Sequential alteration of apoptosis, p53 expression, and cell proliferation in the rat pancreas treated with 4-hydroxyaminoquinoline 1-oxide.

Changes in p53 expression, apoptosis and cell proliferation after treatment with 4-hydroxyaminoquinoline 1-oxide (4HAQO) were investigated in the rat pancreas and liver, target and nontarget organs for tumorigenesis, respectively. Male rats were given a single intravenous injection of 4HAQO at a dose of 20 mg/kg body weight and control rats received vehicle alone and were euthanized after 2-72 hours. Pancreata and livers were removed for histopathological examination, immunohistochemistry for p53 protein, PCNA and Ki-67, and TUNEL labeling and electron microscopic observation for detecting apoptosis. In the pancreas, p53 expression and apoptosis were significantly increased first at 4 and 6 hours, respectively, while no change was evident in the liver. The rates peaked at 24 hours, consistent with the peak for PCNA-labeling, while Ki-67-labeling rates peaked at 72 hours. Electron microscopically, apoptotic changes in pancreatic acinar cells were observed after 2 hours. No significant apoptosis, p53 expression or cell proliferation were noted in the pancreatic tissues of the control rats nor in liver cells regardless of 4HAQO treatment. Taken together with our previous data, the results suggest that apoptosis, p53 expression, and enhanced cell replication are closely related phenomena involved in the carcinogenesis of 4HAQO following DNA adduct formation.

Induction of pancreatic islet cell tumors in rats by repeated intravenous administration of 4-hydroxyaminoquinoline 1-oxide.

The inducibility of pancreatic islet cell tumors by administration of 4-hydroxyaminoquinoline 1-oxide (4HAQO) was investigated in male 6-week-old Sprague-Dawley rats. Rats were given 4HAQO intravenously at a weekly dose of 5 mg/kg 4 times (group 1) or a single dose of 10 mg/kg (group 2). Control rats received the vehicle alone (group 3). Fifty-six weeks after the first 4HAQO administration, all surviving animals were killed and the pancreas was examined histopathologically, immunohistochemically and ultrastructurally. The incidences and multiplicities of islet cell tumors in groups 1, 2, and 3 were 52.3% (p < 0.05 vs group 2, p < 0.01 vs group 3), 19.2% and 0%, and 0.70/animal (p < 0.05 vs group 2, p < 0.01 vs group 3), 0.23 and 0, respectively. Islet cell carcinomas were induced only in group 1, accounting for 6/44 (26%) tumors. Islet cell hyperplasias were found in 61.4% (p < 0.05 vs group 3), 42.3% and 10.0% of groups 1, 2, and 3, with multiplicities of 0.95 (p < 0.05 vs groups 2 and 3), 0.54 and 0.20, respectively. As compared with normal islets from control subjects, islet cell tumors showed an increase in the number of insulin positive cells associated with cytological features indicative of enhanced insulin synthesis and secretion, and a decrease in the number of glucagon positive cells without ultrastructural signs of modified secretory activity. Thus our results indicate that repeated intravenous administration of 4HAQO to rats is useful for the induction of islet cell tumors at high incidence.

Dual effects of prolonged ACTH stimulation on 4-hydroxyaminoquinoline 1-oxide-induced adrenocortical lesions in rats.

The effects of a long-acting synthetic ACTH on 4-hydroxyaminoquinoline 1-oxide (4HAQO)-induced adrenocortical lesions were investigated in female rats. A total of 140 6-week-old rats were divided into 4 equal groups, given a single s.c. injection of 7 mg/kg 4HAQO or vehicle, followed by repeated sc administration of the synthetic ACTH or no further treatment. Subgroups of 10 rats in each group were sequentially sacrificed at weeks 20, 30, and 40. Adenomas and adenomatous nodules developed in the adrenal cortex of animals receiving 4HAQO and the chronic ACTH stimulation. Both lesions were located in the deeper zones of the adrenal cortex adjacent to the medulla and were composed of large-sized, clear-type cells. From week 20, middle zone, cortical cystic degeneration, which mimics the age-associated degenerative change named adrenal peliosis, was frequently observed in the adrenal glands of animals treated with 4HAQO alone. Its development was inhibited by ACTH. In the control animals, peliotic changes occurred at low incidence and only at the termination of experiment. These results indicate that long-term stimulation of ACTH promotes the development of adrenocortical tumors but suppresses the occurrence of adrenal peliosis in rats treated with 4HAQO.

