Carbamazepine Mitigates Total Parenteral Nutrition Associated Liver Disease in a Novel Ambulatory Piglet Model

Abstract

Background:

Total parenteral nutrition (TPN) remains a critical therapeutic option in patients who cannot tolerate enteral feeding. However, while lifesaving, TPN is associated with significant side effects including liver injury, the etiology of which is multifactorial. Carbamazepine (CBZ), an anti-epileptic medication, is known to modulate hepatic fibrosis and hepatocellular injury in a variety of liver diseases. We hypothesized that CBZ could prevent parenteral nutrition associated liver disease (PNALD), which we tested using our novel ambulatory TPN piglet model.

Methods:

Piglets were fitted with jugular catheters and infusion pumps for TPN and randomized to enteral milk feeding (EN, n=7), TPN (TPN, n=6), or TPN with parenteral CBZ (CBZ, n=6) for 2 weeks. Serum and liver tissue were analyzed via light microscopy, quantification of serum liver injury markers, Ki67 and CK-7 indexing, and RT-qPCR.

Results:

TPN-fed piglets in our model developed manifestations of PNALD, particularly increased serum bilirubin, γ-glutamyl transferase, liver cholestasis, and Ki67 expression compared to EN animals (p<0.03). CBZ therapy in TPN-fed animals led to significant reduction in these markers of injury (p<0.05). Investigation into the mechanism of these therapeutic effects revealed increased expression of SREBP-1, PPAR-α, and FABP in TPN-fed animals receiving CBZ (p<0.03). Further investigation revealed increased LC3 expression and decreased LAMP1 expression with CBZ therapy (p<0.03).

Conclusion:

CBZ administration mitigates PNALD severity, suggesting a novel therapeutic strategy targeting TPN associated side effects, and may present a paradigm change to current treatment options.

1. Introduction

Total parenteral nutrition (TPN) is an alternative to enteral feeding that supplies all nutritional needs intravenously and is clinically indicated for patients with impaired or lost gut function (1, 2, 3). While TPN therapy has many benefits, it is known to cause a myriad of injuries to the liver, including parenteral nutrition associated liver disease (PNALD), which is characterized by steatosis, cholestasis, dyslipidemia, hepatic inflammation, and fibrosis. Additionally, gut atrophy is also known to occur during TPN (4, 5). The mechanisms driving these detrimental changes are multifactorial. Prematurity, central line infections, TPN constituents such as lipids, altered gut microbiota, and gut-systemic signaling have all been implicated in the etiology of TPN-associated injury (6, 7).

Recent literature also indicates that dysregulation of autophagy may be a driver for PNALD and several other hepatic inflammatory diseases, including alpha-1 antitrypsin deficiency (8, 9). Indeed, data have shown that in alpha-1 antitrypsin deficiency, functional loss of autophagy results in increased inclusion body-like structures and increased cellular inflammation (9). While the mechanisms underlying whether and how TPN modulates autophagy are unclear, TPN administration has been shown to cause changes in mTOR signaling, a key regulator of protein translation and lysosomal degradation, and may overall dysregulate normal autophagic homeostasis (10, 11). Because of this finding that TPN may cause autophagic dysregulation, attention has been turned towards pharmaceuticals that may modulate autophagy, such as carbamazepine (CBZ), as a therapeutic target to mitigate PNALD.

CBZ, like several other neuroactive drugs in its class, is thought to deplete inositol, which in turn stimulates autophagy through decreasing levels of myo-inositol-1,4,5-triphosphate (12). Excitingly, CBZ was tested on an alpha-1 antitrypsin deficient model and was shown to reduce hepatic fibrosis as well as the intracellular load of the misfolded ATZ protein (13). Similarly, CBZ was shown to reduce hepatocellular death in fibrinogen storage disease, a closely related hepatic inflammatory disorder (14). With this in mind, we tested whether TPN would cause dysregulation of common autophagy genes and whether CBZ could reduce PNALD either through modulation of autophagy, or through a different mechanism, using our ambulatory piglet model (3).

