Use of Steatotic Grafts in Liver Transplantation: Current Status

It is well known that the demand for liver allografts greatly exceeds the supply. Indeed, while organ shortage remains a major limitation in liver transplantation (LT), the demand for organs continues to increase worldwide. Aimed at expanding the liver donor pool, recent reports suggested the use of livers previously discarded for LT that were positive for hepatitis B virus and hepatitis C virus.1, 2 Liver steatosis is another factor to take into consideration when evaluating donor livers because it is highly frequent (30% in cadaveric and 20% in living donors) and may influence recipient outcomes.3 This is the case with nonalcoholic fatty liver disease (NAFLD) being a common cause of organ rejection. Ischemia/reperfusion injury (IRI) is another important cause of liver damage during LT, resulting in the induction of an acute inflammatory response that can lead to significant tissue damage and organ dysfunction.4

NAFLD grafts show increased vulnerability to ischemia/reperfusion (I/R) when they are transplanted. Moreover, different mechanisms of cell death in fatty versus nonfatty livers point to potential differences in the mechanisms involved in hepatic IRI. Accordingly, strategies that are effective in the healthy liver may not be useful in the presence of steatosis.5

In the present review, we summarize the current understanding of I/R pathophysiology in steatotic livers and describe those studies aimed at specifically developing therapies for this type of liver graft.

NAFLD, a Common Disease Affecting the Use of Steatotic Grafts for LT: Current Knowledge, Pharmacological Strategies, and Future Perspectives

Because of the increasing prevalence of NAFLD, this pathology has become a major focus of interest for researchers in the liver field. Previous reviews have comprehensively described NAFLD pathophysiology, potential treatment strategies, and ongoing clinical trials,6-8 and the main findings are summarized in Fig. 1. In the scenario of LT, treatments for NAFLD donors may aim to reduce fatty infiltration, ameliorate injury, and/or improve the profile of several metabolic risk factors before the transplant procedure, especially in living donor liver transplantation (LDLT). Nevertheless, current therapies for NAFLD require longterm intervention (ranging from 6 to 72 weeks),9-11 which represents a limitation in the case of clinical LT from cadaveric donors. Thus, further investigations to optimize therapies and reduce the pretreatment times during this emergency procedure are necessary.

There is now evidence in LDLT showing that bariatric surgery performed in recipients before, during, or after LT improves survival by reducing obesity-related comorbidities as well as reducing the incidence of recurrent NAFLD/nonalcoholic steatohepatitis (NASH).12 Among the different effects, there is a decrease in portal influx of free fatty acids (FAs) following surgery, which may be related both to feeding restrictions and to metabolic changes in visceral insulin sensitivity.13 On the other hand, adhesions, distortions of bowel continuity complicating biliary reconstruction, and nutritional deficiencies perpetuated by a bypass procedure have also been reported. A longer-term concern is the prevalence of weight regain, which may also have a significant effect on NAFLD.14 The pharmacological treatments of NAFLD as well as bariatric surgery may provide a solution for patients with steatosis to be successful liver donors. This would not only benefit the recipients, who would receive better functioning livers, but it would also reduce the metabolic syndrome in the donors, allowing them to have a faster recovery from the transplant procedure.9 A remedy for steatotic livers in living donor candidates will also increase the organ pool. Indeed, the first successful case of a right lobe living donor hepatectomy in a patient who previously underwent bariatric surgery for obesity was recently published.15

