The molecular bases underlying the alterations in the circulating PON1 activity together with its synthesis and secretion in chronic liver diseases are far from being completely understood, despite these changes being of crucial importance in the pathophysiology of chronic liver impairment. As illustrated in the present investigation, patients with liver cirrhosis had an increased serum concentration of total peroxides (a marker of oxidative stress), MCP-1 (an index of inflammation), and P-III-P (a marker of liver fibrogenesis). All these changes were observed to be strongly associated with a decrease in serum PON1 activity. Although a direct functional relationship between serum PON1 activity alterations, inflammation and fibrogenesis cannot be deduced from the present investigations, these data provide further support for studies showing that oxidative stress activates hepatic stellate cell and macrophages in vitro and in experimental models of liver impairment [32, 36–38]. A possible answer to the question of why serum PON1 activity is decreased is that PON1 is inactivated by oxidized lipids, as was shown by Aviram et al.  who demonstrated that the incubation of PON1 in vitro with oxidized palmitoyl arachidonoyl phosphatidylcholine, lysophosphatidylcholine, oxidized cholesteryl arachidonate and oxidized LDL resulted in inactivation of PON1 arylesterase activity. Indeed, in a previous study, we observed that alcoholic patients with a normal, or with minimally-affected liver function, already had a decreased serum PON1 activity that was associated with an increase in the circulating concentration of malondialdehyde . These data support the hypothesis of a direct inhibition by lipid peroxides on the PON1 enzyme active site. However, liver cirrhosis is a chronic and diffuse disease that is usually protracted in its development, and in which multiple metabolic derangements appear progressively, including an altered capacity of pan-protein synthesis by the liver. A likely possibility (and one which does not exclude an inhibition by lipid peroxidation products) is that, in advanced liver impairment, changes in HDL structure and composition influence PON1 activity. This appears feasible since it is well documented that PON1 activity is profoundly dependent on the lipid and protein compositional environment of the HDL particles .
Results from the present study suggest that PON1 activity and mass are associated, both in control subjects as well as in cirrhotic patients, with a specific sub-population of HDL particles containing apo A-I, but with not apo A-II and apo E. These results agree with a study by Cabana et al.  that showed a similar distribution of PON1 in human HDL sub-fractions obtained by FPLC and with experimental studies in apo A-II transgenic mice in which apo A-II enrichment was found to displace PON1 from HDL and, as well, to impair the antioxidant and athero-protective function of this particle [42, 43]. However, our results differ from other reports in which HDL was isolated by other methods. Bergmeier et al.  observed a high PON1 activity in HDL3 particles isolated by ultracentrifugation, and which contained apo E. Moren et al.  observed considerable PON1 activity in apo A-II-containing HDL isolated by immunoprecipitation. Methodological differences probably account for the observed discrepancies, but we believe that HPLC or FPLC fractionation methods are less physico-chemically aggressive and, as such, are less likely to alter HDL composition in the process of isolation.
Non-denaturing gel electrophoresis showed a considerable increase in very small and very large HDL particles from cirrhotic patients compared to the control subjects. These changes are probably secondary to alterations in CETP, PLTP, and LCAT levels in these patients. Although reduced LCAT activity is common in a variety of liver diseases , the current report is, to the best of our knowledge, the first to identify increased CETP and PLTP activities as being associated with chronic liver impairment. These findings will require confirmation in much larger studies of cases versus controls. The mechanisms leading to increased CETP and PLTP activities in chronic liver disease are also unclear at present. The outcome observed could be due to increased synthesis and secretion, decreased degradation (or both) or an entirely different mechanism such as an effect of HDL composition on enzyme activity. Again, further studies are required. The combination of decreased LCAT and increased CETP activities in the patients probably resulted in the increase in small HDL particles. However, the greater PLTP activity would also have resulted in increased fusing of small HDL into the larger HDL particles found in the patients [47–49]. Therefore changes in the activities of the lipid transfer proteins could explain the different pattern of HDL particles found in the patients. An increase in PLTP would lead to larger HDL particles, while an enhancement in CETP and a decrease in LCAT would result in smaller HDL particles with an increase in the HDL3 sub-fraction. While the impact of a larger particle size in relation to HDL function remains unclear, extensive data from the literature show that small HDL particles lose most of their antioxidant properties, and may become pro-oxidant and pro-inflammatory instead [50, 51]. In addition, results from the present study suggest that these small particles do not carry PON1 and, as such, would contribute to the decrease in serum PON1 activity observed. However, it is also possible that changes in HDL size are associated with firmer binding of PON1 but not with changes in activity.
We noted an inverse relationship between phospholipid content and PON1 activity in HDL particles. In addition, considerable amounts of PON1 protein, but low enzyme activity, were observed in particles eluting at HPLC fractions #31-35 in which phospholipids concentrations were below the detection limit of the assay. These results suggest that an appropriate balance between phospholipids and PON1 is important in maintaining the enzyme's activity. Indeed, the proportional distribution of phospholipids and apo A-I in HDL has been shown to play an important role in PON1 secretion, stability and activity. Sorenson et al.  showed that PON1 was associated with HDL more via the binding to phospholipids than to apo A-I. Nevertheless, apo A-I stabilized the enzyme activity more than did phospholipids alone. These results were confirmed by Deakin et al. . Hence the hypothesis follows that an increased phospholipids to apo A-I ratio, as observed in cirrhotic patients, contributes to PON1 instability. Previous studies from our group [22, 23, 31] showed that PON1 activity in patients with liver disease is inversely related to PON1 mass. These results prompted us to investigate the nature of PON1 protein in patients with cirrhosis. One of the more interesting findings from the present investigation has been the identification by Western blot of an abnormal protein that reacted against the anti-PON1 antibody, had a higher molecular weight than PON1, and was present in the HDL from cirrhotic patients but not in control subjects. This protein was also present in the livers of patients with steatosis (a more benign form of liver disease) and in rats with CCl4-induced cirrhosis. Additionally in these animals, the amount of abnormal PON1 increased along the time-course of the disease development. These results highlight the presence of this protein in liver impairment per se, irrespective of the etiology or, even, the animal species. The incubation of HDL with PNGase F strongly suggests that this protein is, indeed, a highly glycosylated form of PON1. Interestingly, Liu et al.  observed that when they produced recombinant human PON1 in a baculovirus system in Hi5 insect cells, a significant proportion of PON1 was synthesized as a protein of approximately 54,000 Da, and which they identified as a PON1 multimer. Taking these data together, we can hypothesize that abnormal PON1 molecules are synthesized under diverse altered conditions such as liver disease, or when human PON1 is synthesized by phylogenetically-distant cells. An important finding of the present study is that when a liver homogenate from a patient with steatosis was incubated in vitro with normal HDL, the altered immunoreactive band was no longer observed. These results suggest that the high molecular weight PON1-like protein is actively incorporated into the HDL particles within the hepatocytes, and that it is not a mere artifact or a result of a non-specific adsorption of PON1 fragments by HDL particles in the circulation. The possibility exists that highly glycosylated PON1 is abnormally assembled into HDL particles in liver disease secondarily to alterations in protein synthesis in hepatocytes or to altered hepatic lipid metabolism.