Cytometry 12:302-307 (1991)
0 1991 Wiley-Liss, Inc
Nuclear Polymorphism and Nuclear Size in Precarcinomatous and Carcinomatous Lesions in Rat Colon and Liver I. Zusman,' M. Kozlenko, and A. Zimber Laboratory of Teratology and Experimental Oncology, Koret School of Veterinary Medicine (I.Z., M.K.), and Department of Animal Science, Faculty of Agriculture (A.Z.),Hebrew University of Jerusalem, Rehovot, Israel Received for publication September 4,1990; accepted December 18, 1990
The role of nuclear polymorphism and nuclear size in an analysis of the differences between colon and liver tumors in rats as well as in an analysis of hepatic dysplasia was studied. It was shown that colon tumors which developed following treatment of animals with a direct carcinogen alone or with a carcinogen followed by secondary bile acid were characterized by low incidence of nuclear polymorphism and by an increased nuclear size in epithelial cells. Metastatic liver tumors in rats with colon tumors were characterized by a high value for the coefficient of nuclear form polymorphism and by a significant decrease in nuclear size. Hepatic dysplasia which developed
In spite of the extensive use of histochemistry and immunohistochemistry in clinical histopathological practice, light microscopical morphometry also provides important information (24). For example, parameters such as cellular and nuclear areas, nuclear-cytoplasmic, and nucleolar-nuclear ratios have been utilized to characterize hepatocellular dysplasia (17). It was shown that morphometric examination complements histological evaluation and allows dysplastic cells to be identified (27). A correlation between nuclear size and the type of lung carcinoma (3,10,16) as well a s prognosis i n breast cancer (20) has been shown. Nuclear perimeter and nuclear optical density were shown to have diagnostic significance in identifying transformed cells of cervical carcinomas (51, salivary gland tumors (15), gastric carcinoma (21), or prostate carcinomas (14,191. The determination of shapes of cells and nuclei has been used in a n analysis of human hepatocellular carcinoma (6).It was shown, for example, that nuclear polymorphism is one of the characteristics of liver cell dysplasia (1,2), the first stage of which is characterized by enlarged nuclei and a n in-
as a result of prolonged treatment with secondary bile acids was characterized by high rate of nuclear form polymorphism and by a significant increase in nuclear size. The obtained results suggest that nuclear polymorphism is dependent upon the type of tissue or organ involved in the cancerous transformation and that it may have significance as a diagnostic marker of precarcinomatous and carcinomatous lesions of digestive organs only when used in combination with other analyses. Key terms: Rat colon and liver tumors, hepatic dysplasia, nuclear form polymorphism, coefficient of polymorphism
creased nuclear-cytoplasmic ratio (23). This latter parameter was also used to distinguish normal and perineoplastic skin in women (9). Differences in nuclear geometry and the number of cytoplasmic granules were described in HL-60 cells cultured in medium containing inducers of differentiation such as dimethyl sulfoxide and retinoic acid (11). Cell shape polymorphism was used as a diagnostic parameter to distinguish between two lines of human prostate cancerous cells (7). In this communication we studied the relation of nuclear shape and nuclear size to the type of the affected tissue and to the stage of its transformation. Rat colon and liver tumors, as well as rat livers undergoing pathological lesions and probably precancerous transformation, were used as a model object.
'Address for correspondence: Dr. I. Zusman. Koret School of Veterinary Medicine, Hebrew University of Jerusalem, Rehovot, P.O. Box 12, Israel 76100.
303
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS
MATERIALS AND METHODS Randomly bred 6-week old male Sprague Dawley rats (Anilab Co., Rehovot, Israel) were housed in plastic cages with stainless steel tops, and given a defined diet (610, Esem Hanegev Ltd., Israel) and water ad libitum. Temperature controlled (24 2°C) rooms with 12/12 darWlight schedule were used. Animals were treated intrarectally with deoxycholic secondary bile acid (DCA), with a direct carcinogen, N-methyl-N’-nitro-N-nitrosoguanidine (MNNG) alone, or with subsequent treatment with DCA a s a tumor promoter. Treatment with MNNG was for 2 weeks, while treatment with DCA was administered for 2 months or 12 months. Rats were killed by cervical dislocation 2 weeks and 2 or 12 months following the beginning of the treatment. The caudal segment of the distal colon (including the site of the drugs’ instillation) was opened longitudinally. Lesions a s well as normal appearing regions close to tumors were immediately fixed in phosphatebuffered saline (PBS)-buffered neutral 4% formaldehyde. Areas of the descending colon from DCA treated rats and from saline treated controls were also taken for histological examination. In parallel, the liver was taken for histological examination from experimental and control animals. Histological sections (3 pm) were stained with hematoxylin-eosin. The nuclear form of epithelial cells lining the crypts in the descending colon was studied in normal and adenocarcinomatous tissue according to the method described elsewhere (1,2). The following forms of nuclei were recorded: round and cuboid, oval, irregular, long columnar, and drop-like. These forms of nuclei were chosen based on the results of our previous study of the morphological changes in colon and liver cells obtained from rats chronically treated with secondary bile acids (Zimber and Zusman, unpublished data). Nuclei were studied in 20 randomly chosen crypts in each of the following groups of animals: 10 control colons, 10 colons from rats treated with DCA, 5 tumors obtained after the treatment with MNNG alone, and 5 tumors obtained after treatment with MNNG and DCA. The number of nuclei studied was no less than 3,000 in each group, i.e., the total amount of measured nuclei of colon cells was 12,000. In the liver, 4 nuclear forms of randomly chosen parenchymatous hepatocytes were recorded: round, oval, long columnar, and irregular. In each group of animals no less than 1,000 cells were studied, and the total number of measured hepatocytes was 4,000. A coefficient of polymorphism of forms (K) was calculated according to the equation derived by Kaplan and Kozlenko (13).
