventionally-fixed paraffin-embedded materials [ 171. Immunohistochemical methods of assessing cellular proliferation in routinely processed tissues have been used in a wide variety of pathological conditions, with their particular advantage of maintenance of morphological structure of tissues making retrospective study possible. However, PCNA studies in salivary gland turnouts are very few [l&21] which may be due to the low occurrence of the lesions. To our knowledge, this is the first quantitative PCNA immunohistochemical report on salivary gland neoplasms to include prognostic information. A current histogenetic theory of salivary gland tumorigenesis excludes acinar cells as potential progenitor cells on the basis that functionally mature cells, such as acinar cells, are terminally differentiated and are incapable of further proliferation. Proliferation for the purposes of repair and regeneration is confined to stem cells residing exclusively amongst luminally-located intercalated duct cells or basallylocated excretory duct cells [22, 231. These postulated stem cells form the basis of the semipluripotential bicellular reserve cell hypothesis [24]. However, such a hypothesis has not been rigorously tested with direct experimental evidence. In contrast, previous studies have suggested that cell division occurs in all major cell types of salivary glands [25291. In normal parotid gland, we identified low numbers of PCNA-positive proliferating cells in acinar, ductal luminal cells and myoepithelial cells from both acini or the duct system. Similar results were obtained for normal salivary glands by Yang et al. [18], The location of proliferating cells was
34
Q. Zhu et al. Table
4. PCNA
parameters
and prognosis
D (n = 12)
11.34
9.71
21.05
0.029
s (n = 20)
(i2.08) 4.89
(*Loo) 8.57
(i2.U) 13.46
(*0.004) 0.021
0.025 (+o.oos, 0.035
0.055 (TO.006) 0.052
(k2.22) P < 0.05
(i2.73) PC 0.05
(+0.004) PC 0.05
(*0.009, P < 0.05
(*0.012) PC 0.05
S versus
D
D = died
of tie
(f0.81) Pi 0.05 tumour;
S = sunwed
wthout
recurrence;
k = standard
error
of mean
similar in CS lesions, but was more frequent. Based on these reports and on our own observations, it appears that any of the epithelial cell types present in normal salivary gland are capable of proliferation and thus have the potential to give rise to the variant tumour cell types arising in salivary gland neoplasms. This is in agreement with the multicellular theory, which postulates that differentiated cells at all levels of the gland, including acinar and basal cells, are capable of cell division [26]. In the present study, the significance of PCNA as a proliferation marker in different benign and malignant neoplasms was studied in comparison with normal tissues and chronic sialadenitis. Our data showed that all parameters were highly correlated with the severity of the lesions and in comparisons between groups, more significant differences were detected for strongly-positive parameters than for weakly-positive ones. CS lesions showed significantly higher values for PIs, than in the N group, whereas the benign tumours (BPA and W) showed significantly higher values for PZsp than in the N group. Such differences were not obtained for its weakly-positive counterpart Piw,. Generally, the malignant groups showed the highest values in all parameters when compared with N, CS and benign lesions. The values for PZse and V, sPEP in ACC and CPA were significantly higher when compared with N, CS and BPA, although weakly-positive parameters were less discriminating. The results for PCNA labelling indices in BPA were unexpected. The BPA group always had lower values of P1, f%P, PIWP, than did the CS and W groups and they were significantly different for P&p, P&p, and PI. The values for the volume densities in BPA were also lower than in CS, although significant differences were not obtained. This was contrary to our original postulate that BPA may exhibit higher proliferation in terms of PCNA immunoreactivity than W and CS lesions, because of its high frequency of recurrence and potential malignant transformation. Our data may reflect the clinically slow growth of typical BPA, and suggests that the high recurrence rate of this tumour may be due to incomplete surgical removal rather than to proliferation. The generally elevated PCNA parameters in CS, when compared with N and BPA, suggest that the cells in CS may have a high turn-over rate induced by the active inflammatory reaction which is reflected in the higher values in proliferation markers. This result is different from that of Yang et al. [19] where BPA showed a higher PCNA labelling index than W. We also noted that positive nuclei were mainly localised in the myoepithelial sheets and the ductlike structures, whereas myxoid and the chondroid areas were negative. Similar findings were described by Yang et al. [18, 191 and Ogama et al. [20], suggesting that prolifer-
ation occurs mainly in the epithelial components rather in the myxoid-chondroid areas in BPA. One of the problems confronting pathologists in salivary gland tumours is to distinguish malignant transformation in pleomorphic adenoma. In general, CPAs have been thought to occur following malignant transformation of the epithelial elements in BPA [4, 301. In practice, however, there may be some difficulties in diagnosing CPAs because of their diverse histological appearances and the absence of strict criteria. Yang et a/. [ 191 reported that the PCNA labelling index in CPA was significantly higher than in BPA. In the present study, CPA consistently showed significantly higher values than BPA for all parameters, except for Vv,wru, suggesting that these parameters may be used as diagnostic discriminators in distinguishing CPA and BPA, and may be helpful predictors in detecting malignant transformation in pleomorphic adenomas as a supplement to conventional histopathological criteria. There has always been difficulty in relating the grading of salivary gland carcinomas to their biological activities and subsequent clinical behaviour when compared with other malignancies, such as squamous cell carcinoma or sarcoma [ 181. Despite the large number of statistical significant differences, we failed to detect clear cut separation between the non-neoplastic, benign and malignant lesions for any single quantitative parameter. This may be a reflection of the heterogeneity of the neoplastic groups. The grade of malignancy between different types and subtypes of salivary gland malignant tumours needs to be further studied by increasing the sample size of each group. This will also reduce the intra-group variation. In the present study, strongly-positive nuclei were both granular and diffusely stained, while the weakly-positive ones were only diffusely stained. Our data showed that more statistically significant differences between groups were detected for strongly-positive parameters than their weakly-positive counterparts. Weakly-positive parameters PIwp failed to discriminate between N, CS and W groups. For PZwp and PI, BPA showed significantly lower values than the N group for P&ID and PI, which was not expected. One possible explanation is that weakly-positive nuclei may include some non-cycling cells because of the long half-life of the PCNA protein and, therefore, may not be a reliable marker of the proliferation status. Theoretically, the granular staining may demonstrate PCNA associated with DNA replication sites, whereas diffuse staining may represent PCNA not associated with replication sites [31]. PCNA levels have been shown to be correlated with survival in patients with several kinds of solid neoplasms. In salivary gland tumours, such correlation was reported in mucoepidermoid carcinoma [21]. In the present study, the strongly-positive parameter Pl,, was able to separate stat-
PCNA
Staining
in Salivary
istically the group who died of the disease from the survivors, suggesting that PCNA has a predictive role in the prognosis of salivary gland tumours. In summary, PCNA immunostaining by PC 10 may be a reliable and quantifiable marker of cell proliferation on formalin-fixed, paraffin-embedded tissues in normal, inflammatory, benign and malignant lesions of parotid gland. Strongly-positive parameters seemed to be better discriminators than weakly-positive ones and, therefore, more reliable diagnostic indicators. The frequency of PCNA-positive cells within a tumour specimen differentiates benign from malignant tumours (between BPA and CPA). All PCNA labelling indices evaluated showed prognostic value for malignant groups. The combination of PCNA immunohistochemistry and morphometry may have potential applications in both diagnosis and prognosis, in conjunction with conventional histopatbological assessment in salivary gland neoplasms.
