Prognostic Significance of Tumor Cell Proliferation Rate ...clincancerres.aacrjournals.org/content/clincanres/2/1/147.full.pdf · Vol. 2, 147-154, January 1996 Clinical Cancer Research

Vol. 2, 147-154, January 1996 Clinical Cancer Research 147

Prognostic Significance of Tumor Cell Proliferation Rate as Determined by the MIB-1 Antibody in Breast Carcinoma: Its Relationship with Vimentin and p53 Protein I

W e n a n c j u s z D o m a g a l a , Maciej M a r k i e w s k i ,

B a r b a r a Harezga , A n d r z e j D u k o w i c z , and

M a r y O s b o r n z Department of Pathology, Medical Academy, 1 Unii Lubelskiej St., 71344 Szczecin [W. D., M. M.] Poland; Department of Oncology, Medical Academy, 4 Paderewskiegu St., 9359 Lodz [B. H., A. D.], Poland; and Department of Biochemistry, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, D-37077 Goettingen, Germany [M. O.]

A B S T R A C T The prognostic value of tumor cell proliferative activity

as measured by the MIB-1 monoclonal antibody in invasive ductal not otherwise specified breast carcinomas was determined for 186 patients, including 111 with no axillary node involvement. The MIB-1 antibody detects the Ki-67 antigen in microwave-processed paraffin sections of the formalin- fixed tumors. The mean MIB-1 score was 16% for all tumors, 16% for the node-negative group, and 15% for the node-positive group. In univariate survival analysis, the MIB-1 score (dichotomized, --10 v e r s u s -->10%) predicted overall 5-year survival in all of these groups. The mean MIB-1 score was significantly higher in vimentin- and p53 protein-positive tumors (P > 0.001) than in negative ones. The impact of vimentin expression and of p53 positivity on the prognostic significance of the tumor cell proliferation rate was assessed. Vimentin was associated significantly with poor 5-year survival in the entire cohort, and a particularly strong association was found in the node-negative group. p53 had a weak but statistically nonsignificant influence on survival. In a multivariate analysis using the Cox proportional hazards model, vimentin (P = 0.0002) was the only independent prognostic factor in node-negative patients. In contrast, the MIB-1 score (P = 0.009) was the only independent prognostic factor in the node-positive group. Therefore, node-negative patients with vimentin-positive tumors and node-positive patients with tumors with high proliferation rates might be appropriate candidates for early adjuvant chemotherapy.

I N T R O D U C T I O N

Axillary lymph node status is regarded as the single most important prognostic indicator in breast cancer. However, additional prognostic factors are needed, because the majority of

breast cancer patients currently diagnosed are axillary node negative, yet 25-35% will relapse (1, 2). Tumor cell proliferation as measured by methods such as tritiated thymidine uptake (3, 4), flow cytometry (5), mitotic counts (6), and assessment of proliferation-associated antigens (7-12) is regarded as a strong indicator of poor prognosis in breast cancer. However, it seems to be interrelated with other prognostic factors. For example, both vimentin positivity (13, 14) and p53 protein positivity (15, 16) hai, e been associated with an increased tumor cell proliferation rate, and all three factors seem to have prognostic significance in breast cancer (4, 16-21). From a practical point of view, it is not known which one or which combination of these markers should be used to obtain maximum reliable prognostic information.

Our interest in this study was 2-fold. First, we wanted to use the MIB-1 antibody in a prognostic study in invasive ductal NOS 3 breast cancer, with the emphasis on node-negative patients. The MIB-1 antibody was generated against recombinant parts of the Ki-67 antigen (22). Ki-67 is a nuclear antigen present in G 1, S 1 G 2, and M phases of the cell cycle but not in resting cells (23). The MIB-1 antibody is particularly useful in the evaluation of the growth fraction of tumor cells, because it immunostains formalin-fixed, paraffin-embedded sections, provided they have been heated in a microwave oven (22). A significant correlation was found between the MIB-1 labeling index and that of Ki-67 on frozen sections (24) and between the MIB-1 and mitotic tumor indices and the Scarff-Bloom-Rich- ardson histological grade of breast carcinomas (25). Our second aim was to determine whether the proliferation rate, vimentin positivity, and p53 positivity of tumor cells are dependent or independent prognostic indicators. Although vimentin (13, 14) and p53 protein (15, 16) are related to the cell proliferation rate, the prognostic effect of p53 protein seems to be related mainly to the proliferation rate of tumor cells (16), whereas vimentin is also associated with increased invasiveness of breast cancer cell lines (26). This is also reflected in vivo, in which, particularly in node-negative patients, those with vimentin-positive tumors have a significantly higher probability of visceral relapse (19). A similar association between p53 status and visceral metastases was not found (19). Therefore, we hypothesized that in node- negative patients, the prognostic effect of vimentin, but not that of p53 protein, might be much stronger than that of the cell proliferation rate. Therefore, we assessed tumor cell proliferation by immunohistochemistry using the novel MIB-1 monoclonal antibody on microwave-treated paraffin sections of formalin-fixed, invasive, ductal NOS breast carcinomas in which the

Received 5/30/95; revised 8/10/95; accepted 8/17/95. 1 This work was supported by European Union Grant CT93-0234. 2 To whom requests for reprints should be addressed. Phone: +49-551- 201-1486; Fax: +49-551-201-1578.

3 The abbreviations used are: NOS, not otherwise specified; ER, estrogen receptor.

Research. on June 21, 2019. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

148 MIB-1, Vimentin, p53, and Prognosis

vimentin and p53 protein statuses of the same tumors were already known. The results were compared with survival curves for a 5-year period.

