Journal of Hepatology, 1992; 16:7-10 7 ©1992 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0168-8278/92/$05.00

HEPAT 01247

Leader

Liver ploidy

G6rard F e l d m a n n

Laboratoire de Biologie Celhdaire, INSERM U-327, FacultO de MOdecine Xavier Bichat (Universitb Paris VII), Paris, France

Definition

According to Lewin (1), ploidy is defined as the number of chromosome sets present in a cell. In mam- mals a chromosome set consists of several autosomes with a sex chromosome. Consequently, a haploid cell is a cell with only one set of chromosomes, a diploid cell (or euploid cell) a cell with two sets of chrosomoses and a polyploid cell, a cell with more than two sets of the haploid genome. The haploid number n is characteristic of the gametes of diploid (2n) organisms. An aneuploid cell is a cell where the chromosomal constitution differs from the usual diploid constitution by loss or duplication of chromosomes or chromosomal segments. Because it is easier to measure DNA content in the nucleus than to count the number of chromosomes, ploidy is very often understood as the quantity of DNA in the cell and accordingly cells are divided into diploid, polyploid and aneuploid cells.

Ploidy and liver cells in the normal state

Liver cells are composed of hepatocytes and sinusoidal and biliary cells. Almost nothing is known about the ploidy of the latter two, but they are usually considered as diploid cells (2). In contrast, ploidy has often been investigated in hepatocytes in several species including man and it is well established that ploidy of mammalian hepatocytes changes from the postnatal period to the adult state. In newborn mammals such as mice or rats, hepatocytes are mainly diploid, whereas in adults poly- ploidization occurs resulting in a mixed hepatocyte population which differs in its nuclear DNA content '3-5). For instance, in the mouse liver (4-6), the percen- Lage of diploid hepatocytes decreases markedly whereas

polyploid cells (tetraploid, octoploid and 16n cells) increase in parallel with the body weight and age of the animal. In the adult mouse with a body weight of 30 g, 2n hepatocytes represent about 25% of the cells, 4n 60%, 8n 12% and 16n less than 3% (4). In adult human liver, diploid hepatocytes represent 55% of the cells, the remainder being tetraploid hepatocytes (7). Estimation of ploidy during the development of the mammalian liver is complicated by the presence of binucleated hepatocytes. The number of these cells is not important in embryos, but it increases rapidly after birth and begins to decrease only in adulthood. According to Doljanski (3), the percentage of binucleated hepatocytes in Wistar rats, jumps from 7 to 47% from the fetal state to 4 weeks after birth, peaks at 5 weeks, and remains around 30% from 6 months to 2 years. Some variations seem to exist in the same strain since Nadal and Zajdela (8) and Sulkin (9) observed lower percentages of binucleated cells in the liver of adult Wistar rats. In young animals, almost all binucleated hepatocytes are diploid (4,8). In older animals, the percentage of diploid binucleated cells decreases as tetraploid binuclear hepatocytes appear, and their evolution is roughly similar to that of mono- nuclear cells.

How ploidy is distributed in the normal adult hepatic lobule is still unclear since there is little information on this point. However, it is known that binucleated cells are located mostly around the portal tracts and it is admitted that diploid cells are localized in the same areas, with polyploidization appearing in the mediolobu- lar zone and reaching a maximum around the centrolob- ular vein (9).

The quantity of DNA in hepatocytes is related to certain morphological and physiological parameters. First, in rats 2n hepatocytes are the smallest cells with a mean diameter of 17.5/~m, whereas the largest cells have a mean diameter of 25/lm and a ploidy of 8n and

7orrespondence to: Dr. G. Feldmann, INSERM U-327, Facult6 de M6decine Xavier Bichat, 16, rue Henri Huchard, 75018 Paris, France.

8 G. FELDMANN

4n hepatocytes have a diameter of about 20pm (10). Second, there is a significant relationship between nuclear surface area and nuclear DNA content (11). Third, cell size and cell volume are directly proportional to ploidy, and the most voluminous hepatocytes are polyploid (4,12). It should be stressed that all these data have been obtained from hepatocytes isolated from rat (10-12) or mouse liver (4). At the present time no information is available on the size of hepatocytes in situ with regard to the distribution of ploidy in the normal hepatic lobule.