1. Bats&is, salivary 2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
J. G. and Luna, M. A., Histopathologic grading of gland neoplasms: mucoepidermoid carcinomas. Annu/.s qf Otology, Rhinology and Larwgology, 1990, 99, 835-838. Dardick, I. and VanNostrand, A. W. I’., Morphogenesis of salivary gland tumours. A prerequisite to improving classification. Pathology Annual, 1987, 22, l-53. Seifert, G., Brocherion, C., Cardesa, A. and Eveson, J. W., WHO international histological classification of tumours: tentative histological classification of salivary gland tumours. Pathology, Research and Practice, 1990, 186, 555-581. Nagao, K., Matsuzaki, O., Saiga, H., et al., Histopathologic studies on carcinoma in pleomorphic adenoma of the parotid gland. Cancer, 1981, 48, 113-121. Yu, C. C. W., Woods, A. L. and Levison, D. A., The assessment of cellular proliferation by immunohistochemistry: a review of currently available methods and their applications. Histochemiral Journal, 1992, 24, 121-131. Yu, C. C. W. and Filipe, M. I., Update on proliferation-associated antibodies applicable to formalin-fixed paraffin-embedded tissue and their clinical applications. Histochemicul Journul, 1993, 25, 843-853. Seifert, G. and Sobin, L. H., The World Health Organisation’s histological classification of salivary gland tumours. A commentary on the second edition. Cuncer, 1992, 70, 379-385. Williams, M. A., Quantitative methods in biology. In Practical Methods in Electron Microscopy, ed. A.M. Glauert. Elsevier!’ North Holland, Amsterdam, pp. 36637. Scott, R. J., Hall, I’. A., Haldane, J. S., et al., A comparison of immunohistochemical markers of cell proliferation with experimentally determined growth fraction. Journcrl o/ Porholog),, 1991, 165, 173-178. DeLellis, R. A., Does the evaluation of proliferative activity predict malignancy or prognosis in endocrine tumours. Humun Palhok,g)!, 1995, 26, 131-l 34. Fujita, S., Takahashi, H. and Okabe, H., Nucleolar organizer regions in malignant salivary gland turnours. ilcla Pathologitri Japonica, 1992, 42, 727-733. Heerden, W. F. van and Raubenheimer, E. J., Evaluation of the nucleolar organizer region associated proteins in minor salivary gland tumours. Journal of Owl Ptrthologv and Mrdicine, 1991, 20, 291-295. Morgan, D. W., Cracker, J., Watts, A. and Shenoi, I’. M., Salivary gland tumours studied by means of the AgNOR technique. Histopathology, 1988, 13, 553-559. Murakami, M., Ohtani, I., Hojo, H. and Wakasa, H., Immunohistochemical evaluation with Ki-67: application to salivary gland turnouts. Journcrl of Ltrrjxgology and Otolog>,, 1992, 106, 35538. Cattoretti, G., Becker, M. H. G., Key, G., et al., Monoclonal antibodies against recombinant parts of the Ki-67 antigen
16
17
18
19
20
21
22.
23.
24. 25.
26.
27.
28. 29.
30.
31.
