The Proliferative Fraction in Lymph Nodes: A Comparison of Proliferating Cell Nuclear Antigen Morphometry to Flow Cytometry f

Transcription

1 ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 22, No. 3 Copyright 1992, Institute for Clinical Science, Inc. The Proliferative Fraction in Lymph Nodes: A Comparison of Proliferating Cell Nuclear Antigen Morphometry to Flow Cytometry f J. S. KRAUSS, M.D., C. G. PANTAZIS, M.D., and F. W. CHANDLER, D.V.M., Ph.D. Department of Pathology, Medical College of Georgia, Augusta, GA ABSTRACT A monoclonal antibody to proliferating cell nuclear antigen (PCNA/ cyclin) has recently becom e available. This antibody, as opposed to Ki-67, can be used on formalin-fixed, paraffin em bedded tissue sections and allows retrospective comparison of PCNA positivity to percent S + G2 + M of the cell cycle. To compare fresh lymphoma deoxyribonucleic acid (DNA) analysis w ith PCNA activity on fixed, paraffin-em bedded sections, prospective flow cytometric studies of cell cycle analysis were performed on lymph nodes removed from 10 patients for diagnosis. Six patients had T-cell lymphoma, two had B-cell lymphoma, and two were benign. Using the peroxidase-antiperoxidase method, the nuclear positivity of archival lymphoma cases was also examined. To quantify PCNA positivity, a unique m orphom etric m ethod was em ployed that utilized digital imaging by high definition television and ELAS (Earth Land Application Software), a geoscience software used extensively for color quantitation of remote sensed data. The immunologic percent PCNA positivity was 26.1 ± 20 vs. percent S + G2 + M by flow cytometry of 22.4 ± 10 with a correlation coefficient (r) of This r-value com pared favorably to data generated for Ki-67 in solid m alignant neoplasm s.13 The six more concordant cases had a percent PCNA positivity of 26.5 ± 10.0 and a percent S + G2+ M of 27.3 ± 8.6, r = Our study is unique in that it compared fresh lymphoma DNA analysis data with paraffin PCNA data. It is our conclusion that immunologic PCNA positivity in paraffin sections correlates with fresh flow cytometric S + G2 + M in lymph nodes, although careful attention m ust be paid to the area of the node quantified for PCNA. * Supported in part by BRSG Award t Send reprint requests to J. S. Krauss, M.D., Department of Pathology, Medical College of Georgia, BIH222B, th Street, Augusta, GA /92/ $01.20 Institute for Clinical Science, Inc.

2 190 KRAUSS, PANTAZIS, AND CHANDLER Introduction Param ount to accurate pathological grading is th e d e te rm in a tio n of the p ro liferativ e fraction of a n eo p lastic process. T raditionally, this has been done by m itotic counts w hich are subjective and inaccurate. C ell cycle analysis which includes a Gq/G j fraction with 2n DNA, an S-phase of DNA replication, and a G2/M fraction of 4n DNA reflects tum or cell grow th k in etics. In m ore recent tim es, a new technology, flow cytom etry, has allow ed very accurate d e te rm in a tio n of S -phase and G2/M fractio n s. H o w ev er, th is m eth o d of d eterm ining grow th fraction requires fresh tissue, which is usually fixed in e th a n o l, for b e s t re s u lts. F u r th e r m ore, the m ethods of assessing proliferation k in etics are tim e consum ing and complex. R ecently, proliferating cell nuclear antigen (PCN A/cyclin), an evolutionarily conservative 36Kd, intranuclear nonhistone polypeptide w hose proliferation is maximized during S-phase of the cell cycle, has been identified.1,3,6 Similar or related proteins accum ulate during interphase leading to triggering of m itosis, thus governing m aturation prom oting factor activity.10 T he m onoclonal antibody to PCNA is reactive in form alin-fixed, paraffin-em bedded histologic se ctio n s; h e n c e, re tro sp e c tiv e stu d ies com paring im m unohistologic PC N A a c tiv ity to flo w c y to m e tric S + G2 + M fractions on fresh material are possible. In this series, lymph node proliferative activity was studied under ideal conditions, i.e., fresh ethanol-fixed samples for flow cytom etry and fixed paraffin -e m b e d d e d sam ples for PCNA, a monoclonal antibody designed for such fixation. A correlation of the flow cytometric proliferative fraction with percent PCNA positivity eittploying com puterized color image analysis is herew ith reported. M aterials and M ethods Im m u n o h is t o c h e m is t r y T h e PC N A m o n o c lo n a l a n tib o d y (MoAb) is co m m ercially available.* Deparaffinized sections of formalin-fixed tissue w ere studied by immunoperoxid a s e te c h n iq u e s w ith o u t c o u n te r- stain.5,6,15,17,18 Five xm tissue sections w ere incubated with a 1:10 dilution or 10 xg of protein per slide. The samples were th e n v isu a liz e d by th e p e ro x id a seantiperoxidase m ethods. Routine hem a toxylin and eosin (H&E) sections for histo p ath o lo g ic ex am in atio n w ere also studied. Blocks w ere retro sp ectiv ely retrieved and deparaffinized for im m unohistologic studies. Im a g e A n a l y s is Microscope. A R e ic h a rt D ia s ta r equipped with 10x.25 NA, 40x 0.66 NA, and 100 x 1.25NA oil immersion plane achrom atic o b jectiv es was th e m icroscope used in this study. A No Abbé (N.A. 1.25) condenser was used for K öhler illu m in atio n. T his co n d en ser could be used for all objectives from 4x to 100 x. The light source was equipped with a 100 W 12V tungsten-halogen light for bright field microscopy. This system provides uniform light intensity throughout its life at a consistent desirable color tem perature of 3200 K (at 12 V setting). The light source contained a variable transform er for adjusting the intensity from 6V to 12V. C a m e r a The camera used in this study was a N E C t three chip CCD that acquires red, green, blue (RGB) images. T he cam era * Dako-PCNA, PC10, Dakopatts, Glostrup, Denmark. t Nippon Electronics Company, Tokyo, Japan.