2. Experimental Design and Methods

2.1. Study Protocol:

The protocol for CBZ treatment of neonatal pigs (piglets) was approved by the Institutional Animal Care and Use Committee of Saint Louis University (SLU Animal Use Protocol 2346, US Department of Agriculture registration 43-R-011). Our experiments were performed following the Guide for the Care and Use of Laboratory Animals. These piglets were procured from an approved class A vendor and, upon arrival, placed in heated cages. As previously described, 7–10 day old piglets were surgically implanted with jugular and duodenal catheters for administration of TPN and CBZ (15).

2.2. Animal Grouping:

Animals were randomly assigned into groups to receive enteral nutrition (EN) (n=7), IV TPN (TPN) (n=6), or IV TPN with enteral CBZ (CBZ) (n=6). CBZ (Sigma-Aldrich, St. Louis, MO, USA) was administered via slow bolus infusions at 30 mg/kg/day, divided into 2 daily doses, extrapolated from published studies evaluating autophagy responses to CBZ treatment (14, 16, 17). CBZ was delivered via the duodenal catheter after dissolving it in 0.5mL of dimethyl sulfoxide (DMSO). This vehicle control was also given to the control TPN animals.

2.3. Nutrition:

As published previously, the EN control group received a swine replacement formula (LitterLife, Merrick’s Inc., WI) at a rate of 260ml/kg/day with 25 g/kg lactose, 12.4 g/kg protein and 5g/kg fat supplemented with electrolytes, trace minerals and vitamins for a total of 187 kcal/kg/day (18). Piglets on TPN were continuously administered a commercial parenteral nutrition preparation (Clinimix E; Baxter, Deerfield, IL, USA) supplemented with Intralipid (Fresenius Kabi, Hersfeld, Germany), at 260 mL/kg/day. This nutrition provided a cumulative 182 kcal/kg/day, consisting of 26 g/kg dextrose, 11.05 g/kg protein, and 5 g/kg fat in the form of soybean oil consisting of linoleic, oleic, and palmitic acids, as well as electrolytes, trace minerals, and vitamins (3, 18). Isocaloric nutrition (19) was provided to all animals for a period of 2 weeks as described previously (3,18).

2.4. Animal Care, Euthanasia and Tissue Collection:

Animal weights were recorded daily. Continuous animal monitoring was performed in compliance with the Institutional Animal Care and Use Committee. The animals were also assessed daily by the Doctor of Veterinary Medicine (DVM) certified personnel. Post-euthanasia, the liver was removed completely after opening the abdomen. Next, the remaining small intestine was removed as described previously (3,18). Slices were cut from segments of the liver and small intestine to be weighed. Smaller pieces were then cut to be snap-frozen in liquid nitrogen and stored in −80°C for future analysis. Peripheral and portal vein blood was also collected for analysis. The Saint Louis University clinical pathology core laboratory was utilized for serum analysis.

2.5. Histology:

Segments of small intestine and liver tissue (2–3 cm) were fixed in 4% buffered formalin for 24 hours. Segments were then stored for 24 hours in 70% ethanol at room temperature. After storage, samples were embedded in paraffin. Slides were then stained with hematoxylin and eosin (H&E), Ki67, and Cytokeratin-7 (CK-7).

Ki67 is a nuclear protein expressed in all cell cycle stages except that of labile cells (G0), which permits its use as a marker of cellular proliferation. The ratio of Ki67 positive nuclei can serve as an objective quantification of cell proliferation (20). CK-7 is a cytokeratin that is found in the epithelial lining of bile ducts and is used to quantify bile duct proliferation (21, 22). Paraffinized liver tissue from each animal cohort was deparaffinized and subsequently stained with either Ki67 or CK-7 immunohistochemical stain. A pathologist, blinded to the experimental groups, then reviewed the stained liver tissue with light microscopy. The pathologist quantified the immunoreactive cells as a ratio of the total number of hepatocytes in 5 randomly selected, non-overlapping high-power fields. The outcomes were reported as the Ki67 index and CK-7 score, respectively.

The liver H&E slides were examined using light microscopy. A pathologist who was blinded to the group assignments then identified bile deposits to evaluate cholestasis, a known side effect of TPN-mediated liver injury (23). The cholestatic foci were totaled and reported in 3 non-overlapping high-power fields to generate a liver cholestasis score, which was calculated from the number of cholestatic deposits per 20x image.

