Acta Neuropathol (1993) 85:495 -502

Acta Neuropathologtca 9 Springer-Verlag1993

Proliferating cell nuclear antigen expression in brain tumors, and its prognostic role in ependymomas: an immunohistochemical study* D. Schiffer, A. Chi6, M. I". Giordana, T. Pezzulo, and M. C.Vigliani Department of Neurology, University of Turin, Via Cherasco 15, 1-10126 Turin, Italy Received August 18, 1992/Revised, accepted No,~emhe~~27, 1992

Summary. Proliferating cell nuclear antigen (PCNA)/cyclin is currently often investigated immunohistochemically in tumors as a marker of cell proliferation, but many problems remain open concerning its reliability as a prognostic factor. P C N A has been studied in a series of 123 brain tumors using the monoclonal antibody PC10. A clear intra- and inter-tumor variability of PCNApositive nuclei has been found, but taking into account the tumor areas with the highest n u m b e r of positive nuclei, a positive correlation between this n u m b e r and the histological malignancy of tumors has been demonstrated. The staining intensity of nuclei was variable; very-intensely positive nuclei, counted separately, are hypothesized to represent nuclei in S-phase of the cell cycle. In ependymomas the investigation included a quantitative statistical analysis. The n u m b e r of P C N A positive nuclei correlated with cell density and mitotic index, but only very intensely positive nuclei showed a significant statistical correlation with survival. In spite of the many possibilities of wrong interpretation of P C N A expression, the most important of which is its deregulation, the m e t h o d is useful in the practice for prognostic purposes. Its important advantages are the possibility of a retrospective application and a visual analysis of the proliferation potential of tumors. Key words: Proliferating cell nuclear antigen (PCNA) Brain tumors - Cell density - Mitotic index - Prognosis

Proliferating cell nuclear antigen (PCNA)/cyclin is a 36-kDa protein auxiliary of D N A polymerase 6 [4]. It increases in the cell through G1, peaks in GflS, decreases

* Supported by CNR, Rome (Special Project A.C.R.O.); AIRC, Milan; Italian Ministry of Scientific Research and University (Progetto Finalizzato 40 %); and CSI-Piemonte Correspondence to: D. Schiffer (address see above)

in G2/and is virtually negative in the M phase [18]. The gene for human P C N A is transcribed both in quiescent and proliferative cells, but P C N A m R N A normally accumulates only in proliferating cells [6]. P C N A can be recognized immunohistochemically in formalin-fixed and paraffin-embedded material [12], and the distribution in non-neoplastic tissues is consistent with its association with cell proliferation [12]. In non-CNS tumors it has been shown to correlate with Ki67 [12, 14, 19], flow cytometry [11, 29], and tritiated thymidine [2, 12] and B r d U [8, 12] incorporation. Few and contradictory findings are available in CNS tumors [1, 20] because of the many technical problems and difficulties in their interpretation. In this article the distribution of P C N A in 123 brain tumors is reported and compared with histological diagnosis. In addition, a correlation of P C N A with survival has been investigated on 60 ependymomas previously studied for histological prognostic factors [26, 27].

Materials and methods Surgical biopsies of 15 meningiomas (2 malignant), 7 oligodendrogliomas (4 anaplastic), 11 medulloblastomas, 10 well-differentiated astrocytomas, 10 anaplastic astrocytomas, 10 glioblastomas and 60 ependymomas (27 supratentorial, 27 infratentorial and 6 subependymomas) were studied. The histological classification used was that of WHO [16]. Anaplastic ependymomas were recognized on the basis of histological parameters previously identified [27]. Samples of neoplastic tissues (1 cm • 0.5 crn) were routinely fixed in Carnoy solution at 0~ ~ for 24 h and embedded in paraffin. Sections were cut at 6 ~tmand air-dried overnight at room temperature. PCNA monoclonal antibody (mAb) (PC 10; Dako) was used at 1:300, overnight. The peroxidase-labeled streptavidin method was employed. In each case, positive nuclei (PN), independently from their staining intensity, and negative nuclei were counted in ten highpower fields (HPF), corresponding to 0.02 mm2, at 1000 x magnification with oil immersion. The areas showing the highest concentration of PN by visual analysis were chosen for counting. The percentage of PN was regarded as the labeling index (LI). Since variations in staining intensity occurred among nuclei,

