Liver 2000: 20: 104–107 Copyright C Munksgaard 2000Printed in Denmark . All rights reserved

LiverISSN 0106-9543

Expression of gap junction protein connexin32 in chronic liver diseases

Yamaoka K, Nouchi T, Kohashi T, Marumo F, Sato C. Expression of gap Kazuaki Yamaoka1,junction protein connexin 32 in chronic liver disease. Toshihiko Nouchi2,Liver 2000: 20: 104–107. C Munksgaard, 2000 Takahiro Kohashi1,

Fumiaki Marumo3 andAbstract: Background: Gap junctions contain intercellular channels Chifumi Sato3,4through which contacting cells communicate directly. The expression of

1Department of Internal Medicine, Hokushinconnexin 32, a major gap junction protein in the liver, during the pro-General Hospital, Nagano, 2Division ofgression of chronic liver diseases has not yet been clarified. Methods: Im-Gastroenterology, Showa General Hospital,munohistochemical staining was performed using anti-connexin 32 anti- Tokyo, 3Second Department of Internal Medicine,

body on 6 specimens of normal human liver, 7 of chronic viral hepatitis, and 4Department of Health Science, Faculty ofand 7 of liver cirrhosis. Results: The number of gap junction plaques in Medicine, Tokyo Medical and Dental University,chronic viral hepatitis and liver cirrhosis was significantly smaller than that Tokyo, Japanin normal liver (10350∫2180 and 7550∫3040 vs 22560∫3700 spots/mm2,p∞0.01). The number of gap junction plaques tended to be lower in livercirrhosis than in chronic viral hepatitis. Conclusion: These results suggestthat in chronic liver diseases impaired intercellular communication between Key words: chronic hepatitis – chronic liver

disease – connexin 32 – gap junction –hepatocytes occurs.immunohistochemical staining – liver cirrhosis

Chifimi Sato, M. D., Department of HealthScience, Faculty of Medicine, Tokyo Medical andDental University, 1–5-45, Yushima, Bunkyo-ku,Tokyo 113–8519, Japan

Received 10 July 1998, accepted 14 September1999

Gap junctions are intercellular channels that canpermeate inorganic ions and molecules smallerthan 1000 Da, and are considered to play an im-portant role in cell growth, cell differentiation andtissue integration. Gap junctions are collections oftransmembrane channels called connexons that di-rectly connect the cytoplasm of neighboring cells(1). Each connexon is composed of a hexamer ofstructural proteins called connexins (Cx) (1). Con-nexin 32 (Cx32) is a major liver gap junction pro-tein and is evenly distributed in normal liver paren-chyma (2). Recently, we reported that the express-ion of Cx32 in human hepatocellular carcinoma(HCC) decreased significantly compared with thatin the surrounding non-carcinomatous cirrhotictissues, suggesting an impairment of cell-to-cellcommunications in human HCC (3). It is un-known, however, whether the expression of con-nexin 32 is impaired in chronic liver diseases. Toclarify the change of its expression in chronic liverdiseases, we performed immunohistochemicalstaining of Cx32 on human diseased livers.

104

Materials and methods

Three samples of normal livers were obtained atsurgery from two patients with symptoms of gall-bladder carcinomas and one with insulinoma.Three other normal specimens were obtained frompatients with a minimum increase of aminotrans-ferases by echo-guided needle liver biopsy orneedle biopsy under laparoscopy. The histologicalfindings of these biopsy specimens were normal ornon-specific changes. In addition, echo-guidedneedle biopsy specimens of diseased liver were ob-tained from 9 patients for diagnostic purposes, in-cluding 7 with chronic viral hepatitis, and 2 withalcoholic cirrhosis. Five specimens of postnecroticcirrhosis were obtained at the time of surgical re-section of HCC. The etiology and the histologicalactivity of chronic viral hepatitis and liver cirrhosisare shown in Table 1. Histological activity was as-sessed by the histological activity index (HAI), ac-cording to Knodell (4).

The immunohistochemical detection of Cx32

Connexin 32 in chronic liver diseases

Table 1. Etiology and histologic activity in seven patients with chronic viralhepatitis and liver cirrhosis

Necroinflammatory Number ofCase Etiology activity* Fibrosis** Cx 32 spots

Chronic viral hepatitis1 HCV 3 ( I-1, II-1, III-1 ) 1 76402 HCV 2 ( I-0, II-0, III-1 ) 0 86403 HCV 1 ( I-1, II-0, III-0 ) 0 91804 HCV 5 ( I-3, II-1, III-1 ) 1 92205 HCV 5 ( I-0, II-1, III-3 ) 0 121806 HBV 5 ( I-1, II-1, III-3 ) 1 122807 HCVπAIH 8 ( I-4, II-1, III-3 ) 3 13290

Liver cirrhosis8 Alcohol 0 ( I-0, II-0, III-0 ) 4 48609 HCV 5 ( I-3, II-1, III-1 ) 4 4910