The small-bowel epithelium was assessed by quantifying the mean villous height and crypt depth in H&E stained slides looking at well-oriented vertical crypts and columns. The reviewers were blinded to the treatment; results were calculated as villous/crypt (V/C) ratio.

2.6. Tissue RNA extraction and RT-qPCR analysis:

RNA extraction was done with TRIzol (15596018; Thermo Fisher) and Sigma-Aldrich GenElute Mammalian Total RNA Miniprep Kit per protocol (RTN70-1KT) on liver tissue. Complementary DNA was generated from 1 ng of isolated RNA via Verso cDNA Synthesis Kit (AB1453B; Thermo Fisher). All primer sequences were validated for each transcript (Table S1) and ordered through Integrated DNA Technologies. Triplicates of RT-qPCR were performed via the Bio-Rad CFX Connect Real-Time System. Beta-actin was used for the internal control and relative mRNA levels were obtained via comparative threshold cycle method.

2.7. Statistical Analysis:

Statistical analysis was performed on GraphPad Prism version 8.4.2. Median and interquartile ranges (IQR) served as descriptive statistics. Serological markers, relative mRNA expression, and histology reads were analyzed with Mann-Whitney U tests and paired t-tests. All tests were 2-sided utilizing a significance level of 0.05.

3. Results

3.1. Baseline assessment and confirmation of CBZ therapy

A total of 19 piglets were used in this study. 7 were assigned to the EN group, 6 to the TPN group, and 6 to the CBZ group. At baseline mean weight for the animals was 3.2kg (Range: 3.0–3.4kg), with no statistical difference among the groups. Animals did not show statistical difference in daily weight gain over the course of the study, regardless of group (Figure 1a). CBZ intake was confirmed by evaluating serum levels obtained at animal euthanasia (6 hours after the last dose). Animals receiving CBZ had a serum concentration of 15.84 ± 1.54 mg/L, with EN and TPN both having < 0.06 mg/L (p=0.001) (Figure 1b).

Figure 1.

Group Characteristics. (A) Daily weight gain and (B) Serum carbamazepine level. No differences in daily weight gain among groups.

CBZ undetectable in EN and TPN animals. CBZ, carbamazepine; EN, enteral nutrition; TPN, total parenteral nutrition.

3.2. Assessment of markers of hepatic inflammation and injury

Elevation of bilirubin is an important marker of cholestatic hepatic injury. Higher levels of conjugated bilirubin were noted in animals receiving TPN compared to EN animals (0.11±0.09 vs. 0.29±0.08 mg/dL, p=0.003). There was a significant decrease in serum bilirubin in animals receiving CBZ treatment (0.21±0.03 mg/dL) compared to the TPN group (p=0.03) and the EN group (p=0.003) (Figure 2a). Gamma-glutamyl transpeptidase (GGT) is a marker for cholangiocyte injury and was significantly increased with TPN compared to EN (18.4±4.8 vs. 26.3±6.9 IU/L, p=0.028), but decreased with CBZ (19.8±2.7 IU/L, p=0.047) (Figure 2b). Bile acid levels were also assessed and were significantly elevated in piglets receiving TPN (9.5±3.6 vs. 15.4±3.1 μmol/L, p=0.01) relative to the EN group, with no change in the CBZ group (13.3±2.2 μmol/L, p=0.21) (Figure 2c).

Figure 2.

Measures of hepatic injury. (A) Serum bilirubin, (B) Gamma-glutamyltransferase, (C) bile acids, (D) liver cholestasis score, (E) Ki67 score, (F) CK-7. Note significant hyperbilirubinemia and high gamma-glutamyltransferase with TPN and its prevention with CBZ. CBZ prevented an increase in hepatic cholestatic foci as well as hepatocyte proliferation (Ki67).

CBZ, carbamazepine; EN, enteral nutrition; TPN, total parenteral nutrition.

