It well known that the induction of proinflammatory cytokines, such as TNFα and IFNγ, causes the catalytic beta subunits β1, β2 and β5 of the constitutive 20S proteasome to be replaced by the LMP2, MECL-1 and LMP7 subunits (28–31), thus forming the immunoproteasome (). Microarray analysis data mining of control vs. DDC refed comparisons on 3 mice per group showed that DDC refeeding caused up regulation of various immunoproteasome catalytic subunits (). In addition, the PA28 alpha subunit, which is part of the regulatory complex 11S in the immunoproteasome (), was up regulated. It is postulated that there was a proinflammatory response associated with DDC refeeding that stimulated the formation of the immunoproteasome at the expense of the 26S proteasome. MDBs formed because of the loss of the turnover of the cytokeratins by the 26S proteasome.
Figure 1 A: Diagram showing the location of the immunoproteasome subunits, which substitute for the catalytic β1, β 2, and β 5 subunits of the 20S proteasome (26S proteasome); B: Gene microarray data mining showed that the immunoproteasome (more ...)
It has been shown that FAT10 is significantly induced in the liver of mice fed DDC for 10 weeks, as well as in the liver of mice refed DDC for 1 week (Oliva et al., 2008
). To analyze the effect of DDC on the immunoproteasome subunits, liver sections were double stained for ubiquitin and FAT10, as well as for ubiquitin and LMP2. FAT10 and LMP2 were found to be highly over expressed in hepatocytes, forming MDBs when mice were refed DDC (). There was some colocalization of LMP2 and FAT10 in the MDB aggregates (). The confocal microscopy analysis showed that LMP2 and FAT10 co-localized in the cytoplasm of hepatocytes (, merged yellow photograph) in a DDC primed mouse liver. Individual hepatocytes over expressed both LMP2 and FAT10 (: FITC for FAT10 and Texas Red for LMP2). Note that intervening normal hepatocytes that did not form MDBs, and did not over express either LMP2 nor FAT10. Confocal analysis also showed that DDC feeding caused the increase of the other immunoproteasome subunits LMP7 and MECL-1 (), along with the induction of FAT10. Therefore, DDC feeding for 10 weeks, as well as DDC refeeding for 1 week, caused an up regulation of FAT10, as well as an up regulation of each immunoproteasome subunit, LMP2, LMP7, and MECL-1, in hepatocytes forming MDBs. To test the hypothesis that DDC feeding causes a switch from the 26S proteasome to the immunoproteasome, the function of the 26S proteasome and the catalytic subunits of the immunoproteasome were measured in MDB forming livers of drug primed mice. illustrates the effects of both DDC feeding and DDC refeeding. It shows that there was no difference between DDC feeding and DDC refeeding. The immunoproteasome subunits were up regulated and the beta 5 subunits that carry the chymotrypsin-like activity for the 26S proteasome was down regulated, whereas there was no significant change in the alpha type subunits. The rationale of using DDC refeeding was to demonstrate that the hepatocytes have a memory, and that the same phenotype, obtained in 10 weeks of DDC feeding, was again obtained in 1 week of re-exposure to the drug. These results indicated that an epigenetic memory is formed in the hepatocytes following the first exposure, and that SAMe is a powerful modulator of this memory (Bardag-Gorce et al., 2007
; Li et al., 2008
Figure 2 Antibodies to FAT10 (green) and Ubiquitin (red) (A, B and C) were used to double stain hepatocytes from DDC treated mice. Note that the MDB forming cells cytoplasm stained positive for FAT10 (green). MDBs were stained with ubiquitin (red). Normal hepatocytes (more ...)
Figure 3 Confocal microscopic analysis of LMP2-FAT10 (A and B), LMP7-FAT10 (C), and MECL-1-FAT10 (D) double antibody staining of hepatocytes from a DDC primed mouse liver. The merged photo (yellow) showed colocalisation of LMP2, LMP7 and MECL-1 with FAT10 in the (more ...)