3 PCNA VERSUS FLOW CYTOMETRY 191 was connected to the microscope through a bayonet/c-m ount connector. T he analog out of the camera RGB lines was connected to the 9-pin RGB connector of the frame grabber in line (AT, T ruevision Vista board with 10 MB of SRAM m em ory). The frame grabber out line was conn e c te d to a 20 inch H igh D efinition (HDTV) Micro Vitec analog/digital m onitor. The band width of this monitor was 20 M Hz with 1280 Horizontal Scan lines and 1024 vertical resolution. The red, green, and blue video frames were individually grabbed with a registration error of less than 0.06 percent. The images captured by this system produced an image of 1024 x 768 x 32 resolution. Each RGB value or primary color was eight bits or 256 illum ination levels/bit (24 bits) with the rem aining eight bits used for image dem ographic data. W ith 24 bits of prim ary color, there can be 16.8 m illion uniq u e color combinations. G raphic image processing was p e r formed on an OPUS Personal Mainframe 5 00t configured w ith th ree M otorola 32 b it RISC (25 M Hz) C P U / FPU processors. The system operated under UNIX System V release 3.2 and contained the B e rk e le y e n v iro n m e n t (system and library calls). A morphometric program em ploying software developed for Landsat and high definition television was used to quantify PCNA positivity. The image processing software used in this study was ELAS and was obtained from the Cosmic Library at the University of Georgia. The ELAS contains a library of 300 specialized modules which provide com plete remote sensing and color image processing capabilities. Color quantitation was achieved producing a classified image that contained the color illum ination intensities (0 to 256) of interest. In im m unoperoxidase, this w ould include t Opus Systems, Inc., Mountainview, CA. red-brow n colors. Approxim ately 2500 cells (two 400x fields) were digitized into 24 bit color images and a color n et of the desired immunoreactive product was formed. From this classified im age, a percent positive of the total num ber of cells could be calculated.11,12,14,16 F l o w C y t o m e t r y Cell cycle analysis was performed with propidium iodide and ribonuclease treatm ent by a modification of the Vindel0v technique on unfixed tissue.2,7 8 9 An Epics 751 was used for flow cytometry. S tatistical data an d c o rrelatio n coefficients were calculated with a TI-81 1calculator. Ten cell cycle analyses of lymph nodes were performed over a four year period. O ne-hundred ml of RNAse A1f solution at a concentration of one mg RNAse A per ml of Dulbecco s phosphate buffered saline (PBS), were added to the centrifuged pellet after removal of the supernatant. After vortexing, one ml of propidium iodide** staining solution at a concentration of 0.05 mg per ml of PBS was added to each tube. The nuclei were incubated for 30 m inutes in the dark at room tem perature. W ithin three hours, analysis of cellular DNA content was perform ed w ith an Epics 751 using a five watt argon laser em itting 150 m illiwatts at 488 nanom e ters. Various filter com binations w ere employed: 488 DI, 560 DI, and a 595 LP. A DNA frequency histogram was generated after 15,000 nuclei were counted, with the exception of one sample which had few nuclei. T he Verity:M odfit software program was used to analyze the histograms. The ratio of aneuploid G0/G 1 m ean peak channels to diploid G0/G 1 m ean peak channels was calculated for Coulter Electronics, Hialeah, FL. 1Texas Instruments, Lubbock, TX. 11Cooper Biomedical, Malvern, PA. ** Calbiochem Behring, Lajolla, CA.

