Protein expression of CD44 (standard and variant isoforms) in hepatocellular carcinoma: relationships with tumor grade, clinicopathologic parameters, p53 expression, and patient survival

Journal of Hepatology2000; 32: 78-84 Printed in Denmark All rights reserved Munksgaard . Copenhagen

Copyright8 EuropeanAssociation for the Studvof the Liver2000

Journalof Hepatology ISSN 0168-8278

Protein expression of CD44 (standard and variant isoforms) in hepatocellular carcinoma: relationshipswith tumor grade, clinicopathologic parameters, p53 expression, and patient survival Kanenori SecondDepartment

Endo and Tadashi Terada

of Pathology, Tottori University, Faculty of Medicine, Yonago, Tottori, Japan

Background/Aims: Members of the CD44 family are transmembrane glycoproteins which act mainly as receptors for hyaluronan. We have examined the expression of CD44s and several CD44v and the relationship between these and hepatocellular carcinoma @ICC) grade, clinicopathological parameters, p53 expression, and patient survival in HCC. Methods: Formalin-fixed, paraffin-embedded tissue sections from 107 surgically resected HCC were examined immunohistochemically using a semiquantitative scoring system to detect the expression of different forms of CD44. Results: The number of CD44s-positive cases was 36 (34%), CD44v5 52 (49%), CD44v6 29 (27%), CD44v7-8 41 (38%), and CD44vlO 26 (24%). Expression of these molecules correlated with high histological grade, being the highest in poorly-differentiated HCC. High CD44v6 expression significantly

correlated with the presence of vascular invasion and p53 overexpression. Kaplan-Meier examination of patient survival revealed that HCC patients with positivity of each of these five molecules had a reduced survival rate, and that HCC patients positive for all the five CD44 molecules had worse survival than HCC patients positive for four or less of these CD44 molecules. In multivariate survival analysis, CD44s positivity was an independent factor. However, positivity for one or more CD44 isoforms was the most useful independent factor for overall survival. Conclusion: These results suggest that up-regulation of CD44 isoforms is associated with poorlydifferentiated HCC and shortened survival.

M

of many normal human epithelial and mesenchymal cells, and is thought to function mainly in lymphocyte recirculation, homing, adhesion and activation, as well as in cell-cell and cell-matrix interactions (1). In liver malignancies, we have recently examined the expression of CD44 in human cholangiocarcinoma (4). CD44 expression in hepatocellular carcinoma (HCC), however, has been reported by only a few authors (57). As there have been no comprehensive studies of CD44s and CD44v in human HCC, we examined the expression of CD44s and several CD44v and the relationship with HCC grade, clinicopathological parameters, p53 expression, and patient survival.

the CD44 family are transmembrane glycoproteins that range from 85 to 250 kilodaltons (1). They act mainly as receptors for hyaluronan (1). CD44 is the product of a single gene located on chromosome 11~13, and is encoded by at least 20 exons (l-3). Ten of these (exons l-5 and 1620) are expressed together on many cell types, and are referred to as the “standard” form of CD44 (abbreviated to CD44s). The other 10 (exons 6-15) are variant exons, called vl-vl0 (l-3). They are alternatively spliced, giving rise to various CD44 variants (abbreviated to CD44v) isoforms (l-3). CD44 is present on the surface EMBERS OF

Key words: CD44; Clinicopathological correlation; Hepatocellular carcinoma; Imnumohistochemistry; Survival.

Received 4 May; revised 4 June: accepted 28 June 1999

Tadashi Terada, Second Department of Pathology, Tottori University, Faculty of Medicine, 86 Nishi-cho, Yonago 683-8503, Tottori, Japan. Tel: 81 859 34 8018. Fax: 81 859 34 8348. Correspondence:

78

Materials

and Methods

Tissue specimens

One hundred and seven archival formalin-tied, paraffin-embedded surgically resected cases of HCC were used. Informed consent was obtained from every patient. Ages ranged from 17 to 80 years, with