*
where K = the coefficient (factor) degree of polymorphism of forms (0 5 K 5 1); n = the number of forms
C
D
E
FIG.1. Distribution of colon epithelial cell nuclei according to their form in control (1) and tumorous tissues which developed following treatment with MNNG (2) or with MNNG and DCA (3). Types of nuclear forms: A, round and cuboidal; B, oval; C, irregular; D,long columnar; E,drop-like.
in a given system (cells, nuclei, etc.); i = a series of form numbers (i = 1, 2 . . . n); and d, = the specific weight of the form i for the given system (Cd, = 1). Karyometric measurements were made on the same cells on which nuclear polymorphism was determined. The largest nuclear diameter was measured utilizing light microscopy a t a magnification of x 1,500, according to the method described elsewhere (16). At least 1,000 colon cells and at least 500 hepatocytes were counted in each group of animals. Ecosoft computerizing system (USA) was used for statistical analysis of experimental data using t-test, chi-squared-test, and one-way analysis of variance (ANOVA) methods.
RESULTS Colon The morphometric studies showed that the most abundant form of normal colon epithelial cell nuclei was the oval form (Fig. l ) , and that the mean size of the nucleus was 7.9 rt 0.01 pm (Table 1). Following treatment with DCA, the colon tissue was not changed significantly as compared to controls (Table 1).In tumorous colon tissue obtained following treatment of animals with MNNG, a change in the polymorphism of nuclear shape was found: the incidence of long columnar nuclei increased significantly (Fig. 1). In parallel,
304
ZUSMAN ET AL.
Table 1 Mean Nuclear Diameter in Colon Epithelial Cells Following Different Treatments Treatment Control DCA (12 months) MNNG (tumorous tissue) MNNG + DCA (tumorous tissue)
01
Nuclear greater diameter (pm) 7.9 2 0.01* 7.8 2 0.01 8.6 k 0.01** 8.6 k 0.01**
*Mean k SE (n = 1,000). **Significantly different from control (Pt 5 0.01) and from P 5 0.01). DCA values (probability, 1.65 X
the mean nuclear size increased significantly (Table 1). In tumors obtained after treatment with MNNG and DCA, the number of different nuclear forms changed similarly to those observed after treatment with MNNG alone (Fig. 1). Similarly, the mean nuclear size in tumorous cells was the same as the following treatment with MNNG alone (Table 1).The coefficient of polymorphism (K) was very similar between all studied groups. The coefficient varied from 0.95 in tumors which were obtained after treatment with MNNG and DCA to 0.97 in tumors which were obtained after treatment with MNNG alone. In controls, the value of K was 0.96.
Liver
Nuclear form. In control animals, the main form of hepatocytes was round, while the other forms were found in very small percentages (Fig. 2). In livers which were obtained 12 months after treatment of the rats with MNNG, diminution in the percent of round nuclei and a significant increase in the percent of oval nuclei were seen (Fig. 2). The presence or absence of a tumor in the colon did not significantly alter the liver nuclear polymorphism for up to 2 months following treatment with MNNG: the coefficient of nuclear form polymorphism was not significantly different between the experimental groups. However, the coefficient calculated for each experimental group was significantly different from control values (Table 2). After prolonged treatment with DCA following previous treatment with MNNG, a significant increase was seen in the number of irregular nuclei (Fig. 3). There were no significant differences in the distribution of hepatocytic nuclei in animals without or with colon tumors (as a result of treatment with MNNG and DCA) (Fig. 3 ) . The coefficient of polymorphism in all these cases was significantly different not only from control values but from the value in livers which were obtained from rats after treatment for 12 months with DCA alone (Table 2). In the metastatic liver tumors, epithelial and mesenchymal-like cells showed a similar distribution of nuclear polymorphism and were significantly different from hepatocytes which were obtained from normal areas of the same livers (Fig. 4).The coefficient of polymorphism was similar in both types of liver tumorous
s
in A
13
B
c
n
FIG. 2. Distribution of hepatocytic nuclei according to their form in control (1)and dysplastic livers which were examined 12 months after treatment of animals with MNNG. 2, the presence of tumor in colon; 3, no tumor in colon. Types of nuclear forms: A, round; B, oval; C, long columnar; D, irregular.
cells (epithelial and connective) and significantly different from its value in normal areas and from control values (Table 2). Nuclear size. This parameter did not change as compared to control values after short term treatment of animals with MNNG with or without DCA, or with a bile acid alone (Table 3). After prolonged treatment with DCA, the mean nuclear size increased significantly, as compared to controls or to experimental groups after short term treatment, as well as to metastatic liver tumors. In metastatic liver tumors, a significant decrease in nuclear size was found in all areas studied (normal and tumorous) and in all types of cells studied (epithelial and mesenchymal-like) (Table 3).