4 192 KRAUSS, PANTAZIS, AND CHANDLER each aneuploid population observed, hence the DNA Index (DI) was generated. In addition, the percentage of cells in the S-phase was also calculated. St a t is t ic a l A n a l y s is Linear regression analysis was used to assess the correlation betw een PCNA labeling and percent S-phase/G2 + M by flow cytometry. The data were analyzed using STATGRAPHICS, a biom edical statistics software. Results Lymph nodes of 10 patients w ere studied by flow cytometry and immunohistologic techniques. E ight lym ph nodes were m alignant (six T-cell lymphomas and two B-cell lym phom as), and two w ere b enign (table I). T he im m unostained lymph node sections w ere not counterstained in order to develop very accurate color analysis of positive nuclei. The staining for PCNA was confined to nuclei in all cases with little if any background staining. Of the m alignant cases, five were high grade and three were interm ediate grade. The average percent S + G2 + M was 22.4 ± 10 and percent PCNA positivity was 26.1 ± 20; the correlation coefficient was r = 0.55 and the linear regression curve w as y = l. l x + 1 w h e re p e rc e n t S + G2 + M = x (figure 1 and table I). W hen the six more concordant results w ere studied, p ercen t S + G2 + M was 27.3 ± 8.6 and percent PCNA positivity was 26.5 ± 10.0 (y = l.lx 3.9; r = 0.96) (table II). For the four less concordant results, percent S + G2 + M was 15 ± 4.6 and the percent PCNA positivity was 25.5 ± 27.9 (y = 5.5x ; r = 0.92) (table III). The histologic grade of the lymphomas correlated highly w ith the calculated proliferative activity by flow cytometry and PCNA positivity. H igh grade lym phomas, as assessed by the International Formulation scheme, were consistently the most positive by PCNA. Shown in figure 2 are a photomicrograph of a peripheral T cell lymphoma and its PCNA immunohistological pattern. In figure 3 is a black and white photo of color image analysis of the im m unocytochem ical reaction product. Note the precise localization (in red pseudocolor) of red-brow n nuclear staining. The red m apping of the digi- TABLE I Comparison of Percent of Proliferating Cell Nuclear Antigen Positivity to Percent of S+G2+M Patient PCNA S+G2+M Diagnosis Peripheral T-cell lymphoma Lymphoblastic lymphoma B-cell lymphoma mixed small and large cell Malignant lymphoma, diffuse, mixed CO Lymphoblastic lymphoma Sarcoid Immunoblastic lymphoma Lymph node, follicular hyperplasia Lymphoblastic lymphoma JQ m 21 Small cell cleaved, nodular lymphoma n= ± ± 10 PCNA = proliferating cell nuclear antigen y = l.lx (r = 0.55)