The cholestasis score (foci/field) was significantly increased in TPN animals relative to EN animals (9.8±1.3 vs. 15.7±2.2, p<0.001) while subsequent CBZ treatment resulted in a similarly reduced score (12.5±2.4, p=0.038) (Figure 2d). Histological images of liver tissue reflect these increased areas of cholestatic foci in the TPN group (Figure 3). Evaluation of hepatocyte proliferation was examined through quantification of Ki67 stained histological slides, and a score was calculated using the total number of Ki67 positive nuclei per total nuclei as detailed in the histology section. A significant increase in the Ki67 score was noted in the TPN vs EN groups (11.9±2.1 vs. 21.3±2.9, p<0.001), while a significant decrease was found when comparing CBZ (15.2±3.2) to TPN (p=0.006) (Figure 2e). Our final measurement for hepatic injury was through quantification of immunohistology for CK-7, a marker for bile duct proliferation. There was a significant increase in hepatocytes positive for CK-7 in animals on TPN when compared to EN (15.6±4.1 vs. 44.3±7.1, p<0.001) but no significant change when comparing the TPN to CBZ group (39.8±8.5, p=0.172) (Figure 2f).

Figure 3.

Liver Histology. Microscopic images of liver tissue from (A) EN, (B) TPN, and (C) CBZ animals. Arrows highlight cholestatic foci.

CBZ, carbamazepine; EN, enteral nutrition; TPN, total parenteral nutrition.

3.3. Analysis of Gut Morphology

Several measures of gut health were compared between the EN, TPN, and CBZ animals to determine the extent to which carbamazepine treatment affects gut atrophy. In order to assess mucosal changes, we determined the small bowel villous/crypt ratio (V/C) (mean villous height/crypt depth) for each group. An observed decrease in V/C ratio was statistically significant between the EN and TPN group. (2.11±0.22 vs. 1.55±0.22, p<0.001). However, the difference between the CBZ group (1.64±0.32) relative to the TPN group was not statistically significant (p=0.28), indicating that CBZ did not prevent gut atrophy (Figure 4a).

Figure 4.

Gut Morphology. (A) The ratio of villous height to crypt depth was quantified as the V/C ratio. (B) Linear gut mass (LGM) of the (B) proximal and (C) distal small bowel. Note significant reduction in V/C ratio as well as proximal and distal LGM with TPN and CBZ.

CBZ, carbamazepine; EN, enteral nutrition; TPN, total parenteral nutrition.

The linear density (g/cm) of the proximal and distal small bowel was also evaluated as previously published (15). Proximal linear density was lower in TPN compared to EN (0.19±0.02 vs. 0.13±0.02 g/cm, p<0.001), with no change with CBZ (0.13±0.03 g/cm, p=0.67) (Figure 4b). Similarly for the distal small bowel, the TPN group demonstrated lower linear density than EN (0.31±0.02 vs. 0.19±0.03 g/cm, p<0.001) (Figure 4c).

3.4. Determination of Effect on Lipid Profile

Analysis of the serum lipid profile was conducted. There were no statistically significant differences in serum cholesterol levels among any of the treatment groups, EN, TPN or CBZ (128.1±12.8 vs. 141.3±24.0 vs. 139.7±22.0 mg/dL, p>0.05) (Figure 5a). A similar analysis was conducted for low-density lipoproteins (LDLs), which distribute lipids throughout the systemic circulation, and for triglycerides, an essential fasting energy source in the fasting state. No differences were observed between the three groups in LDL or triglyceride levels (73.6±18.9 vs. 83.2±15.0 vs. 76.7±21.2 mg/dL, p >0.05) (139.1± 27.9 vs. 154.8±18.2 vs. 144.5±20.7 mg/dL, p>0.05) (Figure 5b, 5c).

Figure 5.

Lipid Profile. (A) Serum cholesterol, (B) low-density lipoproteins, and (C) triglycerides. No differences were noted among the groups.

CBZ, carbamazepine; EN, enteral nutrition; TPN, total parenteral nutrition.

3.5. Quantitative PCR Results of Liver Genes

Quantitative PCR for key hepatobiliary receptor and transporter genes was conducted to evaluate their transcriptional expression. There was a significant decrease (1.00±0.31 vs. 0.64±0.25, p= 0.04) in fold change with TPN relative to EN in the bile salt export pump (BSEP) gene (Figure 6a). Gene expression of sterol regulatory binding protein 1 (SREBP-1) was stimulated by CBZ, with a significant increase in expression compared to TPN (1.07±1.01 vs. 4.82±2.76, p=0.025) (Figure 6b). The gene encoding for fatty acid binding protein (FABP), which is integral for lipid signaling pathways, demonstrated a similar trend (0.79±0.84 vs. 4.61±3.73, p=0.025) (Figure 6f). Additionally, a decrease in the expression of peroxisome proliferator activated receptor α (PPAR-α), a gene involved in fatty acid oxidation, was observed in animals on TPN compared to EN (1.08±0.39 vs. 0.24±0.16, p=0.002), which was reversed by the administration of CBZ (1.12±0.36, p=0.004) (Figure 6c). Expression of CAR, MRP2, HNF4, FGFR4, SHP, and HPRT1 did not differ between EN, TPN, and CBZ groups (Figure 6d, 6e, 6g–j).