496 very-intenselypositive nuclei (VIPN) were counted separately and evaluated as percentages of the total number of nuclei. As VIPN were considered nuclei in which the PCNA staining masked the counterstaining with hematoxylin, The evaluation and the counting of nuclei were performed simultaneouslyby two observers with a microscope equipped with a discussion tube. The intra-observer variability was checked with a random recounting of the same tumor areas. Its range was between 3 % and 14 %. Inflammatory and vessel cells were not counted. In parallel sections stained with hematoxylin-eosin(H&E), the mitotic index (MI) and cell density were calculated by counting the number of mitoses and tumor ceils per ten HPF at 1000 x magnification, corresponding to those with the highest number of PN.

Statistical evaluation Internal correlation between histological parameters were made using Pearson product-moment coefficient [7]. In all figures the straight line was calculated by means of the least square method [7]. A P of 0.05 was considered statistically significant. Survival analysis was performed by the actuarial method [15] and the significance was tested by the log-rank test [23]. The calculations were made using the BMDP statistical package programs 1L and 2L [10].

Results T h e r e was a v a r i a b i l i t y of t h e p e r c e n t a g e s of P N f r o m o n e t u m o r a r e a to t h e other, w i t h t h e h i g h e s t v a l u e s i n g l i o b l a s t o m a s a n d e p e n d y m o m a s a n d t h e lowest i n m e n i n g i o m a s . I n Table 1 t h e m e a n , t h e m e d i a n a n d t h e r a n g e of t h e h i g h e s t p e r c e n t a g e s of P N a n d V I P N for e a c h t u m o r t y p e are i n d i c a t e d . T h e first two v a l u e s i n c r e a s e d w i t h t h e histological m a l i g n a n c y of t u m o r s a n d c o r r e l a t e d w i t h M I . V I P N r e p r e s e n t e d a b o u t o n e i n five of P N , w i t h t h e e x c e p t i o n of w e l l - d i f f e r e n t i a t e d astrocyt o m a s . I n astrocytic t u m o r s , t h e p e r c e n t a g e s of P N a n d VIPN increased from well-differentiated astrocytomas to g l i o b l a s t o m a s , t h r o u g h a n a p l a s t i c a s t r o c y t o m a s , b u t t h e ratio V I P N / P N was h i g h e r (0.41) i n a s t r o c y t o m a s than in anaplastic astrocytomas and glioblastomas. E n d o t h e l i a l p r o l i f e r a t i o n s c o n t a i n e d less f r e q u e n t P N

and VIPN than tumor parenchyma and circumscribed n e c r o s e s were l o c a t e d i n areas w i t h a high n u m b e r of P N a n d V I P N (Fig. l a - c ) . I n o l i g o d e n d r o g l i o m a s , t h e n u m b e r of P N a n d V I P N was m u c h h i g h e r i n a n a p l a s t i c t h a n in classic cases (Fig. l d ) , w i t h a slight c h a n g e of t h e r a t i o V I P N / P N . I t is n o t e w o r t h y t h a t P N a n d V I P N v a l u e s were m u c h h i g h e r t h a n i n a s t r o c y t o m a s . A s t r o c y t i c areas a n d e n d o t h e l i a l proliferations showed less-frequent PN and V I P N than typical o l i g o d e n d r o g l i a l areas. I n m e d u l l o b l a s t o m a s , p e r c e n t a g e s of P N a n d V I P N were high (Fig. 2a), a n d p o s i t i v e l y c o r r e l a t e d w i t h M I , i n spite of t h e difficulties i n c o u n t i n g m i t o s e s in m e d u l l o b l a s t o m a s . Pale islands a n d folliculi in t h e d e s m o p l a s t i c v a r i a n t were m o s t l y n e g a t i v e (Fig. 2b) or c o n t a i n e d a l o w e r n u m b e r of P N t h a n classic m e d u l l o b l a s t o m a t o u s areas; i n s o m e cases, o n l y t h e i r p e r i p h e r a l parts cont a i n e d PN. I n areas of n e u r o n a l d i f f e r e n t i a t i o n t h e n u c l e i w e r e n e g a t i v e (Fig. 2c). I n m e n i n g i o m a s , P N were v e r y few, w i t h a low local v a r i a t i o n i n each case, r a n g i n g f r o m 0 % to 14 % . T h e ratio V I P N / P N was 0.14. I n m e n i n g i o m a s , d i a g n o s e d as atypical or m a l i g n a n t , t h e m e a n p e r c e n t a g e of b o t h P N a n d V I P N d o u b l e d , w i t h a ratio V I P N / P N of 0.20 (Fig. 2d).