Nucleolar segregation as an early marker for DNA damage; an experimental study in rats treated with 4-hydroxyaminoquinoline 1-oxide.

Male 6-week-old Sprague Dawley rats were given a single intravenous injection of 4-hydroxyaminoquinoline 1-oxide (4HAQO) at a dose of 20 mg/kg in order to produce ultrastructural changes as possible morphological biomarkers for toxicity. Immunohistochemically demonstrated formation of 4HAQO-DNA adduct was correlated with the changes found. Nucleolar alteration, demonstrable by electron microscopy as segregation of nucleolar components into granular and fibrillar compartments, was evident in cells of the target organs, exocrine pancreas and adrenocortex, but not of the non-target liver parenchyma. Sequential observation clarified that such alteration was highest in frequency 6 h and 4 h after 4HAQO administration in pancreatic acinar cells and adrenocortical cells respectively. Electron microscopically, apoptotic changes of acinar cells were evident 2 h after injection of 4HAQO. DNA adduct formation was consistently demonstrated in the same target organs showing nucleolar segregation, the highest frequency being noted 4 h after 4HAQO treatment in both pancreatic acinar cells and adrenocortical cells. Our results thus indicate an identity of the target cells for nucleolar segregation and 4HAQO-DNA adduct formation which correlates with 4HAQO-toxicity. We suggest that nucleolar segregation occurs subsequent to the generation of DNA damage.

Site-specific DNA damage and 8-hydroxydeoxyguanosine formation by hydroxylamine and 4-hydroxyaminoquinoline 1-oxide in the presence of Cu(II): role of active oxygen species.

Mutagenic hydroxylamine (NH2OH) and 4-hydroxyamino-quinoline 1-oxide (4-HAQO), a carcinogenic metabolite of 4-nitroquinoline 1-oxide (4-NQO), cleaved isolated DNA in the presence of Cu(II), but not in the presence of Mn(II), Mn(III), Fe(II) or Fe(III). The Cu(II)-mediated DNA damage by NH2OH was inhibited by catalase and bathocuproine, a Cu(I)-specific chelator, but not by scavengers of hydroxyl free radical. With the Cu(II)-mediated DNA damage by 4-HAQO, similar scavenger effects were observed. It is suggested that free .OH is not the main active species causing the DNA damage in both the cases. The predominant cleavage sites were thymine residues, especially the thymine residue of 5'-GTC-3' sequence. Since the cleavage pattern was similar to that induced by Cu(I) plus H2O2 but not to that induced by Cu(II) plus H2O2, it is speculated that the copper-oxygen complex derived from the reaction of H2O2 with Cu(I) participates in the DNA damage. 8-Hydroxydeoxyguanosine (8-OH-dG) residues were efficiently formed in calf thymus DNA treated with NH2OH plus Cu(II) or 4-HAQO plus Cu(II). The role of Cu(II)-mediated DNA damage and 8-OH-dG formation in the genotoxicity of NH2OH, 4-HAQO and 4-NQO is discussed.

4-Hydroxyaminoquinoline 1-oxide can convert the Z-form of poly(dG-m5dC).poly(dG-m5dC) back to B-form-DNA.

Reactions between the chemical carcinogen 4-hydroxyaminoquinoline 1-oxide (4HAQO) and poly(dG-dC).poly(dG-dC), poly(dG-m5dC).poly(dG-m5dC) and poly(dG-br5dC).poly(dG-br5dC) were studied under B-form and Z-form conformation conditions. 4HAQO bound covalently to both B-form and Z-form-DNA. The extent of the B-form to Z-form-DNA transition was diminished by 4HAQO, as inferred by spectroscopic methods. Moreover, conversion of the Z-form-DNA back to a right handed DNA close to the B-form was observed with this carcinogen.

Temporal dissociation between cell proliferation and eosinophilic foci development in the exocrine pancreas of rats initiated with 4-hydroxyaminoquinoline 1-oxide and administered soybean trypsin inhibitor.