5 PCNA VERSUS FLOW CYTOMETRY 193 TABLE III Discordant Data S + G2 + M () FIGURE 1. Flow Cytometry S + G2 + M vs. percent PCNA S + G2 + M is X-axis; PCNA percent is y-axis. Patient PCNA S+G2+M c25.5 ±27,9 15 ±4.6 II PCNA = proliferating cell nuclear antigen y = 5.5x-57.5 (r = 0.92) tiz e d im age c o n s is te n tly id e n tifie d stained nuclei, and PCNA staining provided valuable data on the proliferative fraction in lymphom as. Comment Both Ki-67 and PCNA are proteins associated with the S-phase; hence, they can be utilized to identify the rate of proliferation of cells which can be prognostically useful in neoplasia. The PCNA identified by the PC-10 monoclonal antibody is not altered by formalin fixation, thus having an advantage over Ki-67 which m ust be performed on frozen sections. Our study compared PCNA positivity in histologic sections to flow cytom- Patient TABLE II Concordant Data PCNA S+G2+M Jfl IS 21 n = ± ±8.6 PCNA = proliferating cell nuclear antigen y = l.lx -3.9 (r = 0.96) etry, w hich is, at p re se n t, th e b e st m ethod for cell cycle analysis. It is our belief that PCNA immunohistology has certain advantages over flow cytometry, including less expensive equipm ent and the ability to examine nuclear positivity in tissue to identify proliferating cells in the affected areas as opposed to identifying the proliferating fraction of a group of cells, some of which may not be neoplastic. T he r value for all 10 sam ples com pared favorably to the r value for Ki-67 vs. flow cytom etric S + G2 + M of a large series of solid m alignant neoplasm s where r was At lower levels of proliferation (percent S + G2 + M = 15 ± 4.6), there was slightly less correlation (r = 0.92; table III) than at higher levels of proliferation (percent S + G2 + M = 27.3, r = 0.96), as determ ined for Ki The disparity betw een the relatively high S + G2 + M of 14 percent by flow cytometry and the low PCNA of one percent in the T-cell lymphoma of patient #4 suggests that these two procedures may be complementary (see tables I and III). The probable underestim ation of the proliferative fraction by PCNA in this instance may reflect prolonged fixation which can abolish PCNA detection by this antibody,6 a heterogeneous tum or population which could cause variable PCNA positivity in different microscopic fields,5 or, as probably occurred here, a s a m p lin g e r r o r r e f l e c t e d b y th e

6 194 KRAUSS, PANTAZIS, AND CHANDLER 2V«i t * Z* V i 1 «** 4 i i r f / ' *0 V V * * - «L*/*6*v -«?r * & * b 8u *tp -,-t -*<t» Vti isl i m 5 / 2 * 0 I- #. i.y j. v- 9 * >1 ^ 4#;v < & **& / * '* - * *j 3*? «'Jt,.- <s t> # #' v ' lk L * * "AS.*.<$ * Wp J j# a»:*.. & 1* #.# & <c» 4&,sL *»!* ',4 ^ * ' SPl ;# ' H. * \; ;# *- m &- V «: K*W F i g u r e 2. T op: Hematoxylin and eosin sectio n of peripheral T -cell lymphoma com posed of small, medium sized and large atypical lym phocytes. Bottom: PCNA of this section from patient # 1, 22 percent positive. (Original magni fication 100 x) * # i.» } S i * JF ; decreased size of the recut section for PCNA vs. the size of the original H + E section. Thus, the precise relationship of PCN A p o s itiv ity to flow c y to m e tric S + G2 + M rem ains to be established. Our correlation was not as good as that of Woods et a l17 for gastrointestinal lym phom a (r = 0.79), b u t it was superior to *1 that of Yu et a l18 for hem angiopericytom a in w h ich PC N A p o s itiv ity a n d flow cytometry did not correlate. O ur m eth ods, however, w ere different in that com p u te riz e d m o rp h o m e try was u se d as opposed to a subjective PC-10 index for m anual cell counting, and fresh unfixed flow cytom etric m aterial was u sed as

7 PCNA VERSUS FLOW CYTOMETRY 195 F i g u r e 3. Black & w hite reproduction of 35mm photographs from HDTV. Top: PCNA. Bot tom: D ig it iz e d RGB image of top photo. Light is pseudored; PCNA posi tive. Note the excellent correlation. opposed to fixed deparaffinized blocks for w hich flow cytom etry is less precise.9 M oreover, the p resen t series of lym ph nodes stu d ie d was less hom ogeneous than previous studies w hich in clu d ed only lymphom as, because two of the 10 specim ens w ere benign (see table I). T h e c u rre n t a u th o rs c o n c lu d e th at PCNA im m unohistologic positivity can be u se d to e v a lu a te p ro life ra tio n in ly m p h n o d e s, a n d t h a t in c e r t a i n instances im m unohistologic evaluation of S phase may be superior to flow cytom etry. Furtherm ore, the m easurem ent of the proliferative fraction in lym phoid neoplasm s is im portant because of its c lo s e c o r r e l a t i o n w ith h i s t o l o g i c grade.2,4,5 Finally, it is our b e lie f that