Fig. 1. a Anaplastic astrocytoma: proliferating cell nuclear antigen (PCNA)-positive nuclei, b Glioblastoma: many nuclei are positive, but only a few in the endothelial proliferations, c Circumscribed necroses develop in areas with high cell density and many positive nuclei, d Anaplastic oligodendroglioma: many nuclei are positive, a-d Counterstaining with hematoxylin; a,d x 400, b,c • 200

Fig. 2. a-c Medulloblastoma. a Area with many positive nuclei; b pale islands are almost negative; c areas with neuronal differentiation are negative, d malignant meningioma: many nuclei are positive, a-d Counterstaining with hematoxylin; a,b,d • 400; e x 200

Table 1. Percentage of proliferating cell nuclear antigen (PCNA) positive and intensively positive nuclei and range of mitotic index in various oncotypes PN

VIPN

Oncotype

(Number Mean + SD Median of cases)

Range

Astrocytomas Anapl. Astrocytomas Glioblastomas Oligodendrogliomas Anapl. Oligodendr. Medulloblastomas Meningiomas Anapl. Meningiomas Ependymomas Anapl. Ependymomas Subependymomas

(10) (10) (10) (3) (4) (11) (13) (2) (44) (10) (6)

3.4 13.2 58.0 8.4 41.6 33.4 0 9.1 3.7 63.6 2.1

4.8 33.7 65.1 19.3 43.7 56.6 8.3 18,2 37.1 74.4 3.7

+ 1.5 + 14.2 _+ 6.5 + 10.7 + 4.4 + 15.2 + 4,4 + 12.8 + 19.4 + 6.5 + 2.3

4.7 38.9 63.2 19.5 43.4 58.0 8.7 39.8 72.7 4.0

-

6.4 52.9 74.1 29.8 45.5 84.7 14.0 27.3 54.1 87.6 5.6

VIPN/PN MI

Mean + SD Median

Range

2.0 5.4 12.7 2.8 9.3 10.4 1.2 3.7 5.5 13.5 1.0

1.4 2.4 7.0 1.4 7.7 4.2 0 1.8 0 8.5 0

_+ 0.6 + 2.2 + 3.5 + 2.5 + 1.8 _+ 2.9 + 0,5 + 2.6 + 3.7 + 2.9 _+ 0.7

1.9 5.0 12.7 1.5 8.1 11.4 1.2 5.4 13.1 0.6

-

PN, PCNA-positive nuclei; VIPN, very intensely positive nuclei; MI, mitotic index; Anapl., anaplastic

3.1 8.5 16.7 5.6 9.5 13.7 3.4 5.5 17.5 22.4 1.3

(Range)

(0.41) (0.16) (0.20) (0.15) (0.21) (0.18) (0.14) (0.20) (0.15) (0.18) -

0 0.13 0.18 0 0.18 0.09 0 0.2 0 0.27 0

-

0.10 0.67 1.20 0.31 0.51 0.82 0.20 0.8 0.49 1.10 0

497

Fig. la-d.

498

Fig. 2a-d.

499

Fig. 3a-d. (for legend see next page)

5OO In e p e n d y m o m a s there was a striking inter- and intra-tumor variability of PN and V I P N percentages (Fig. 3a,b). The range of percentages was much wider in classic cases than in anaplastic ones. T h e ratio VIPN/PN was similar to that of the other tumors. T h e percentages of PN and V I P N increased with cell density (Fig. 3c) and were not influenced by perivascular pseudorosettes, rosettes, canals and papillary structures (Fig. 3d). Circumscribed necroses developed in areas of very-high cell density with a high n u m b e r of P N and VIPN. In comparison with t u m o r cells, hyperplastic endothelial cells showed a n u m b e r of PN much lower, even in multilayered proliferations. Nuclear inclusions were negative. Normal ependymal lining, occasionally found in the tumor periphery, was mostly negative with the exception of the cases in which it appeared in continuity with the t u m o r proliferation. A direct correlation was found between cell density, which ranged between 30 and 149/HPF at 1000 x magnification and the percentage of PN (Fig. 4a). T h e percentage of V I P N correlated with that of PN and both correlated with the MI (Fig. 4 b - d ) . Life table analysis showed that survival did not correlate with the percentage of PN, but with that of V I P N in supratentorial cases (Fig. 5a,b). In subependymomas PN were very rare, with the values similar to those in astrocytomas. Associated ependymomatous areas showed similar values to classic ependymomas.

oell denelty

180

a

140 lO0 100

,.