Male Sprague-Dawley rats received a single i.v. injection of 4-hydroxyaminoquinoline 1-oxide (HAQO) 7 mg/kg body weight or vehicle alone, and starting 7 days thereafter, were then fed basal diet with or without a 5% soybean trypsin inhibitor (SBTI) supplement. Subgroups were sequentially killed after 2, 4, 7, 14, 30, 60 and 100 days on this regimen, in each case 1 h after injection of bromodeoxyuridine (BrdU). In the HAQO/SBTI and SBTI alone groups, 2 days after the SBTI treatment the labeling indices of acinar cells were increased approximately 12- and 11-fold respectively, dropping rapidly thereafter and returning to the control value by day 30. The earliest eosinophilic foci were noted in the HAQO/SBTI group 60 days after HAQO initiation, with the component cells demonstrating markedly increased labeling indices in contrast to the completely normalized levels observed in the surrounding exocrine tissue. On the other hand, eosinophilic foci were scarcely induced in the HAQO or SBTI alone group throughout the experiment. These results thus indicate a clear temporal dissociation between initial proliferation in parenchymal pancreatic tissue caused by SBTI and subsequent development of eosinophilic foci in rats initiated with HAQO.

Modifying effects of soybean trypsin inhibitor on development of eosinophilic nodules and basophilic foci in the exocrine pancreas of male Sprague-Dawley rats treated with 4-hydroxyaminoquinoline 1-oxide.

Administration of 4-hydroxyaminoquinoline 1-oxide (HAQO) to rats results in development of 2 types of pancreatic acinar lesions, namely eosinophilic nodules and basophilic foci. To cast light on the biological character of these lesions, 5-week-old male Sprague-Dawley rats were given a single intravenous injection of HAQO at a dose of 7 mg/kg and, thereafter, fed soybean trypsin inhibitor (SBTI) at dose levels of 10% and 5%. At week 57, all rats were killed for pathological examination of pancreatic tissue. The incidence of eosinophilic nodules was significantly higher in the HAQO/SBTI group than in the HAQO-alone group, whereas the basophilic acinar foci were observed to occur less frequently and to be smaller in the HAQO/SBTI-treated animals. Administration of SBTI is known to increase the blood level of cholecystokinin, a trophic factor for pancreatic acinar cells. Thus, the present findings suggest that long-term elevation of this endocrine factor can affect the two types of pancreatic acinar lesions in essentially different ways, namely enhancing development of eosinophilic nodules, while suppressing the occurrence of basophilic foci.

Morphological, biochemical and autoradiographic studies of pancreatic acinar cell necrosis and its regeneration induced by 4-hydroxyaminoquinoline-1-oxide in rats.

A single intravenous injection of 4-hydroxyaminoquinoline-1-oxide (4-HAQO) induced a dose-dependent diffuse pancreatic acinar cell necrosis within 48 hours after the administration in rats. Regeneration following the necrosis ensued by 72 hours. The serum amylase levels were markedly elevated within 24 hours, which paralleled a decrease of digestive enzyme contents in the pancreas. The change of these biochemical parameters corresponded well with the morphological change of the pancreas. Autoradiographic studies with 3H-thymidine revealed that the labeling index of the regenerating acinar cells was considerably higher, reaching a peak value of nearly 18.2% at 72 hours after injection of 4-HAQO and it still remained high in 2.67% even at 168 hours after injection.

5-Chloro-7-iodo-8-hydroxyquinoline (clioquinol) inhibits the nerve growth factor-induced stimulation of RNA synthesis in neonatal rat superior cervical ganglion, in vitro--comparison with effects of methylmercuric chloride and 4-hydroxyaminoquinoline-N-oxide.

The inhibitory effects of 5-chloro-7-iodo-8-hydroxy-quinoline (clioquinol), methylmercuric chloride and 4-hydroxyaminoquinoline-N-oxide(4-HAQO) on DNA, RNA and protein syntheses in the neonatal rat superior cervical ganglion (SCG) were studied in relation to the action of mouse 2.5S nerve growth factor (NGF), using organ cultures. RNA and protein syntheses in SCG were stimulated approximately 3- and 2-fold, respectively, by NGF (1 microgram/ml), but the DNA synthesis was only slightly or not at all stimulated. Methylmercuric chloride and 4-HAQO dose-dependently inhibited DNA, RNA and protein syntheses, either in the presence or in the absence of NGF. On the other hand, clioquinol (up to 100 microM) slightly or not at all inhibited RNA synthesis in the absence of NGF; however, it did abolish the NGF-induced stimulation of RNA synthesis in the presence of NGF. The DNA and protein syntheses were dose-dependently inhibited by clioquinol, either in the presence or in the absence of NGF. We conclude from this study that the interaction between clioquinol and the functions of NGF raises the question of a possible toxicity of the drug on specific neurons.