Tumours
35
(MIB 1 and MIB 3) detect proliferating cells in microwaveprocessed formalin-tixed paraffin sections. Journal qf‘ Pathology, 1992, 168, 357-363. Bravo, R., Synthesis of the nuclear protein cyclin (PCNA) and its relationship with DNA replication. E.xperimental Cell Resrarch, 1992, 163, 287-293. Waseem, N. H. and Lane, D. I’., Monoclonal antibody analysis of the proliferating cell nuclear antigen (PCNA). Structural conservation and the detection of a nucleolar form. Journal of Cell Scienw, 1990, 96, 121-129. Yang, L., Hashimura, K., Qin, C., Shrestha, I’., Sumitomo, S. and Mori, M., Immunoreactivity of proliferating cell nuclear antigen in salivary gland tumours: an assessment of growth potential. Virchows Archiv A Pathologicul Anutomy und Histopatholog~~, 1993, 422, 481-486. Yang, L., Liu, B., Qin, C., et al., Comparison of proliferating cell nuclear antigen index in benign and malignant salivary gland pleomorphic adenoma. European Journal qf Cancer, Owl Oncology, 1994, 30B, 56-60. Ogama, I., Miyauchi, M., Takata, T., Vuhahula, E., Ijuhin, N., Nikai, N. and Nikai, H., Proliferative activity of salivary gland pleomorphic adenomas and myoepitheliomas as evaluated by the proliferating cell nuclear antigen (PCNA) labeling index (LI). Journal of Oral Puthology und Medicine, 1993, 22, 4477450. Batsakis, J., Staging of salivary gland neoplasms: role of histopathologic and molecular factors. American Journal cf Surgery, 1994, 168, 386-390. Batsakis, J. G., Salivary gland neoplasis: an outcome of modified morphogenesis and cytodifferentiation. Oral Surgery, Oral M&icine. Oral Pathology, 1980, 49, 2299232. Batsakis, J. G., Regezi, J. A., Luna, M. A. and el-Naggar, A., Histogenesis of salivary gland neoplasms: a postulate with prognostic implications. Journal qf Laryngology rend Otologj,, 1989, 103, 9399944. Eversole, L. K., Histogenetic classification of salivary gland tumors. Archives of Pathology, 1971, 92, 433-443. Dardick, I., Current status of histogenetic and morphogenetic concepts of salivary gland tumorigenesis. Critical Rrviebts in Owl Bmlogy und Mrdicine, 1993, 4, 639-677. Dardick, I., Byard, R. W. and Carnegie, J. A., A review of the proliferative capacity of major salivary glands and the relationship to current concepts of neoplasmia in salivary glands. Oral Surgrr~~. Oral Medicine. Oral Puthologx, 1990, 69, 53-67. Denny, l’. C., Chai, Y., Klauser, D. K. and Denny, I-‘. A., Parenchymal cell proliferation and mechanisms for maintenance of granular duct and acinar cell populations in adult male mouse submandibular gland. Anuromical Record, 1993, 235, 475-485. Redman, R. S., Myoepithelium of salivary glands. Microscopy Resrurch and Twhniyues, 1994, 27, 25-45. Burford-Mason, A. I’., Cummins, M. M., Brown, D. H., Mackay, A. J. and Dardick, I., Immunohistochemical analysis of the proliferative capacity of duct and acinar cells during ligation-induced atrophy and subsequent regeneration of rat parotid gland. Journul of’ Oral Parhology and Medicine, 1993, 22, 440-446. Luna, M. A., Batsakis, J. G., Tortoledo, M. E. and Deljunco, G. W., Carcinomas ex monomorphic adenoma of salivary glands. Journal of Lurvngolog~ and O/oh/g>,, 1989, 103, 756759. Bravo, R. and Macdonald-Bravo, H., Existence of two populations of cycliniproliferating cell nuclear antigen during the cell cycle. Association with DNA replication sites. Journul of Cell Bio&, 1987, 105, 154991554.
Acknowledgements-The authors would like to thank Miss Car-man Leung for her skillful technical assistance and Mr Johnny Leung for his help with the photography. This work was supported by a grant from the Committee on Research and Conference Grants of The University of Hong Kong.
Copyright
PII:
The Assessment Immunostaining
Oral Oncolo~ Vol 33, No. I, pp. 29-35, 1997 < ,997 Elsewer Saence Ltd. All rights reserved Printed m Great Bntam 0964-,955:97 $17 oo+o.oo
SO964-1955(96)00038-3
of Proliferating Cell Nuclear Antigen (PCNA) in Human Benign and Malignant Epithelial Lesions of the Parotid Gland Q. Zhu,
Department
F.H.
of Anatomy, The
White
and G.L.
University
of Hong
Tipoe Kong,
Hong
Kong
Immunoreactivity of proliferating cell nuclear antigen (PCNA) was assessed in formalin-fixed, paraffin-embedded sections from human normal parotid gland (N; n = 12), chronic sialadenitis (CS; n = S), Warthin’s tumour (W; n = lo), benign pleomorphic adenoma (BPA; n = ll), mucoepidermoid carcinoma (MEC; n = 14), carcinoma in pleomorphic adenoma (CPA; n = 10) and adenoid cystic carcinoma (ACC; n = 12) of the parotid gland, using the monoclonal antibody PC 10. The morphometric parameters measured comprised PCNA labelling indices (PI = the numerical percentage of PCNA positive nuclei) and volume densities of PCNA positive nuclei ( Vv, PEP = the relative volume of positive nuclei per unit volume of reference epithelium). All parameters were expressed in relation to total positive, as well as to strongly- and weakly-positive nuclei. In general, the values of PCNA parameters increased progressively in benign lesions in comparison with the N group, and in malignant neoplasms in comparison with non-neoplastic groups and benign lesions. The stronglypositive parameters showed more statistically significant differences than weakly-positive ones, suggesting that weakly-stained nuclei may include some non-cycling cells and, therefore, that weakly-positive parameters may not be reliable proliferation markers. Values for all parameters in CPA were significantly higher than those in BPA, suggesting that these parameters may be used as diagnostic discriminators. Spearman rank correlation analysis showed a highly positive correlation between the morphometric parameters and the severity of the lesions. Furthermore, the mean values of PIsp were significantly higher in patients who died of the malignant tumours than in those patients who survived. Our results indicate that PCNA indices might be useful markers for discriminating between benign (BPA) and malignant tumours of the parotid gland and that the parameter PZsp may have prognostic applications. I’ 1997 Elsevier Science Ltd. All rights reserved. Key words: PCNA,
salivary
Oral Oncology, Vol. 33, No.
gland, proliferation,
1, pp. 29-35,
morphometry,
prognosis, diagnosis, neoplasia
1997
INTRODUCTION
Salivary gland neoplasms exhibit a wide spectrum of histopathological features and diverse biological behaviour which does not follow survival patterns of other tumours in the head and neck region [ 1, 21. It can be difficult to separate salivary gland neoplasms into benign and malignant categories because of their characteristically bland histomorphological appearances and incomplete diagnostic criteria. Furthermore, malignant transformation may occur in some benign lesions such as benign pleomorphic adenoma, especially when they are left untreated [3, 41. Correspondence to F. H. White. Received 10 May 1996; provisionally accepted 23 May 1996; revised manuscript received 14 June 1996.
Reliable quantitative indicators would enhance objectivity and improve cancer diagnostic and prognostic decision making. Proliferating cell nuclear antigen (PCNA) has been demonstrated to be a useful proliferation marker. In some neoplasms, PCNA immunoreactivity has been found to correlate well with the malignant grade of the lesions, with the outcome of the patients, and with other proliferation markers of the cell cycle, e.g. bromodeoxyuridine (BrdU) incorporKi-67 expression and ation, thymidine labelling, S + G*M phase fraction by flow cytometry [5, 61. The aim of this study was to evaluate cellular proliferation with PCNA, using quantitative methods in human parotid gland lesions, in order to determine whether these methods can provide useful information for both diagnosis and prognosis.