80 (SO

9

"

40 2O i

0 0

i

i

i

20

% O| PCINA posltl~

nuclei

% of PCNA intensely positive nuolel

JIO

b 20

10

B 0

=~- " 0

1,2

i 20

i, 40 % of PCNA

i

i

60 poeltlve nuolel

80

i 100

40 % of PCNA

60 positive nuolel

80

100

20 nuclei

26

m l t o t ~ index

c

Discussion P C N A immunohistochemistry can be employed in retrospective studies and it can be applied to material e m b e d d e d in paraffin. It has b e e n shown that the P C N A LI is a m a r k e r of the proliferative potential of tumors and correlates with other LI. Type and duration of fixation are limiting factors of the reliability of the findings [12]; short fixations in formalin or in alcoholic solutions, like Carnoy, however, do not seem to affect the demonstration of P C N A , which is confirmed by present results. T h e most important problems in fixed and e m b e d d e d material remain the variation of the total amount of P C N A in the cell cycle phases and the differences in detectable P C N A levels in different cell types. The f o r m e r refers to the possibility that positive S phases go unrecognized, because the immunohistochemical system has been titered to high P C N A contents [8]. This does not seem to have occurred in our experiments, since the values of PN are slightly higher than those found by others [1, 20]. V I P N represent a relatively constant fraction of PN in all the t u m o r types, except astrocytomas; the hypothesis might be put for-

1

0,8 0,6 0.4 0.2 0

20

0

mltotlo index

d

1,2 1 0,8 0,6 0.4

0.2

0

0

" 6 % of PCNA

10 Intensely

16 poWtlve

Fig. 4a-d. Ependymomas. a Correlation between cell density per Fig. 3a-d. Ependymoma. a Positive nuclei greatly vary from one area to the other; b different staining intensities of nuclei in an area; c many positive nuclei are found in areas with high cell density; d positive nuclei are found in different structures, a-d Counterstaining with hematoxylin; a x 100; b-d x 400

high-power field at 1000 x and PCNA-positive nuclei (P < 0.001); b correlation between PCNA-positive nuclei and PCNA intensely positive nuclei (P < 0.001); c correlation between positive nuclei and mitotix index (P < 0.01); d Correlation between PCNA intensely positive nuclei and mitotic index (P < 0.01)

501 percent survival

100 p = n.s,

80

I

60

~

percent

> 50 percent 4O

2O

a i 500

0 0

i 1000

i 1500

i 2OO0

time after surgery (days) )ercant survival 100

80

l.,

60

I

~ 7 percent

40 p ( 0.05

20

I

> 7 percent I

b

i 500

I

1000

i 1500

time after surgery (days)

2000

Fig. 5a,b. Supratentorial ependymomas. Life tables according to the percentage of (a) PCNA-positive nuclei (P = n.s.) and (b) PCNA intensely positive nuclei (P < 0.05)

ward that they correspond to the S phase, where PCNA has been observed to increase by two-to threefold [22]. No acceptable explanations could be found for the peculiar behavior of astrocytomas. Other possible misinterpretations of positive staining include the detection of PCNA when the cell has already left the cell cycle, since the half-life of PCNA is 20 h [3], and mainly the deregulation of PCNA expression. This could be due to growth factors, such as Epidermal Growth Factor (EGF), which may stimulate the accumulation of PCNA m R N A as well as the synthesis of the protein, dissociated from that of D N A [12, 13]. The most important advantage of the PCNA method is the possibility of a visual analysis. The PN are not uniformely distributed, prevailing in florid tumor areas and being almost absent in quiescent or degenerated tumor areas. Considering the tumor areas with the highest value of PN-VIPN in each tumor, a correlation between this value and the histological malignancy is evident from our results. Cell density, the number of PN and number of mitoses correlate both by visual analysis and by statistical evaluation in tumors such as gliomas and ependymomas, where cell density may be a prognostic factor [9, 25], and not in tumors where cell density is an intrinsic characteristic, such as meningiomas and