It has been hypothesized that when a cell such as the 4n tetraploid hepatocyte contains 4 sets of chromosomes, it should produce twice as many enzymes or proteins as a diploid hepatocyte. Several researchers have investi- gated the metabolic behaviour of these two classes of hepatocytes with, however, discrepant results. For instance, Tulp et al. (13) noted that the quantity of succinate dehydrogenase and NADPH cytochrome c reductase was equal in diploid and tetraploid hepato- cytes, whereas that of lactate dehydrogenase was signifi- cantly lower in diploid than in tetraploid cells. Bernaert et al. (10) observed no difference in glucose-6-phospha- tase activity between 2n and 4n hepatocytes. Moreover, the location of glucose-6-phosphatase, which is preferen- tially periportal in situ, does not correspond to the distribution of 2n hepatocytes which, as we have seen above, are mainly located around the portal tracts. Discrepant results have also been obtained for plasma proteins secreted by hepatocytes. Foucrier et al. (11) measured the secretion of transferrin by isolated rat hepatocytes and observed that a two-fold secretion ratio was not found between tetraploid and diploid hepato- cytes. These results are contradictory to those reported by Le Rumeur et al. (14) who found a direct relationship between the degree of ploidy in hepatocytes and the amount of albumin secreted by these cells.

How can the polyploidization of hepatocytes be explained? The explanation, now generally accepted (5,15), was provided several years ago by Nadal and Zajdela (8) and is based on two cellular phenomena: mitosis of diploid hepatocytes without separation of daughter cells (acytokinetic mitosis) leading to the forma- tion of binucleated diploid hepatocytes, and mitosis of the latter cells with fusion of their mitotic nuclei giving mononuclear daughter tetraploid cells. The latter can form binucleated tetraploid hepatocytes by an acytoki- netic mitosis and so on... The increasing number of binucleated ceils from the postnatal period to the adult state and the decreasing number of these binucleated cells from the adult state constitute the two points which

strongly support the explanation provided by Nadal and Zajdela.

With regard to normal liver ploidy, several other facts must be emphasized. Polyploidization is an irreversible phenomenon (5) and hepatocyte DNA content increases with the age of the animal (5,16). After partial hepatec- tomy in animals, there is a decrease in diploid hepato- cytes, an increase in polyploid cells, particularly 4n and 8n cells, and a reduction in the relative number of binucleated cells (5).

Measurement of ploidy in hepatocytes

The best way to measure hepatocyte ploidy would be to count the number of chromosomes in the different classes of hepatocytes. Since this is not possible in practice, DNA content is usually measured in the nuclei of hepatocytes after labelling their DNA by the Feulgen method or with more recently proposed markers such as ethidium bromide or propidium iodide. In animals, hepatocytes are firstly isolated by collagenase perfusion and DNA content measured either by microspectropho- tometry (4) or by flow cytometry (17-19). In humans, DNA content is measured directly by microspectropho- tometry of liver sections prepared for histological exami- nation (20-22) or by flow cytometry of nuclei isolated from fresh or frozen liver fragments and labelled with specific stains (23,24). Flow cytometry is also used to analyze nuclei extracted from thick paraffin-embedded liver sections (23,25,26). The advantages and disadvan- tages of microspectrophotometry versus flow cytometry cannot be discussed in detail in this paper. Briefly, with the first method, only a few number of nuclei are observed (200-300 per liver section), whereas with the second several thousand nuclei are analyzed. In contrast with the first method, it is possible to choose the nuclei to be investigated, and also to mark abnormal nuclei and distinguish hepatocyte nuclei from nuclei of other hepatic cells. Although flow cytometry does not have these advantages, this technique is being used more and more to determine DNA content in normal and patho- logical tissues (27). In addition, with flow cytometry, it is now possible to obtain information on the cell cycle, particularly on the S-phase, the DNA-synthetising phase, which could be of interest in the prognosis of cancer (23,28). Finally, working on paraffin-embedded tissues (29) allows retrospective studies, even if there has been criticism about the use of flow cytometry on paraffin- embedded tissue (23,30).