DISCUSSION Our previous studies showed that short (2 weeks) intra-rectal treatment of rats with MNNG alone resulted in the development of colon carcinomatous tumors, and that subsequent treatment of animals with secondary bile acids could induce the development of metastatic liver tumors (Zimber, unpublished data). More recently, we found that relatively short (up to 6 months) treatment of rats with DCA following a 2week treatment with MNNG resulted in colon tumors and in dysplasia of the liver (Zusman, unpublished data). Moreover, liver dysplasia was found after feeding rats DCA (26). All these processes were accompa-
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS
305
Table 2 Coefficient of Nuclear Form Polymorphism i n Liver Hepatocytes Following Different Treatments
I
Treatment Control DCA (12 months) MNNG, no tumor in colon MNNG, tumor in colon MNNG + DCA (12 months), no tumor in colon MNNG + DCA (12 months), tumor in colon MNNG + DCA (12 months), metastatic tumor in liver Normal area Tumorous area, epithelial cells Tumorous area, mesenchymal cells
Coefficient (K) 0.48 0.71* 0.79* 0.72* 0.80“ 0.89**
0.51 0.98** 0.93”*
“Significantly different from control values (P5 0.05). **Significantly different from values in normal area (P 5 0.01).
marker for human hepatocellular carcinoma (6). A high correlation was found between the increase in nuclear polymorphism in transformed rat liver epithelial cells cultured in vitro and their tumorigenicity when injected into nude mice (25). A B C D The high correlation between nuclear size and clinFIG.3. Distribution of hepatocytic nuclei according to their form in ical prognosis was shown in patients with breast candysplastic livers which were obtained after prolonged (12 months) cer: carcinomas with large nuclear size (more than 146 treatment of animals with DCA alone (1)or following treatment with wm2) were lethal in 35% of patients, while among paMNNG without the development of colon tumor (2) or with induction tients with small nuclear size carcinomas (98 Fm2) of colon tumor (3). Types of nuclei as in Figure 2. death was found only in 6% (20). In another breast cancer study it was shown that the mean nuclear area and the number of mitoses per slide provided optimal nied by different changes in the form and size of the discrimination between different grades of cancer (10). Changes in nuclear size may be considered as charnuclei. Previously, it was shown that neoplastic lesions in acteristic for tumorous transformation, as was previdimethylhydrazine-induced colon tumors in mice ex- ously shown in studies of colon carcinogenesis (18,221. hibited varying degrees of dysplastic changes, which In our experiments, the carcinomatous changes of colon initially were characterized by the appearance of a ho- epithelial cells were accompanied by a significant inmogeneous population of atypical cells in the proximal crease in their mean size. Nuclear size in hepatocytes parts of crypts (8). Our morphometric studies showed decreased significantly in liver tumorous tissues, but that, in rats, large differences in cell shapes were increased significantly in livers obtained from rats present in control colon epithelial cells, and that this which showed signs of liver dysplasia (as a result of variability in nuclear form was not changed signifi- long treatment with a secondary bile acid alone or tocantly in tumors. Subsequently, the coefficient of nu- gether with a carcinogen) but without development of a clear form polymorphism was almost unchanged in co- tumor. This confirms the observations in the literature lon tumorous tissues as compared t o controls. In cases that development of liver dysplasia is accompanied by when hepatocellular dysplasia was caused by treat- an increase in nuclear size in hepatocytes in an early ment with secondary bile acids, the rate of nuclear stage in the development of liver carcinoma (1,2,23). polymorphism was higher as compared t o controls. The decrease in nuclear cell size in metastatic liver Similar observations of the increase of polymorphism tumors has been suggested to be a morphometric of cytological characteristics of malignant cells were marker which is specific to hepatocellular carcinomas described in other organs (4,12) and in cells cultured in (17). In conclusion, we suggest that nuclear form polymorvitro (11). It appears that nuclear polymorphism in hepatocytes phism may serve as a marker for pathological changes is more sensitive than nuclear size as a parameter for only in organs or in tissues which are characterized by evaluation of dysplastic changes: it changed signifi- low rate of polymorphism in the normal situation. Furcantly even in those cases when nuclear size was not thermore, changes in nuclear form polymorphism are changed relative t o controls. An increase in liver cell not specific for precancerous lesions in digestive organs polymorphism has been described as a characteristic but are indicative of their pathological lesions in gen-
ZUSMAN ET AL.
306
sions of digestive organs unless i t is combined with other analyses. Objective information about precancerous changes may be obtained only with a combined analysis using morphometric, histochemical, and immunocytochemical methods, as was postulated by Weber et al. (24).