Figure 6.

Quantitative PCR of Liver. (A) BSEP, (B) SREBP-1, (C) PPAR, (D) CAR, (E) MRP2, (F) FABP, (G) HNF4, (H) FGFR4, (I) SHP and (J) HPRT1. Statistical differences with CBZ were noted for SREBP-1, PPAR and FABP.

BSEP, bile salt export protein; CAR, constitutive androstane receptor; CBZ, carbamazepine; EN, enteral nutrition; FABP6, fatty acid binding protein 6; FGFR4, fibroblast growth factor receptor 4; HNF4, hepatocyte nuclear factor 4; HPRT1, hypoxanthine guanine phosphoribosyltransferase 1; MRP2, multidrug resistance-associated protein 2; PPAR, peroxisome proliferator-activated receptor; SHP, small heterodimer partner; SREBP-1, sterol regulatory element-binding protein 1; TPN, total parenteral nutrition.

3.6. Quantitative PCR Results of Autophagy Genes

Given carbamazepine’s previously demonstrated ability to induce autophagy in the liver, we further investigated autophagy as the mechanism through which CBZ mitigates PNALD. We assessed markers for autophagy in the hepatocytes for each group, testing for the expression of genes involved in the formation of autophagic structures and which induce autophagic flux.

Lysosomal associated membrane protein, or LAMP1, is correlated with the presence of lysosomes in the cell (Figure 7a). There were no statistically significant differences between the EN and TPN groups (1.05±0.35 vs. 1.06±0.43, p=0.943), and a decrease of expression in the CBZ animals (0.54±0.14) was statistically significant relative to the TPN group (p=0.017).

Figure 7.

Quantification of Autophagy Markers. (A) LAMP1, (B) LC3, (C) p62 and (D) BECN1. Note the significant decrease in LAMP1 and increase in LC3 with CBZ.

BECN1, Beclin-1; CBZ, carbamazepine; EN, enteral nutrition; LAMP1, lysosomal-associated membrane protein 1; LC3, microtubule-associated protein 1A/1B-light chain 3; p62, ubiquitin-binding protein p62; TPN, total parenteral nutrition.

Light chain 3, or LC3, is a marker for autophagic activity. Levels of LC3 mRNA were significantly increased in the CBZ animals relative to the TPN animals (1.07±0.57 vs. 1.80±0.51, p=0.028) (Figure 7b). Notably, LC3 fold change trended downwards with TPN treatment but rose with CBZ administration. However, there was no statistical significance between the EN (1.17±0.57) and TPN groups (p=0.78).

The p62 protein is involved in transporting protein aggregates to the autophagosome and is an indicator of later stages in autophagy (24). The downregulation observed in TPN animals as well as the upregulation in the CBZ animals were not statistically significant (1.28±0.70 vs. 0.87±0.68 vs. 1.89±1.04, p>0.05) (Figure 7c).

Lastly, BECN1, or Beclin-1, is a gene which regulates the induction of autophagosome formation (25). TPN resulted in upregulation of mRNA expression relative to EN animals (1.04±0.11 vs. 2.36±1.13, p=0.032), but CBZ (1.72±0.70) did not have a significant effect on expression of BECN1 (p=0.26) (Figure 7d).

4. Discussion

In this study, we assess CBZ’s ability to mitigate PNALD in our novel ambulatory piglet model and investigate a mechanism through which this occurs. Providing TPN to piglets induced increases in several key markers of hepatic injury and cellular proliferation, particularly serum bilirubin, GGT, the cholestasis score, and the Ki67 score. Subsequently, these changes were mitigated in animals that received CBZ. Indeed, CBZ administration reduced serum markers of injury, decreased cellular injury on histology, and regulated the expression of genes involved in hepato-biliary metabolism. A previously proposed mechanism through which these effects occur is through the modulation of autophagy with CBZ. However, our investigation suggests that although autophagy may play a role in CBZ’s ability to diminish PNALD, it is not likely a main contributor to CBZ’s favorable effects in this setting.