10 Unknown 7 ( I-3, II-3, III-1 ) 4 540011 Alcohol 0 ( I-0, II-0, III-0 ) 4 555012 HCV 3 ( I-1, II-1, III-1 ) 4 972013 HBV 8 ( I-4, II-1, III-3 ) 4 1040014 HCV 7 ( I-3, II-1, III-3) 4 12010

*Knodell score I (periportal/bridging necrosis), II (intralobular degenerationand focal necrosis) and III (portal inflammation), ** Knodell score IV; HBVΩhepatitis B virus; HCVΩhepatitis C virus; AIHΩautoimmue hepatitis; Cx 32Ωconnexin 32.

was performed using an avidin-biotin complex per-oxidase technique (3, 5). Cryostat sections of athickness of 7 mm were fixed in cold acetone for10 min. The sections were washed three times inphosphate-buffered saline (PBS). A mouse mono-clonal anti-Cx32 antibody (clone 6–3G11, suppliedby Nippon Shinyaku Co., Ltd.), which reactsspecifically with intact Cx32, was diluted 1:1000 inPBS containing 5% goat serum. The sections werethen incubated with the primary antibody for 1 hat room temperature. After three washings in PBS,biotin-conjugated goat antimouse IgG (1:100,Sigma) was applied on the sections for 30 min.After three washings in PBS, an avidin-biotin com-plex (Vectastain, Burlingame, California) wasadded for 30 min. The color was developed for 2min in freshly prepared Tris buffer (pH 7.6) con-taining 0.003% H2O2, 0.02% 3,3-diaminobenz-idine, and 0.04% nickel chloride. The sections werethen counterstained with methyl green. Negativecontrols, with the application of control mouse as-cites in place of the primary antibody, were em-ployed, and these uniformly demonstrated absenceof a reaction. The number of Cx32-positive spots/mm2 in normal and diseased liver tissues wascounted at random in 15 photographic fields (mag-nification ¿400).

Values are expressed as means∫SD. Statisticalanalysis was performed using one way analysis ofvariance with Student-Newman-Keuls test formultiple comparisons.

105

Results

Cx32-positive spots were observed at intercellularborders of hepatocytes. In normal livers, manyCx32-positive spots were widely distributedthroughout parenchymal areas. In contrast, Cx32-positive spots were irregularly distributed inchronic viral hepatitis and in liver cirrhosis (Fig.1). We did not find pre-neoplastic lesions in any ofthe specimens we studied, and there was no differ-ence in the distribution of Cx32 in liver cirrhosiswith or without hepatocellular carcinoma. Therewas no close relationship between the number ofgap junctions and necroinflammatory activity(Table 1). The number of gap junction plaques inchronic viral hepatitis (nΩ7) and liver cirrhosis(nΩ7) was significantly less than that in normallivers (nΩ6) (10350∫2180 and 7550∫3040 VS22560∫3700 spots/mm2, p∞0.01). The number ofgap junction plaques was lower in liver cirrhosisthan in chronic viral hepatitis, but the differencedid not reach statistical significance (Fig. 2).

Discussion

Altered expression of Cx during liver regenerationand hepatocarcinogenesis has been thoroughlystudied in rat livers or isolated rat hepatocytes. Inrats, decreased levels of liver gap junction proteinshave been observed after partial hepatectomy andin acute liver injury by hepatotoxic chemicals, sug-gesting a close association between the reduced ex-pression of gap junction proteins and liver re-generation (6–9). Cx32 mRNA and Cx32 proteindecreased significantly during chemical hepatocar-cinogenesis in rats (10–12). On the other hand, afew studies have been carried out on gap junctionsin human HCC. In an immunohistochemical study,Oyamada et al. reported that there was no decreasein Cx32-positive spots in HCC (13). In contrast,we demonstrated that Cx32-positive spots de-creased significantly in HCC (3). Recently, Krutov-skikh et al. reported that Cx32-positive spots weretranslocated to the cytoplasm, while by Westernblot analyses there was no decrease in Cx32 pro-tein in HCC (14).

In the present study, we showed that the ex-pression of Cx32 decreased significantly in chronicviral hepatitis and liver cirrhosis compared withnormal liver, and tended to decrease with the pro-gression of liver injury, although the difference didnot reach statistical significance. Liver regenerationmay explain to some extent the reduction of gapjunction numbers in chronic viral hepatitis. Hyper-plastic nodules in rat chemical hepatocarcinogen-esis have been reported to show a reduced level ofCx32 mRNA, and altered gap junction communi-

Yamaoka et al.

cation capacity is considered to be a relativelyearly event in multistage carcinogenesis (9). Therewas no evidence of the relationship between he-patic carcinogenesis and gap junctions in our im-munohistochemical study.

The other liver gap junction protein is connexin

Fig. 1. Immunoperoxidase staining of connexin 32 in normal liver (A, B), chronic viral hepatitis (C), and liver cirrhosis (D). (A)Many connexin 32-positive spots are distributed evenly in liver parenchyma, but not on the sinusoidal surface (¿370). (B) Gapjunction plaques are observed at intercellular border of hepatocytes (¿1000). (C, D) Gap junction plaques are distributed irregularlyin liver parechyma, and are absent adjacent to the expanded portal zone (PZ) and fibrous band (FB). The number of connexin 32-positive spots is decreased compared with normal liver.