M A T E R I A L S AND M E T H O D S Patients. Formalin-fixed and paraffin-embedded tumor

samples from 186 unselected female patients with primary invasive ductal NOS breast cancer who underwent mastectomy with lymph node dissection between 1980 and 1987 were examined. These were retrieved from the files of the Department of Oncology, Medical Academy of Lodz. Clinical follow-up was available for at least 60 months. Histological typing was performed as described by Millis and Girling (27), and histological grading was according to Bloom and Richardson (28). The computerized database contained the age of the patient, the number of positive axillary lymph nodes, the size of the tumor, histological type and grade, the stage of the disease at diagnosis, the treatment protocol, and the date of the last checkup or death. In addition, vimentin and p53 protein statuses of the tumor were also on file. Some patients analyzed in this study were included in previous studies on vimentin expression and survival (18) and on p53, vimentin, survival, and sites of metastases (19). There were 111 axillary lymph node-negative and 75 lymph node- positive patients. Lymph node-negative patients underwent either only mastectomy with removal of the axillary lymph nodes (45 patients) or mastectomy with removal of the nodes followed by radiation therapy (19 patients), chemotherapy (22 patients), hormone therapy (3 patients), radiation therapy and chemotherapy (7 patients), or combinations of radiation therapy, chemotherapy, and hormone therapy (11 patients). Four patients were treated by quadrantectomy with removal of axillary nodes followed by radiation therapy. Lymph node-positive patients underwent either only mastectomy with removal of the axillary lymph nodes (6 patients) or mastectomy with removal of the nodes followed by radiation therapy (22 patients), chemotherapy (4 patients), hormone therapy (2 patients), radiation therapy and chemotherapy (22 patients), or combinations of radiation therapy, chemotherapy, and hormone therapy (18 patients). One patient underwent quadrantectomy with removal of axillary nodes followed by radiation therapy. As in all studies of this type, the use and kind of adjuvant therapy have to be taken into account in interpreting the results.

Immunohistochemistry. Sections were deparaffinized, immersed in citrate buffer, and heated in a microwave oven according to Cattoretti et al. (22). The sections were incubated with MIB-1 monoclonal antibody (Dianova, Hamburg, Ger- many) diluted 1:30 for 45 min at room temperature. Sections were washed and reacted with biotinylated rabbit antimouse antibody and streptavidin peroxidase (Histostain-SP kit; Zymed Laboratories, Inc., San Francisco, CA). The sections were washed and then lightly counterstained with hematoxylin. Areas with the highest numbers of MIB-l-positive cells were identi- fied by scanning sections at low magnification. These areas occurred mostly at the periphery of tumors. Counts were started in these areas, and then additional contiguous fields were selected randomly until at least 500 tumor cells were counted using a 10 X 10 square ocular grid and ×400 magnification. The results were scored as the percentage of tumor cell nuclei

reactions with the MIB-1 antibody. The MIB-1 score was assessed independently by two pathologists (W. D. and M. M.) without knowledge of other data of the patient. The results were combined, and the mean was used. The MIB-1 scores determined by W. D. correlated highly with those determined by M. M. (linear regression analysis, r = 0.85; P > 0.0001; 95% confidence interval).

Statistical Analysis. The association between MIB-1 score and vimentin and p53 statuses was assessed by the Student t test. Univariate survival analysis was based on the Kaplan- Meier product limit estimates of survival distribution. Differ- ences between survival curves were tested using the ×2. The Cox proportional hazards model was used in multivariate analysis of the survival data and for assessment of relative risk ratios. All data were analyzed with the SAS computer program PHREG (version 6.08; SAS Institute, Inc., Cary, NC). The multivariate analysis included only those cases in which complete sets of data were available, i.e., 177 for all, 106 for node-negative, and 71 for node-positive patients.

RESULTS

MIB-1 Staining in Breast Carcinomas All tumor cells in which red color was found in the

nuclei or nucleoli were scored as positive regardless of staining intensity (which in general was strong). MIB-1 positivity was found exclusively in the nuclei or nucleoli, with some variation in intensity in different cells within the same section. Mitotic figures were stained strongly in all cases. Al- though the distribution of positive nuclei was heterogenous, immunoreactive tumor cells could be distinguished easily from nonreactive tumor cells and benign cells. All 186 tumors were positive for MIB-1, but considerable intertumoral and intratumoral variations of MIB-1 scores were found. The mean MIB-1 score for the whole series of tumors was 16% (SD, 14%; range, 0.1-60.4%; median, 12%); for tumors from axillary lymph node-negative patients, the mean score was 17% (SD, 14%; range, 0.1-52.7%; median, 14%); and for tumors from node-positive patients, the mean score was 15% (SD, 15%; range, 0.4-60.5%; median, 10%). The mean MIB-1 score in our study, i.e., 16.2%, falls within the range of values reported with the use of the MIB-1 antibody (15.2- 24%; Refs. 24, 25, and 29) as well as the Ki-67 antibody on frozen sections of breast carcinomas, i.e., between 7.2 and 22% (12, 30-32). The considerable variability in MIB-1 and Ki-67 scores in the different studies has been attributed to the use of different evaluation methods and to the inherent variability between individual patients and also between distinct high-power fields from the same tissue section (12). Never- theless, Wintzer et al. (12) found a high degree of correlation between results obtained by two different counting methods, i.e., 10 random high-power counts versus the area with the highest labeling density count.