ACKNOWLEDGMENTS We thank Dr. D.A. Beckman for his advice during revision of the manuscript and Dr. D. Kozlenko for his valuable assistance. LITERATURE CITED 1. Anthony PP: Precursor lesions for liver cancer in humans. Cancer Res 36:2579-2583, 1976. 2. Anthony PP, Vogel CL, Barker LF: Liver cell dysplasia: A premalignant condition. J Clin Pathol 26:217-223, 1973. 3. Aru A, Nielsen K: Stereological estimates of nuclear volume in primary lung cancer. Pathol Res Pract 185:735-739, 1989. 4. Benture Remacha ML, Colmenar Rodriguez L: Fundamentos y principios de 10s metodos planimetricos y citofotometricos. Oncologia 9:30-38, 1987. 5. Bibbo M, Montag AG, Lerma-Puertas E, Dytch HE, Leelakusolvong S, Bartels PH: Karyometric marker features in tissue adjacent to invasive cervical carcinomas. Anal Quant Cytol Histol 11:281-285, 1989. A B c 1) 6. Bottles K, Cohen MG, Holly EA, Chiu S-H, Abele JS, Cello JP, Lim RC, Miller T R A step-wise logistic regression analysis of FIG.4. Distribution of hepatocytic nuclei according to their form in hepatocellular carcinoma. An aspiration biopsy study. Cancer 62: different parts of a metastatic liver tumor which developed after 558-563, 1988. treatment of rats with MNNG and DCA for 12 months. 1, Normal 7. Carruba G, Pavone C, Pavone-Macaluso M, Mesiti M, d'Aquino A, area. Tumorous area: hepatocytes (2) and mesenchyme-like cells (3). Vita G, Sica G, Castagnetta L: Morphometry of in vitro systems. Types of nuclei as in Figure 2. An image analysis of two human prostate cancer cell lines (PC3 and DU-145). Pathol Res Pract 185:704-708, 1989. 8. Chang WWL: The mode of formation and progression of chemiTable 3 cally induced colonic carcinoma. In: Carcinoma of the Large Nuclear Diameter in Liver Hepatocytes Following Bowel and Its Precursors, Ingall JRF, Mastromarino AJ (eds). Different Treatments Alan R. Liss, Inc., New York, 1985, pp 217--235. 9. Dalrymple JC, Brough AK, Monaghan JM: Morphometric analNuclear ysis of nuclearicytoplasmic ratios in normal and perineoplastic Treatment diameter (pm) vulvar skin. Histopathology 14:645-653, 1989. Control 9.9 0.01* 10. Diest van PJ, Risse EKJ, Schipper NW, Baak JPA, Mouriquand MNNG DCA (12 months), metastatic J : Comparison of light microscopic grading and morphometric tumor in liver features in cytological breast cancer specimens. Pathol Res Pract Normal area 8.2 ? 0.01** 185:612-616, 1989. Tumorous area, epithelial cells 7.4 2 0.01** 11. Dufer J, Biaken D, Joly P, Benoist H, Carpenter Y, Desplaces A: Tumorous area, mesenchymal cells 7.7 & 0.01** Quantitative morphological aspects of granulocytic differentiaDCA (2 months) 9.4 ? 0.02 tion in HL-60 cells by dimethylsulfoxide and retinoic acid. Leuk 11.2 ? 0.01**~*** DCA (12 months) Res 13:621-627, 1989. 10.6 0.01**.*** 12. Gavin FM, Gray C, Sutton J , Clayden AD, Banks RI, Bird CC: MNNG, no tumor in colon 9.4 ? 0.02 MNNG, tumor in colon Morphometric differences between cytologically benign and maDCA (2 months), MNNG lignant serous effusions. Acta Cytol 32:175-182, 1988. no tumor in colon 9.5 ? 0.02 13. Kaplan BL, Kozlenko MD: On the method of a unitized measureMNNG DCA (12 months), ment of the cellular polymorphism degree in an early diagnosis of no tumor in colon 10.4 0.01**,*** cancer. Sov Med 10233-86, 1978. MNNG DCA (12 months), 14. Kelemen PR, Buschmann RJ, Weisz-Carrington P: Nucleolar tumor in colon 10.4 k 0.01**~*** prominence as a diagnostic variable in prostatic carcinoma. Cancer 65:1017-1020, 1990. *Mean 2 SE (n = 500). ""Significantly different from control values (P 2 5 0.01). 15. Layfield LJ,Hall TL, Fu YS: Discrimination of benign versus malignant mixed tumors of the salivary gland using digital im***Significantly different from experimental data from age analysis. Cytometry 10:217-221, 1989. groups with metastatic tumor in liver (P, 1.54 x lop7;P 5 16. Lee TK, Horner RD, Silverman JF, Jackson DV, Anderson-Goetz 0.01). D, Scarantino CW: Implications of nuclear diameter in small cell lung carinoma. Anal Quant Cytol Histol 12:78-84, 1990. 17. Matturri L, Bauer D: Morphometric characteristics of hepatocellular dysplasia. Anal Quant Cytol Histol 10:339-341, 1988. eral. These results lend support to the hypothesis that 18. Proudlock RJ, Allen JA: Micronuclei and other nuclear anomalies nuclear polymorphism cannot be utilized as a diagnosinduced in various organs by diethylnitrosamine and 7,12tic marker of precarcinomatous and carcinomatous ledimethylbenz(a)anthracence.Mutat Res 174:141-143, 1986.
+
*
*
+ +
+
*
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS 19. Robutti F, Pilato FP, Betta P - G A new method of grading malignancy of prostate carcinoma using quantitative microscopic nuclear features. Pathol Res Pract 185:701-703, 1989. 20. Schondorf H, Bastert G, Lobet M, Naujoks H: The nuclear area size as a parameter of malignancy in breast cancer. Geburtschilfe Frauenheilkd 49:272-276, 1989. 21. Sugar J, Molnar B, Szentirmay 2: DNA cytometry and morphometry by TV based image analysis system (Tas) in the diagnosis of gastric carcinoma. Anticancer Res 10:237-239, 1990. 22. Wargovich MJ, Goldberg MT, Newmark HL, Bruce W R Nuclear aberrations as a short-term test for genotoxicity t o the colon: Evaluation of nineteen agents in mice. J Natl Cancer Inst 71: 133-137, 1983. 23. Watanabe S, Ckita K, Harada T, Kodama T, Numa Y, Takemoto T, Takahashi T Morphological studies of the liver cell dysplasia. Cancer 51:2197-2205, 1983. 24. Weher T, Saeger W, Ludecke DK: Light microscopical morphom-
307
etry, immunocytochemistry, and clinical correlations of primary adenomas a t various stages of oncocytic transformation. Acta Endocrinol (Copenh) 116:489-496, 1987. 25. Worland PJ, Hampton LL, Thorgeirsson SS, Huggett AC: Development of an in vitro model of tumor progression using v-raf and v-rafiv-myc transformed rat liver epithelial cells: Correlation of tumorigenicity with the downregulation of specific proteins. Mol Carcinog 390-29, 1990. 26. Zimher A, Zusman I, Bentor R, Pinus H: Effects of lithocholic acid exposure throughout pregnancy on late prenatal and early postnatal development in rats. Teratology (in press). 27. Zusman I, Laschenko S, Pisarenko E: On the factors determinig the nucleus volume in early avian embryos. Arch Anat Histol Embryo1 64:50-55, 1973.
