Biomarkers in nonalcoholic fatty liver disease

1In Vitro Drug Safety and Biotechnology; 2Department of Pharmacology and Toxicology, Faculty of Medicine, University of Toronto; 3Division of Gastroenterology, Sunnybrook Health Sciences Centre; 4Department of Internal Medicine, University of Toronto, Toronto, Ontario Correspondence: Dr Manuela G Neuman, Department of Pharmacology and Toxicology, In Vitro Drug Safety and Biotechnology, Banting Institute, 100 College Street, Lab 217, Toronto, Ontario M5G 0A3. Telephone 416-398-4880, fax 416-398-4880, e-mail m_neuman@rogers.com Received for publication June 29, 2014. Accepted October 9, 2014 The obesity epidemic has begun to compromise the health of the population by promoting the premature development of the metabolic syndrome (MS), which significantly increases the risk for liver disease early in life. Approximately 30% to 40% of patients with nonalcoholic fatty liver disease (NAFLD) develop nonalcoholic steatohepatitis (NASH). It is estimated that 10% to 30% of patients with NAFLD develop cirrhosis after 10 years, with NAFLD believed to be the most common cause of cryptogenic cirrhosis. A diet rich in saturated fats and refined carbohydrates leads to hyperinsulinemia and fatty liver. Dietary intervention remains the current standard of care for NAFLD and NASH; however, this intervention often fails to control the disease. NASH, defined as the advanced end of the spectrum of chronic NAFLD, is emerging as an important cause of liver disease. The pathogenesis of NAFLD/NASH and its natural history is captured in liver disease clinics, liver transplantation, diabetes, lipid disorders and obesity. NAFLD/NASH is further studied in pediatric liver and nutrition clinics. Described by Adler and Schaffner (1) and Ludwig et al (2), NASH is a common manifestation of liver cell injury of various etiologies and of metabolic disorders of fatty acid metabolism. NASH is a chronic liver condition, and can progress to cirrhosis and end-stage liver disease. As the most aggressive form of NAFLD, NASH carries the Review

Several secondary causes of the typical histological lesions of NASH are described, including total parenteral nutrition; hypobetalipoproteinemia; jejuno-ileal bypass; certain forms of bariatric surgery; genetic-metabolic diseases such as hereditary tyrosemia type 1, galactosemia, Prader-Willi syndrome, glycogen storage disease, lypodystrophy syndromes, sialidosis, fucosidosis, mannosidosis, citrulinemia, argininemia and porphyria cutanea tarda; and malnutrition and acute starvation (12).Natural toxins, environmental toxic substances, as well as several drugs, including amiodarone, didanosine, hydralazine, methotrexate, tamoxifen, perhexiline, coralgil, corticosteroids and calcium channel blockers, can produce macrovesicular steatosis (12).Gut microbiota, producing lipopolysaccharides, may also sensitize the liver to tumour necrosis factor (TNF)-alpha, leading to hepatotoxicity (12).Toll-like receptor-4 causes nuclear factorkappa B activation and results in TNF-alpha production (12).Histological changes in the liver during NASH are nearly identical to those induced by excessive alcohol intake, as seen in patients with alcoholic steatohepatitis.However, NASH is observed in patients without any evidence of alcohol abuse, which is a flexible definition (13).The separation of NASH from alcoholic liver disease remains based on a history of minimal or no alcohol use.Integral to this definition is the clear identification of a cut-off value for significant alcohol use in light of evidence that even 20 g/day of alcohol can cause hepatic steatosis (14).In addition, underlying host factors may further confer a greater susceptibility for the development of significant liver disease in the setting of chronic alcohol exposure.
It is believed that NAFLD may be the most prevalent form of liver disease in developed countries (15).Fatty liver not associated with alcohol consumption is now recognized as possibly the most common cause of chronic, asymptomatic liver enzyme elevation in the United States and Europe (16).NASH patients may be asymptomatic or present with mild abdominal pain (17).The liver damage observed in NASH has been well described even though the pathogenesis of the disease remains uncertain.Macrovesicular and/or microvesicular steatosis, ballooning degeneration of hepatocytes, lobular inflammation and, occasionally, cirrhosis characterize the histology of this condition.Steatosis is observed in acinar zone 3, along with zone 3 Mallory bodies and/or acinar zone 3 sinusoidal fibrosis.Morphologically, the mitochondria are swollen, and paracrystalline inclusion bodies can be visualized using electron microscopy.The diagnosis of NASH requires additional morphological evidence of hepatic injury, ranging from inflammation and hepatocellular ballooning to Mallory's hyaline and fibrosis, the latter ranging from minimal to cirrhosis.The histological features of NAFLD/NASH are identical to those of alcoholic liver injury (18).
Biomarkers can be used as unbiased differential indicators of illness onset, aid in the classification of a diseased or nondiseased state, provide the ability to stage disease progression and/or offer insight into its relative severity.An individual's risk of developing a disorder may also be obtained from biomarker research.As such, a prognostic indicator could be used for risk stratification of the general population.In addition to identifying illnesses, the efficacy of clinical or therapeutic interventions aimed toward these disorders may also be obtained.
Figure 1 illustrates a possible strategy of identifying noninvasive biomarkers based on the methodology used and pathophysiology pathway.