To correlate the MIB-1 score with vimentin and p53 statuses and to analyze the relationship to survival, a 10% cutoff level for the MIB-1 score was chosen to separate the high-risk group from the low-risk breast cancer. In previous survival studies using the Ki-67 antibody, various cutoff levels were used, e.g., 7.5% (9), 9% (8), 10% (10), 13% (11),



Clinical Cancer Research 149

A 100

60

A- 4O

0 0

B 100

.-_- 60 t-J Q ~. 40

20

0 0

C 100

80

-- 60

o

A- 4O

20

all

" % ' - ~ , ~ ~<10% (n=80)

X 2= 8.6 p< 0.01

I I

1'o 3'0 4'0 50 Survival (month)

node negative

MIBl~<10% (n--45~

MIBI>I 0% ( n ~

X 2= 3.9 p< 0.05

I I

1'0 210 3'0 40 50 Survival (month)

node positive

~ 1 ~< 10% (n=35)

MIB1>10% (n:40) \ \

\

X2= 6.7 p< 0.01

00 1'0 210 310 410 50

Fig. 1 Effect of low (-----10%) versus high (>10%) MIB-1 score on overall survival of all (A), node-negative (B), and node-positive (C) patients with invasive ductal NOS breast carcinoma, n, number of patients at risk. Patients with tumors with low MIB-1 scores have a significantly better 5-year survival.

20% (33), 5 and 20% (7), and 16 and 12% (12). In some reports, the median was used as the cutoff level (8, 12). Although we did not examine the effect of varying the MIB-1

cutoff level in this study, the choice of cutoff level for Ki-67

is probably not too critical, as shown by Wintzer et al. (12) ,

who reported significant differences of Kaplan-Meier sur-

vival curves irrespective of whether 12% (i .e. , the median value of the random count; P < 0.05) or 16% (i .e. , the median value of the highest labeling density; P < 0.001) was used as the cutoff level for Ki67. In addition, the 10% cutoff level chosen in this study shows a clear correlation with survival (Fig. 1; also see below).

Table 1 MIB-1 score in relation to vimentin and p53 status in patients with invasive ductal NOS breast carcinomas

MIB-1 n (%; mean + SD) P

All patients 186 16.18 _+ 14.19 Vimentin

Negative 145 13.26 _+ 11.69 Positive 41 26.52 _+ 17.30 <0.001

p53 Negative 129 12.44 + 10.81 Positive 50 26.62 _+ 16.89 <0.001

Node negative 111 16.75 + 13.86 Vimentin

Negative 80 12.99 _+ 11.20 Positive 31 26.41 _+ 15.49 <0.001

p53 Negative 76 13.19 + 11.44 Positive 32 25.07 _+ 15.67 <0.001

Node positive 75 15.35 + 14.73 Vimentin

Negative 65 13.57 _+ 12.35 Positive 10 26.84 +_ 23.00 <0.01

p53 Negative 53 11.35 _+ 9.84 Positive 18 29.36 +_ 19.03 <0.001

Correlation of MIB-1 Scores with Vimentin and p53 Protein

The mean MIB-1 score was significantly higher in vimentin-positive than in vimentin-negative tumors (P < 0.001 for all and for axillary node negative-patients; P < 0.01 for node- positive patients), p53-positive tumors also had a significantly

higher mean MIB-1 score than those cancers that did not express this protein (P < 0.001 for all and for node-negative and node-positive patients; Table 1).

Prognostic Significance of MIB-1 Score with Respect to Vimentin and p53 Protein Status

All Patients. One hundred sixteen (61%) patients survived 5 years. The mean follow-up time for these patients was 88 months. In the univariate survival analysis, the MIB-1 score (dichotomized, -<10 versus >10%) and vimentin predicted 5-year survival (Table 2). Classical prognostic factors such as lymph node status, tumor size, and histological grade were also strong indicators of prognosis (Table 2). Survival curves also show that a high MIB-1 score is associated with poor survival of patients at 5 years in the entire cohort (P < 0.01), in node- negative patients (P < 0.05), and in node-positive patients (P <

0.01; Fig. 1, A-C). When all the factors listed in Table 2 were tested in the Cox

proportional hazards model for their relationship with overall survival, the four factors selected by the model as independent

predictors of survival at 5 years were node status, tumor size, MIB-1 score (dichotomized, -<10 versus >10%) and vimentin

status (Table 3). Axil lary Node-negative Patients. The results of univa-

riate survival analysis indicate that vimentin status, tumor size, and MIB-1 score (dichotomized) predict 5-year survival (Table 2). Survival curves for node-negative patients also show that a high MIB-1 score was associated significantly with poor sur-




Table 2 Prognostic factors predicting 5-year survival of patients

5-yr survival

n n % P

All patients MIB-1

--<10% 80 59 >10% 106 57

Vimentin Negative 145 96 Positive 41 20

p53 Negative 129 82 Positive 50 29

Grade I + II 92 66 III 94 50

Size (mm) --<30 57 48 >30 127 68

Age (yr) <50 60 41 >--50 126 75

Nodes Negative 111 80 Positive 75 36

Node-negative patients MIB- 1

-< 10% 45 37 > 10% 66 43



Grade I + II 58 45 III 53 35

Size (nma) -<30 44 39 >30 65 41

Age (yr) <50 35 27 ->50 76 53

Node-positive patients MIB-1

-< 10% 35 22 >10% 40 14



Grade I + II 34 21 III 41 15

Size (mm) -<30 13 9 >30 62 27

Age (yr) <50 25 14 >-50 50 22

74 54 0.005

66 49 0.04

64 58 NS ~

72 53 0.009

84 54 0.00006

68 59 NS

72 48 0.0008

82 65 0.049

30 52 0.002

77 59 NS

78 66 NS

89 63 0.003

77 70 NS

63 35 0.015

49 40 NS

45 56 NS

62 37 0.029

69 44 NS

56 44 NS

NS, not significant.