medulloblastomas. However, in the latter a correlation is also found between number of PN and histological malignancy, as in meningiomas, or MI, as in medulloblastomas. In oligodendrogliomas, the number of PN seems to be too high in comparison with astrocytomas. An explanation may be that in this tumor type mitoses are more common than in astrocytomas. In some studies [21], the number of mitoses does not play a prognostic role, but in other investigations it correlates with survival [5, 17]. Even though the number of PN strongly increases in anaplastic oligodendrogliomas,VIPN seem to be more representative of the proliferation potential of the tumor, since it is also found in ependymomas. In supratentorial ependymomas, but not in infratentorial ones, the number of mitoses and the cell density have been previously demonstrated to be prognostic factors permitting identification of malignancy [27]. In the same tumors the percentage of VIPN can be regarded as a further marker of malignancy when its value is higher than 7 %. PN, even though correlating generally with histological signs of malignancy, are not predictive of outcome. PCNA expression in normal cells adjacent to tumors was found in breast carcinomas [12] and attributed to PCNA deregulation. We observed in glioblastomas peritumoral PCNA-positive cells which have been interpreted as infiltrating cells or as reactive astrocytes. Where ependymomas reached the normal ependymal covering, this contained PN; very likely these cells took part in the tumor proliferation. The low number of PCNA-positive nuclei of pale islands of medulloblastomas is consistent with their interpretation as sites of neuronal differentiation, in line with the negative PCNA reaction of cells showing a neuronal differentiation. The crowding of PN around circumscribed necrotic areas of ependymomas and glioblastomas is consistent with the interpretation of such areas as developing in sites of high cell density, through the imbalance between the proliferative capacity of endothelial cells and tumor cells [24, 28]. This finding is confirmed by the observation that endothelial proliferations contain a lower number of PN than tumor cells. A general finding of paramount importance is that in all the tumors, but especially in the malignant ones, the number of PN seems to be too high. If it represents the number of cells in the proliferating pool, the growth fraction of tumors would be terribly high and this does not correspond to the present knowledge on the proliferation potential of brain tumors. VIPN, therefore, correspond more closely to the real proliferation potential of tumors, as the comparison with survival in ependymomas demonstrates. Some mechanisms must intervene in producing a so-widespread PCNA expression. Whatever the case, PCNA staining represents a good prognostic marker, as it parallels the histological malignancy, even though it does not add anything new to what is already known from proliferation indices at present employed.

502

References 1. Allegranza A, Girlando S, Arrigoni GL,Veronese S, Mauri FA, Gambacorta M, Pollo B, Dalla Palma P, Barbareschi M (1991) Proliferating cell nuclear antigen expression in central nervous system neoplasms. Virchows Arch [A] 419:417-423 2. Battersby S, Anderson TJ (1990) Correlation of proliferative activity in breast tissue using PCNA/cyclin. Hum Pathol 21:781 3. Bravo R, MacDonald-Bravo H (1987) Existence of two populations of cyclin proliferating cell nuclear antigen during the cell cycle: association with DNA replication sites. J Cell Biol 105:1549-1554 4. Bravo R, Frank R, Blundell PA, MacDonald-Bravo H (1987) Cyclin/PCNA is the auxiliary protein of DNA-polymerase delta. Nature 326:515-517 5. Burger PC, Rawlings CE, Cox EB, McLendon RE, Schold SC Jr, Bullard DE (1987) Clinicopathologic correlations in the oligodendroglioma. Cancer 59:1345-1352 6. Chang CD, Ottavio L,Travalli S, Lipson KE, Baserga R (1990) Transcriptional and posttranscriptional regulation of the proliferating cell nuclear antigen gene. Mol Cell Biol 10: 3289-3296 7. ColtonT (1974) Statistics in medicine. Little, Brown and Co., Boston 8. Coltrera MC, Gown AM (1991) PCNA/cyclin expression and BrdU uptake define different subpopulations in different cell lines. J Histochem Cytochem 39:23-30 9. Daumas-Duport C, Scheithauer B, O'Fallon J, Kelly P (1988) Grading of astrocytomas. A simple and reproducible method. Cancer 62:2152-2165 10. Dixon WJ (1990) BMDP Statistical Software Manual. University of California Press, Berkeley, California 11. Garcia RL, Coltrera MD, Gown AM (1989) Analysis of proliferative grade using anti-PCNA/cyclin monoclonal antibodies in fixed, embedded tissues. Comparison with flow cytometric analysis. Am J Pathol 134:733-739 12. Hall PA, Levison DA, Woods AL, Yu CC-W, Kellock DB, Watkins JA, Barnes DM, Gillett CE, Camplejon R, Dover R, Waseem NH, Lane DP (1990) Proliferating cell nuclear antigen (PCNA) immunolocalization in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasms. J Pathol 162:285-294 13. Jaskulski D, Gatti C, Travali S, Calabretta B, Baserga R (1988) Regulation of the proliferating cell nuclear antigen cyclin and thymidine kinase mRNA levels by growth factors. J Biol Chem 263:10175-10179 14. Kamel O, LeBrun D, Davis R, Berry G, Warnke R (1991) Proliferative activity of malignant lymphomas in formalinfixed paraffin-embedded tissue using anti-PCNA/cyclin 19A2 (abstract). Lab Invest 64:75A 15. Kaplan EL, Meier P (1958) Non-parametric estimation from incomplete observation. J Am Stat Assoc 53:457-481