CD44 in hepatocellular

an average of 60 years. The male to female ratio was 87:20. The background liver showed cirrhosis in 67 (63%) cases and chronic hepatitis in 40 (27%). The aetiology of the background liver disease was hepatitis B virus in 38 cases, hepatitis C virus in 44, and cryptogenic in 25. Multiple 3-pm serial sections were obtained from each paraffin block, one was stained with hematoxylin and eosin and one with elastic van Gieson for assessment of vascular invasion. The rest were subjected to immunohistochemical staining. Immunohistochemistry

Sections were immunohistochemically stained for CD44s, CD44v5, CD44v6, CD44v7-8, CD44v10, and ~53, using the standard avidinbiotin-peroxidase complex (ABC) method. In brief, the sections were processed to unmask antigens by microwave oven heating in 10 rim/l citric acid buffer (pH 6.0) and subsequent detergent treatment using polyoxyethylene sorbitan monolaurate (Tween 20) for 30 min. The sections were then treated with normal serum for 20 min, followed by the application of primary antibodies shown in Table 1 at 4°C overnight. The sections were then treated with biotinylated secondary antibodies (Vector Lab, Burlingame, CA, USA), and with the avidinbiotin-peroxidase complex (Vectastein ABC kit, Vector Lab) for 1 h each. The reaction products were developed by 3.3’-diaminobenxidine tetrahydrochloride solution containing 0.03% H202. Nuclei were lightly counterstained with hematoxylin. No staining was obtained when nonimmune serum or phosphate-buffered saline was used instead of the primary antibodies. Scoring of immunoreactivity

We used a semi-quantitative scoring system to evaluate the results. As shown in Table 2, immunoreactivity was assessed by the percentage of positive cells and strength of staining, each being scored 0 to +3. The scoring was done in HCC tissue and in the surrounding nonHCC hepatocytes; the former was categorised as “HCC score” and the latter as “non-HCC score”. The final score obtained was “HCC

TABLE 1 Primary monoclonal antibodies used in this study Antibody to

Clone

Class

source

Dilution

CD44s

SSF-2

IgG2a

1:200

CD44v5 CD44v6 CD44v7-8 CD44vlO P53

VFF-8 VFF-7 VFF-17 VFF-14 DO-7

IgGl IgGl IgG2b IgGl IgG2b

Bender MedSystem, Vienna, Austria Bender MedSystem Bender MedSystem Bender MedSystem Bender MedSystem Dakopatts, Glostrup, Denmark

1:150 1:20 1:20 1:20 1:75

score” minus “non-HCC score”. The score >O was termed “positive”, and that 50 “negative”. For assessment of the positivity of p53 protein immunostaining, positive nuclei of more than 10% of total cells was regarded as “positive“. Correlations with clinicopathological parameters

Immunohistochemical results were correlated with HCC histological grades, size, capsular invasion, vascular invasion, presence of satellite nodules and p53 expression. HCC grading was performed according to the Liver Cancer Study Group of Japan (8) and classified into well-, moderately- and poorly-differentiated. Vascular invasion was assessed on elastica-van-Gieson staining, and categorised as either mild, moderate or severe. Correlation to patient survival

Follow up varied between 5 and 86 months, with the mean and standard deviation being 25.5220.0 months. Survival curves were computed according to the method of Kaplan and Meier. Statistical analysis

Scores were expressed as meantstandard deviation. The ANOVA test, Spearman rank correlation test, and Mann-Whitney’s U test were employed with a significant level of pcO.05. For survival analysis, the log-rank test was used with a significant level of pcO.05. Univariate and multivariate survival analyses were also performed using Cox’s proportional hazards method to assess prognostic significance of staining patterns in relation to other pathological variables.

Results General observations

CD44s, CD44v5, CD44v6, CD44v7-8, and CD44vlO immunoreactivities were observed in the cell membrane of HCC cells in an identical pattern (Fig. 1A). In nontumorous areas of 10 (9%) cases, very weak expression of CD44v5, CD44v7-8, and CD44vlO was infrequently found (Fig. 1B). In all cases, “HCC score” was over “non-HCC score”. In the 107 HCC, positive cases of CD44s, CD44v5, CD44v6, CD44v7-8, and CD44vlO expression were 36 (34%), 52 (49%), 29 (27%), 41 (38%), and 26 (24%), respectively (Table 3). Six cases were positive for all of these five molecules (6%), and 27 cases were negative for all the molecules (25%). Relationship with HCC histological grade