Table 3 Relationship of prognostic factors to reduced 5-year overall survival

Results of the Cox proportional hazards model analysis for all (n = 177) and for node-negative (n = 106) and node-positive (n = 71) patients. Other factors tested in the Cox model were histologic grade, p53 status, and age of the patient.

All patients Node negative Node positive

RR ~ P RR P RR P

MIB-1 2.06 0.005 NS 2.43 0.009 Vimentin 1.88 0.016 3.55 0.0002 NS Lymph nodes 2.40 0.0003 NA NA Tumor size 2.28 0.007 NS NS

RR, relative risk of death; NS, not significant; NA, not applicable.

vival at 5 years (P < 0.05; Fig. 1B). Sixty-five % of patients with tumors with MIB-1 scores > 10% survived compared with 82% with MIB-1 scores -<10%.

When lymph node-negative patients were first divided by vimentin status, and then survival curves were plotted by MIB-1

score, there was no significant difference in survival for patients with tumors with high versus low proliferative activity either in

the vimentin-positive (Fig. 2a, curve B versus C) or vimentin- negative subgroup, although there was a trend in the latter for highly proliferating tumors to indicate worse prognoses (curve D versus A). However, if node-negative patients were first

divided according to the MIB-1 scores (high versus low) of their tumors, and then survival curves were plotted by vimentin status, vimentin expression stratified the patients significantly (curves C versus D and A versus B). Patients with vimentin- positive tumors had worse survival irrespective of low (curve B, P < 0.01) or high (curve C, P < 0.05) proliferation rate. Thus, analysis of survival curves indicates that in node-negative patients, vimentin is a stronger indicator of poor prognosis than proliferative activity of tumor cells. At 5 years, only about 50% of patients with vimentin-positive tumors survive, irrespective of whether the patients had tumors with high or low MIB-1 scores (curves C and B). In contrast, the proliferation rate only seems to play a role in the survival of node-negative patients with vimentin-negative tumors. Here, about 75% of patients with fast-proliferating tumors, i.e., MIB-1 scores > 10%, survive 5 years, compared with 90% of those with MIB-1 scores -< 10%, although the difference does not reach statistical significance (×2 = 2.157).

Node-negative patients with p53-negative and slowly proliferating (MIB-1 score, --<10%) tumors had better prognoses

than those with high MIB-1 scores (Fig. 2b, curve A versus D)

but the trend was not significant (×2 = 3.146). If node-negative

patients with p53-positive tumors were divided by MIB-1 score (curve B versus C), survival curves differed slightly, in that

patients with highly proliferating tumors had worse 5-year survival, but the difference did not reach statistical significance. When the node-negative patients were divided by MIB-1 score and p53 status (curves C versus D and A versus B), patients with p53-positive tumors had worse survival in both subgroups, but, again, the trends did not reach statistical significance.

In the Cox proportional hazards model, of all the factors listed in Table 2, only vimentin status was selected as an independent predictor of survival in node-negative patients (Table 3).




a oo

80

60

o ~_ 40

20

V-MIBI~<IO% n=39 ~ - - " - " - " " - - M

. . . . . . . V-M~l~O~.n=41 %_

" - - "~--" "~ V+MIBI>IO% \.n:25 V+MIB1~<10% n=6

I I

°o 2o 3'o ,'o 5o Survival (months)

a 80

"=- 60 0 ~_ 40

~';'~,1"- ~'~"L~ ~ - ~ V-MIBI~<IO% n=31 = . ° . . . • • = . . . . . ,.= " ~

~=-4 . - . - - . a i. ~ . . - "4 v÷MIBl>l°% n:6

I V-MIBI>10% n=34\

0 0 llO

" V+MIBI~<IO%n=4 D ' ' ' ' ' ' ' ' ' ' ' ' ' . . . . . . B

I / I I

20 30 40 50 Survival (months)

b OO 80

~ 60 0 ~- 4O

20

p53-MIB1~<10% n=38 4- "

" "

053+MIB1>10% n=26

°o ;o 2'o 3'o .'o Survival (months)

Fig. 2 Kaplan-Meier overall survival curves by vimentin (a) or p53 (b) status and MIB-1 score (high versus low) in node-negative patients with invasive NOS breast carcinoma. Patients with vimentin-positive tumors (a, curves B and C) have worse survival irrespective of low (curve B) or high (curve C) proliferation rate (B versus A, P < 0.01; C versus D, P < 0.05). In the vimentin-negative subgroup (curves A and D), there is a trend (A versus D, not significant) for highly proliferating tumors to indicate worse prognosis. Patients with p53-positive tumors (b, curves B and C) have slightly worse survival, but the trends did not reach statistical significance (A versus D, B versus C, A versus B, and D versus C, not significant).

Axillary Node-positive Patients. The univariate survival analysis shows that a high MIB-1 score and high histological grade are associated significantly with 5-year survival of patients (Table 2). Survival curves also show a significant prognostic difference when node-positive patients were divided according to proliferation rate (P < 0.01; Fig. 1C).

When this group of patients was divided by MIB-1 score and survival curves plotted by vimentin status, a significant negative correlation of vimentin expression and survival was seen only in patients with slowly proliferating tumors (P < 0.05, Fig. 3a, curve A ver sus B; compare with curve C ver sus D) .

