**4. Zonation of VLDL secretion during sepsis**

It has been suggested that parenchymal capacity for VLDL secretion is zonated. Zonation refers to a phenotypic heterogeneity that is well established in many essential liver functions (Jungermann & Katz, 1989). While some authors suggested that VLDL secretion might be higher in perivenous (PV) hepatocytes because of their higher capacity for fatty acid synthesis (Guzman & Castro, 1989), others proposed that it could be concentrated in the periportal (PP) area (Kang & Davis, 2000) since higher expression of the cholesterol synthesis rate-limiting enzyme 3-hydroxy-3-methylglutaryl-CoA reductase was found (Singer et al., 1984).

Zonation has also been evidenced in non parenchymal liver cells. Kupffer cells are more abundant and larger in the PP than in the PV zone (Bouwens et al., 1992), and expression of IL-6, a key cytokine acting on hepatocytes in response to endotoxin, occur preferentially in the PP region (Fang et al., 1998).

After 18 h of endotoxin treatment, highly pure rat PP and PV hepatocyte subpopulations, assessed by cytometry, were maintained in suspension for 2 h. Endotoxin treatment provoked zonation of VLDL secretion. The induction in VLDL-apoB secretion was markedly higher in PP hepatocytes (~90%) than in PV cells (~38%). In addition, the increase in the VLDL associated lipid, particularly in triglycerides, was lower than the enhance in apoB output, consequently producing changes in VLDL features which were triglyceride poor (Aspichueta et al., 2005). Endotoxin doubled apoB mRNA and increased by 50% MTP mRNA in PP hepatocytes when compared to their fasted controls, the increase in apoB

we detected an elevation of 5 fold in the number of circulating VLDL particles, measured as apoB quantities, at 8 h from LPS administration, without any modification in apoB transcript level. The increment in VLDL-TG is of greater magnitude (8 fold), indicating that during the first phase of the septic response TG-rich VLDL particles accumulate in the circulation

Different mechanisms seem to be involved in the second phase of septic response. The serum fatty levels drop below controls, which would suggest a lower availability of fatty

Endotoxic rats showed a higher number (10 fold) of circulating VLDL particles in rats at 18 h, but the content of the lipid in each VLDL is reduced (Aspichueta et al., 2006; Bartolome et al., 2010). This was accompanied by high levels of *apob* gene transcript, which could provide for high apoB availability increasing the secretion of lipid poor VLDL particles. Using intact rats in the fasted state, injected with the LPL inhibitor Triton WR-1339, we have shown that the TG and cholesterol secreted into VLDL released by the liver to the blood in 2 h was not enhanced by LPS administration to the same extent as the VLDL-apoB production was (Aspichueta et al., 2005). In this way, hepatocytes isolated from 18 hour LPS-treated rats secreted TG-poor VLDL, and although secretion was highly stimulated, global triglyceride secretion in VLDL remained unchanged. This was related to unchanged rates of fatty acid esterification, measured as

In septic hypertriglyceridemic rats, 24 h after sepsis induction, the increase in plasma TG was associated to a decrease in VLDL-TG clearance rate, due to suppressed mRNA levels, protein mass and activity of LPL in peripheral tissues (Lanza-Jacoby et al., 1997; Lanza-Jacoby & Tabares, 1990). Thus, in the early phase of the septic reaction hypertriglyceridemia is mostly due to high VLDL secretion driven by availability of lipids in the hepatocyte; and during the second phase, hypertriglyceridemia would be the result of LPL inhibition, and the increase in apoB transcription would be responsible for the increased secretion of VLDL particles (Fig.3).

It has been suggested that parenchymal capacity for VLDL secretion is zonated. Zonation refers to a phenotypic heterogeneity that is well established in many essential liver functions (Jungermann & Katz, 1989). While some authors suggested that VLDL secretion might be higher in perivenous (PV) hepatocytes because of their higher capacity for fatty acid synthesis (Guzman & Castro, 1989), others proposed that it could be concentrated in the periportal (PP) area (Kang & Davis, 2000) since higher expression of the cholesterol synthesis rate-limiting

Zonation has also been evidenced in non parenchymal liver cells. Kupffer cells are more abundant and larger in the PP than in the PV zone (Bouwens et al., 1992), and expression of IL-6, a key cytokine acting on hepatocytes in response to endotoxin, occur preferentially in

After 18 h of endotoxin treatment, highly pure rat PP and PV hepatocyte subpopulations, assessed by cytometry, were maintained in suspension for 2 h. Endotoxin treatment provoked zonation of VLDL secretion. The induction in VLDL-apoB secretion was markedly higher in PP hepatocytes (~90%) than in PV cells (~38%). In addition, the increase in the VLDL associated lipid, particularly in triglycerides, was lower than the enhance in apoB output, consequently producing changes in VLDL features which were triglyceride poor (Aspichueta et al., 2005). Endotoxin doubled apoB mRNA and increased by 50% MTP mRNA in PP hepatocytes when compared to their fasted controls, the increase in apoB

enzyme 3-hydroxy-3-methylglutaryl-CoA reductase was found (Singer et al., 1984).