Q. Zhu et al.
30
MATERIALS Tissue
AND
METHODS
determined by replacing a negative control.
antibody
with PBS as
specimens
All tissues used in the present study were formalin-fixed, paraffin-embedded blocks obtained from 57 patients who underwent primary surgery between 1977 and 1987 in the Stomatological Hospital, West China University of Medical Sciences, People’s Republic of China. They comprised histologically-normal tissues (N; n = 12), chronic sialadenitis (CS; n = 8), Warthin’s tumour (W; n = lo), benign pleomorphic adenoma (BPA; IZ = 1 l), carcinoma in pleomorphic adenoma (CPA; n = lo), mucoepidermoid carcinoma (MEC; n = 1 l), carcinoma in pleomorphic adenoma (CPA; n = 12). All cases were from the parotid gland. CPAs were characterised by coexistence of both benign and malignant elements, with their histopathological subtypes comprising mucoepidermoid carcinoma (n = l), adenoid cystic carcinoma (n = 2), squamous cell carcinoma (n = 2), myoepithelial carcinoma (n = 2), adenocarcinoma (n = 1) and undifferentiated carcinoma (n = 2). All the lesions were classified according to the WHO classification [7]. Four micrometre haematoxylin and eosin stained sections were reviewed to confirm the original diagnosis by the authors. Only lesions in which diagnoses were unanimous were included in the study. The normal tissues, which showed apparently normal histological features on qualitative examination, were obtained from BPAs and were at least 1 cm distant from the capsule of the tumour. Immunohistochemisty
Immunohistochemical studies were performed with the avidin-biotin-peroxidase complex (ABC) method. Four micrometre sections were dewaxed, rehydrated and treated with 0.75% hydrogen peroxide in phosphate saline solution (PBS) for 20 min at room temperature to block endogenous peroxidase activity. The following procedure was performed with ABC kits (Dakopatts, Carpinteria, California, U.S.A.). The sections were incubated with normal horse serum for 20 min at 37°C to block non-specific antibody binding. This was followed by incubation overnight at 4’C with mouse antihuman monoclonal antibody PC 10 (Dako) diluted to 1: 100 in PBS. The sections were then incubated with biotinylated antimouse immunoglobulin at 37’C for 30 min and subsequently for 30 min with the avidin-biotin complex at 37°C. Antigen localisation was visualised with a peroxidase substrate solution containing 0.05% 3,3’-diaminobenzidine tetrahydrochloride and 0.03% hydrogen peroxide in Tris buffer solution. Th.e sections were lightly counterstained with Harris’s haematoxylin and mounted in Permount (Fisher Scientific, Fairlawn, New Jersey, U.S.A.). The specificity of the immunohistochemical reactions was
Table No. patients
No. blocks/ patient
77 *Samples
the primary
comprised
1. Sampling No.
sections/ block
1 nuclei
of all normal
procedure
1 and pathological
epithelial
Assessment
and quantification
PI = Y?
staining
x 100.
NEN
(ii) PZsr was determined by the ratio of strongly-positive nuclei and the total number of nuclei in the reference epithelium multiplied by 100. NspN is the number of stronglypositive nuclei in the reference epithelium P&P
NSPN = _I NEN
x 100.
by the ratio of weakly-positive (iii) PZT was determined nuclei and the total number of nuclei in the reference epithelium multiplied by 100; NW,, is the number of weaklypositive nuclei in the reference epithelium NWPN
PIWP = 7
x 100
1vEN
Volume densities. (i) The volume density of positive nuclei in the reference epithelium (V,, pEp) was defined as the ratio of the number of uoints falling on both stronglv-_ and
for PCNA
quantification
No. fields/ section 10
cell types.
of PCNA
PCNA counting was performed with a light microscope (Olympus) equipped with a 36 point eyepiece graticule. The total side length of the lattice was 0.25 mm. The total area of the lattice was 0.0625 mm2 and each individual square of the graticule was 0.0025 mm2 at the final magnification of 400 x Microscopical fields were systematically sampled from each section. The sampling procedure is summarised in Table 1. The minimal sample size for each parameter was determined by progressive mean plot analysis [8] from one representative section for each patient which yielded a percentage error of the final mean of <5%. All epithelial cells which contained reaction product in their nuclei were considered to be positive. Positive nuclei were further subclassified as strongly positive, which showed intense brown staining with either a granular or diffuse pattern and weakly positive, which possessed light diffuse staining. No attempt was made to distinguish between the different types of the epithelial cells such as duct-like cells, myoepithelial cells, fibroblast-like cells, plasma-like cells, and clear cells in either control or neoplastic lesions. All counting was performed blind, without prior knowledge of clinical and follow-up data. The following parameters were determined. PCNA labelling indices. (i) PI was defined as the percentage of both strongly- and weakly-positive nuclei in the reference epithelium. NPN is the number of positive nuclei in the reference epithelium and NEN is the total number of nuclei in the reference epithelium
No. nuclei*/ field 200
Total no. nuclei/patient 2000
PCNA weakly-positive nuclei (Z’PN) and the number falling on the reference epithelium (Pan) h'.PEP
=-.
Staining
in Salivary Tumours
of points
(ii) The volume density of strongly-positive nuclei (Vv, sFEP) was defined as the ratio of the number of points falling on strongly-positive nuclei Z’s,,, and the number of points falling on the reference epithelium PSPN = ~ PEP
(iii) The volume density of weakly-positive nuclei (Vv, wPEF) was defined as the ratio of the number of points falling on weakly-positive nuclei (PwPN) and the number of points falling on the reference epithelium (Pa,,) h.X'PEP
PWPIG
= ~.