When the node-positive patients were divided by vimentin status and survival curves were plotted by proliferation rate, there was a negative prognostic significance of high MIB-1 score for patients with vimentin-negative tumors (P < 0.01, curve A ver sus D) . An even stronger poor prognostic effect of high MIB-1 score was seen in node-positive patients with p53-negative tumors (Fig. 3b, curve A ver sus D; P < 0.001). When node-positive patients were divided by MIB-1 score and survival curves were plotted by p53 status, the trend emerged for

blOO 80

\ ~ \ o " ~ "~ , ' ' ' * , .... ,..p53-MIB1~<10 '/o n=27 " :

~ 60 p53+MIB1~<lO% n=5

\ p53+MIB1>10% C Y 40 - \

p~3-MJBl>lO;/oT--~8\ 2O D

I I

°o ;o 20 3o ;o so Survival (months)

Fig. 3 Overall survival of patients with node-positive invasive ductal NOS breast carcinoma. Kaplan-Meier survival curves by vimentin (a) or p53 (b) and MIB-1 score (high versus low). Note a negative prognostic effect of high MIB-1 score for patients with vimentin-negative (a, curve A versus D, P < 0.01) or p53-negative (b, curve A versus D, P < 0.001) tumors. A significant negative correlation of vimentin expression and survival is seen only in patients with slowly proliferating tumors (a, curve A versus B, P < 0.05; C versus D and C versus B, not significant).

worse survival of patients with p53 positive, slowly proliferating tumors and better survival of patients with p53-positive, rapidly proliferating tumors; however, both trends did not reach statistical significance (curves A ver sus B and C versus D) .

A multivariate analysis using the Cox model showed that in axillary node-positive patients, only the MIB-1 score was selected as an independent prognostic indicator (Table 3).

D I S C U S S I O N

Our data show that the tumor cell proliferation rate as determined by MIB-1 activity on paraffin sections is associated significantly with 5-year overall survival in the entire cohort and in the node-negative and node-positive subsets of patients. Our results confirm reports that demonstrated a significant correlation between the proliferation rate of cancer cells as determined by Ki-67 immunostaining in frozen sections (7, 9-12, 33-35) or by the Ki-S1 antibody in paraffin-embedded tissue (36) and survival of breast cancer patients. Most of these studies report on short-term, disease-free survival. Of 12 studies published to date on Ki-67 and prognosis, only two examined more than 100 node-negative patients. Although Railo et al. (10) and Gasparini et al. (9) examined 196 and 164 such patients, respectively, they reported on shorter survival (4 and 3 years, respectively). There-




fore, our group of 111 node-negative patients with 5-year survival, although relatively small, is comparable in size to that used in previous studies with Ki-67 antibodies. The only report on Ki-67 and 5-year survival (11) is based on 42 node-negative patients, whereas a very recent article also examines MIB-1 staining and long-term survival (37). Our data are also in agreement with studies documenting the significant prognostic value of the proliferation rate measured by thymidine labeling (3, 4) or flow cytometry (5). However, because of the relatively small size of the study population, and because the results were not validated in an independent set of patients, our results should be viewed as a pilot study.

Using a multivariate Cox analysis of survival we found that the tumor cell proliferation rate, i.e., MIB-1 score, provided independent significant prognostic information in the entire cohort of patients as well as in the node-positive subset. In the latter group, it was the only independent predictor of overall survival at 5 years.

In contrast, for node-negative patients, the MIB-1 score was not an independent predictor of overall survival. The proliferation rate of tumor cells ceased to be a significant factor after vimentin was taken into account in the Kaplan-Meier survival curves (Fig. 2a, curves B versus C and A versus D) and did not enter the Cox model. Vimentin status was the only independent predictor of overall survival at 5 years in this group.

Attempts to use other proliferation-associated factors, such as the ER or epidermal growth factor receptor, to substratify node-negative patients have been relatively unsuccessful. For instance, Railo et al. (10) saw no difference in disease-free survival between node-negative patients in the categories ER negative and Ki-67 >10% and ER positive and Ki-67 --<10%, although they were able to demonstrate a 35% difference in 4-year survival (P < 0.005) between stage II patients in these two groups. Toi et al. (38) also were unable to stratify node- negative patients into high- and low-risk groups using the epidermal growth factor receptor, another marker linked closely to cell proliferation.

Attempts to stratify node-negative patients using p53 have yielded conflicting results. Using immunohistochemistry, a significant association with shortened survival (16, 17, 21), a small but statistically nonsignificant negative effect (19, 39), and a positive effect (40) of p53 positivity on survival of node-negative patients have been reported. No significant difference in survival between patients with mutant and wild-type p53 tumors has been found (41). In this report, we show that for our series of breast carcinomas, p53 positivity is not an independent predictor of 5-year survival when tested against the proliferation rate of tumor cells and vimentin expression. This is in agreement with the results of Isola et al. (16), who showed that p53 had independent prognostic value in a multivariate analysis only in the absence of the S-phase fraction. If the S-phase fraction was included in the analysis, only S-phase and tumor size emerged as independent predictors of overall survival. However, Allred et al. (17) reported that positive p53 and S-phase were independent predictors of reduced disease-free survival despite the strong direct correlation between p53 positivity and tumor proliferation rate. p53 protein is associated strongly with proliferation but not necessarily with invasion. Recently, Hsiao et al.

(42) have shown that in acute lymphocytic leukemias, some

specific p53 mutations were associated with increased invasiveness, and others were not. Thus, tumors determined by immunohistochemistry to be p53 positive are most likely a heterogenous mixture of growth-promoting and ineffective mutations, and, thus, it may be important to determine the precise p53

mutations present in each tumor rather than only p53 positivity or negativity.