16. Kleihues E Burger PC, Scheithauer BW (1993) Histological typing of tumours of the central nervous system. WHO Blue Book, 2nd edn. Springer, Berlin Heidelberg New York Tokyo, pp 1-37 17. KrosJM,van Eden CG,Vissers CJ, Mulder AH,van der Kwast TH (1992) Prognostic relevance of DNA flow cytometry in oligodendroglioma. Cancer 69:1791-1798 18. Kurki P, Vanderlaan M, Dolbeare F, Gray J, Tan EM (1986) Expression of proliferating cell nuclear antigen (PCNA)/cyclin during the cell cycle. Exp Cell Res 166:209-219 19. Levison DA, Hall PA,Woods AL,Yu C, Barnes DM,Wasseem N, Lane DP (1990) Evaluation of PCNA (proliferating cell nuclear antigen) immunostaining as a marker of cell proliferation in formalin-fixed paraffin-embedded tissues. J Pathol 161:340 20. Louis DN, Edgerton S, Thor AD, Hedley-Whyte ET (1991) Proliferating cell nuclear antigen and Ki-67 immunohistochemistry in brain tumors: a comparative study. Acta Neuropathol (Berl) 81:675-679 21. M6rk SJ, Halvorsten TB, Lindegard KE Eide GE (1986) Oligodendroglioma. Histologic evaluation and prognosis. J Neuropathol Exp Neurol 45:65-78 22. Morris GF, Mathews MB (1989 Regulation of proliferating cell nuclear antigen during cell cycle. J Biol Chem 264:13856-13864 23. Peto R, Pike MG, Armitage R Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG (1977) Design and analysis of randomized clinical trails requiring prolonged observations of each patient: analysis and examples. Br J Cancer 35:1-39 24. Schiffer D, Giordana MT, Soffietti R, Sciolla R (1982) Histological observations on the regrowth of malignant gliomas after radiotherapy and chemotherapy. Acta Neuropathol (Berl) 58:291-299 25. Schiffer D, Chi6 A, Giordana MT, Leone M, Soffietti R (1988) Prognostic value of histologic factors in adult cerebral astrocytoma. Cancer 61:1368-1393 26. Schiffer D, Chi6 A, Giordana MT, Migheli A, Palma L, Polio B, Soffietti R,Tribolo A (1991) Histologic prognostic factors in ependymoma. Childs Nerv Syst 7:177-182 27. Schiffer D, Chi6 A, Cravioto H, Giordana MT, Migheli A, Soffietti R,Vigliani MC (1991) Ependymoma: internal correlations among pathological signs: the anaplastic variant. Neurosurgery 29:206-210 28. Tannock IF (1970) Population kinetics of carcinoma cells, capillary endotelial ceils and fibroblasts in a transplantated mouse mammary tumor. Cancer Res 30:2470-2476 29. Woods AL, Hanby AM, Hall PA, Wasseem N, Lane DR Levison D (1990) The prognostic value of PCNA (proliferating cell nuclear antigen) immunostaining in gastro-intestinal lymphomas. J Pathol 161:342