LIVER PLOIDY 9

Pioidy and liver cancer

It is generally accepted that the nuclear DNA content of cancerous cells constitutes an important parameter of malignancy. Measurement of DNA is often performed in human tumors since karyotyping of tumor cells, particularly solid tumors, is difficult. As with normal cells, it is assumed that chromosomal abnormalities, frequently observed in malignant cells (27), are directly reflected by the DNA content of their nuclei. Moreover, it has been suggested that measurement of DNA can give information on both the diagnosis and prognosis of cancer (27).

Ploidy of cancerous hepatocytes has recently been measured in several series of human hepatocarcinomas, (7,20-26,31) most often from Far-East countries (20- 23,25,26) with two objectives: to evaluate the percentage of diploid versus aneuploid liver cancers, the diploid cancers being considered to have a better prognosis than aneuploid forms; to find correlations between degree of ploidy and several other parameters, such as age, sex, presence of HbsAg, liver cirrhosis, tumor size and patient survival.

In relation to the first point, the percentage of diploid and aneuploid liver cancers varies markedly from one series to another. For instance, Ezaki et al. (20) observed that 38% of their 60 cases presented a diploid state while 62% were aneuploid. No difference was reported by Fujimoto et al. (25) in their series of 157 cases in the percentage of diploid and aneuploid cancers. In contrast, Chen et al. (26) noted that a large majority of the tumors (78% of their 50 cases) were aneuploid. Finally, Chiu et al. (23) recently reported that diploid liver cancers represented about 65% of their 130 cases. In France, F16jou et al., in a preliminary communication (24), found a percentage of 53% of aneuploid liver cancers in a series of 58 cases; in the United States, McEntee et al. (31) found a percentage of 67% of non-diploid tumors in a series of 46 patients. If we total the largest series reported from Far-East countries (20,23,25,26), we obtain about 400 cases of primary liver cancer with 56% aneuploid having, in principle, a worse prognosis.

The classification of liver cancer into two groups - - diploid and aneuploid - - does not give clear indications about the evolution of these cancers. Discrepant findings have been reported. For instance, while several authors (20,23,25) reported that DNA aneuploidy increased with tumor size, this relationship was not found by others (26,31). Four groups (23,25,26,31) did not observe a relationship between DNA ploidy and the age or sex of patients, positivity of HbsAg, or presence of cirrhosis.

One group (20) noted that grades II and III of the morphological classification of liver cancer proposed by Edmonson and Steiner correlated significantly with the DNA pattern (high percentage of aneuploidy), whereas another group (31) did not find this correlation. The much debated point of survival rate is not different in diploid and aneuploid liver cancers according to several authors (20,26,31). However, aneuploid tumors are con- sidered by some authors to have a bad prognosis (25). This has been recently confirmed by Chiu et al. (23) who suggested analyzing liver cancers by flow cytometry with a new classification of the DNA pattern which takes into account the subtle modifications of ploidy observed in tumor cells (see Ref. 23 for details) and a calculation of the S-phase as employed by Clark et al. for the prognosis of breast cancer (28). In 130 cases, Chiu et al. (23) observed that patients with large liver tumors and an aneuploid pattern have a significantly lower overall survival and a higher recurrence rate than patients with diploidy.

Conclusions

It would appear that the investigation of liver ploidy has not yet reached its plateau. At the present time, many experimental results have been published on ani- mals. Since it is easy to confirm the quantity of DNA recorded by morphological means using biochemical methods (14,17), the results obtained on ploidy, e.g. during liver development or after partial hepatectomy, are strongly supported. The functional relationship between hepatocyte DNA content and the amount of protein or enzymes produced by the cell awaits further results. The investigation of ploidy in the human liver, particularly in primary liver cancer, depends on the techniques used to measure DNA in the liver cells. The problems of analysis of paraffin-embedded sections are obvious (23,29,30). Flow cytometry is not easy to control and analysis of the cell cycle, particularly of the S-phase, can be made using several different mathematical formu- las (32). If these remarks are taken into account, the originality of the work presented by Nagasue et al. (33) must be emphasized. The authors observed that the degree of ploidy was comparable in synchronous liver cancers (two hepatic tumors at the same time) but different in metachronous liver cancers (primary and recurrent tumors). If, as the authors postulate, DNA ploidy reflects clonal origin, their results suggest that most recurrent cancers develop from different cell clones whereas this is the case for only a minority of tumors