We have shown that vimentin expression is a strong indicator of poor prognosis in node-negative patients as well as in the whole cohort of patients with invasive ductal NOS breast carcinoma (18). The prognostic value of vimentin in the whole cohort of patients has been confirmed in two reports (20, 43).

Why is the proliferation rate of tumor cells a less useful prognostic indicator than vimentin positivity in node-negative patients? This may be due to the association of vimentin expression not only with proliferation but also with the invasiveness of tumor cells both in vitro (26, 44-47) and in vivo (19). Poor prognoses in a subset of node-negative patients depends on hematogenous micrometastases, particularly on visceral ones i.e., to the brain, lungs, and liver, that were not detected at the time of the first operations. During tumor progression, breast carcinoma cells probably go through several stages in a complex sequence of events that lead ultimately to metastasis. One event leading to metastasis also may induce vimentin expression (48), although vimentin is probably associated with only one of several different pathways leading to the metastatic phenotype, because vimentin-negative tumors also metastasize. Direct sup- port for the idea that vimentin-positive tumors may be more invasive is given by data showing that in node-negative patients, vimentin-positive primary invasive ductal NOS breast carcinomas metastasized 3.5 times as often to the lung, liver, and brain as did vimentin-negative primary carcinomas (19). Thus, vimentin expression in breast cancer might be associated with the expression of genes involved in the metastatic process. How- ever, the molecular basis of the selective advantage offered to tumor cells by events that are associated with vimentin expression is not known. Although an increased proliferation rate of tumor cells also might contribute indirectly to metastasis, e.g.,

because of evolving clones that have a higher probability of acquiring a metastatic phenotype, our data suggest that this is not sufficient for proliferation to become a prognostic factor independent of vimentin expression in node-negative patients.

The node-positive patients are a mixed group of those who have only regional and those who may have widespread although undetected dissemination of tumor cells at the time of surgery. One may expect that in this group, the higher the proliferation rate of tumor cells, the worse the prognosis. Our results show that, indeed, it is the node-positive group in which the determination of tumor cell proliferation rate is important, because tumors with high proliferation rates signify poor prognoses irrespective of vimentin or p53 status.

The proliferation rate, vimentin status, p53 status, histological grade, and size of tumors were tested in a multivariate Cox analysis for their relationship with 5-year overall survival of patients with invasive ductal NOS breast carcinoma. Although the study population is relatively small (186 patients), the results identify the vimentin status in node-negative patients and the proliferation rate of tumor cells in node-positive patients as important prognostic factors.




REFERENCES 1. Adair, F., Berg, J., Joubert, L., and Robbins, G. F. Long-term follow up of breast cancer patients: the 30-year report. Cancer (Phila.), 33: 1145-1150, 1974.

2. DeVita, V. T., Hellman, S., and Rosenberg, S. A. Cancer Principles and Practice of Oncology. Philadelphia: J. B. Lippincott Co., 1989.

3. Silvestrini, R., Daidone, M. G., and Gasparini, G. Cell cycle kinetics as a prognostic marker in node-negative breast cancer. Cancer (Phila.), 56: 1982-1987, 1985.

4. Tubiana, M., Pejovic, M. H., Chavaudra, N., Contesso, G., and Malaise, E. P. The long-term prognostic significance of the thymidine labelling index in breast cancer. Int. J. Cancer, 33: 441-445, 1984.

5. Kallioniemi, O. P., Hietanen, T., Mattila, J., Lehtinen, M., Lauslahti, K., and Koivula, T. Aneuploid DNA content and high S phase fraction of tumour cells are related to poor prognosis in patients with primary breast cancer. Eur J. Cancer & Clin. Oncol., 23: 277-282, 1987.

6. Laroye, G. J., and Minkin, S. The impact of mitotic index on predicting outcome in breast carcinoma: a comparison of different counting methods in patients with different lymph node status. Mod. Pathol., 4: 456-460, 1991.

7. Bouzubar, N., Walker, K. J., Griffiths, K., Ellis, I. O., Elston, C. W., Robertson, J. F. R., Blarney, W. W., and Nicholson, R. I. Ki-67 immunostaining in primary breast cancer: pathological and clinical associations. Br. J. Cancer, 59: 943-947, 1989.

8. Gaglia, P., Bernadi, A., Venesio, T., Caldarola, B., Lauro, D., Cappa, A. P., Calderini, P., and Liscia, D. S. Cell proliferation of breast cancer evaluated by anti-BrdU and anti-Ki-67 antibodies: its prognostic value on short-term recurrences. Eur. J. Cancer, 29A: 1509-1513, 1993.

9. Gasparini, G., Bevilacqua, P., Pozza, F., Meli, S., Boracchi, P., Marubini, E., and Sainsbury, J. R. C. Value of epidermal growth factor receptor status compared with growth fraction and other factors for prognosis in early breast cancer. Br. J. Cancer, 66: 970-976, 1992.

10. Railo, M., Nordling, S., von Boguslawsky, K., Leivonen, M., Kyll6nen, L., and von Smitten, K. Prognostic value of Ki-67 immuno- labelling in primary operable breast cancer. Br. J. Cancer, 68: 579-583, 1993.

11. Sahin, A. A., Ro, J., Ro, J. Y., Blick, M. B., el-Naggar, A. K., Ordonez, N. G., Fritsche, H. A., Smith, T. L., Hortobagyi, G. N., and Ayala, A. G. Ki-67 immunostaining in node-negative stage I/II breast carcinoma. Significant correlation with prognosis. Cancer (Phila.), 68: 549-557, 1991.