Atypical acinar cell lesions of the pancreas in mice induced by 4-hydroxyaminoquinoline-1-oxide.

The carcinogenic effect of 4-hydroxyaminoquinoline-1-oxide (4-HAQO) on exocrine pancreas of mice after a single intravenous (i.v.) injection was investigated. At a dose of 24 mg/kg body weight, 100% of the mice killed at the end of 40 weeks had developed atypical acinar cell foci (AAF) in the pancreas. Each pancreas contained an average of 18.3 +/- 2.7 AAF. Cells in AAF were arranged as acini and their cytoplasm contained eosinophilic zymogen granules. The nuclei were large, round to oval, basally located and contained 1-2 prominent nucleoli. Cells in AAF showed a markedly increased mitotic activity with an index of 4 +/- 0.7/1000 cells as compared to a value of 0.25 +/- 0.25 in the acinar cell population of control pancreas. Autoradiographic studies showed a labeling index of 23 +/- 3 in AAF as compared to only 1.5 +/- 0.29 in the pancreatic acinar cells of controls. These findings indicate that 4-HAQO induces pancreatic lesions in mice which are quite similar to the acidophilic foci observed in rat pancreas and may serve as another useful model for studies of pancreatic carcinogenesis.

Spontaneous neurinoma in an African lungfish, Protopterus annectens, and DNA repair studies on normal and neoplastic tissues.

Neurinomas developed in an African lungfish (Protopterus annectens), living in an aquarium in Western Japan. The 2 tumors, measuring 7.5 X 9.0 X 6.5 and 13 X 4 X 6 cm, were located on the skin. As shown by light microscopy, tumor cells were composed of spindle-shaped cells with huge pleomorphic nuclei, which were arranged in parallel rows or whorls in interlacing connective tissue. Long-term culture of these tumor cells was achieved in vitro at 25 degrees C with use of conditioned medium over a period of more than 4 months. The nuclear DNA contents of erythrocytes (normal diploids, 2C) and tumor cells dispersed from the fixed tumor tissues were measured by 4',6-diamidino-2-phenylindole hydrochloride-DNA microfluorometry by using mouse cerebellar small granule cells (normal diploids, 2C) as a reference. The 2C value of the lungfish was approximately 28-fold greater than that of the mouse. Furthermore, consistent with the nuclear pleomorphism observed by light microscopy, the nuclear DNA contents of tumor cells showed a wide distribution from hypo-2C to hyper-4C. DNA repair synthesis was measured autoradiographically in organ cultures of the tumor, lung, and skin, exposed to chemical carcinogens or UV radiation. Considerable repair was observed in the tumor and skin cells exposed to 1-methyl-1-nitrosourea (CAS: 684-93-5), N-methyl-N'-nitro-N-nitrosoguanidine (CAS: 70-25-7), or 254-nm or sunlamp UV light. Only traces of repair synthesis were detected in lung exposed to 1-methyl-1-nitrosourea or N-methyl-N'-nitro-N-nitrosoguanidine. 4-Hydroxyaminoquinoline 1-oxide (CAS: 4637-56-3) did not induce repair in any of the three tissues. The observed values for repair, relative to the amount of DNA, were similar to those in other fishes.

4-Hydroxyaminoquinoline 1-oxide metabolism and DNA adducts in the early stage of tumorigenesis in rats: comparison of target organ pancreas with non-target organ liver.

In order to probe key early molecular events which might be responsible for the initiation of rat pancreatic tumorigenesis by 4-hydroxyaminoquinoline 1-oxide (4-HAQO), the uptake and metabolism of carcinogen and the formation and subsequent repair of DNA adducts were monitored under conditions of high and low tumorigenicity, respectively in partially pancreatectomized and non-operated animals, and in the liver, a non-target organ for this carcinogen. Although uptake of radioactively labelled 4-HAQO was higher in the liver than in the pancreas, generation of DNA adducts was 20 times greater in the latter organ. This discrepancy was probably due to a difference in the metabolic profile of 4-HAQO. The spectrum of the adducts was qualitatively similar in both organs. No qualitative or quantitative differences could be established under the high and low tumorigenicity conditions with regard to DNA adduct formation or persistence. The major difference was the presence of a relatively large extent of pancreatic DNA replication under the high tumorigenic condition. The results indicated that metabolic profile of 4-HAQO, quantity of DNA adducts and levels of DNA replication are key factors involved in initiation of tumorigenesis.