10 G. FELDMANN

in synchronous cancers. More generally these results open

up a new dimension in the analysis of carcinogenesis.

However, it should be noted that some years ago K u o

et al. (34) while investigating ploidy in pr imary liver

cancer and in recurrent or metastatic lesions did no t find

a different D N A pat tern in most of the cases. They

concluded that liver cancer may come from a single done.

Acknowledgement

The au thor wishes to thank Miss Sonya Z a b a t h for

her help in the p repara t ion of the manuscr ip t .

References

1 Lewin B. Genes IV. Cambridge, MA: Oxford University Press, 1990.

2 Medvedev ZA. Age-related polyploidization of hepatocytes: the cause and possible role. Exp Gerontol 1986; 21: 277-82.

3 Doljanski F. The growth of the liver with special reference to mammals. Int Rev Cytol 1960; 17: 217-41.

4 Epstein CJ. Cell size, nuclear content, and development of poly- ploidy in the mammalian liver. Proc Natl Acad Sci USA 1967; 57: 327-34.

5 Brodsky WY, Uryvaeva IV. Cell polyploidy: its relation to tissue growth and function, lnt Rev Cytol 1977; 50: 275-332.

6 Shima A, Sugahara T. Age-dependent ploidy class changes in mouse hepatocytes as revealed by Feulgen-DNA cytofluorometry. Exp Gerontol 1976; 11: 193-203.

7 Saeter G, Lee C-Z, Schwarze PE, et al. Changes in ploidy distri- butions in human liver carcinogenesis. J Natl Cancer Inst 1988; 80: 1480-5.

8 Nadal C, Zajdela F. Polyploidie somatique dans le foie de rat. I. I.,¢ r61e des cellules binucl6es dans la g6n~se des cellules polyplo~des. Exp Cell Res 1966; 42:99-116.

9 Sulkin NM. A study of the nucleus in the normal and hyperplastic liver of the rat. Am J Anat 1943; 73: 107-25.

10 Bernaert D, Wanson JC, Mosselmans R, De Paermentier F and Drochmans P. Separation of adult rat hepatocytes into distinct subpopulations by centrifugal elutriation. Morphological, morpho- metrical and biochemical characterization of cell fractions. Biol Cell 1979; 34: 159-74.

11 Foucrier J, P6chinot D, Rigaut JP, Feldmann G. Transferrin secretion and hepatocyte ploidy: analysis at the single cell level using a semi-automatic image analysis method. Biol Cell 1988; 62: 125-31.

12 Sweeney GD, Cole FM, Freeman KB, Patel HV. Heterogeneity of rat liver parenchymal cells. Cell volume as a function of DNA content. J Lab Clin Med 1979; 94: 718-25.

13 Tulp A, Welagen JJMN, Emmelot P. Separation of intact rat hepatocytes and rat liver nuclei into ploidy classes by velocity sedimentation at unit gravity. Biochim Biophys Acta 1976; 451: 567-82.

14 Le Rumeur E, Beaumont C, Guillouzo C, Rissel M, Bourel M, Guillouzo A. All normal rat hepatocytes produce albumin at a rate related to their degree of ploidy. Biochem. Biophys Res Commun 1981; 101: 1038-46.

15 James J. The genesis of polyploidy in rat liver parenchymal cells. CytoBiol (Eur J Cell Biol) 1977; 15: 410-9.

16 Zajicek F, Schwartz-Arad D, Bartfeld E. The streaming liver. V. Time and age-dependent changes of hepatocyte DNA content, following partial hepatectomy. Liver 1989; 9: 164-71.