PEP
Statistical
analyses
The mean for each parameter was calculated per individual patient and the standard error of the mean was estimated for each group. The Chi-square test for goodness of fit was used to determine the normal distribution of the data. Log transformation was performed if the data were found to be not normally distributed. One-way analysis of variance and Duncan’s multiple range test were performed to detect the differences between the groups. The Spearman rank test was used to detect the correlation between the PCNA parameters and the severity of the disease. Statistical analyses were performed using Statgraphics Plus software. RESULTS Qualitative
were
detected
scattered
throughout
the neoplastic
PPN PEP
h.SPEP
nuclei cells.
31
observations
All PCNA staining was confined to nuclei. In normal tissues, PCNA-positive nuclei were occasionally found in acinar, ductal and all myoepithelial cells associated with both acinar and ductal systems (Fig. 1). The frequency of positive nuclei in the CS group was slightly increased when compared with the N group (Fig. 2). In Warthin’s turnours, positive nuclei were observed in both inner and outer epithelial layers as well as in the germinal centres of the lymphoid follicles (Fig. 3). In the BPA group, positive nuclei were infrequent and were located in myoepithelial sheets and both the luminal and outer layer cells of the duct-like structures, whereas the nuclei in myxoid and chondroid areas were negative (Fig. 4). The positive reaction was apparently increased in the three malignant groups. In the MEC group, positive reactions were present in epidermoid and intermediate cells (Fig. 5), but mucous cells were all negative. In the ACC group, PCNA-positive nuclei were scattered throughout the neoplastic epithelial components in both solid and cribriform areas (Fig. 6). The distribution of PCNA-positive nuclei in the CPA group varied according to the histological subtypes (Fig. 7). The distribution of positive nuclei in the subtypes of ACC and MEC were similar to those mentioned above. In the squamous cell carcinoma subtype, positive nuclei were localised mainly in basal cells and occasional parabasal cells of the epithelial islands. In the myoepithelial carcinoma subtype, positivity was shown in all neoplastic myoepithelial nuclei. In adenocarcinoma, positive
Quantitative
observations
The results of the quantitative data and statistical analyses are summarised in Tables 2 and 3. Generally, the values for the majority of parameters increased in malignant lesions when compared to non-neoplastic groups and benign lesions. Z’CNA labelling indices. I&, was significantly higher in CS and W groups than in the N group and was significantly higher in malignant lesions (CPA, MEC and ACC) than in N and BPA groups. The value of PZsr in BPA was significantly higher than in N, but was lower than in all the other groups. The CPA group showed significantly the highest value when compared with N, CS, W and BPA, but no significant differences were detected in comparisons between other malignant groups. The mean values of P&, in MEC and ACC groups were higher than in CS and W groups, but a statistically significant difference was only detected between CS and ACC. Although the values of PZw, increased in malignant groups in comparison with N, CS and benign lesions, the only significant differences detected for this parameter were between BPA and ACC groups, which had significantly lowest and highest values, respectively. The trend for PI was similar to that for PZwp, except that the value for BPA was significantly different compared with all other groups. Volume densities. The mean values of Vv,,,E, increased significantly in benign lesions when compared with N, and in malignant groups when compared with N, CS and benign lesions. The values in CPA and ACC were significantly higher than in the other groups. The MEC value was significantly higher than those for N, W and BPA. V,,,,,, failed to detect differences between each group except for ACC, where the value was consistently higher than all the other groups. The trend for V,,,,, was similar to that for V,,,,,, and statistically significant differences were detected between ACC and all other groups, between CPA and N, between CPA and CS, between CPA and benign groups and between CPA and ACC. Speaman
rank
correlation
analyszs
Highly positive correlations were observed between all parameters and all groups, indicating that the increasing values of each parameter strongly correlated with increasing severity of the lesions. Prognostic
value
of the PCNA
quantitative
methods
Follow-up data were collected from all 36 patients with malignant tumours (Table 4). Amongst them, 1 patient with CPA died on a date unknown to us, 1 patient with CPA was followed-up for only 12 months, 1 patient with ACC survived with a coexisting carcinoma of the urinary bladder and another patient with MEC died of diabetes. These patients were excluded from the analysis. Of the remaining patients, 12 died of the tumour between 5 and 132 months after the operation and 20 patients survived to the end of the follow-up date. Of these 20 patients, the survival time ranged from 60 to 156 months and amongst them, 7 patients had one or more recurrent tumours and 13 patients were without any recurrence between the operation
32
Q. Zhu
Fig.
1. One
PCNA-positive Fig.
Fig. Fig. Fig.
Fig. Fig.
3. Nuclei
from
basal
nucleus
nuclei
6. Numerous 7. Positive
is localised
2. PCNA-positive
4. A positive 5. Positive
nucleus
nuclei epithelial
is located
are present PCNA-positive
nuclei
are present
et al.
in an acinar cell in normal parotid gland (x400). t = strongly-positive fl= weakly-positive nucleus. are distributed in the intercalated duct cells in chronic sialadenitis (x400).
cells and
Corn
in a myoepithelial
in epidermoid nuclei
lymphocytes
of Iymphoid
cell in the myoepithelial
follicles nest
and
stain
intermediate cells in mucoepidermoid are negative (x400). are scattered in myoepithelial cells in cribriform
in the neoplastic
epithelial
cells
in an adenocarcinoma
intensely
in benign
with
pleomorphic
carcinoma, adenoid
adenoma but
cystic
in pleomorphic
PCNA
mucous
nucleus;
(x400). (x400). cell nuclei
carcinoma
(x400).
adenoma
(x200).