TABLE 2 Scoring system used in this study 1. Evaluation of immunoreactivity Strength score Percentage score 0: no staining + 1: l-33% positive +2: 3466% positive +3: 67-100% positive

carcinoma

0: no staining + 1: mild immunoreaction +2: moderate immunoreaction +3: marked immunoreaction

2. HCC score and non-HCC score HCC score=percentage score+strength score in HCC cells Non-HCC score=percentage score+strength score in nonHCC hepatocytes 3. Final score Score=HCC score - non-HCC score HCC, Hepatocellular carcinoma.

Fig. 2 shows the mean and standard deviation of the CD44 scores. In total, CD44s, CD44v5, CD44v6, CD44v7-8, and CD44vlO mean scores were 0.942 1.42, 1.35k1.60, 0.6521.11, 0.9121.26, and 0.5721.05, respectively. CD44s, CD44v6, CD44v7-8, and CD44vlO scores significantly correlated positively with HCC grade, though the CD44v5 score was not significant (Fig. 2). The scores were highest in poorly-differentiated HCC and lowest in well-differentiated HCC. In addition, tumour grade was significantly higher in HCC patients with more positive expressions of all five molecules than in those with lesser positive expressions. 79

X End0 and T. Terada

Fig. I. Positive expression of CD44s (A), (CD44v5, CD44v6, CD44v7-8, CD44vlO were identical), and positive expression of CD44s in non-tumorous areas (B) in hepatocellular carcinoma. The expression in all instances was in a membranous pattern. Immunoperoxidase, ~450.

Correlations

with clinicopathological

parameters

There were no differences in the scores of the five molecules in relation to HCC size, and the presence or absence of capsular invasion. There were no significant differences of vascular invasion in CD44s, CD44v5, and CD44v7-8 scores, although CD44v6 and CD44vlO scores were significantly higher in HCC with severe vascular invasion than in HCC with mild invasion. No significant differences were noted in the presence of satellite nodules. Expression of p53 showed no significant correlation with expression of CD44 isoforms, with the exception of CD44v6 which showed a positive correlation with p53 over-expression. Survival

Fig. 3 and Fig. 4 represent Kaplan-Meier curves. HCC patients with positivity of each of the CD44 molecules had significantly reduced survival compared to those with negativity of each of these molecules (Fig. 3). Cumulative survival was significantly worse for HCC patients with the more positive expressions of the five molecules than HCC patients with the lesser positive

expressions (Fig. 4). In the multivariate survival analysis with Cox’s proportional hazard model, the presence of satellite nodules was identified as an independent factor for overall survival (p=O.O39) (Table 4), but other clinicopathological parameters such as sex, age, aetiology, tumour histological grade, tumour size, presence of capsular invasion, vascular invasion, and p53 expression were not. Table 4 shows the hazard ratios of all CD44 parameters which were related to other pathological variables according to Cox’s proportional hazards analysis. Univariate analysis showed that all CD44 isoforms were determinants of overall patient survival. In the multivariate survival analysis, however, CD44s was identified as the only independent factor for overall survival (p=O.O48). However, more than one CD44 immunoreactive positivity in the five parameters was identified as the most useful independent factor for overall survival (p=O.OlO).

Discussion There have been conflicting results for CD44 expression and tumour behaviour. Although most reports

TABLE 3 Frequency and scores of CD44 isoforms in the 107 cases of HCC cD44s Number of positive cases Score (m?SD) HCC, Hepatocellular carcinoma. m?SD,

80

36 (34%) 0.9421.42

mea&standard

CD44v5 52 (49%) 1.35t1.60 deviation.

CD44v6

CD44v7-8

CD44vlO

29 (27%) 0.6551.11

41 (38%) 0.91k1.26

26 (24%) 0.57+1.05

CD44 in hepatocellular carcinoma

CD446

score

well

moderate Differentiation

moderate Differentiation

well

Poor

Differentiation

Poor

Differentiation

Differentiation

Fig. 2. Mean and standard deviation of scores of CD44s, CD44v5, CD44v6, CD44v7-8, and CD44vlO in relation to histological differentiation of hepatocellular carcinoma. Scores of CD44 are correlated with tumour grade. * Indicates statistical signtficance (pCO.05).