Hypertrophic foci of pancreatic acinar cells in rats.

A morphologically distinctive type of pancreatic acinar cell foci, different from hyperplastic nodules and adenomas, in rats has been recognized for two decades. The lesions have been observed to occur spontaneously and to be induced experimentally. They consist of enlarged acinar cells with abundant cytoplasm of altered staining characteristics and prominent nuclei. There is, however, a wide divergence of opinion among investigators regarding the nature of the lesions. As a result of different interpretations and classifications, many terms have been given to them. Based on morphologic characteristics, the author has designated the lesions as hypertrophic foci, a descriptive morphological term. The biologic significance with particular reference to age and the relationship with acinar cell neoplasia is discussed. Also included in the review are similar lesions in other rodent species.

Effect of norharman on DNA strand breaks and mutation of Chinese hamster V79 cells by chemical carcinogens.

The effect of norharman, which shows a comutagenic activity in the Salmonella mutation assay, was examined on the action of mutagens towards Chinese hamster V79 cells. Norharman reduced the DNA strand breaks by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), but enhanced the 4-hydroxyaminoquinoline l-oxide ( 4HAQO )-induced DNA strand breaks. Norharman also reduced the cytotoxicity and the mutagenicity of MNNG but enhanced the cytotoxicity of 4HAQO to V79 cells. In the Salmonella mutation assay, norharman showed no effect on the mutagenic activity of MNNG but reduced the mutagenic activity of 4HAQO . The effect of norharman on the action of mutagens to V79 cells appeared dependent on mutagens and did not correlate with the effect observed in Salmonella mutation assay.

Proliferative lesions of the exocrine pancreas in male F344/N rats.

While the rat pancreas is susceptible to experimental cancer induction, the spontaneous incidence of pancreatic cancer in this species is reported to be very low. However, we observed unusually high incidences of focal acinar hyperplasia and acinar adenoma in vehicle control male F344/N rats of some NCI/NTP 2-year toxicological studies. The vehicle in these studies was corn oil given by gavage. Focal acinar hyperplasia, acinar adenoma, and acinar carcinoma (found rarely) represent a continuous spectrum of proliferative lesions of the exocrine pancreas. While the carcinomas have clear morphological indications of malignancy, the biological behavior of focal acinar hyperplasia and acinar adenoma is not known. Although induction of acinar carcinomas is considered clear evidence of carcinogenicity of a test chemical, significantly increased incidences in treated rats of acinar adenomas but not carcinomas provides some evidence of carcinogenicity. The association of acinar hyperplasia and adenoma with vegetable oil gavage complicates the interpretation when marginally elevated incidences of these lesions are observed in rats administered the test chemical in vegetable oil vehicle. Studies of the biological behavior of exocrine pancreatic lesions in male rats would be helpful in assessing the significance of their presence when found after test compound administration.

Arrest of DNA elongation by DNA polymerases at guanine adducts on 4-hydroxyaminoquinoline 1-oxide-modified DNA template.

In vitro modification of M13 phage single-stranded DNA with 4-hydroxyaminoquinoline 1-oxide (4HAQO) resulted in four kinds of adducts: three guanine adducts, QGI, QGII, and QGIII; and one adenine adduct, QA, at ratios of 16.4 47.3, 13.7, and 22.6, respectively. The carcinogen-modified DNA, initiated with a sequence-defined oligodeoxynucleotide primer, was replicated in vitro with Escherichia coli DNA polymerase I (Klenow fragment) and calf thymus DNA polymerases alpha and beta. The reaction products were analyzed on a DNA-sequencing gel. DNA elongation by DNA polymerase I was arrested at putative guanine adducts on the template in three ways: at one base prior to guanine; at positions opposite to guanine; and at one base beyond guanine. Similar patterns of elongation arrest were also obtained with the mammalian DNA polymerases alpha and beta. In contrast to guanine adducts, the adenine adduct, QA, might lack the capacity to arrest DNA chain elongation by DNA polymerases.