PCNA Table
Groups N(n
W(n
0.29 (f0.05)
= 8)
BPA(n = MEC(n = CI’A(n = ACC(n =
14)
9.93 (*2.60) 5.96
12)
(+1.16) f = standard
error
3.83 (~0.6 1 J 5.75 (iO.52, 7.09 (&0.71)
4.25 (+0.59)
4.69 (iO.80)
10)
labelling
PI
3.54
(f0.34) 0.82 (+o. 17)
11)
PCNA
(f0.60) 3.36 (k0.32)
2.40 (kO.24) 2.83
= 10)
results:
1.51 (t-0.5.5) 4.82
2.32 (f0.67) 9.51
(iO.57) 5.03 (+I .20)
(fl.26) 14.95 (f3.55)
14.38 (k3.44)
.20.33 (k3.98)
DISCUSSION The proliferative activity of a tumour is defined as the growth fraction, the number of cycling cells relative to the total number of cells [9]. Estimation of growth fraction has been used to determine the aggressiveness and progression of various tumour types [lo]. Currently, the techniques for analysis of proliferative activity include mitotic counting, AgNOR DNA flow cytometry, counting, BrdU (bromodeoxyuridine) incorporation and thymidine labelling. Each method measures a different parameter of proliferative activity and has its limitations [lo]. Using the technique of AgNOR counting, highly significant differences have been detected between the number of AgNORs in benign and malignant salivary gland tumours [ll-141. Although the technique is easy to carry out, the nature and significance of the measurement remains uncertain [5]. More recently, Ki67, a proliferation-associated antigen has attracted much attention. Murakami et al. [14] reported a significant differences in the frequency of Ki-67-positive cells between carciTable
3. Results
of statistical
mm
PI
Z’Zsr
analysis
vV,SPEP
vV,WPEP
h,PEP
*
*
*
*
cs
*
*
**
*I
*
*
W BPA MEC CPA ACC
**
*
* *
* * * *
* * ** *
N
indices
and volume
vV.WPEP
VV,SPEP
0.0002 (*o.ooo 0.0082 (*0.0015) 0.0032 (+o.ooos, 0.0020 (~0.0008, 0.0142 (*0.002x) 0.0299 (+0.0058) 0.0274 (~0.0054,
densities
1)
VV,PEP
0.0034 (~o.ooox,
0.0035 (+0.000x,
0.0089 (+0.0015, 0.0058
0.0171 (*0.0027) 0.0090
(~0.0018, 0.0070 (*0.0028)
(f0.0024) 0.0090 (iO.0032)
0.0137 (+0.003x)
0.0279 (+0.0057)
0.0166 @0.0043) 0.0538
0.0465 (kO.0097) 0.0813
(+0.0127)
(kO.0164)
of mean.
and the end of the follow-up date. These 20 patients were classified under the surviving group. The values for PZsp and PI in the group that died of the tumour were significantly higher than those in the group that survived. There was a significant difference in the parameter PZsP between the group that died of the tumour and the group that survived, with values for the former being higher than those of the latter (Table 4). PZwP and volume density parameters showed no statistically significant differences between the two groups (Table 4).
Group
33
in Salivary Tumours
m-w
PkP
= 12)
CS(n
2. Quantitative
Staining
*
*
*
* *** * **
One-way ANOVA followed down the column indicates nificant at 95% confidence
* * * * by Duncan’s that group interval.
* ** * multiple comparisons
** * *
*
*
range test; Overlap are not statistically
* of (*) sig-
noma and benign pleomorphic adenoma and between benign pleomorphic adenoma and normal tissues. However, Ki-67 requires fresh frozen tissues. A new monoclonal antibody, MIB 1, has been raised against the recombinant Ki67 which can be used on conventional paraffin-embedded tissues [ 151. PCNA is a highly conserved 36 kd non-histone nuclear protein that functions as an accessory protein to DNA polymerase delta in DNA synthesis. It plays a role in the initiation of cell proliferation, necessary for DNA replication. It is expressed in late Gi and increases through S and Gs, being lowest during mitosis and, hence, is considered as a cellular proliferation marker [ 161. PC 10 monoclonal antibody against PCNA is designed to recognise PCNA in conventionally-fixed paraffin-embedded materials [ 171. Immunohistochemical methods of assessing cellular proliferation in routinely processed tissues have been used in a wide variety of pathological conditions, with their particular advantage of maintenance of morphological structure of tissues making retrospective study possible. However, PCNA studies in salivary gland turnouts are very few [l&21] which may be due to the low occurrence of the lesions. To our knowledge, this is the first quantitative PCNA immunohistochemical report on salivary gland neoplasms to include prognostic information. A current histogenetic theory of salivary gland tumorigenesis excludes acinar cells as potential progenitor cells on the basis that functionally mature cells, such as acinar cells, are terminally differentiated and are incapable of further proliferation. Proliferation for the purposes of repair and regeneration is confined to stem cells residing exclusively amongst luminally-located intercalated duct cells or basallylocated excretory duct cells [22, 231. These postulated stem cells form the basis of the semipluripotential bicellular reserve cell hypothesis [24]. However, such a hypothesis has not been rigorously tested with direct experimental evidence. In contrast, previous studies have suggested that cell division occurs in all major cell types of salivary glands [25291. In normal parotid gland, we identified low numbers of PCNA-positive proliferating cells in acinar, ductal luminal cells and myoepithelial cells from both acini or the duct system. Similar results were obtained for normal salivary glands by Yang et al. [18], The location of proliferating cells was
34
Q. Zhu et al. Table
4. PCNA
parameters
and prognosis
D (n = 12)
11.34
9.71
21.05
0.029
s (n = 20)
(i2.08) 4.89
(*Loo) 8.57
(i2.U) 13.46
(*0.004) 0.021
0.025 (+o.oos, 0.035
0.055 (TO.006) 0.052
(k2.22) P < 0.05
(i2.73) PC 0.05
(+0.004) PC 0.05
(*0.009, P < 0.05
(*0.012) PC 0.05
S versus
D
D = died
of tie
(f0.81) Pi 0.05 tumour;
S = sunwed
wthout
recurrence;
k = standard
error
of mean