12. Wintzer, H. O., Zipfel, I., Schulte-M6ntung, J., Hellerich, U., and von Kleist, S. Ki-67 immunostaining in human breast tumors and its relationship to prognosis. Cancer (Phila.), 67: 421-428, 1991.

13. Domagala, W., Lasota, J., Bartkowiak, J., Weber, K., and Osborn, M. Vimentin is preferentially expressed in human breast carcinomas with low estrogen receptor and high Ki-67 growth fraction. Am. J. Pathol., 136: 219-227, 1990.

14. Raymond, W., and Leong, A-Y. Vimentin--a new prognostic pa- rameter in breast carcinoma? J. Pathol., 158: 107-114, 1989.

15. Cattoretti, G., Rilke, F., Andreola, S., and D'Amato, D. P53 expression in breast cancer. Int. J. Cancer, 41: 178-183, 1988.

16. Isola, J., Visakorpi, T., Holli, K., and Kallioniemi, O-P. Association of overexpression of tumor suppressor protein p53 with rapid cell proliferation and poor prognosis in node-negative breast cancer patients. J. Natl. Cancer Inst., 84:1109-1114, 1992.

17. Allred, D., Clark, G., Elledge, R., Fuqua, S., Brown, R., Chamness, G., Osborne, C., and McGuire, W. Association of p53 protein expression with tumor cell proliferation rate and clinical outcome in node- negative breast cancer. J. Natl. Cancer Inst., 85: 200-206, 1993.

18. Domagala, W., Lasota, J., Dukowicz, A., Markiewski, M., Striker, G., Weber, K., and Osborn, M. Vimentin expression appears to be associated with poor prognosis in node-negative ductal NOS breast carcinomas. Am J Pathol., 137: 1299-1304, 1990.

19. Domagala, W., Striker, G., Szadowska, A., Dukowicz, A., Harezga, B., and Osborn, M. p53 protein and vimentin in invasive ductal NOS

breast carcinoma--relationship with survival and sites of metastases. Eur. J. Cancer, 30A: 1527-1534, 1994.

20. Russo, A., Bazan, V., Morello, V., Tralongo, V., Nagar, C., Nuara, R., Dardanoni, G., Bazan, P., and Tomasino, R. M. Vimentin expression, proliferating cell nuclear antigen and flow cytometric factors. Prognostic role in breast cancer. Anal. Quant. Cytol. Histol., 16: 365- 174, 1994.

21. Thor, A., Moore, D. I., Edgerton, S., Kawasaki, E., Reihsaus, E., Lynch, H., Marcus, J., Schwartz, L., Chen, L-C., Mayall, B., and Smith, H. Accumulation of p53 tumor suppressor gene protein: an independent marker of prognosis in breast cancers. J. Natl. Cancer Inst., 84: 845- 855, 1992.

22. Cattoretti, G., Becker, M. H. G., Key, G., Duchrow, M., Schltiter, C., Galle, J., and Gerdes, J. Monoclonal antibodies against recombinant parts of the Ki-67 antigen (MIB 1 and MIB3) detect proliferating cells in microwave-processed formalin-fixed paraffin sections. J. Pathol., 168: 357-363, 1992.

23. Gerdes, J., Lemke, H., Baisch, H., Wacher, H. H., Schwab, U., and Stein, H. Cell cycle analysis of a cell proliferation-associated nuclear antigen defined by the monoclonal antibody Ki-67. J. Immunol., 133: 1710-1715, 1984.

24. Barbareschi, M., Girlando, S., Mauri, F. M., Forti, S., Eccher, C., Mauri, F. A., Togni, R., Palma, P., and Doglioni, C. Quantitative growth fraction evaluation MIB1 and Ki-67 antibodies in breast carcinomas. Am. J. Clin. Pathol., 102: 171-175, 1994.

25. Weidner, N., Moore, D. H., II, and Vartanian, R. Correlation of Ki-67 antigen expression with mitotic figure index and tumor grade in breast carcinomas using the novel "paraffin"-reactive MIB 1 antibody. Hum. Pathol., 25: 337-342, 1994.

26. Thompson, E., Soonmyoung, P., Brunner, N., Sommers, C., Zug- maier, G., Clarke, R., Shima, T., Torri, J., Donahue, S., Lippman, M., Martin, G., and Dickson, R. B. Association of increased basement membrane invasiveness with absence of estrogen receptor and expression of vimentin in human breast cancer cell lines. J. Cell. Physiol., 150: 534-544, 1992.

27. Millis, R., and Girling, A. The breast. In: S. S. Sternberg (ed.), Diagnostic Surgical Pathology, pp. 253-313. New York: Raven Press, Ltd., 1989.

28. Bloom, H., and Richardson, W. Histologic grading and prognosis in breast cancer. Br. J. Cancer, 11: 359-377, 1957.

29. Rajakariar, R., and Walker, R. A. Pathological and biological fea- tures of mamographically detected invasive breast carcinomas. Br. J. Cancer, 71: 150-154, 1995.

30. Gerdes, J., Lell6, R. J., Pickartz, H., Heidenreich, W., Schwarting, R., Kurtsiefer, L., Stanch, G., and Stein, H. Growth fractions in breast cancers determined in situ with monoclonal antibody Ki-67. J. Clin. Pathol., 39: 977-980, 1986.

31. McGurrin, J. F., Doria, M. I., Dawson, P. J., Karrison, T., Stein, H. O., and Franklin, W. A. Assessment of tumor cell kinetics by immunohistochemistry in carcinoma of breast. Cancer (Phila.), 59: 1744-1750, 1987.