On the other hand, in our view, further research is required to select drugs that can regulate hepatic fatty infiltration with minimal adverse effects and to optimize such potential treatments (ie, dose and pharmacokinetics) before being translated into treatments for human disease. In our opinion, drugs or surgical procedures that reduce hepatic steatosis should be considered with caution in clinical liver surgery due to the potential adverse effects of these drugs or the potential risk associated with bariatric surgery. In this sense, bariatric surgery is not currently recommended for NAFLD/NASH patients because mortality in those patients is dictated by cardiovascular events.16 In addition, different results have been reported on the role of lipids in liver regeneration. Data based on genetic or pharmacological approaches indicate that the presence of steatosis is beneficial17, 18 or at least may not impact19, 20 regeneration following partial hepatectomy (PH). However, the interruption of the development of steatosis during the regenerative process negatively impacts liver regeneration.21, 22 In our view and in line with that reported previously by different authors,19, 20 fat accumulation alone is not sufficient to inhibit liver regeneration. Differences in the experimental surgical conditions, the approaches evaluated to modify lipid accumulation, and the forms of steatosis, which may result in different regenerative signaling pathways, should also be considered. In this sense, recent clinical experience also supports this idea, showing similar postoperative outcomes in steatotic and nonsteatotic LT, suggesting that additional genetic or environmental factors impact the effects of steatosis on hepatic regeneration.23 Preclinical studies and larger clinical trials should be carried out that are aimed at evaluating whether the reduction in fat infiltration affects both the donor’s and the recipient’s postoperative outcomes.

Alternatively, new therapeutic options to reduce IRI in NAFLD donors are being investigated. This review will discuss reported studies in this sense.

In this regard, the limited donor population is a current concern that requires further characterization to improve the procedure for preconditioning steatotic livers and to expand the donor pool to fit the growing demand for LT.25 Indeed, a growing focus in LT research relates to the use of so-called extended criteria donor grafts,26 with steatosis being a frequent reason to define a graft as extended criteria donor. The degree and type of hepatic steatosis, inflammation, fibrosis, and damage should be evaluated before surgery in potential donor grafts as well as in patients in the immediate postoperative period after LT.

At present, there are no known clinical or biochemical markers able to estimate the degree of fatty infiltration of the liver.27, 28 Several reports have demonstrated that liver steatosis is associated with the elevation of certain clinical and biochemical markers28-30 with low predictive potential. Indeed, transaminases and specifically alanine aminotransferase (ALT) commonly used as liver damage markers are not 100% reliable because some patients suffering from NAFLD do not present high levels of ALT.31 Recent studies evaluating other biomarkers including circulating microRNA reported promising results,32 although larger studies are still necessary to validate them in the future.

During procurement of the graft in the donor, the gold standard to assess hepatic steatosis is a histological analysis.33 The practice of transplant programs varies between hospitals because some perform liver biopsies on all potential donors, whereas others perform them only when significant steatosis cannot be ruled out.25 Nevertheless, it is important to note that liver biopsy is an invasive procedure that has a significant interobserver variability among experts for both quantitative and qualitative assessments of the histologic features of liver steatosis and is not suitable for continuous monitoring. Unfortunately, current imaging methods are inaccurate and inadequate for quantification of liver steatosis and do not distinguish clearly between the microvesicular and the macrovesicular types.34 Magnetic resonance techniques can enable more direct assessment of the amount of fat in the liver than can either computed tomography (CT) or ultrasound (US).35 In a study applying state-of-the-art equipment for US, CT, and magnetic resonance imaging, Saadeh et al. reported that only a hepatic fat accumulation above 25%-30% can be reliably detected radiologically.36 Also, none of these modalities was able to either distinguish steatosis from steatohepatitis or to detect individual pathologic features required to establish steatohepatitis such as necroinflammatory changes, hepatocyte ballooning, and fibrosis.29 Other tools, such as biochemical impedance and transient elastography (TE) have been shown to predict steatosis/fibrosis, and the use may be extended in assessments of the donor liver. TE has been validated as a screening tool for advanced fibrosis in NAFLD, though it cannot be used to rule out inflammation of the liver.37 The use of noninvasive, rapidly available, sensitive, specific, and cost-effective biomarkers could simplify liver injury assessment and improve the time to management of postoperative complications in steatotic livers that undergo surgery.

