**3. Gap junctions and connexin expression in normal liver**

The liver is composed of several cell types and consequently it expresses different connexin isoforms. Biliary cells, endothelial cells and oval cells (hepatic stem cells) express mainly Cx43. Hepatocytes express both Cx26 and/or Cx32 depending on the spatial localization in the hepatic lobule. For example, Cx32 is uniformly distributed throughout the liver and therefore it is expressed by hepatocytes in hepatic acinus, while Cx26 is preferentially expressed by hepatocytes localized in the periportal spaces. Thus, most non-parenchymal cells express Cx43 while hepatocytes express Cx26 or Cx32. Considering that hepatic cells establish cell-to-cell communication by channels consisted of different connexins it is not surprising that they present different permeability (permselectivity). It was demonstrated that Cx32 gap junctions had a 10-fold higher relative permeability to adenosine compared with Cx43 channels. By contrast, the phosphorylation status of the adenosine shifted its preferential selectivity from Cx32 channels towards those formed by Cx43. The Cx43 intercellular channels were eight times more permeable to AMP and ADP than Cx32 channels, and the permeability of ATP through Cx43 was more than 300-fold better than that through Cx32 channels. Thus biophysical properties of gap junctional channel depend on the type and phosphorylation status of the connexin that form the channel.

In normal conditions, adult liver display very low proliferative activity. However after injuries or partial hepatectomy, the remaining intact cells start to proliferate. During this proliferation stage, it was observed decreasing of Cx26 and Cx32 expression in hepatocytes. Similar situation was observed when hepatocytes from primary culture were stimulated by mitogen (Kojima et al., 1997). It is important to remember that during liver regeneration process there is not only cell proliferation of hepatocytes but also cell differentiation of oval cells into hepatocytes. Oval cells naturally express Cx43, however along the differentiation program this cell type switch the connexin isoform preferentially express from Cx43 to Cx32

gap junction requires the appropriate cell-to-cell adhesion. There is evidence that interaction of Cx43 with the tight junction protein (ZO1) may play a role in regulating the size of the

In general, the turnover rate of connexin is very fast in relation to other plasma membrane proteins. According to studies performed *in vivo* and *in vitro,* the half-lives of Cx26 and Cx32 are respectively 2 and 3 hours (Traub et al., 1983; 1987). The removal of gap junction from the plasma membrane occurs by endocytosis. During this process, both membranes of the gap junction are internalized into one of the adjacent cells and thereby form a doublemembrane vacuole called annular gap junction. These structures are further degraded by both lysosomes and proteasomes. The preferential degradation pathway is associated to both cell and connexin type ( Laird et al., 2005; 2006). The degradation of Cx32 in the liver occurs mainly via the lysosomal pathway (Rahman et al., 1993). Furthermore the phosphorylation status of connexin is important to regulate its internalization and degradation. *In vitro* studies have been done to understand the mechanism involved in Cx43 internalization and degradation. They showed that the internalization and degradation of Cx43 gap junction is closely related to its hyperphosphorylation (via Mek/Erk pathway)

The liver is composed of several cell types and consequently it expresses different connexin isoforms. Biliary cells, endothelial cells and oval cells (hepatic stem cells) express mainly Cx43. Hepatocytes express both Cx26 and/or Cx32 depending on the spatial localization in the hepatic lobule. For example, Cx32 is uniformly distributed throughout the liver and therefore it is expressed by hepatocytes in hepatic acinus, while Cx26 is preferentially expressed by hepatocytes localized in the periportal spaces. Thus, most non-parenchymal cells express Cx43 while hepatocytes express Cx26 or Cx32. Considering that hepatic cells establish cell-to-cell communication by channels consisted of different connexins it is not surprising that they present different permeability (permselectivity). It was demonstrated that Cx32 gap junctions had a 10-fold higher relative permeability to adenosine compared with Cx43 channels. By contrast, the phosphorylation status of the adenosine shifted its preferential selectivity from Cx32 channels towards those formed by Cx43. The Cx43 intercellular channels were eight times more permeable to AMP and ADP than Cx32 channels, and the permeability of ATP through Cx43 was more than 300-fold better than that through Cx32 channels. Thus biophysical properties of gap junctional channel depend

gap junction plaque (Hunter et al., 2003).

and ubiquitination (Leithe and Rivedal, 2004).

**3. Gap junctions and connexin expression in normal liver** 

on the type and phosphorylation status of the connexin that form the channel.

In normal conditions, adult liver display very low proliferative activity. However after injuries or partial hepatectomy, the remaining intact cells start to proliferate. During this proliferation stage, it was observed decreasing of Cx26 and Cx32 expression in hepatocytes. Similar situation was observed when hepatocytes from primary culture were stimulated by mitogen (Kojima et al., 1997). It is important to remember that during liver regeneration process there is not only cell proliferation of hepatocytes but also cell differentiation of oval cells into hepatocytes. Oval cells naturally express Cx43, however along the differentiation program this cell type switch the connexin isoform preferentially express from Cx43 to Cx32 and/or Cx26 (Zhang et al., 1994). Hepatocytes cultured *in vitro* commonly increase Cx43 expression with the concomitant decrease of Cx32 expression (Figure 3). Thus, the immortalized cell lines derived from liver express Cx43 instead of Cx32. Furthermore, hepatocarcinogenesis process leads the hepatocytes to express again Cx43, event that contributes to lose of the differentiated phenotype.

Fig. 3. Laser scanning confocal microscope images of BRL3A (Normal liver cells) submitted to immunofluorescence reaction with anti-connexin43 antibody: A) TRITC-phalloidin stained microfilaments; B) Cx43 is presented in the cytoplasm and gap junctions (green); C) nuclei in blue and D) merged channels .

Gap Junction Intercellular Communication and Connexin

Expression Profile in Normal Liver Cells and Hepatocarcinoma 281

Fig. 4. Scrape loading/dye transfer\* assay performed in BRL3A (A, C and E) and HTC cells and HTC cells (B, D and F).Differential interference contrast microscopy images in A and B, fluorescent images in C and D, and merge in E and F. BRL3A presented good communication capacity because it possible to observe at least 5 rows of fluorescent cells from scrape. HTC cells was deficient in GJIC, the fluorescent cells were visualized only in areas near to scrape.

\*Scrape loading and dye transfer (SL/DT) is a functional assay widely used to evaluate the level of the intercellular communication and it is based in the introduction of the non-permeable fluorescent dye (Lucifer Yellow, MW= 457,2) into cells of monolayer culture through a transient cut in the cell

membrane. Lucifer Yellow does not diffuse through intact membrane but it is transferred into adjacent

cells via GJ in competent cells. The transference is monitored with fluorescence microscopy.

Although the relationship between GJIC and cell proliferation is well established, its involvement in liver cell apoptosis is not fully understood. It was demonstrated that GJIC is induced in the early phases of apoptosis in serum-deprived rat WB-F344 liver epithelial cells parallel to the increased expression and phosphorylation of Cx43.