0 1991 Wiley-Liss, Inc
Nuclear Polymorphism and Nuclear Size in Precarcinomatous and Carcinomatous Lesions in Rat Colon and Liver I. Zusman,' M. Kozlenko, and A. Zimber Laboratory of Teratology and Experimental Oncology, Koret School of Veterinary Medicine (I.Z., M.K.), and Department of Animal Science, Faculty of Agriculture (A.Z.),Hebrew University of Jerusalem, Rehovot, Israel Received for publication September 4,1990; accepted December 18, 1990
The role of nuclear polymorphism and nuclear size in an analysis of the differences between colon and liver tumors in rats as well as in an analysis of hepatic dysplasia was studied. It was shown that colon tumors which developed following treatment of animals with a direct carcinogen alone or with a carcinogen followed by secondary bile acid were characterized by low incidence of nuclear polymorphism and by an increased nuclear size in epithelial cells. Metastatic liver tumors in rats with colon tumors were characterized by a high value for the coefficient of nuclear form polymorphism and by a significant decrease in nuclear size. Hepatic dysplasia which developed
In spite of the extensive use of histochemistry and immunohistochemistry in clinical histopathological practice, light microscopical morphometry also provides important information (24). For example, parameters such as cellular and nuclear areas, nuclear-cytoplasmic, and nucleolar-nuclear ratios have been utilized to characterize hepatocellular dysplasia (17). It was shown that morphometric examination complements histological evaluation and allows dysplastic cells to be identified (27). A correlation between nuclear size and the type of lung carcinoma (3,10,16) as well a s prognosis i n breast cancer (20) has been shown. Nuclear perimeter and nuclear optical density were shown to have diagnostic significance in identifying transformed cells of cervical carcinomas (51, salivary gland tumors (15), gastric carcinoma (21), or prostate carcinomas (14,191. The determination of shapes of cells and nuclei has been used in a n analysis of human hepatocellular carcinoma (6).It was shown, for example, that nuclear polymorphism is one of the characteristics of liver cell dysplasia (1,2), the first stage of which is characterized by enlarged nuclei and a n in-
as a result of prolonged treatment with secondary bile acids was characterized by high rate of nuclear form polymorphism and by a significant increase in nuclear size. The obtained results suggest that nuclear polymorphism is dependent upon the type of tissue or organ involved in the cancerous transformation and that it may have significance as a diagnostic marker of precarcinomatous and carcinomatous lesions of digestive organs only when used in combination with other analyses. Key terms: Rat colon and liver tumors, hepatic dysplasia, nuclear form polymorphism, coefficient of polymorphism
creased nuclear-cytoplasmic ratio (23). This latter parameter was also used to distinguish normal and perineoplastic skin in women (9). Differences in nuclear geometry and the number of cytoplasmic granules were described in HL-60 cells cultured in medium containing inducers of differentiation such as dimethyl sulfoxide and retinoic acid (11). Cell shape polymorphism was used as a diagnostic parameter to distinguish between two lines of human prostate cancerous cells (7). In this communication we studied the relation of nuclear shape and nuclear size to the type of the affected tissue and to the stage of its transformation. Rat colon and liver tumors, as well as rat livers undergoing pathological lesions and probably precancerous transformation, were used as a model object.
'Address for correspondence: Dr. I. Zusman. Koret School of Veterinary Medicine, Hebrew University of Jerusalem, Rehovot, P.O. Box 12, Israel 76100.
303
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS
MATERIALS AND METHODS Randomly bred 6-week old male Sprague Dawley rats (Anilab Co., Rehovot, Israel) were housed in plastic cages with stainless steel tops, and given a defined diet (610, Esem Hanegev Ltd., Israel) and water ad libitum. Temperature controlled (24 2°C) rooms with 12/12 darWlight schedule were used. Animals were treated intrarectally with deoxycholic secondary bile acid (DCA), with a direct carcinogen, N-methyl-N’-nitro-N-nitrosoguanidine (MNNG) alone, or with subsequent treatment with DCA a s a tumor promoter. Treatment with MNNG was for 2 weeks, while treatment with DCA was administered for 2 months or 12 months. Rats were killed by cervical dislocation 2 weeks and 2 or 12 months following the beginning of the treatment. The caudal segment of the distal colon (including the site of the drugs’ instillation) was opened longitudinally. Lesions a s well as normal appearing regions close to tumors were immediately fixed in phosphatebuffered saline (PBS)-buffered neutral 4% formaldehyde. Areas of the descending colon from DCA treated rats and from saline treated controls were also taken for histological examination. In parallel, the liver was taken for histological examination from experimental and control animals. Histological sections (3 pm) were stained with hematoxylin-eosin. The nuclear form of epithelial cells lining the crypts in the descending colon was studied in normal and adenocarcinomatous tissue according to the method described elsewhere (1,2). The following forms of nuclei were recorded: round and cuboid, oval, irregular, long columnar, and drop-like. These forms of nuclei were chosen based on the results of our previous study of the morphological changes in colon and liver cells obtained from rats chronically treated with secondary bile acids (Zimber and Zusman, unpublished data). Nuclei were studied in 20 randomly chosen crypts in each of the following groups of animals: 10 control colons, 10 colons from rats treated with DCA, 5 tumors obtained after the treatment with MNNG alone, and 5 tumors obtained after treatment with MNNG and DCA. The number of nuclei studied was no less than 3,000 in each group, i.e., the total amount of measured nuclei of colon cells was 12,000. In the liver, 4 nuclear forms of randomly chosen parenchymatous hepatocytes were recorded: round, oval, long columnar, and irregular. In each group of animals no less than 1,000 cells were studied, and the total number of measured hepatocytes was 4,000. A coefficient of polymorphism of forms (K) was calculated according to the equation derived by Kaplan and Kozlenko (13).