Figure 4 DDC feeding for 10 weeks and DDC refeeding for 1 week caused a significant up regulation of the immunoproteasomes subunits LMP2, LPM7 and MECL-1. However the level of 26S proteasome beta5 subunit was decreased indicating there was a switch of the 26S (more ...)
Western blot analysis and RT-PCR showed that LMP2 () and LMP7 () were up regulated in the livers of mice refed DDC. SAMe supplementation in the diet prevented the up regulation of the immunoproteasome subunits (). MECL-1, the 3rd immunoproteasome catalytic subunit, was also up regulated by DDC refeeding, and SAMe also prevented this up regulation ().
Figure 5 Immunoproteasome LMP2, LMP7, and MECL-1 subunits induction in the liver homogenates of mice refed DDC and mice refed DDC + SAMe, as shown by Western blot ((a) and (c), left), and by RT-PCR ((b) and (c), right). (d) is the loading control. The striped (more ...)
The gene expression of TNFα was also analyzed by RT-PCR. It was up regulated by DDC refeeding. This up regulation was prevented by adding SAMe to the DDC diet (). Likewise, TNFα and IFNγ receptors were up regulated, which could enhance the response to TNFα and IFNγ.
Figure 6 The gene expression of the receptors for IFNγ, analyzed by PCR, showed a significant increase of both IFN receptor isoforms R1 and R2 (a,b). They form the IFN heterodimer in the liver of mice refed DDC. SAMe prevented this increase. Tumor necrosis (more ...)
The above results suggested that DDC refeeding caused an up regulation of the proinflammatory cytokine response, which induces the formation of the immunoproteasome. Therefore, to further investigate the effects of DDC feeding on the proteasome activity, proteasome chymotrypsin-like activity was measured, and the results showed that the 26S proteasome activity was significantly decreased by DDC refeeding (), while the 20S proteasome activity showed no change. These results indicated that DDC refeeding specifically affected the polyubiquitin 26S proteasome pathway responsible for the degradation of ubiquitinated proteins (). In addition, when mice were withdrawn from the DDC treatment, 26S proteasome activity recovered. However, when the mice were re-exposed to DDC for 7 days, the 26S proteasome activity was, again, significantly decreased (). These results indicated that there was a cellular memory of the first DDC exposure in the hepatocytes. An epigenetic mechanism was involved in the DDC-induced proinflammatory response and accumulation of polyubiquitinated proteins. shows that the addition of SAMe to the DDC diet, a major methyl donor in the remethylation pathway, prevented the inhibition of the 26S proteasome. It also prevented the accumulation of polyubiquitinated proteins caused by the inhibition of the 26S proteasome both by feeding DDC () and by refeeding DDC (), as determined by Western blot analysis. The proteasome subunit composition strongly influences the proteolytic functions of the proteasome complex. The types of proteins digested by the two types of proteasome are different.
Figure 7 (a) 26S proteasome activity in the cytosolic fraction from liver homogenates of mice fed DDC for 1 month (DDC), mice fed DDC 1 month, then withdrawn for 1 month (DDC Wd), and mice fed DDC, withdrawn, and then refed DDC for 7 days (Refed DDC). Control (more ...)
To determine the mechanism of MDB formation and test the role of IFNγ induction in MDB formation, primary cell cultures, isolated from 1 month withdrawn DDC primed mice, were treated with IFNγ (200 U/ml) (Shenandoah, Warwick, PA) (). The cultures were incubated for 6 days, and the cells were double stained for ubiquitin (Texas red) and FAT10 (FITC green). MDBs formed in gigantic hepatocytes (120 µm in diameter) after IFNγ treatment, and stained positive for ubiquitin (red) (). The number of the MDBs in the hepatocytes treated with IFNg was 4 fold higher than the number of MDBs in the hepatocytes with no treatment ().
Figure 8 Double stain for ubiquitin (Red) and CK-8 (Green) was performed on primary liver cells cultures grown for 6 days with or without IFNγ (100 U/ml) added. The cells were isolated from the liver from a DDC-primed mouse and a normal control mouse. (more ...)