In our model, TPN induced elevated levels of serum bilirubin, GGT, bile acids, cholestasis, Ki67, and CK-7 compared to EN animals. These markers have been shown to be indicative of TPN-induced liver injury (26). Other studies examining PNALD in neonatal piglets, utilizing varying nutritional regimens and timeframes, have reported higher baseline bilirubin levels in neonatal piglets, however many have not differentiated between total and conjugated bilirubin unlike our current study (3, 15, 18, 27, 28). Other studies involving neonatal piglet models of PNALD have noted that TPN leads to increased values of bilirubin at around 2 weeks (27, 28). Despite our model’s modest elevation in bilirubin in animals on TPN, perhaps driven by testing lab differences, variable nutritional approach, length of treatment, or piglet variety, the multifold increase in conjugated bilirubin from 0.11 mg/dl in EN piglets to 0.29 mg/dl in TPN piglets nonetheless signifies that TPN contributes towards the progression of PNALD. Similarly, GGT levels in neonatal piglet PNALD models have been previously shown to range from around 40 to over 200 IU/L (27, 28). Despite our lower baseline of 18.4 IU/L in EN animals, our observed increase to 26.2 IU/L with TPN further supports the development of liver injury. Taken together, our observed elevations in bilirubin, GGT, and other markers of hepatocyte injury indicate that TPN nonetheless advances the development of PNALD in our model system. Our future studies will also explore higher lipid loads at 10mg/kg/day to exacerbate the resulting liver injury.

Pertinently, the administration of CBZ decreased TPN-induced hepatic injury, specifically reducing GGT, Ki67, conjugated bilirubin, and the cholestasis score. The decreased GGT and Ki67 levels with CBZ suggest that CBZ may reduce cellular damage and the resulting hepatocellular proliferation that occurs from that damage. Additionally, decreased conjugated bilirubin and cholestasis scores with CBZ imply that CBZ may play a role in increasing bile clearance. Histological analysis of tissues from each group confirmed a significant increase in cholestasis with TPN, which was mitigated with CBZ treatment. Thus, further investigation into potential mechanisms through which CBZ decreases these measures remains a focus of future studies.

Intestinal mucosal atrophy of the gut is a consequence of a lack of enteral stimulation and was revealed in post-sacrifice gut tissue collection. Consistent with our published work, piglets provided intravenous nutrition demonstrated gut atrophy relative to their enterally fed counterparts (29). The statistically significant drop in V/C ratio in the TPN and CBZ animals can be attributed to the reduced intraluminal enteral load (29, 30). The lack of improvement with CBZ suggests that liver injury mitigation is not accompanied by improvement in histological markers of gut atrophy and that differential pathways account for liver and gut injury mechanisms. Despite this limitation of CBZ as a therapeutic agent, its effects on liver were novel and pronounced.

Several hepatocyte-derived markers of gut health and gut-liver crosstalk were assessed, as the pathogenesis of PNALD has been tightly linked to downstream signaling from intestinal changes that affect hepatic metabolism. We noted that many important genes for hepatocyte metabolism and physiology were modulated in TPN animals. These genes are related to the secretory abilities and metabolism by hepatocytes. The reduction in gene expression of bile salt export protein (BSEP) in TPN animals parallels the noticeable impaired metabolism and cholestasis relative to EN animals. BSEP is expressed in hepatocytes to aid in the excretion of bile acids in the liver (6). Thus, its reduction leads to an accumulation of bile in the liver which may contribute to the cholestatic foci seen in the TPN animals.

Hepatic injury mitigation could also be correlated to the increased expression of the sterol regulatory element binding protein gene (SREBP-1), the peroxisome proliferator-activated receptor family (PPAR) and fatty acid binding protein (FABP) genes in CBZ treated animals. Given that SREBPs are integral transcription factors for lipid metabolism in the context of nutritional, growth, and other cellular processes, and that FABPs play a role in intracellular lipid transport, both pose key pathways for mitigating the effect of TPN on the liver (31, 32). Similarly, PPARs are responsible for the transcription of key genes related to the modulation of triglycerides, glucose, and fatty acid metabolism (33). Particularly, PPAR-α is implicated in fatty acid oxidation and energy production through the generation of ketones and has been shown to play a role in reducing dyslipidemia through decreasing circulating triglyceride levels (33). Notably, genetic expression of SREBP-1, FABP, and PPAR-α were increased in CBZ animals and decreased in TPN animals.