8 196 KRAUSS, PANTAZIS, AND CHANDLER th e p o sitiv ity of PCNA in ro u tin e ly processed tissues confirms its superiority to Ki-67 w hich is only useful in frozen sections. Acknowledgment Thanks are extended to Janet Hobbs and William Wansley for their technical assistance, to Gary West of Crystal DATA Systems, Huntsville, AL, for hardware support, and to Steve Bong of Crystal Image Technologies, for software and programming support. References 1. Br a v o, R., F r a n k, R., B l u n d e l l, P. A., et al.: Cyclin/PCNA protein is the auxiliary protein of DNA polymerase alpha. Nature 326: , Br a y l a n, R. C., B e n s o n, N. A., a n d N o u r s e, V. A.: C e llu l a r D N A o f h u m a n n e o p la s tic B -cells m e a s u re d b y flo w c y to m e try. C a n c e r R es. 44: , C e l i s, J. E., Br a v o, R., L a r s e n, P. M., et al.: A nuclear protein whose level correlates directly with the proliferative state of normal as well as transformed cells. Leuk. Res. 8: , D ia m o n d, L. W., N a t h w a n i, B. N., and Ra p - PAPORT, H.: Flow cytometry in the diagnosis and classification of malignant lymphoma and leukemia. Cancer 50: , G a r c ia, R. I., C o l t e r a, M. D., and G o w n, A. M.: Analysis of proliferative grade using anti PCNA/Cyclin monoclonal antibodies in fixed, embedded tissues. Comparison with flow cytometric analysis. Am. J. Path. 124: , H a l l, P. A., L e v is o n, D. A., W o o d s, A. L., et al.: Proliferating cell nuclear antigen (PCNA) immunolocalization in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasms. J. Pathol. 262: , Koss, L. G., C z e r n i a k, B., H e r z, F., and WERSTO, R.: Flow cytometric measurements of DNA and other cell components in human tumors. A critical appraisal. Human Pathol. 20: , Look, A. T., M e l v in, S. L., W il l ia m s, D. L., et al.: Aneuploidy and percentage of S-phase cells determined by flow cytometry correlate with cell phenotype in childhood acute leukemia. Blood 60: , M c I n t i r e, T. L., G o l d e y, S. H., B e n s o n, N. A., and B r a y l a n, R. C.: Flow cytometric analysis of DNA in cells obtained from deparaffinated formalin-fixed lymphoid tissues. Cytometry 8: , M u r r a y, A. W. and Ki r s c h n e r, M. W.: What controls the cell cycle. Sci. Am. 264:56 63, O d u m, L. D., B a r r e t t, J. M., P a n t a z is, C. G., S t o d d a r d, L. S., and M c D o n o u g h, P. G.: Immunocytochemical study of ras and myc proto-oncogene polypeptide expression in the human menstrual cycle. Am. J. Obstet. Gynecol. 161: , P a n t a z is, C. G., Ya g h m a i, F., and B o n g, S.: Microscope-based Image Archiving and Analysis System: RISC-based X Windows Workstation. 6th Annual Informative Technology in the Health Sciences Conference. Memphis, TN, May 15, 1991, p S a h in, A. A., Ro, J. Y., A del, K. E.-N., et al.: Tumor proliferative fraction in solid malignant neoplasms. A comparative study of Ki-67 immunostaining and flow cytometric determinations. Am. J. Clin. Pathol. 96: , Sa m u e l s, V., B a r r e t t, J. M., B o c k m a n, D. E., P a n t a z is, C. G., and A l l e n, M. B.: Immunocytochemical study of transforming growth factor expression in benign and malignant gliomas. A m. J. Pathol. 134: , v a n D ie r e n d o n c k, J. H., W ijs m a n, J. H., K e ij- ZER, R., et al.: Cell cycle-related staining patterns of anti-proliferating cell nuclear antigen monoclonal antibodies. Comparison of BrdUrd labeling and Ki-67 staining. Am. J. Pathol. 238: , W ie d, G. L., Ba r t e l s, P. H., B ib b o, M., and D y t c h, H. E.: Image analysis in quantitative cytopathology and histopathology. Human Pathol. 20: , W o o d s, A. L., H a l l, P. A., S h e p h e r d, N. A., e t a l.: T h e a s s e s s m e n t o f p r o lif e r a tin g c e ll n u c le a r a n tig e n (PCNA) im m u n o sta in in g in p rim ary g a stro in te stin a l ly m p h o m a s a n d its re la tio n sh ip to h isto lo g ic a l g ra d e, S + G 2+ M p h a s e frac tio n (flow c y to m e tric a n a ly sis) a n d p ro g n o sis. H isto p a th o lo g y 29:21 27, Yu, C. C.-W., H a l l, P. A., F l e t c h e r, C. D. W., et al.: Hemangiopericytomas: The prognostic value of immunohistochemical staining with a monoclonal antibody to proliferating cell nuclear antigen (PCNA). Histopathology 29:29-33, 1991.