CD44 s

!‘i ‘%_

CD44v5 c

IL :-t..

‘L

-----._

II

-‘---I_.,__

1,--‘-

negative 1.,________________~___-..

L, negative

positive L...... I

, ,p,
_

I

I

I

I

I

0

10

20

30

40

7

50

I

60

*

I

!

80

90

I

70

D

10

20

30

40

50

so

70

80

so

survival (months)

survival (months)

CD44v7-8 , / /, I 1 I 4. I

CD44v6

CD44vlO

I ( I’ I

p
10

20

30

40

50

60

70

survival (months) Fig. 3. Kaplan-Meier CD44v6, CD44v7-8,

80

90

0

10

20

30

40

50

60

70

survival (months)

80

so

0

10

20

30

40

50

60

70

80

so

survival (months)

curves of patients with hepatocellular carcinoma. Patients with positive expression of CD44s, CD44v5, and CD44vlO had a worse prognosis than patients with negative expression.

81

L Endo and T Terada

0

10

20

30

40

50

60

70

60

been linked to poor prognosis in non-Hodgkin’s lymphoma (20), and gastric (18,21,22), breast (16) and colorectal carcinomas (23-25). The association between altered CD44 expression and altered cell proliferation has also been described in several carcinomas, such as colon (2628), ovary (29), and skin cancers (30,31). In human HCC, Mathew et al. (5) showed that up-regulation of cell surface CD44 (CD44s, CD44v3, and CD44v6) expression by HCC cells was significantly related to the presence of vascular invasion, but not to tumour grade or tumour proliferation indices. Terris et al. (6) found that immunoreactive CD44s was not a predictor of outcome in patients with HCC. On the other hand, Washington et al. (7) showed that CD44 expression in HCC was associated with a shorter survival. Previous studies have indicated that normal and non-tumorous hepatocytes are negative for any isoforms of CD44 (4,5,7). In the present study, however, non-HCC hepatocytes in cirrhosis and chronic hepatitis were infrequently positive for CD44v5, CD44v78, and CD44v10, suggesting that hepatocytes in cirrhosis and chronic hepatitis complicated by HCC may aberrantly express these CD44 molecules. Because cirrhosis and chronic hepatitis are preneoplastic conditions, these CD44 molecules may play a role in the early stage of HCC carcinogenesis. In addition, the present study revealed that 2449% of HCC cases aberrantly expressed CD44 isoforms, suggesting that hepatocytes newly express CD44 isoforms after neoplastic transformation, probably due to genetic alterations of the CD44 gene. To the best of our knowledge, there have been no comprehensive reports of the relationship between aberrant CD44 expression and HCC histological grades. Mathew et al. (5) showed that there was no significant correlation between CD44s expression and tumour grade. However, they used a small number of samples

SO

survival (months) Fig. 4. Kaplan-Meier curves of patients with hepatocellular carcinoma (dotted line=all types of CD44 negative; broken line=one or two types of CD44 positive; thin line= three or four types of CD44 positive; thick line=aU types of CD44 positive). Cumulative survival was sigr$cantly reduced for patients with the more positive expressions of the fivemolecules than with the lesserpositive expressions of these molecules.

have suggested that up-regulation of CD44s and CD44v is associated with tumour progression, invasion and metastasis, as well as with patient survival (9-l l), some reports did not find any relationship (12,13). In animal experiments, transfection of complementary DNA (CD44v6) into a non-metastasising rat pancreatic carcinoma cell line conferred metastasising properties to it (14); pretreatment of anti-(variant) CD44 was subsequently shown to retard or prevent metastasis (15). In humans, several clinical investigators have reported that expression of CD44v, particularly CD44v6, is associated with high invasive and metastatic potential in breast (16), gastric (17,18), and colorectal carcinomas (19). Increased expression of various CD44 has

TABLE 4 Association of CD44 immunoreaction with overall survival analysed bythe univariate and multivariate Cox proportional hazards model Variable

Univariate analysis Hazards ratio

95% CI

Satellite nodules 2.41 1.28455 cD44s 2.26 1.264.07 CD44v5 1.91 1.05-3.48 CD44v6 2.66 1.484.79 CD44v7-8 2.50 1.38454 CD44vlO 1.92 1.05-3.52 CD44 3.23 1.83-5.71 (one or more positive vs. all negative) CI, confidence interval.