Noting the improvement in the piglets with CBZ, we explored whether autophagy may be another mechanism implicated in the pathogenesis of PNALD and whether CBZ may modulate autophagy. Briefly, autophagy is a cell-survival and cytoplasmic recycling process characterized by lysosomal degradation that is initiated in states of cellular stress, including starvation (11, 34). Autophagic function in the setting of parenteral nutrition, however, is an area not well-studied, though it has been speculated to be dysregulated with TPN, which rodent and rabbit studies have noted (11, 35).

In this study, we chose to examine LAMP1, p62, LC3, and BECN1 gene expression as measures of autophagic induction. Their roles in this process are well-characterized and are reflective of the degree of ongoing autophagy (36). Briefly, LAMP1 encodes proteins expressed on the autolysosome (37). Expression correlates to the number, though not activity, of lysosomes (37). P62 similarly is a chaperone for lysosomal traffic and is transcriptionally upregulated in autophagy (24). Microtubule-associated light chain protein, or LC3, is key to creation, transport, and fusion of the autophagosome and lysosome in autophagy and also correlates with autophagic activity (38). Lastly, BECN1 is a component of the complex leading to formation and induction of the auto-phagolysosome (25). Its expression is another indicator of autophagy (34).

Despite speculation and a handful of prior studies in other mammals, no change was observed in LAMP1, LC3, or p62 with TPN administration in our model, suggesting that the injury observed from TPN is likely independent from the process of autophagy.

Though we did not notice a difference in autophagy from TPN administration alone, CBZ has been well-characterized as an inducer of autophagy and has been shown to reduce liver injury in other pathological states. In a model of alpha-1 antitrypsin deficiency, CBZ reduced hepatic fibrosis and aided in the clearance of misfolded proteins through the enhancement of autophagy (13). In another study, CBZ alleviated hepatic steatosis secondary to ethanol consumption through a similar mechanism (35). Thus, further studies should examine the degree of hepatic fibrosis and steatosis as manifestations of PNALD that could potentially be mitigated with CBZ administration. Additionally, studies should be performed to evaluate the protein expression of various key autophagic genes with Western blots or immunohistochemistry. A limitation to our study is that we examined only gene expression through quantitative PCR, which may not reflect differences in the downstream expression of autophagy proteins due to translational regulation. These nevertheless pose exciting targets for future studies and for the interrogation of additional mechanistic pathways.

Our study concludes that although CBZ may affect autophagy in the setting of PNALD, the mechanism of its improvement of hepatic injury may not be driven by regulation of autophagy alone. Further analysis of autophagy modulation could be useful in drawing further conclusions, and non-autophagic mechanisms through which CBZ improves PNALD should be explored.

5. Conclusions

This study explored the potential therapeutic benefits of administration of carbamazepine in a novel TPN piglet model. We hypothesized that the previously observed autophagy-inducing effects of carbamazepine may be reproduced in our ambulatory piglet model and may mitigate the symptoms of parenteral nutrition associated liver disease. We demonstrated that such hepatic injury prevention occurred when piglets were given carbamazepine. The results of analysis of autophagic structures and flux demonstrated a possible correlation between autophagic induction and the observed improved liver health of the piglets, however, further in-depth exploration of the autophagic abilities of carbamazepine as well as other mechanisms for how hepatic rescue may have occurred are warranted.

Clinical Relevancy Statement:

Although total parenteral nutrition may provide short-term nourishment in patients with gut injury, long-term dependence is known to cause significant injury to the liver. Using an ambulatory piglet model, our study identifies carbamazepine as a therapeutic agent that mitigates the development of parenteral nutrition associated injury. While carbamazepine may play a role in modulating autophagy, a mechanism through which it is known to reduce liver injury in other disease settings, the mechanism in parenteral nutrition associated liver injury warrants further elucidation.