[3H]oleate incorporation into TG (Aspichueta et al., 2005; Aspichueta et al., 2006).

acids of extrahepatic origin for hepatic VLDL-TG secretion.

**4. Zonation of VLDL secretion during sepsis** 

the PP region (Fang et al., 1998).

(Bartolome et al., 2010).

genetic expression was of a lesser extend in PV cells. Regarding to de novo synthesis of lipids for VLDL assembly, the incorporation of [3H]acetate into TG and cholesterol did not change by endotoxin challenge.

We concluded that periportal and perivenous hepatocytes exhibited similar capabilities for VLDL assembly and secretion in normal conditions; and, only the endotoxic condition led PP hepatocytes to a marked increase in TG-poor VLDL secretion (Fig 4).

Fig. 3. Proposed model for the biphasic response to endotoxin in VLDL metabolism. In the first phase the stimulation of lipolysis provides fatty acids that are taken by the liver and esterified to be secreted into TG-rich VLDL. In the second phase apoB mRNA levels are increased providing the apolipoprotein for secretion of TG-poor VLDL.

Fig. 4. Model of VLDL secretion by periportal and perivenous hepatocytes in fed state and 18 h after fasting or endotoxin treatment. Endotoxin effect when compared with the fasted state is marked with

Disrupted VLDL Features and Lipoprotein Metabolism in Sepsis 207

2008; Perez et al., 2006). IL-1β was the most potent and was the only one presenting a doseresponse effect. The effect of the three cytokines was redundant, as the increase was not additive when they were combined. However, none of the treatments with cytokines modified the amount of TG and total lipids secreted as components of VLDL, suggesting

We conclude that Kupffer cells play a role in the rise of VLDL secretion detected during the inflammatory processes and that the three cytokines TNF-α, IL-6 and IL-1β may be involved, nevertheless other Kupffer cells mediators are necessary to accomplish increased

As stated before, the assembly of VLDL is a complex process that depends on the

Since we have found that under LPS treatment VLDL-apoB secretion was always increased, and given that not always enhanced apoB secretion is linked to high levels of apoB transcript, we hypothesized that during the acute phase response, transcriptional or posttranscriptional regulation affecting apoB mRNA levels might occur supplying more apoB

During the first phase of the septic response we detected elevated circulating VLDL-apoB and -TG after 8 h of LPS treatment without altered apoB transcript levels. Taking into account that at this time point of the septic response, circulating fatty acid levels were elevated, we propose that fatty acid uptake by the liver is increased and large amounts of TG are synthesized. Since the N-terminus of apoB acquires neutral lipids in the endoplasmic reticulum membrane (Hussain et al., 2008), more nucleation sites are expected to be generated in apoB leading to increased apoB secretion. This could result in an increased hepatic secretion of triglycerides in VLDL particles, which would accumulate in the circulation, even in the absence of augmented levels of hepatic apoB mRNA (Bartolome et

In the second phase of the septic response, after 18 h of LPS challenge, enhanced VLDLapoB secretion is accompanied by increased apoB mRNA levels (Bartolome et al., 2010). In addition, hepatocytes isolated 18 h after LPS administration presented higher levels of apoB transcript and secreted more VLDL particles, been this effect more marked in PP cells. At this time point of the septic response, lipid poor VLDL particles are secreted (Aspichueta et al., 2005) and lipid synthesis is not modified (Aspichueta et al., 2006). Therefore, the increase in *apob* gene transcript would provide the additional apoB necessary to enhance VLDL-apoB secretion. Similarly, the inflammatory cytokines TNF-α (Bartolome et al., 2007), IL1-β (Bartolome et al., 2010) and IL-6 (Perez et al., 2006) augmented the levels of apoB mRNA and secretion of VLDL particles without changing

Our hypothesis was that endotoxin-enhanced VLDL-apoB secretion was driven by higher transcription rates. However, we did not find a rise in transcription rate of *apob* gene when we measured the incorporation of 5'-[α-32P]-UTP into newly synthesized RNA in liver nuclei from 16 h LPS-treated rats (Bartolome et al., 2010). We reported that global transcription rate in endotoxic liver was nearly two times higher than in control rats as expected in the acute phase response for up-regulating the positive proteins. However, the transcription rate of

that these particles are lipid poor.

**6. Higher apoB availability within the hepatocytes** 

availability of lipids and apoB (Davidson & Shelness, 2000).

the amounts of lipid secreted in the VLDL.

lipid association.

for VLDL assembly.

al., 2010).