82

Multivariate analysis p

Hazards ratio

95% CI

p

0.007 0.007 0.034 0.001 0.002 0.033 0.0001

2.06 1.95 1.67 1.58 1.66 1.32 2.34

1JW4.07 1.00-3.78 0.87-3.21 0.77-3.25 0.79-3.47 0.67-2.58 1.224.46

0.039 0.048 0.126 0.214 0.178 0.422 0.010

CD44 in hepatocellular carcinoma

(17 HCC cases), and their materials were biopsies which could not examine the whole of the tumour (5). In other studies, CD44 expression was found to increase with tumour grade in breast (16,32) and colorectal cancers (24,25). The present study showed that the expression of all five CD44 molecules correlated with HCC grade, being highest in poorly-differentiated HCC, intermediate in moderately-differentiated HCC and lowest in well-differentiated HCC. These findings suggest that aberrant CD44 expression is related to dedifferentiation of HCC. In addition, since the grade of HCC becomes high with progression (33-36), CD44 expression may be closely associated with the grade of HCC. In the present study, CD44v6 and CD44vlO were significantly related to vascular invasion, although this was not the case with CD44s, CD44v5 and CD44v78. In this respect, Mathew et al. (5) also showed that expression of CD44s, CD44v3 and CD44v6 in HCC was significantly related to the presence of vascular invasion. In addition, Terris et al. (6) reported a positive correlation between CD44s expression and HCC vascular invasion. Our finding of CD44v6 expression is compatible with that of Mathew et al. (5) suggesting that CD44v6 is closely related to vascular invasion and ultimately to metastasis. However, CD44s expression in our study is not compatible with previous reports (5,6). It has been thought that CD44s adheres to hyaluronan when it is activated, but the adhesiveness is not dependent on CD44 expression status (37-39). This may be a reason for the negative findings in our study, Our results also suggest that CD44vl0, but not CD44v5 and CD44v7-8, is associated with vascular invasion. There were no significant differences in the five CD44 molecules in relation to HCC size, capsular invasion, and the presence of satellite nodules, suggesting that these parameters are independent of the expression of the five CD44 molecules. In the present study, expression of p53 showed no significant correlation with expression of CD44 isoforms, with the exception of CD44v6. Darai et al. (29) found that CD44v6 expression was correlated with a higher incidence of p53 expression in ovarian cancers. Kim et al. (40) showed that CD44 expression in colorectal adenomas is an early event occurring prior to K-ras and p53 gene mutations. These findings suggest that aberrant p53 expression may be dependent on upregulation of CD44v6, and may play an important role in tumour progression. In the present study, CD44v6 expression was associated with high p53 expression, suggesting a close link between CD44v6 and p53 in HCC. Finally, the present study is the first report of a cor-

relation between the expression of CD44 isoforms and patient survival in HCC. We found that patients with expression of CD44 had worse prognosis than patients with no CD44 expression. Therefore, our results indicated that aberrant expression of CD44s and CD44v (CD44v5, CD44v6, CD44v7-8, CD44vlO), and the presence of satellite nodules are indicative of poor survival in HCC patients.

Acknowledgements The authors are grateful to Prof. Nobuaki Kaibara and Drs. Yasunori Hirooka and Kiyoshi Kishi for providing liver tissue specimens.