Various studies have evaluated the possible association between the percentage of steatosis and/or the pattern of fat distribution and the postoperative outcomes after LT. In microvesicular steatosis, the cytoplasm of the hepatocytes contains tiny lipid vesicles without nuclear dislocation.38 However, macrovesicular steatosis is characterized by a single, bulky fat vacuole in the hepatocytes, which displaces the nucleus to one side. On the other hand, in microvesicular steatosis, multiple small droplets are finely dispersed in the cytoplasm without nuclear displacement. Livers with macrovesicular steatosis are more intolerant to ischemic injury than those with microvesicular steatosis.33 The triglycerides deposition as intracytoplasmic fat droplets is associated with an increase in hepatocellular volume, which may induce distortion and narrowing of the sinusoid, with a reduction in the luminal diameter by up to 50% compared with that in the healthy liver.39 Increases in liver fat infiltration are consistent with the appearance of severe alterations in liver blood flow and hepatic microcirculation, respectively. In addition, macrovesicular steatosis is associated with increased risk of acute cellular rejection and bile duct loss suggestive of chronic ductopenic rejection in allograft recipients.40 These observations are in accordance with previous reports, indicating that transplantation outcomes are not affected by hepatic microvesicular steatosis regardless of its severity. Therefore, such livers can be safely used, assuming there are no other donor or recipient risk factors.41 In contrast, some transplant programs exclude donors with macrovesicular steatosis exceeding 10%-15%.25 The use of grafts with moderate steatosis (>30% and 60%) is indeed controversial because these may impose a relative risk on posttransplant outcomes. Previous reports have shown an increased incidence of primary nonfunction after LT from donors with moderate steatosis compared with nonsteatotic livers,42 whereas other authors have reported excellent results.43 Donor livers with severe macrovesicular steatosis (>60%) are typically ruled out for transplantation because of an expected high risk of graft failure.33 Consequently, from our point of view, and as reported by other authors,24, 25 it should be considered that transplant dysfunction is multifactorial, depending on factors like advanced donor age, cold ischemia time (CIT), and intensive care stay, suggesting that isolated graft steatosis is not the only cause of dysfunction or initial nonfunction after LT.24 In fact, a large study (5051 patients) showed that when cold ischemia extends beyond 11 hours, a low degree of macrovesicular steatosis (20%) is associated with an increased risk of graft loss.44 In this sense, it has been recommended that organs with >30% steatosis may only be used if other factors are controlled (ie, donor age <40 years, short CIT of <5 hours, and noncirculatory cause of death).25, 45

In summary, in the absence of clear evidence on the postoperative outcomes when using steatotic liver grafts and considering the difficulties involved in assessing the degree of steatosis among potential liver donors, the utilization of steatotic donor livers has been highly variable among surgeons, centers, and countries, with some surgeons discarding livers that may have had acceptable outcomes.46 Instead of discarding steatotic grafts for transplant, the community should look for new therapeutic strategies in order to minimize or modify factors that could further endanger graft quality and increase the potential use of steatotic livers for transplant. In this context, pharmacological or surgical strategies aimed at reducing the vulnerability of steatotic liver grafts to I/R should be considered. This could increase the availability in the organ pool and consequently reduce the waiting list for transplantation. It should be noted, however, that evaluation of the percentage of steatosis using a noninvasive method is difficult and the criteria for the use of livers for transplantation differ between surgeons, centers, and countries.

Animal Models of LT From Steatotic Grafts

Multiple experimental studies have been performed with the aim of increasing the number of available liver grafts and improving the postoperative outcomes in steatotic LT. In our view, animal models of LT should be reproducible, reliable, simple, affordable, and technically available with minimal disadvantages.

In the specific situation of NAFLD and given the complexity of human behavior and biology, key factors, such as physical activity, social environment, psychological stress factors, and genetics, should be kept in mind when selecting an animal model to study NAFLD (because they may also be important for the development of the disease in humans) and also when interpreting data obtained from these experimental models.47 In this sense, an ideal animal model of the disease should reflect the hepatic histopathology and pathophysiology of human NAFLD/NASH. Some animal models of NASH recapitulate the liver injury observed in humans but in a different metabolic context (nutritional deficiencies or insulin sensitivity), whereas others were developed by considering the metabolic context of NAFLD but without the full development of steatohepatitis. More recently, models of overnutrition have been achieved through high-fat feeding, eg, forced caloric overload in genetically hyperphagic mice fed a high-fat diet (HFD). These models appear most suited to study the complex biological interactions that are involved in the development of fatty liver disease.48 However, there is not an agreement on what comprises steatohepatitis in experimental models, thus hampering the research done in NASH and limiting the interpretation of results about therapeutic approaches and metabolic pathways.