*
where K = the coefficient (factor) degree of polymorphism of forms (0 5 K 5 1); n = the number of forms
C
D
E
FIG.1. Distribution of colon epithelial cell nuclei according to their form in control (1) and tumorous tissues which developed following treatment with MNNG (2) or with MNNG and DCA (3). Types of nuclear forms: A, round and cuboidal; B, oval; C, irregular; D,long columnar; E,drop-like.
in a given system (cells, nuclei, etc.); i = a series of form numbers (i = 1, 2 . . . n); and d, = the specific weight of the form i for the given system (Cd, = 1). Karyometric measurements were made on the same cells on which nuclear polymorphism was determined. The largest nuclear diameter was measured utilizing light microscopy a t a magnification of x 1,500, according to the method described elsewhere (16). At least 1,000 colon cells and at least 500 hepatocytes were counted in each group of animals. Ecosoft computerizing system (USA) was used for statistical analysis of experimental data using t-test, chi-squared-test, and one-way analysis of variance (ANOVA) methods.
RESULTS Colon The morphometric studies showed that the most abundant form of normal colon epithelial cell nuclei was the oval form (Fig. l ) , and that the mean size of the nucleus was 7.9 rt 0.01 pm (Table 1). Following treatment with DCA, the colon tissue was not changed significantly as compared to controls (Table 1).In tumorous colon tissue obtained following treatment of animals with MNNG, a change in the polymorphism of nuclear shape was found: the incidence of long columnar nuclei increased significantly (Fig. 1). In parallel,
304
ZUSMAN ET AL.
Table 1 Mean Nuclear Diameter in Colon Epithelial Cells Following Different Treatments Treatment Control DCA (12 months) MNNG (tumorous tissue) MNNG + DCA (tumorous tissue)
01
Nuclear greater diameter (pm) 7.9 2 0.01* 7.8 2 0.01 8.6 k 0.01** 8.6 k 0.01**
*Mean k SE (n = 1,000). **Significantly different from control (Pt 5 0.01) and from P 5 0.01). DCA values (probability, 1.65 X
the mean nuclear size increased significantly (Table 1). In tumors obtained after treatment with MNNG and DCA, the number of different nuclear forms changed similarly to those observed after treatment with MNNG alone (Fig. 1). Similarly, the mean nuclear size in tumorous cells was the same as the following treatment with MNNG alone (Table 1).The coefficient of polymorphism (K) was very similar between all studied groups. The coefficient varied from 0.95 in tumors which were obtained after treatment with MNNG and DCA to 0.97 in tumors which were obtained after treatment with MNNG alone. In controls, the value of K was 0.96.
Liver
Nuclear form. In control animals, the main form of hepatocytes was round, while the other forms were found in very small percentages (Fig. 2). In livers which were obtained 12 months after treatment of the rats with MNNG, diminution in the percent of round nuclei and a significant increase in the percent of oval nuclei were seen (Fig. 2). The presence or absence of a tumor in the colon did not significantly alter the liver nuclear polymorphism for up to 2 months following treatment with MNNG: the coefficient of nuclear form polymorphism was not significantly different between the experimental groups. However, the coefficient calculated for each experimental group was significantly different from control values (Table 2). After prolonged treatment with DCA following previous treatment with MNNG, a significant increase was seen in the number of irregular nuclei (Fig. 3). There were no significant differences in the distribution of hepatocytic nuclei in animals without or with colon tumors (as a result of treatment with MNNG and DCA) (Fig. 3 ) . The coefficient of polymorphism in all these cases was significantly different not only from control values but from the value in livers which were obtained from rats after treatment for 12 months with DCA alone (Table 2). In the metastatic liver tumors, epithelial and mesenchymal-like cells showed a similar distribution of nuclear polymorphism and were significantly different from hepatocytes which were obtained from normal areas of the same livers (Fig. 4).The coefficient of polymorphism was similar in both types of liver tumorous
s
in A
13
B
c
n
FIG. 2. Distribution of hepatocytic nuclei according to their form in control (1)and dysplastic livers which were examined 12 months after treatment of animals with MNNG. 2, the presence of tumor in colon; 3, no tumor in colon. Types of nuclear forms: A, round; B, oval; C, long columnar; D, irregular.
cells (epithelial and connective) and significantly different from its value in normal areas and from control values (Table 2). Nuclear size. This parameter did not change as compared to control values after short term treatment of animals with MNNG with or without DCA, or with a bile acid alone (Table 3). After prolonged treatment with DCA, the mean nuclear size increased significantly, as compared to controls or to experimental groups after short term treatment, as well as to metastatic liver tumors. In metastatic liver tumors, a significant decrease in nuclear size was found in all areas studied (normal and tumorous) and in all types of cells studied (epithelial and mesenchymal-like) (Table 3).