ABNORMAL LIVER FUNCTION TEST RESULTS
Elevated alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels reflect nonspecific hepatocellular damage.In NAFLD/NASH, aminotransferase levels may be elevated two to four times over the upper limit of normal (19), with ALT being higher than AST, in contrast to alcoholic steatohepatitis.However, in the absence of advanced disease, routine liver function tests are either normal or typically show only mild elevations in aminotransferase levels, with alkaline phosphatase and gamma-glutamyl transferase (GGT) 1.5 to three times the upper limit of normal.
Several studies involving hepatology clinic patients undergoing liver biopsy and morbidly obese individuals undergoing bariatric surgery have found ALT levels to be higher in the presence of NASH than in those with simple steatosis, although this has not been universally observed (19).However, close to 80% of patients with fatty liver in cohort studies have shown ALT levels within normal limits (20).Adams et al (21) reported that aminotransferase levels fall over time as hepatic steatosis and inflammation improve.Aminotransferase levels do not correlate with the degree of fibrosis (22).
The diagnostic accuracy of ALT cut-offs for diagnosing NASH was examined in a series of women undergoing bariatric surgery.Reducing the cut-off from 30 IU/L to 19 IU/L improved sensitivity for the diagnosis of NASH from 42% to 72%.However, this was at the expense of specificity, which fell from 80% to 42% (23).Patients with NAFLD who have high levels of ALT are, therefore, more likely to have inflammation that may lead to NASH.