A variety of animal models of liver steatosis can be applied when investigating the impact of steatosis on liver surgery. Mice and rats have mostly been used as experimental animals in models of NAFLD/NASH, which can be classified into genetic, nutritional, and a combination of genetic and nutritional models.49 Large animals are more expensive and less available because of the requirements of special breeding facilities. In addition, studies have demonstrated problems in inducing reliable steatosis in pigs. For instance, the model described by Lee et al. did not show high levels of triglycerides, and the livers did not present macrovesicular steatosis.50 Rodent hepatic steatosis can be induced by a genetic leptin mutation (ie, Zucker rats, ob/ob mice) or by modulation of nutritional factors. The genetically modified rodents overeat due to the lack of the controlling effect of leptin and, consequently, develop combined microvesicular and macrovesicular steatosis without inflammatory changes. The nutritional models are based either on diets with a high percentage of saturated fat or with amino acid deficiency (ie, choline and methionine), which is essential for hepatic lipid excretion.51 We recently reported a new rat model of NASH that fulfills the major characteristics of human disease.52 However, its applicability in the field of LT remains unexplored.

To our knowledge, only a few studies have combined the use of steatotic grafts with orthotopic LT and were mainly developed in rat models of either nutritionally or genetically induced moderate steatosis.53 Additionally, livers procured following brain death (BD) currently constitute the majority of transplanted organs,54 and steatosis occurs in 9%-26% of biopsied donor livers,55 which are both characteristic risk factors for LT. Thereby new experimental conditions should focus not only on liver graft damage associated with transplantation but also on BD donors to most closely mimic the clinical situation of LT and ultimately develop effective therapeutic strategies in this setting.56

Experimental findings support clinical evidence that BD impairs the viability of organs for transplantation, triggering hemodynamic, hormonal, and inflammatory responses. BD itself induces hypoperfusion in the mesenteric microcirculation, thereby leading to increased local inflammation and organ dysfunction.56, 57 The effects of BD in steatotic LT have been evaluated with the finding that liver damage and inflammatory response were increased in steatotic livers compared with nonsteatotic ones.56

Mechanisms Involved in IRI in Steatotic LT

From a microcirculatory point of view, now we know that all sinusoidal cells suffer significant deregulations leading to their dysfunction, ultimately contributing to IRI. Indeed, liver sinusoidal endothelial cells (LSEC) are seriously compromised in both warm and cold ischemia associated with LT. The lack of biomechanical stimuli occurring during procurement and preservation deteriorates the protective phenotype of LSEC by down-regulating the expression of the transcription factor Kruppel-like factor 2 (KLF2), which orchestrates the transcription of a variety of protective genes including the endothelial nitric oxide synthase (eNOS), the antithrombotic molecule thrombomodulin (TM)m or the nuclear erythroid 2 p45-related factor 2 (NRF2).58 Impairment of the microcirculation is considered a major cause of reperfusion injury in steatotic livers.59 Fatty accumulation in the cytoplasm of hepatocytes is associated with an increase in cell volume that reduces the size of the hepatic sinusoid space and may contribute to a partial or complete obstruction of the hepatic sinusoid space, which causes a reduction in hepatic blood flow.59 An imbalance between vasoconstrictors (ie, endothelin [ET] 1) and vasodilators (ie, nitric oxide [NO]) negatively affects the hepatic microcirculation in steatotic livers (Fig. 2). Blebs and solidified fat globules released into the sinusoidal space further compromise the microvascular space and impair hepatic microcirculation.59