DISCUSSION Our previous studies showed that short (2 weeks) intra-rectal treatment of rats with MNNG alone resulted in the development of colon carcinomatous tumors, and that subsequent treatment of animals with secondary bile acids could induce the development of metastatic liver tumors (Zimber, unpublished data). More recently, we found that relatively short (up to 6 months) treatment of rats with DCA following a 2week treatment with MNNG resulted in colon tumors and in dysplasia of the liver (Zusman, unpublished data). Moreover, liver dysplasia was found after feeding rats DCA (26). All these processes were accompa-
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS
305
Table 2 Coefficient of Nuclear Form Polymorphism i n Liver Hepatocytes Following Different Treatments
I
Treatment Control DCA (12 months) MNNG, no tumor in colon MNNG, tumor in colon MNNG + DCA (12 months), no tumor in colon MNNG + DCA (12 months), tumor in colon MNNG + DCA (12 months), metastatic tumor in liver Normal area Tumorous area, epithelial cells Tumorous area, mesenchymal cells
Coefficient (K) 0.48 0.71* 0.79* 0.72* 0.80“ 0.89**
0.51 0.98** 0.93”*
“Significantly different from control values (P5 0.05). **Significantly different from values in normal area (P 5 0.01).
marker for human hepatocellular carcinoma (6). A high correlation was found between the increase in nuclear polymorphism in transformed rat liver epithelial cells cultured in vitro and their tumorigenicity when injected into nude mice (25). A B C D The high correlation between nuclear size and clinFIG.3. Distribution of hepatocytic nuclei according to their form in ical prognosis was shown in patients with breast candysplastic livers which were obtained after prolonged (12 months) cer: carcinomas with large nuclear size (more than 146 treatment of animals with DCA alone (1)or following treatment with wm2) were lethal in 35% of patients, while among paMNNG without the development of colon tumor (2) or with induction tients with small nuclear size carcinomas (98 Fm2) of colon tumor (3). Types of nuclei as in Figure 2. death was found only in 6% (20). In another breast cancer study it was shown that the mean nuclear area and the number of mitoses per slide provided optimal nied by different changes in the form and size of the discrimination between different grades of cancer (10). Changes in nuclear size may be considered as charnuclei. Previously, it was shown that neoplastic lesions in acteristic for tumorous transformation, as was previdimethylhydrazine-induced colon tumors in mice ex- ously shown in studies of colon carcinogenesis (18,221. hibited varying degrees of dysplastic changes, which In our experiments, the carcinomatous changes of colon initially were characterized by the appearance of a ho- epithelial cells were accompanied by a significant inmogeneous population of atypical cells in the proximal crease in their mean size. Nuclear size in hepatocytes parts of crypts (8). Our morphometric studies showed decreased significantly in liver tumorous tissues, but that, in rats, large differences in cell shapes were increased significantly in livers obtained from rats present in control colon epithelial cells, and that this which showed signs of liver dysplasia (as a result of variability in nuclear form was not changed signifi- long treatment with a secondary bile acid alone or tocantly in tumors. Subsequently, the coefficient of nu- gether with a carcinogen) but without development of a clear form polymorphism was almost unchanged in co- tumor. This confirms the observations in the literature lon tumorous tissues as compared t o controls. In cases that development of liver dysplasia is accompanied by when hepatocellular dysplasia was caused by treat- an increase in nuclear size in hepatocytes in an early ment with secondary bile acids, the rate of nuclear stage in the development of liver carcinoma (1,2,23). polymorphism was higher as compared t o controls. The decrease in nuclear cell size in metastatic liver Similar observations of the increase of polymorphism tumors has been suggested to be a morphometric of cytological characteristics of malignant cells were marker which is specific to hepatocellular carcinomas described in other organs (4,12) and in cells cultured in (17). In conclusion, we suggest that nuclear form polymorvitro (11). It appears that nuclear polymorphism in hepatocytes phism may serve as a marker for pathological changes is more sensitive than nuclear size as a parameter for only in organs or in tissues which are characterized by evaluation of dysplastic changes: it changed signifi- low rate of polymorphism in the normal situation. Furcantly even in those cases when nuclear size was not thermore, changes in nuclear form polymorphism are changed relative t o controls. An increase in liver cell not specific for precancerous lesions in digestive organs polymorphism has been described as a characteristic but are indicative of their pathological lesions in gen-
ZUSMAN ET AL.
306
sions of digestive organs unless i t is combined with other analyses. Objective information about precancerous changes may be obtained only with a combined analysis using morphometric, histochemical, and immunocytochemical methods, as was postulated by Weber et al. (24).
