What is the optimal lipid emulsion for preventing intestinal failure-associated liver disease following parenteral feeding in a rat model of short-bowel syndrome?

The survival prognosis of patients with intestinal failures, such as short-bowel syndrome (SBS), intestinal motility disorder, and inflammatory bowel disease, has dramatically improved due to the development of total parenteral nutrition (TPN). Many patients with intestinal failure require long-term PN, which can cause complications, such as catheter-related blood stream infection (CRBSI) and intestinal failure-associated liver disease (IFALD). Kelly et al. stated that IFALD is characterized by a progressive pathology of cholestasis and steatosis inducing hepatic fibrosis and ultimately leading to liver cirrhosis [1]. Cholestasis occurs in 15–80% of IFALD neonates and infants, and steatosis, also known as non-alcoholic fatty liver disease (NAFLD), occurs in 40–60% of IFALD adults.

Several studies have implicated soybean oil lipid emulsion (SOLE) in the occurrence of hepatic cholestasis and steatosis of IFALD for two reasons. First, SOLE is rich in ω-6 polyunsaturated fatty acids (PUFAs), such as linoleic acid. Arachidonic acid, a metabolite of linoleic acid, is a precursor of inflammatory cytokines. Second, phytosterols contained in SOLE may exacerbate the development of cholestasis through the inhibition of bile acid transporters.

Fish oil lipid emulsion (FOLE) is rich in ω-3 PUFAs, such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). EPA is a substrate of precursors with anti-inflammatory effects. Since the ω-6 PUFA linoleic acid and the ω-3 PUFA α-linolenic acid are metabolized by the same enzymes, the metabolisms of these molecules compete. Furthermore, since FOLE contains less phytosterol than SOLE, it has been used to prevent and treat IFALD [2, 3].

However, as FOLE contains an insufficient amount of ω-6 PUFAs for nutritional needs, concerns about essential fatty acid deficiency (EFAD) and growth failure, especially for children, have arisen. Composite lipid emulsion (CLE), also known as SMOFlipid^Ⓡ (Fresenius Kabi Australia Pty Ltd.), is composed of soybean oil (SO) (30%), medium-chain triglycerides (30%), olive oil (25%), and fish oil (FO) (15%). It may be a better lipid emulsion than SOLE and was recently developed and used as a first lipid emulsion due to its nutritional benefits. Composite lipid emulsion has features such as a reduced amount of phytosterols and a different fatty acid composition, including EPA and DHA, compared with SOLE. However, while CLE has been used clinically [4], Lee et al. showed that long-term CLE resulted in cholestasis of IFALD, and changing to FOLE monotherapy improved IFALD [5]. Furthermore, dyslipidemia has been reported as an adverse event in children [6]. The effects of CLE and FOLE in SBS patients are thus unclear at present.

No experimental studies so far have evaluated the effect and mechanisms of lipid emulsions for hepatic steatosis of IFALD in a parenterally fed model following massive bowel resection. In addition, pediatric surgeons sometimes clinically encounter IFALD following massive bowel rection and parenteral nutrition.

We, therefore, compared the effect of three different lipid emulsions (SOLE, CLE and FOLE) for inducing hepatic steatosis of IFALD in a parenterally fed rat model of SBS.

Patients with intestinal failure who require long-term PN are at risk of developing IFALD. Lauriti et al. reported that 50% of patients undergoing long-term PN were afflicted with this disease [11]. IFALD is associated with cholestasis and steatosis and progresses to liver cirrhosis and liver failure. Among potential risk factors, the intravenous administration of SOLE may be involved in the development of IFALD. Several studies have shown that optimal lipid emulsion might improve the prognosis of SBS patients [2, 12, 13]. We conducted our study to compare the efficacy of three lipid emulsions for hepatic steatosis of IFALD using a parenterally fed rat model following massive bowel resection.

The major findings of the current study were as follows: (1) the serum ALT level was not associated with histological hepatic steatosis; (2) the serum T-CHO levels in the SOLE group were significantly higher than in the FOLE group; (3) the serum β-sitosterol and campesterol levels in the SOLE group were significantly higher than in the other three groups; (4) regarding hepatic histology, SOLE and CLE induced severe and moderate hepatic steatosis, respectively, while FOLE did not steatosis at all; (5) the levels of hepatic inflammatory cytokines, especially TNF-α, were significantly lower in the FOLE group than in the SOLE group; and (6) PN with FOLE did not induce EFAD based on the TT ratio.

The serum ALT level in the control group was significantly higher than in the SOLE group, but this value slightly exceeded the upper limit of the normal range. The most important point of our serum biochemical tests was that the serum ALT level did not reflect the histological severity of hepatic steatosis. The severity of hepatic steatosis was not necessarily associated with the hepatic transaminase level, so we must be alert for hepatic steatosis during PN, even when hepatic transaminase levels are within normal limits clinically.

In our experiments, the serum T-CHO level in the SOLE group was significantly higher than in the FOLE group. A recent clinical study for pediatric patients with IFALD showed that the increase in serum plant sterol levels by SOLE augmented cholesterol synthesis [14]. In our study, the β-sitosterol and campesterol levels in the SOLE group were significantly higher than in the other three groups. The augmentation of plant sterol levels due to SO in the SOLE group may induce significant hypercholesterolemia.