32. Wrba, F., Chott, F., Reiner, A., Reiner, G., Markis-Ritzinger, E., and Holzner, J. H. Ki-67 immunoreactivity in breast carcinomas in relation to transferrin receptor expression, estrogen receptor status and morphological criteria. An immunohistochemical study. Oncology (Basel), 46: 255-259, 1989.

33. Veronese, S. M., Gambacorta, M., Gottardi, O., Scanzi, F., Ferrari, M., and Lampertico, P. Proliferation index as a prognostic marker in breast cancer. Cancer (Phila.), 71: 3926-3931, 1993.

34. Locker, A. P., Birrell, K., Bell, J. A., Nicholson, R. I., Elston, C. W., Blamey, W. R., and Ellis, I. O. Ki-67 immunoreactivity in breast carcinoma: relationships to prognostic variables and short term survival. Eur. J. Surg. Oncol., 18: 224-229, 1992. 35. Weikel, W., Beck, T., Mitze, M., and Knapstein, P. G. Immuno- histochemical evaluation of growth fraction in human breast cancers using monoclonal antibody Ki-67. Breast Cancer Res. Treat., 18: 149- 154, 1991.




36. Kreipe, H., Alm, P., Olsson, H., Hauberg, M., Fischer, L., and Parwaresch, R. Prognostic significance of a formalin-resistant nuclear proliferation antigen in mammary carcinomas as determined by the monoclonal antibody Ki-S1. Am. J. Pathol., 142: 651-657, 1993. 37. Pinder, S. E., Wencyk, P., Sibbering, D. M., Bell, J. A., Elston, C. W., Nicholson, R., Robertson, J. F. R., Blamey, R. W., and Ellis, I. O. Assessment of the new proliferation marker MIB 1 in breast carcinoma using image analysis--associations with other prognostic factors and survival. Br. J. Cancer, 71: 146-149, 1995.

38. Toi, M., Osaki, Y., Yamada, H., and Toge, T. Epidermal growth factor receptor expression as a prognostic indicator in breast cancer. Eur. J. Cancer, 27: 977-980, 1991.

39. Bosari, S., Lee, A. K. C., ViNe, G., Heatley, G. J., and Coggi, G. Abnormal p53 immunoreactivity and prognosis in node-negative breast carcinomas with long-term follow-up. Virchows Arch. A. Pathol. Anat. Hist., 421: 291-295, 1992.

40. Lipponen, P., Aaltomaa, S., Syrj~inen, S., and Syrj~inen, K. p53 protein expression in breast cancer as related to histopathological char- acteristics and prognosis. Int. J. Cancer, 55: 51-56, 1994.

41. Caleffi, M., Teague, M. W., Jensen, R. A., Vnencak-Jones, C. L., Dupont, D., and Parl, F. F. p53 gene mutations and steroid receptor in breast cancer. Cancer (Phila.), 73: 2147-2156, 1994.

42. Hsiao, M., Low, J., Dorn, E., Ku, D., Pattengale, P., Yeargin, J., and Haas, M. Gain-of-function mutations of the p53 gene induce lympho- hematopoietic metastatic potential and tissue invasiveness. Am. J. Pathol., 145: 702-714, 1994.

43. Tomasino, R. M., Russo, A., Bazan, V., Morello, V., Tralongo, V., Nagar, C., Salvato, M., Taormina, P., Marrazzo, A., and Nuara, R. Evaluation of integrated morpho-biological indicators in breast cancer. In Vivo (Athens), 7: 601-605, 1993.

44. Thompson, E. W., Torri, J., Sabol, M., Sommers, C. L., Byers, S. W., Valverius, E. M., Martin, G. R., Lippman, M. E., Stampfer, M. R., and Dickson, R. B. Oncogene-induced basement membrane invasiveness in human mammary epithelial cells. Clin. & Exp. Metas- tasis, 12: 181-194, 1994.

45. Sommers, C. L., Thompson, E. W., Torri, J. A., Kemler, R., Gelm, E. P., and Byers, S. W. Cell adhesion molecule uvormorulin expression in human breast cancer cell lines: relationship to morphology and invasive capacities. Cell Growth & Differ., 2: 365-372, 1991.

46. Stover, D. M., Carey, I., Garzon, R. I., and Zehner, Z. E. A negative regulatory factor is missing in human metastatic breast cancer cell line. Cancer Res., 54: 3092-3095, 1994.

47. Azzam, H. S., Arand, G., Lippman, M. E., and Thompson, E. W. Association of MMP-2 activation potential with metastatic progression in human breast cancer cell lines independent of MMP-2 production. J. Natl. Cancer Inst., 85: 1758-1765, 1993.

48. Sommers, C. L., Walker-Jones, D., Heckford, S. E., Worland, P., Valverius, E., Clark, R., McCormick, F., Stampfer, M., Abularach, S., and Gelmann, E. P. Vimentin rather than keratin expression in some hormone-independent breast cancer cell lines and in oncogene-trans- formed mammary epithelial cells. Cancer Res., 49: 4258-4263, 1989.



1996;2:147-154. Clin Cancer Res W Domagala, M Markiewski, B Harezga, et al. relationship with vimentin and p53 protein.determined by the MIB-1 antibody in breast carcinoma: its Prognostic significance of tumor cell proliferation rate as

Updated version

http://clincancerres.aacrjournals.org/content/2/1/147

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/2/1/147To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Prognostic Significance of Tumor Cell Proliferation Rate ...clincancerres.aacrjournals.org/content/clincanres/2/1/147.full.pdf · Vol. 2, 147-154, January 1996 Clinical Cancer Research