Effects of caffeine on pancreatic tumorigenesis by 4-hydroxyaminoquinoline 1-oxide in partially pancreatectomized rats.

The effects of caffeine on pancreatic tumorigenesis by 4-hydroxyaminoquinoline 1-oxide (4-HAQO) and on pancreatic DNA synthesis were studied in partially pancreatectomized male Wistar rats. 4-HAQO was injected i.v. as a single dose of 7 mg/kg body weight 3 days after partial pancreatectomy. Caffeine was injected s.c. every 12 h at the maximum tolerated dose (m.t.d.) of 120 mg/kg body weight, half the m.t.d., and one quarter the m.t.d. from 12 to 72 h before and 0 to 72, 72 to 132, and 0 to 132 h after 4-HAQO treatment. Post-treatment with caffeine from 0 to 132 h had a dose-dependent biphasic effect on pancreatic tumorigenesis: post-treatment with the m.t.d. of caffeine decreased the total number of nodules, whereas treatment with one quarter the m.t.d. of caffeine increased their number. Decrease in the number of nodules was also observed on post-treatment with the m.t.d. of caffeine from 0 to 72 or from 72 to 132 h. Pretreatment with the m.t.d. of caffeine had no significant effect on the number of nodules. Recovery of pancreatic DNA synthesis was slower after simultaneous treatment with the m.t.d. of caffeine and 4-HAQO than after treatment with 4-HAQO alone. The possible mechanism of the effect of caffeine on pancreatic tumorigenesis induced by 4-HAQO in rats is discussed.

Malignant fibrous histiocytomas induced by 4-(hydroxyamino)quinoline 1-oxide in rats.

Subcutaneous and bone malignant fibrous histiocytomas (MFH) were induced in high incidence in rats by 4-(hydroxyamino)-quinoline 1-oxide (4-HAQO). Subcutaneous MFH were induced locally by repeated weekly injections of 1 mg 4-HAQO/rat for 4 weeks. Between 16 and 48 weeks after the final treatment, 13/15 (87%) male noninbred Wistar rats developed tumors. The histologic subtypes of these tumors were as follows: 11 fibrous, 1 myxoid, and 1 giant cell. Bone MFH were induced between 18 and 25 weeks by implantation of 4-HAQO (8 mg/rat) into the bone marrow of the tibia in 12/14 (86%) male noninbred Fischer 344 rats. The histologic subtypes of these tumors were as follows: 6 fibrous, 4 myxoid, and 2 giant cell. Morphologically, these MFH appeared similar to human MFH.

Two populations of cells with differing proliferative capacities in atypical acinar cell foci induced by 4-hydroxyaminoquinoline-1-oxide in the rat pancreas.

A single intravenous injection of 4-hydroxyaminoquinoline-1-oxide in male Wistar rats at a dose of 6 mg. per kg. of body weight induced atypical acinar cell foci in 100 per cent of the animals. Atypical acinar cell foci could be classified histologically as basophilic and acidophilic foci and acidophilic nodules. Cells in baseophilic foci were large, contained an irregular nucleus and a markedly basophilic cytoplasm with a few to small number of zymogen granules (zymogen-poor cells). By transmission electron microscopy, these cells showed markedly irregular plasma membranes, a well-developed rough endoplasmic reticulum and a few zymogen granules. Cells in acidophilic foci and nodules contained an intensely eosinophilic granular cytoplasm (zymogen-rich cells) and a large oval to round nucleus. By transmission electron microscopy, the cells showed zymogen-rich cytoplasm and irregular lateral plasma membranes. Mitotic activity was completely absent or very rarely observed in normal pancreas or basophilic foci, in contrast to acidophilic foci and nodules in which a mean value of 2.75 +/- 1.27 per 1000 cells was found. Autoradiography confirmed these differences between the proliferative capacity of cells in basophilic foci (1 +/- 1 labeled nuclei per 1000 cells) and acidophilic foci (23.2 +/- 3.15 labeled nuclei per 100 cells). These studies indicate that 4-hydroxyaminoquinoline-1-oxide induces two types of atypical acinar cell foci with different morphologic features and proliferative capacity.