ACKNOWLEDGMENTS We thank Dr. D.A. Beckman for his advice during revision of the manuscript and Dr. D. Kozlenko for his valuable assistance. LITERATURE CITED 1. Anthony PP: Precursor lesions for liver cancer in humans. Cancer Res 36:2579-2583, 1976. 2. Anthony PP, Vogel CL, Barker LF: Liver cell dysplasia: A premalignant condition. J Clin Pathol 26:217-223, 1973. 3. Aru A, Nielsen K: Stereological estimates of nuclear volume in primary lung cancer. Pathol Res Pract 185:735-739, 1989. 4. Benture Remacha ML, Colmenar Rodriguez L: Fundamentos y principios de 10s metodos planimetricos y citofotometricos. Oncologia 9:30-38, 1987. 5. Bibbo M, Montag AG, Lerma-Puertas E, Dytch HE, Leelakusolvong S, Bartels PH: Karyometric marker features in tissue adjacent to invasive cervical carcinomas. Anal Quant Cytol Histol 11:281-285, 1989. A B c 1) 6. Bottles K, Cohen MG, Holly EA, Chiu S-H, Abele JS, Cello JP, Lim RC, Miller T R A step-wise logistic regression analysis of FIG.4. Distribution of hepatocytic nuclei according to their form in hepatocellular carcinoma. An aspiration biopsy study. Cancer 62: different parts of a metastatic liver tumor which developed after 558-563, 1988. treatment of rats with MNNG and DCA for 12 months. 1, Normal 7. Carruba G, Pavone C, Pavone-Macaluso M, Mesiti M, d'Aquino A, area. Tumorous area: hepatocytes (2) and mesenchyme-like cells (3). Vita G, Sica G, Castagnetta L: Morphometry of in vitro systems. Types of nuclei as in Figure 2. An image analysis of two human prostate cancer cell lines (PC3 and DU-145). Pathol Res Pract 185:704-708, 1989. 8. Chang WWL: The mode of formation and progression of chemiTable 3 cally induced colonic carcinoma. In: Carcinoma of the Large Nuclear Diameter in Liver Hepatocytes Following Bowel and Its Precursors, Ingall JRF, Mastromarino AJ (eds). Different Treatments Alan R. Liss, Inc., New York, 1985, pp 217--235. 9. Dalrymple JC, Brough AK, Monaghan JM: Morphometric analNuclear ysis of nuclearicytoplasmic ratios in normal and perineoplastic Treatment diameter (pm) vulvar skin. Histopathology 14:645-653, 1989. Control 9.9 0.01* 10. Diest van PJ, Risse EKJ, Schipper NW, Baak JPA, Mouriquand MNNG DCA (12 months), metastatic J : Comparison of light microscopic grading and morphometric tumor in liver features in cytological breast cancer specimens. Pathol Res Pract Normal area 8.2 ? 0.01** 185:612-616, 1989. Tumorous area, epithelial cells 7.4 2 0.01** 11. Dufer J, Biaken D, Joly P, Benoist H, Carpenter Y, Desplaces A: Tumorous area, mesenchymal cells 7.7 & 0.01** Quantitative morphological aspects of granulocytic differentiaDCA (2 months) 9.4 ? 0.02 tion in HL-60 cells by dimethylsulfoxide and retinoic acid. Leuk 11.2 ? 0.01**~*** DCA (12 months) Res 13:621-627, 1989. 10.6 0.01**.*** 12. Gavin FM, Gray C, Sutton J , Clayden AD, Banks RI, Bird CC: MNNG, no tumor in colon 9.4 ? 0.02 MNNG, tumor in colon Morphometric differences between cytologically benign and maDCA (2 months), MNNG lignant serous effusions. Acta Cytol 32:175-182, 1988. no tumor in colon 9.5 ? 0.02 13. Kaplan BL, Kozlenko MD: On the method of a unitized measureMNNG DCA (12 months), ment of the cellular polymorphism degree in an early diagnosis of no tumor in colon 10.4 0.01**,*** cancer. Sov Med 10233-86, 1978. MNNG DCA (12 months), 14. Kelemen PR, Buschmann RJ, Weisz-Carrington P: Nucleolar tumor in colon 10.4 k 0.01**~*** prominence as a diagnostic variable in prostatic carcinoma. Cancer 65:1017-1020, 1990. *Mean 2 SE (n = 500). ""Significantly different from control values (P 2 5 0.01). 15. Layfield LJ,Hall TL, Fu YS: Discrimination of benign versus malignant mixed tumors of the salivary gland using digital im***Significantly different from experimental data from age analysis. Cytometry 10:217-221, 1989. groups with metastatic tumor in liver (P, 1.54 x lop7;P 5 16. Lee TK, Horner RD, Silverman JF, Jackson DV, Anderson-Goetz 0.01). D, Scarantino CW: Implications of nuclear diameter in small cell lung carinoma. Anal Quant Cytol Histol 12:78-84, 1990. 17. Matturri L, Bauer D: Morphometric characteristics of hepatocellular dysplasia. Anal Quant Cytol Histol 10:339-341, 1988. eral. These results lend support to the hypothesis that 18. Proudlock RJ, Allen JA: Micronuclei and other nuclear anomalies nuclear polymorphism cannot be utilized as a diagnosinduced in various organs by diethylnitrosamine and 7,12tic marker of precarcinomatous and carcinomatous ledimethylbenz(a)anthracence.Mutat Res 174:141-143, 1986.
+
*
*
+ +
+
*
NUCLEAR POLYMORPHISM IN RAT COLON AND LIVER TUMORS 19. Robutti F, Pilato FP, Betta P - G A new method of grading malignancy of prostate carcinoma using quantitative microscopic nuclear features. Pathol Res Pract 185:701-703, 1989. 20. Schondorf H, Bastert G, Lobet M, Naujoks H: The nuclear area size as a parameter of malignancy in breast cancer. Geburtschilfe Frauenheilkd 49:272-276, 1989. 21. Sugar J, Molnar B, Szentirmay 2: DNA cytometry and morphometry by TV based image analysis system (Tas) in the diagnosis of gastric carcinoma. Anticancer Res 10:237-239, 1990. 22. Wargovich MJ, Goldberg MT, Newmark HL, Bruce W R Nuclear aberrations as a short-term test for genotoxicity t o the colon: Evaluation of nineteen agents in mice. J Natl Cancer Inst 71: 133-137, 1983. 23. Watanabe S, Ckita K, Harada T, Kodama T, Numa Y, Takemoto T, Takahashi T Morphological studies of the liver cell dysplasia. Cancer 51:2197-2205, 1983. 24. Weher T, Saeger W, Ludecke DK: Light microscopical morphom-
307
etry, immunocytochemistry, and clinical correlations of primary adenomas a t various stages of oncocytic transformation. Acta Endocrinol (Copenh) 116:489-496, 1987. 25. Worland PJ, Hampton LL, Thorgeirsson SS, Huggett AC: Development of an in vitro model of tumor progression using v-raf and v-rafiv-myc transformed rat liver epithelial cells: Correlation of tumorigenicity with the downregulation of specific proteins. Mol Carcinog 390-29, 1990. 26. Zimher A, Zusman I, Bentor R, Pinus H: Effects of lithocholic acid exposure throughout pregnancy on late prenatal and early postnatal development in rats. Teratology (in press). 27. Zusman I, Laschenko S, Pisarenko E: On the factors determinig the nucleus volume in early avian embryos. Arch Anat Histol Embryo1 64:50-55, 1973.