References 1. Rudzki 2, Jothy S. CD44 and the adhesion of neoplastic cells. J Path01 Mol Path01 1997; 50: 57-71. 2. Screaton GR, Bell MV, Jackson DG, Cornelis FB, Gerth U, Bell JI. Genomic structure of DNA encoding the lymphocyte homing receptor CD44 reveals at least 12 alternatively spliced exons. Proc Nat1 Acad Sci USA 1992; 89: 1216&4. 3. Tolg C, Hofmann M, Herrlich P, Ponta H. Splicing choice from ten variant exons establishes CD44 variability Nucleic Acids Res 1993; 21: 1225-9. 4. Ashida K, Terada T, Kitamura Y, Kaibara N. Expression of Ecadherin, alpha-catenin, beta-catenin, and CD44 (standard and variant isoforms) in human cholangiocarcinoma; an immunohistochemical study. Hepatology 1998; 27: 97482. 5. Mathew J, Hines JE, Obafunwa JO, Burr AW, Tooe K, Burt AD. CD44 is expressed in hepatocellular carcinoma showing vascular invasion. J Path01 1996; 179: 74-9. 6. Terris B, Puig PL, Belghiti J, Degott C, Henin D, Flejou JE Prognostic influence of clinicopathologic features, DNA-ploidy, CD44H and p53 expression in a large series of resected hepatocellular carcinoma in France. Int J Cancer 1997; 74: 614-9. 7. Washington K, Telen MJ, Gottfried MR. Expression of cell adhesion molecule CD44 in primary tumors of the liver: an immunohistochemical study. Liver 1997; 17: 17-23. 8. Liver Cancer Study Group of Japan. The General Rules for the Clinical and Pathological Study of Primary Liver Cancer, 3rd ed. Tokyo: Kanehara Press; 1992. 9. Albelda SM. Role of integrins and other cell adhesion molecules in tumor progression and metastasis. Lab Invest 1993; 68: 4-17. 10. Matsumura Y, Talin D. Significance of CD44 gene products for cancer diagnosis and disease evaluation. Lancet 1992; 340: 10538. 11. Speiser P, Wanner C, Tempfer C, Mittelbock M, Hanzal E, Todesca DB, et al. CD44 is an independent prognostic factor in earlystage cervical cancer. Int J Cancer 1997; 74: 185-8. 12. Charpin C, Garcia S, Bouvier C, Devictor B, Andrac L, Choux R, et al. Automated and quantitative immunocytochemical assays of CD44v6 in breast carcinoma. Hum Path01 1997; 28: 28996. 13. Fujita N, Yaegashi N, Ide Y, Sato S, Nakamura M, Ishikawa I, et al. Expression of CD44 in normal human versus tumor endometrial tissue: possible implication of reduced expression of CD44 in lymph-vascular space involvement of cancer cells. Cancer Res 1994; 54: 3922-8. 14. Gunthert U, Hofmann M, Rudy W, Reber S, Zoller M, Hausmann 1, et al. A new variant of glycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell 1991; 65: 13-24. 15. Seiter S, Arch R, Reber S, Komitowski D, Hofmann M, Ponta H, et al. Prevention of tumor metastasis formation by arlti-variant CD44. J Exp Med 1993; 177: 443-55. 16. Kaufmann M, Heider K, Sinn H, von Minckwitz G, Ponta H,

83

L Endo and T. Terada

17.

18.

19.

20.

21.

22.

23.

24.

25.

Herrlich P. CD44 variant exon epitopes in primary breast cancer and length of survival. Lancet 1995; 345: 615-9. Harn HJ, Ho LI, Chang JY,Wo CW, Jiang SY, Lee HS, et al. Differential expression of the human metastasis adhesion molecule CD44V in normal and carcinomatous stomach mucosa of Chinese subjects. Cancer 1995; 75: 106571. Saito H, Tsujitani S, Katano K, Ikeguchi M, Maeta M, Kaibara N. Serum concentration of CD44 variant 6 and its relation to prognosis in patients with gastric carcinoma. Cancer 1998; 83: 109&101. Wielenga V, Heider KH, Offerhaus G, Adolf G, van den Berg F, Ponta H, et al. Expression of CD44 variant proteins in human colorectal cancer is related to tumor progression. Cancer Res 1993; 53: 47546. Horst E, Meijer CJLM, Radaszkiewicz T, Ossekoppele GJ, Van Krieken JHJM, Pals ST, et al. Adhesion molecules in the prognosis of diffuse large-cell lymphoma: expression of lymphocyte homing receptor (CD44), LFA-1 (CDlla/18), and ICAMl(CD54). Leukemia 1990; 4: 595-9. Washington K, Gottfried MR, Telen MJ. Expression of the cell adhesion molecule CD44 in gastric adenocarcinomas. Hum Path01 1994; 25: 1043-9. Mayer B, Jauch KW, Gunthert U, Figdor CG, Schildber Fw, Funke I. De-novo expression of CD44 and survival in gastric cancer. Lancct 1993; 342: 1019-22. Mulder JR, Kruyt PM, Sewnath M, Oosting J, Seldenrijk CA, Weidema WF, et al. Colorectal cancer prognosis and expression of exon-v6containing CD44 proteins. Lancet 1994; 344: 147&2. Ropponen KM, Eskelinen MJ, Lipponen PK, Alhava E, Kosma V-M. Expression of CD44 and variant proteins in human colorcctal cancer and its relevance for prognosis. Stand J Gastroenterol 1998; 33: 301-9. Wielenga VJM, van der Voort R, Mulder JWR, Kruyt PM, Weidema WF, Oosting J, et al. CD44 splice variants as prognostic markers in color&al cancer. Stand J Gastroenterol 1998; 33: 82-