17 Digernes V, Bolund L. The ploidy classes of adult mouse liver cells. A methodological study with flow cytometry and cell sorting. Virchow's Arch Cell Pathol 1979; 32: 1-10.

18 Yim APC. Some flow-cytofluorometric studies of the nuclear ploidy

of mouse hepatocytes. III. Further observations on early changes in nuclear ploidy of mouse hepatocytes following various experi- mental procedures. Br J Exp Pathol 1982; 63: 458-61.

19 Danielsen HE, Steen HB, Lindmo T, Reith A. Ploidy distribution in experimental liver carcinogenesis in mice. Carcinogenesis 1988; 9: 59-63.

20 Ezaki T, Kanematsu T, Okamura T, Sonoda T, Sugimachi K. DNA analysis of hepatocellular carcinoma and clinicopathologic implications. Cancer 1988; 61: 106-9.

21 Yoshida Y, Okamura T, Kanematsu T, Kakizoe S, Sugimachi K. Comparison between microspectrophotometry and cytofluorome- try in measurements of nuclear DNA in human hepatocellular carcinomas. Cancer 1988; 62: 755-79.

22 Hoso M, Nakanuma Y. Cytophotometric DNA analysis of hepato- cellular carcinoma with Mallory bodies. Virchow's Arch Pathol Anat 1989; 416: 51-5.

23 Chiu J-H, Kao H-L, Wu L-H, Chang H-M, Lui W-Y. Prediction of relapse of survival after resection in human hepatomas by DNA flow cytometry. J Clin Invest 1992; 89: 539-45.

24 Fl6jou JF, Belghiti J, Degott C, H6nin D, Potet F. Analyse de I'ADN en cytom6trie de flux dans une s6rie de 58 carcinomes h6patocellulaires. Gastroent6rol Clin Biol 1992; 16: A-239.

25 Fujimoto J, Okamoto E, Yamanaka N, Toyosaka A, Mitsunobu M. Flow cytometric DNA analysis of hepatocellular carcinoma. Cancer 1991; 67: 939-44.

26 Chen M-F, Hwang T-L, Tsao K-C, Sun C-F, Chen T-J. Flow cytometric DNA analysis of hepatocellular carcinoma; preliminary report. Surgery 1991; 109: 455-8.

27 Merkel DE, Dressler LG, McGuire WL. Flow cytometry, cellular DNA content and prognosis in human malignancy. J Clin Oncol 1987; 5: 1690-1703.

28 Clark GM, Dressier LG, Owens MA, Pounds G, Oldaker T, McGuire WL. Prediction of relapse or survival in patients with node-negative breast cancer by DNA flow cytometry. N Engl J Med 1989; 320: 627-33.

29 Hedley DW, Friedlander ML, Taylor IW, Rugg CA, Musgrove EA. Method for analysis of cellular DNA content of paraffin- embedded pathological material using flow cytometry. J Histochem Cytochem 1983; 31: 1333-5.

30 Becker RL Jr, Mikel UV. Interrelation of formalin fixation, chroma- tin compactness and DNA values as measured by flow and image cytometry. Anal Quant Cytol Histol 1990; 12: 333-41.

31 McEntee GP, Batts KA, Katzmann JA, Ilstrup DM, Nagorney DM. Relationship of nuclear DNA content to clinical and patho- logic findings in patients with primary hepatic malignancy. Surgery 1992; 111: 376-9.

32 Frierson HF Jr. Ploidy analysis and S-phase fraction determination by flow cytometry of invasive adenocarcinomas of the breast. Am J Surg Pathol 1991; 15: 358-67.

33 Nagasue N, Kono H, Chan Y-C et al. DNA ploidy pattern in synchronous and metachronous hepatocellular carcinomas. J Hepatol 1992; 16: 208-14.

34 Kuo S-H, Sheu J-C, Chen D-S, Sung J-L, Lin C-C, Hsu H-C. DNA Clonal heterogeneity of hepatocellular carcinoma demon- strated by Feulgen-DNA analysis. Liver 1987; 7: 359-63.