According to the histological liver injury grading, hepatic steatosis in the SOLE groups was significantly more severe than in the C and FOLE groups, and that in the CLE group was also more severe than in the C group. Moderate and severe hepatic steatosis was observed in both the SOLE and CLE group, which included SO. Regarding the lobular inflammatory grading, surprisingly, the CLE and FOLE groups had statistical significance compared with the C group. It should be noted that there was no significant difference between the C and SOLE groups, but instead that the lobular inflammation in the SOLE and CLE groups alone showed moderate grade damage. The hepatocyte ballooning in the SOLE group was significantly more severe than that in the FOLE group, despite there being no significant difference from the C group. The results of hepatocyte ballooning are considered to be statistical interpretations based on differences in sample size. These findings suggest that the histological liver injury caused by TPN in the SBS rat model may be associated with SO. Nandivada et al. previously reported that parenteral SO with parenteral FO, even in a 1:1 ratio, induced hepatic steatosis [15]. Their experimental SO fat dose included 1.2 g/kg/day. In our experiment, the SO fat dose in the CLE group was only 0.6 g/kg/day. Based on these results, even a small amount of SO fat may strongly affect hepatic steatosis. In addition, the findings of lobular inflammation and hepatocyte ballooning were not concomitant with the level of inflammatory cytokines. Hence further studies are required to clarify the relationship between the histological changes, such as lobular inflammation and hepatocyte ballooning, and the cytokine levels.

SOLE supplies abundant ω-6 PUFAs, such as linoleic acid, which is metabolized into arachidonic acid. Arachidonic acid is a substrate of pro-inflammatory eicosanoids. In contrast, FOLE supplies abundant ω-3 PUFAs, such as EPA and DHA. EPA is a substrate of anti-inflammatory eicosanoids. Because ω-3 and ω-6 PUFAs have competing metabolisms. In our study, according to its fatty acid profile, nevertheless SOLE itself did not include arachidonic acid, while the amount of arachidonic acid in the SOLE group was significantly higher than in the other three groups. Since SOLE contained five times linoleic acid of ω-6 PUFAs as much as α-linolenic acid of ω-3 PUFAs, this result implied that SOLE would metabolize ω-6 PUFAs more significantly than ω-3 PUFAs. Several studies have also suggested that hepatic steatosis may be induced by SOLE and attenuated by FOLE [2, 16], with similar results shown here.

In our study, however, moderate steatosis was observed in the CLE group. According to the fatty acid profile, arachidonic acid levels in the CLE group were significantly lower than in the SOLE group but higher than in the FOLE group. Furthermore, EPA and DHA levels in the CLE group were significantly lower than in the FOLE group but higher than in the SOLE group. Given the interaction between ω-6 and ω-3 PUFAs, we speculate that an imbalance of EPA, such as an anti-inflammatory effect, and arachidonic acid, such as a pro-inflammatory effect, may have been associated with the occurrence of hepatic steatosis in the CLE group. The composition of CLE may, therefore, require some modification to prevent hepatic steatosis.

Many studies have explored the relationship between steatosis and inflammatory cytokines. We evaluated hepatic IL-6 and TNF-α levels. Given the relationship between the severity of hepatic steatosis and cytokines, the expression of TNF-α was more concomitant with the severity of hepatic steatosis than that of IL-6. Hotamisligil et al. also stated that TNF-α was a key factor in the development of NAFLD [17]. In the present study, hepatic TNF-α levels were significantly higher in the SOLE group and non-significantly higher in the CLE group than in the FOLE group. Although we speculated that TNF-α in the liver would play a pivotal role in the development of hepatic steatosis in our model, however, we have not measured anti-inflammatory cytokines in this study, and further studies are needed to clarify the mechanism. Our experimental model closely reflects clinical IFALD because of the PN feeding following a massive bowel resection, making it useful for further studies.

Hepatic autophagy is impaired in NAFLD, and Tanaka et al. suggested that Rubicon, an autophagy-regulating protein, played a pathogenetic role in NAFLD by inducing lipid accumulation via the inhibition of autophagy [18]. We, therefore, suspected that autophagy was also related to steatosis in IFALD and analyzed the Rubicon antibody levels using Western blotting. The expression of Rubicon antibody in the SOLE group was highest, and that in the CLE group was higher than in the control and FOLE groups. Given our result, we speculated that Rubicon might be involved in the development of hepatic steatosis in our model. Further studies will be required to clarify this mechanism. If autophagy is indeed related to the onset of IFALD, as with NAFLD, this finding may help establish prevention and therapeutic methods in the future.

The current study showed that FOLE attenuated steatosis, just as was shown in previous reports [13]. However, given that this emulsion contains < 7% linoleic acid, there is a risk of EFDA with the standard dose of FOLE, so manufacturers do not recommend FOLE monotherapy. The TT ratios in the FOLE and CLE groups in our study were increased compared with the C group but still well below 0.2, and clinical symptoms of EFAD, such as hair loss and dry skin, were not observed. Given the TT ratio in our study, FOLE and CLE therapy appear safe and acceptable without risk of EFAD. However, this result is only based on the findings of a 2-week animal experiment, and it is, therefore, necessary to carefully observe the clinical symptoms and measure the fatty acid fraction levels regularly in actual clinical practice.

Hepatic steatosis of IFALD, known as NAFLD, was observed after the administration of SOLE and CLE using a parenterally fed rat model with SBS, which is an authentic IFALD model, as these surgical and nutritional conditions closely resemble those of clinical cases of SBS. Hepatic steatosis was not observed on the administration of FOLE. Based on the histological findings of hepatic steatosis, FOLE attenuated hepatic steatosis without EFAD and CLE may require modification for IFALD patients with long-term PN. Furthermore, inflammatory cytokines, especially TNF-α, and Rubicon may be associated with the development of hepatic steatosis. Further studies are needed to clarify (1) the mechanism underlying the onset of hepatic steatosis using lipid emulsion with SO and (2) the optimal composition of CLE that will not induce hepatic steatosis.