26. Mulder Jw, Wielenga V, Polak M, van den Berg F, Adolf G, Herrlich P et al. Expression of mutant p53 protein and CD44 variant proteins in colorectal tumorigenesis. Gut 1995; 36: 76 80. 27. Abbasi AM, Chester KA, Talbot IC. CD44 is associated with proliferation in normal and neoplastic human colorectal epithelial cells. Eur JCancer 1993; 29A: 1995-2002.

84

28. Furuta K, Zahurak M, Yang X-L. Relationship between CD44 expression and cell proliferation in epithelium and stroma of colorectal neoplasms. Am J Path01 1996; 149: 1147-55. 29. Darai E, Combrouze Fw, Fauconnier A, Madelenat E Potet F, Scoazec JY. Analysis of CD44 expression in serous and mutinous dorderline tumors of the ovary; comparison with cystadenomas and overt carcinomas. Histopathology 1998; 32: 151-9. 30. Penney NS, CD44 expression in normal and inflamed skin. J Cutan Path01 1993; 20: 25&3. 31. Seelentag WKF, Gunthert U, Saremaslani F! CD44 standerd and variant isoform expression in human epithermal skin tumors is not correlated with tumor aggressiveness but down-regulated during proliferation and tumor de-differentiation. Int J Cancer (Pred Oncol) 1996; 69: 25&3. 32. Joensuu H, Klemi PJ, Toikkanen S, Jalkanen S. Glycoprotein CD44 expression and its association with survival in breast cancer. Am J Path01 1993; 143: 867-74. 33. Terada T, Nakanuma Y. Expression of ABH blood group antigens, Ulex europaeus agglutinin I, and type IV collagen in the sinusoids of hepatocellular carcinoma. Arch Path01 Lab Med 1991; 115: 5&5. 34. Terada T, Nakanuma Y. Expression of ABH blood group antigens, receptor of CJlex europaeus agglutinin, and factor VIII-related antigen in adenomatous hyperplasia in human cirrhotic livers. Hum Path01 1991; 22: 48693. 35. Terada T, Nakanuma Y. Arterial elements and perisinusoidal cells in borderline hepatocellular nodules and small hepatocellular carcinomas: an immunohistochemical study using anti-asmooth muscle actin antibody. Histopathology 1995; 27: 333-9. 36. Sakamoto M, Hirohashi S, Shiiosato Y. Early stage of multistep hepatocarcinogenesis: adenomatous hyperplasia and early hepatoccllular carcinoma. Hum Path01 1991; 22: 172-8. 37. Haynes BF, Liano H, Patton KL. The transemembrane hyaluronate receptor (CD44): multiple functions, multiple forms. Cancer Cells 1991; 3: 347-50. 38. Underhill C. CD44: the hyaluronane receptor. J Cell Sci 1992; 103: 293-8. 39. Lesley J, Hyman R, Kincade PW. CD44 and its interaction with extracellular matrix. Adv Immunol 1993; 54: 271-335. 40. Kim H, Yang X-L, Rosada C, Hamilton SR, August JT. CD44 expression in colorectal adenomas is an early event occurring prior to K-ras and p53 gene mutation. Arch Biochem Biophys 1994; 310: 5047.