106

26 (Cx26), which is not evenly distributed in nor-mal liver tissues, and is unsuitable for semiquan-titative analysis in terms of the progression ofchronic liver disease. We carried out immunohisto-chemical detection of Cx26 in chronic viral hepa-titis in a few cases. Expression of Cx26 was less

Connexin 32 in chronic liver diseases

Fig. 2. A quantitative analysis of connexin 32-positive spots innormal liver, chronic viral hepatitis, and liver cirrhosis. Thenumber of gap junction plaques in chronic viral hepatitis andliver cirrhosis was significantly lower than that in normal liver(10350∫2180 and 7550∫3040 vs 22560∫3700 spots/mm2,p∞0.01).

than that of Cx32, and, at least, the compensatoryover-expression of Cx26 for the decrease of Cx32was not observed (data not shown). Cx43, a maincardiac gap junction protein, was not detected inthe few cases we studied (data not shown).

Details of regulation mechanisms of gap junc-tion proteins are unknown. It is generally acknowl-edged that gap junction plaques on the cell surfaceare not stable ( half-lives ∞5 h in rats ) (15). In ratand human liver tumors, it has been reported thatC32-positive spots are detected mainly either intra-cytoplasmically or in plasma membranes free fromcontact with other cells, and it has been suggestedthat the aberrant localization of Cx32 in tumorcells is due to the disruption of the establishmentof Cx32 into gap-junction plaques (14, 16).Whether Cx32-positive spots were functional ornot could not be established in the present study.

In conclusion, the expression of Cx32, a majorliver gap junction protein, was decreased signifi-cantly in chronic viral hepatitis and liver cirrhosiscompared with normal liver. The significance ofthe decrease of the number of gap junctions inchronic liver disease remains to be elucidated.

107

References

1. K N, G N. The gap junction communicationchannel. Cell 1996; 84: 381–8.

2. N B, D R, T D, T O, W- K, R J. Two homologous protein components ofhepatic gap junctions. Nature 1987; 329: 732–4.

3. Y K, N T, T J, H S, MF, S C. Expression of gap junction protein connexin32 and E-cadherin in human hepatocellular carcinoma. JHepatol 1995; 22: 536–9.

4. K RG, I KG, B WC, et al. Formulationand application of a numerical scoring system for assessinghistological activity in asymptomatic chronic active hepa-titis. Hepatology 1981; 4: 431–5.

5. Y K, N T, M F, S C. Alpha-smooth-muscle actin expression in normal and fibrotic hu-man livers. Dig Dis Sci 1993; 38: 1473–9.

6. M T, T A, I M, S T. Single ad-ministration of hepatotoxic chemicals transiently decreasesthe gap-junctional-protein levels of connexin 32 in rat liver.Eur J Biochem 1991; 196: 37–42.

7. H E, L C, W W, et al. Quantitativeimmunoassay of total cellular gap junction protein connex-in 32 during liver regeneration using antibodies specific tothe COOH-terminus. Tissue Cell 1992; 24: 61–73.

8. F K, G B, M A, et al. Gap junctionsand growth control in liver regeneration and in isolated rathepatocytes. Hepatology 1997; 25: 847–55.

9. K T, Y M, M C, et al. Differentchanges in expression and function of connexin 26 and con-nexin 32 during DNA synthesis and redifferentiation in pri-mary rat hepatocytes using a DMSO culture system. Hep-atology 1997; 26: 585–97.

10. F D, M M, O M, T H, I N,Y H. Changes in gap junction protein (Connexin32) gene expression during rat liver carcinogenesis. J CellBiochem 1989; 41: 97–102.

11. N M, H J, P D, P H. Reversible alterationin the expression of the gap junctional protein connexin 32during tumor promotion in rat liver and its role during cellproliferation. Cancer Commun 1990; 2: 21–31.

12. K V, O M, Y H. Sequentialchanges of gap-junctional intercellular communicationsduring multistage rat liver carcinogenesis. Carcinogenesis1991; 12: 1701–6.

13. O M, K V, M M, P C,F B, Y H. Aberrant expression of gap junc-tion gene in primary human hepatocellular carcinomas: in-creased expression of cardiac-type gap junction gene con-nexin 43. Mol Carcinog 1990; 3: 273–8.

14. K V, M G, M N, et al. Alteredhomologous and heterologous gap-junctional intercellularcommunication in primary human liver tumors associatedwith aberrant protein localization but not gene mutation ofconnexin 32. Int J Cancer 1994; 56: 87–94.

15. L D. The life cycle of a connexin: gap junction forma-tion, removal, and degradation. J Bioenerg Biomembr1996; 28: 311–8.

16. K V, M M, M G, Y H.Inhibition of rat liver gap junction intercellular communi-cation by tumor-promoting agents in vivo. Association withaberrant localization of connexin proteins. Lab Invest 1995;72: 571–7.