similar in CS lesions, but was more frequent. Based on these reports and on our own observations, it appears that any of the epithelial cell types present in normal salivary gland are capable of proliferation and thus have the potential to give rise to the variant tumour cell types arising in salivary gland neoplasms. This is in agreement with the multicellular theory, which postulates that differentiated cells at all levels of the gland, including acinar and basal cells, are capable of cell division [26]. In the present study, the significance of PCNA as a proliferation marker in different benign and malignant neoplasms was studied in comparison with normal tissues and chronic sialadenitis. Our data showed that all parameters were highly correlated with the severity of the lesions and in comparisons between groups, more significant differences were detected for strongly-positive parameters than for weakly-positive ones. CS lesions showed significantly higher values for PIs, than in the N group, whereas the benign tumours (BPA and W) showed significantly higher values for PZsp than in the N group. Such differences were not obtained for its weakly-positive counterpart Piw,. Generally, the malignant groups showed the highest values in all parameters when compared with N, CS and benign lesions. The values for PZse and V, sPEP in ACC and CPA were significantly higher when compared with N, CS and BPA, although weakly-positive parameters were less discriminating. The results for PCNA labelling indices in BPA were unexpected. The BPA group always had lower values of P1, f%P, PIWP, than did the CS and W groups and they were significantly different for P&p, P&p, and PI. The values for the volume densities in BPA were also lower than in CS, although significant differences were not obtained. This was contrary to our original postulate that BPA may exhibit higher proliferation in terms of PCNA immunoreactivity than W and CS lesions, because of its high frequency of recurrence and potential malignant transformation. Our data may reflect the clinically slow growth of typical BPA, and suggests that the high recurrence rate of this tumour may be due to incomplete surgical removal rather than to proliferation. The generally elevated PCNA parameters in CS, when compared with N and BPA, suggest that the cells in CS may have a high turn-over rate induced by the active inflammatory reaction which is reflected in the higher values in proliferation markers. This result is different from that of Yang et al. [19] where BPA showed a higher PCNA labelling index than W. We also noted that positive nuclei were mainly localised in the myoepithelial sheets and the ductlike structures, whereas myxoid and the chondroid areas were negative. Similar findings were described by Yang et al. [18, 191 and Ogama et al. [20], suggesting that prolifer-
ation occurs mainly in the epithelial components rather in the myxoid-chondroid areas in BPA. One of the problems confronting pathologists in salivary gland tumours is to distinguish malignant transformation in pleomorphic adenoma. In general, CPAs have been thought to occur following malignant transformation of the epithelial elements in BPA [4, 301. In practice, however, there may be some difficulties in diagnosing CPAs because of their diverse histological appearances and the absence of strict criteria. Yang et a/. [ 191 reported that the PCNA labelling index in CPA was significantly higher than in BPA. In the present study, CPA consistently showed significantly higher values than BPA for all parameters, except for Vv,wru, suggesting that these parameters may be used as diagnostic discriminators in distinguishing CPA and BPA, and may be helpful predictors in detecting malignant transformation in pleomorphic adenomas as a supplement to conventional histopathological criteria. There has always been difficulty in relating the grading of salivary gland carcinomas to their biological activities and subsequent clinical behaviour when compared with other malignancies, such as squamous cell carcinoma or sarcoma [ 181. Despite the large number of statistical significant differences, we failed to detect clear cut separation between the non-neoplastic, benign and malignant lesions for any single quantitative parameter. This may be a reflection of the heterogeneity of the neoplastic groups. The grade of malignancy between different types and subtypes of salivary gland malignant tumours needs to be further studied by increasing the sample size of each group. This will also reduce the intra-group variation. In the present study, strongly-positive nuclei were both granular and diffusely stained, while the weakly-positive ones were only diffusely stained. Our data showed that more statistically significant differences between groups were detected for strongly-positive parameters than their weakly-positive counterparts. Weakly-positive parameters PIwp failed to discriminate between N, CS and W groups. For PZwp and PI, BPA showed significantly lower values than the N group for P&ID and PI, which was not expected. One possible explanation is that weakly-positive nuclei may include some non-cycling cells because of the long half-life of the PCNA protein and, therefore, may not be a reliable marker of the proliferation status. Theoretically, the granular staining may demonstrate PCNA associated with DNA replication sites, whereas diffuse staining may represent PCNA not associated with replication sites [31]. PCNA levels have been shown to be correlated with survival in patients with several kinds of solid neoplasms. In salivary gland tumours, such correlation was reported in mucoepidermoid carcinoma [21]. In the present study, the strongly-positive parameter Pl,, was able to separate stat-
PCNA
Staining
in Salivary
istically the group who died of the disease from the survivors, suggesting that PCNA has a predictive role in the prognosis of salivary gland tumours. In summary, PCNA immunostaining by PC 10 may be a reliable and quantifiable marker of cell proliferation on formalin-fixed, paraffin-embedded tissues in normal, inflammatory, benign and malignant lesions of parotid gland. Strongly-positive parameters seemed to be better discriminators than weakly-positive ones and, therefore, more reliable diagnostic indicators. The frequency of PCNA-positive cells within a tumour specimen differentiates benign from malignant tumours (between BPA and CPA). All PCNA labelling indices evaluated showed prognostic value for malignant groups. The combination of PCNA immunohistochemistry and morphometry may have potential applications in both diagnosis and prognosis, in conjunction with conventional histopatbological assessment in salivary gland neoplasms.