Steatotic livers are more susceptible to lipid peroxidation because of either their lower antioxidant defenses, superoxide dismutase (SOD), and glutathione (GSH), or their greater production of reactive oxygen species (ROS) from the mitochondria and/or xanthine/xanthine oxidase (XOD) system or both.60 Neutrophils and endoplasmic reticulum (ER) stress have been implicated in the increased vulnerability of steatotic livers to IRI.61 Differences were also observed when we analyzed the role of the renin-angiotensin (Ang) system, as the angiotensin-converting enzyme inhibitor (ACE)–AngII–angiotensin II receptor (ATR) and angiotensin-converting enzyme 2 (ACE2)–Ang–(1-7)–Mas receptor (Mas) axis plays a major role in nonsteatotic and steatotic grafts, respectively.5 It is well known that Kupffer cells (KCs) suffer from a profound activation process that is promoted by the release of damage-associated molecular patterns (DAMPs) from neighboring necrotic hepatic cells and pathogen-associated molecular patterns (PAMPs). Toll-like receptors (TLRs) recognize both PAMPs and DAMPs, leading to activation of downstream signaling cascades.42 Reduced retinol-binding protein 4 (RBP4) and TLR4 levels and increased peroxisome proliferator-activated receptor (PPAR) γ levels were observed in steatotic livers compared with nonsteatotic livers.53, 62 The vulnerability of steatotic livers subjected to ischemia is also associated with reduced adiponectin and resistin levels.63 It has been reported that hepatocytes with fatty infiltration develop massive necrosis after IRI, rather than apoptosis observed in nonsteatotic livers. This may be due to low adenosine triphosphate (ATP) production and dysfunction of regulators of apoptosis (B cell lymphoma 2, B cell lymphoma extra large, and B cell lymphoma 2–associated X protein).60 The up-regulation of sterol regulatory element-binding protein 1c (SREBP-1c) and the transcription factor liver X receptor (LXR) α in steatotic liver grafts contributes to the increased de novo synthesis of FA. It has been also reported that the hepatic uptake of FAs is facilitated by cell surface receptors, including CD36.64 Finally, it is worth considering recent findings indicating that BD exacerbates hepatic I/R damage in steatotic liver grafts when compared with nonsteatotic livers.57 The injurious effects of BD are mainly mediated through KC activation, leading to tumor necrosis factor α (TNF-α) amplification. BD induction resulted in a reduction in adrenocorticotropic hormone (ACTH), cortisol, high-mobility group box 1 (HMGB1), and TLR4 in steatotic LT compared with nonsteatotic LT.53, 57, 65, 66 All of these mechanisms increased the vulnerability of steatotic livers to IRI.

Conclusion

However, much needs to be done to transfer experimental results to clinical practice. Indeed, we still do not know precisely how steatosis (type, grade, and so on) affects the postoperative results and graft survival. In addition, the success or dysfunction of liver grafts may be affected by other factors, including donor age, CIT, and length of intensive care stay. Furthermore, a method for determining the extent of steatosis remains imprecise and inconsistently reported. All these premises imply that new preclinical studies in the field of LT must be performed in experimental models as close as possible to the real clinical conditions (ie, coming from cadaveric or living donors, with proper ischemia times, presence of steatosis, age of the donor, or type of death). Indeed, the extent and time of ischemia, the type of donor and liver submitted to I/R, and the presence of liver regeneration all influence the mechanisms of hepatic IRI and, thus, the effects of therapeutic strategies. However, few advances merging the presence of steatosis and the different conditions of donors have been made in recent years. Future preclinical studies and larger clinical studies aimed at evaluating whether the reduction in fatty infiltration affects both the donor and the recipient postoperative outcomes should be examined. Nevertheless, drugs or surgical procedures that reduce hepatic steatosis should be considered with caution in clinical liver surgery by the potential adverse effects of these drugs or the potential risks associated with bariatric surgery. On the other hand, there is a high heterogeneity between methods and therapeutic options for reducing IRI, the majority of them being administered to donors before IRI and not to recipients, which would facilitate their clinical applicability. In addition, there is no consensus on which experimental model should be used, especially given that the effectiveness of a certain strategy could differ depending on the surgical conditions evaluated. Thus, as researchers, we should focus on the use of standardized experimental models of LT with steatotic grafts that resemble clinical practice as much as possible in order to consider and evaluate the relevance of all of the variables involved. Of course, the selection of an appropriate experimental model of steatosis is required if the aim is to evaluate the effects of the presence of fatty infiltration (by itself without associated inflammation) on the underlying mechanisms responsible for either inflammation associated with I/R or regeneration. Only under these premises will we be able to solve the current clinical problem with steatotic grafts and transfer experimental results to clinical practice.