1. Bats&is, salivary 2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
J. G. and Luna, M. A., Histopathologic grading of gland neoplasms: mucoepidermoid carcinomas. Annu/.s qf Otology, Rhinology and Larwgology, 1990, 99, 835-838. Dardick, I. and VanNostrand, A. W. I’., Morphogenesis of salivary gland tumours. A prerequisite to improving classification. Pathology Annual, 1987, 22, l-53. Seifert, G., Brocherion, C., Cardesa, A. and Eveson, J. W., WHO international histological classification of tumours: tentative histological classification of salivary gland tumours. Pathology, Research and Practice, 1990, 186, 555-581. Nagao, K., Matsuzaki, O., Saiga, H., et al., Histopathologic studies on carcinoma in pleomorphic adenoma of the parotid gland. Cancer, 1981, 48, 113-121. Yu, C. C. W., Woods, A. L. and Levison, D. A., The assessment of cellular proliferation by immunohistochemistry: a review of currently available methods and their applications. Histochemiral Journal, 1992, 24, 121-131. Yu, C. C. W. and Filipe, M. I., Update on proliferation-associated antibodies applicable to formalin-fixed paraffin-embedded tissue and their clinical applications. Histochemicul Journul, 1993, 25, 843-853. Seifert, G. and Sobin, L. H., The World Health Organisation’s histological classification of salivary gland tumours. A commentary on the second edition. Cuncer, 1992, 70, 379-385. Williams, M. A., Quantitative methods in biology. In Practical Methods in Electron Microscopy, ed. A.M. Glauert. Elsevier!’ North Holland, Amsterdam, pp. 36637. Scott, R. J., Hall, I’. A., Haldane, J. S., et al., A comparison of immunohistochemical markers of cell proliferation with experimentally determined growth fraction. Journcrl o/ Porholog),, 1991, 165, 173-178. DeLellis, R. A., Does the evaluation of proliferative activity predict malignancy or prognosis in endocrine tumours. Humun Palhok,g)!, 1995, 26, 131-l 34. Fujita, S., Takahashi, H. and Okabe, H., Nucleolar organizer regions in malignant salivary gland turnours. ilcla Pathologitri Japonica, 1992, 42, 727-733. Heerden, W. F. van and Raubenheimer, E. J., Evaluation of the nucleolar organizer region associated proteins in minor salivary gland tumours. Journal of Owl Ptrthologv and Mrdicine, 1991, 20, 291-295. Morgan, D. W., Cracker, J., Watts, A. and Shenoi, I’. M., Salivary gland tumours studied by means of the AgNOR technique. Histopathology, 1988, 13, 553-559. Murakami, M., Ohtani, I., Hojo, H. and Wakasa, H., Immunohistochemical evaluation with Ki-67: application to salivary gland turnouts. Journcrl of Ltrrjxgology and Otolog>,, 1992, 106, 35538. Cattoretti, G., Becker, M. H. G., Key, G., et al., Monoclonal antibodies against recombinant parts of the Ki-67 antigen
16
17
18
19
20
21
22.
23.
24. 25.
26.
27.
28. 29.
30.
31.
Tumours
35
(MIB 1 and MIB 3) detect proliferating cells in microwaveprocessed formalin-tixed paraffin sections. Journal qf‘ Pathology, 1992, 168, 357-363. Bravo, R., Synthesis of the nuclear protein cyclin (PCNA) and its relationship with DNA replication. E.xperimental Cell Resrarch, 1992, 163, 287-293. Waseem, N. H. and Lane, D. I’., Monoclonal antibody analysis of the proliferating cell nuclear antigen (PCNA). Structural conservation and the detection of a nucleolar form. Journal of Cell Scienw, 1990, 96, 121-129. Yang, L., Hashimura, K., Qin, C., Shrestha, I’., Sumitomo, S. and Mori, M., Immunoreactivity of proliferating cell nuclear antigen in salivary gland tumours: an assessment of growth potential. Virchows Archiv A Pathologicul Anutomy und Histopatholog~~, 1993, 422, 481-486. Yang, L., Liu, B., Qin, C., et al., Comparison of proliferating cell nuclear antigen index in benign and malignant salivary gland pleomorphic adenoma. European Journal qf Cancer, Owl Oncology, 1994, 30B, 56-60. Ogama, I., Miyauchi, M., Takata, T., Vuhahula, E., Ijuhin, N., Nikai, N. and Nikai, H., Proliferative activity of salivary gland pleomorphic adenomas and myoepitheliomas as evaluated by the proliferating cell nuclear antigen (PCNA) labeling index (LI). Journal of Oral Puthology und Medicine, 1993, 22, 4477450. Batsakis, J., Staging of salivary gland neoplasms: role of histopathologic and molecular factors. American Journal cf Surgery, 1994, 168, 386-390. Batsakis, J. G., Salivary gland neoplasis: an outcome of modified morphogenesis and cytodifferentiation. Oral Surgery, Oral M&icine. Oral Pathology, 1980, 49, 2299232. Batsakis, J. G., Regezi, J. A., Luna, M. A. and el-Naggar, A., Histogenesis of salivary gland neoplasms: a postulate with prognostic implications. Journal qf Laryngology rend Otologj,, 1989, 103, 9399944. Eversole, L. K., Histogenetic classification of salivary gland tumors. Archives of Pathology, 1971, 92, 433-443. Dardick, I., Current status of histogenetic and morphogenetic concepts of salivary gland tumorigenesis. Critical Rrviebts in Owl Bmlogy und Mrdicine, 1993, 4, 639-677. Dardick, I., Byard, R. W. and Carnegie, J. A., A review of the proliferative capacity of major salivary glands and the relationship to current concepts of neoplasmia in salivary glands. Oral Surgrr~~. Oral Medicine. Oral Puthologx, 1990, 69, 53-67. Denny, l’. C., Chai, Y., Klauser, D. K. and Denny, I-‘. A., Parenchymal cell proliferation and mechanisms for maintenance of granular duct and acinar cell populations in adult male mouse submandibular gland. Anuromical Record, 1993, 235, 475-485. Redman, R. S., Myoepithelium of salivary glands. Microscopy Resrurch and Twhniyues, 1994, 27, 25-45. Burford-Mason, A. I’., Cummins, M. M., Brown, D. H., Mackay, A. J. and Dardick, I., Immunohistochemical analysis of the proliferative capacity of duct and acinar cells during ligation-induced atrophy and subsequent regeneration of rat parotid gland. Journul of’ Oral Parhology and Medicine, 1993, 22, 440-446. Luna, M. A., Batsakis, J. G., Tortoledo, M. E. and Deljunco, G. W., Carcinomas ex monomorphic adenoma of salivary glands. Journal of Lurvngolog~ and O/oh/g>,, 1989, 103, 756759. Bravo, R. and Macdonald-Bravo, H., Existence of two populations of cycliniproliferating cell nuclear antigen during the cell cycle. Association with DNA replication sites. Journul of Cell Bio&, 1987, 105, 154991554.
Acknowledgements-The authors would like to thank Miss Car-man Leung for her skillful technical assistance and Mr Johnny Leung for his help with the photography. This work was supported by a grant from the Committee on Research and Conference Grants of The University of Hong Kong.