MARKERS OF APOPTOSIS
Cytokeratin (CK)-18 is the major intermediate filament protein of the liver.Caspases cleave CK-18 during hepatocyte apoptosis and create CK-18 fragments that can be detected by immunoassay (24).A recent meta-analysis found a wide range of cut-off values used in M30 assays, based on whether the study authors aimed for 'best sensitivity', 'best specificity' or 'best balance between sensitivity and specificity' to diagnose NASH.Based on the overall analysis, Kwok et al (24) conclude that M30 provides moderate accuracy due to a high variability between cut-offs and respective diagnostic accuracy among studies.Table 1 summarizes recent studies measuring CK-18 and markers of apoptosis (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36).
Three ELISA-based assays have been recently described to measure CK-18.The M30 assay detects hepatocyte apoptosis through the Figure 1) Noninvasive biomarkers.Biomarkers can be measured using several technologies including glycomics, proteomics and/or metabolomics, and imaging techniques such as magnetic resonance (MR) and elastography.Pathophysiology: oxidative stress, lipid peroxidation and activity of microbiota can lead to an increase in mitochondrial permeability and caspase activation.At the tissue level, this leads to steatosis, and death by apoptosis and necrosis.An increase in interleukin (IL)-6 and transforming growth factor (TGF)-beta levels leads to transglutamination, which may induce formation of Mallory-Denk bodies in the cells.An increase in TGF-beta, IL8 and RANTES levels result in the recruitment of neutrophils, leading to cellular inflammation.An increase in TGF-beta levels activates stellate cells and leads to collagen secretion as well as a reduction in extracellular matrix degradation.This phenomenon ultimately results in histological fibrosis.FasL Fas ligand; TNF Tumour necrosis factor Continued on next page identification of a caspase-cleaved fragment of CK-18.In contrast, both M65 (M6 capture and M5 detection antibodies) and M65ED (M5 capture and M6 detection antibodies) assays detect total cell death through the identification of both caspase-cleaved and uncleaved CK-18 (29).Measuring serum biomarkers of cell death is considered to be a noninvasive assessment of fibrosis stages in patients with chronic liver disease because apoptosis is involved in liver fibrosis.While CK-18 M30 was unable to differentiate between low and moderate fibrosis stages, CK-18 M65 and CK-18 M65ED showed better predictive value for fibrosis scores ≥2.All three markers functioned similarly in predicting precirrhotic fibrosis scores ≥5.Overall, all three assays could be used effectively to differentiate between NAFLD and NASH; however, only the CK-18 M65 and CK-18 M65ED assays could differentiate between NAFLD and healthy controls (29).CK-18 M30 was repeatedly found to differentiate between NASH and non-NASH in NAFLD samples.While a clear trend was observed for higher CK-18 levels in patients with more advanced stages of inflammation, steatosis and fibrosis, CK-18 M30 was repeatedly found to be less sensitive, with its usefulness limited by an inability to adequately differentiate between healthy individuals and NAFLD patients.In a large sample of overweight/obese subjects, Cusi et al (34) found significant differences in plasma CK-18 levels using M30 ELISA among all of non-NAFLD, simple steatosis and NASH patients.Differences between controls and non-NASH patients were also found in some smaller samples (26,30,31), while they were absent in other samples (25,29).CK-18 M30 further failed to differentiate between fibrosis stages 1 and 2. In the same study, M65 and M65ED could be used to predict fibrosis stages 1 or 2 (31).
Two models combining CK-18 and hyaluronic acid, and CK-18 and soluble Fas, respectively, showed better predictive value for NASH versus non-NASH in patients with NAFLD (27,28).Soluble Fas ligand was elevated in NAFLD patients with fibrosis compared with NAFLD without fibrosis (37).Elsewhere, double immunohistochemistry staining for CK-8/18 and ubiquitin improved detection of hepatocyte damage in NAFLD.This assay was used to detect ballooned hepatocytes, because ballooned hepatocytes lack CK-8/18 immunostaining on biopsy.Staining was correlated with fibrosis stage, especially advanced fibrosis, and steatohepatitis in 40 adult NAFLD core liver biopsies (38).
The number of hepatic progenitor cells and ductular reaction, assessed by CK-7 immunostaining, was used as an indicator of different fibrosis patterns in 38 pediatric NASH patients (39).Although the incidence of CK-7-positive centrilobular hepatocytes using immunostaining was 64.3% in 14 needle biopsy liver specimens belonging to NASH patients, CK-7 immunostaining was not associated with the stage of fibrosis or the grade of steatosis.CK-7 immunostaining was more common in patients with elevated serum AST and GGT (40).
Significantly higher caspase 3 and 8 activity was observed in patients with NASH than in simple steatosis in a small sample of 50 NAFLD patients.Mean caspase 3 and 8 activity scores were comparable between patients with normal and patients with elevated ALT levels (41).

Continued on next page
Mean (± SD) adiponectin levels were low (6.88±4.50µg/mL) in a small sample of 12 NAFLD patients (67).Serum adiponectin levels were lower in patients with NASH than in those without NASH in small samples of patients with biopsy-proven NAFLD (33,35,60,64,68).Based on findings from a recent meta-analysis (69), adiponectin is a biomarker of NAFLD progression to NASH because differences in serum adiponectin levels between controls and NAFLD patients are less pronounced.Adiponectin is a negative predictor of NASH in NAFLD patients at a cut-off of 29.16 µg/L (70).Furthermore, adiponectin levels were decreased in NAFLD patients compared with patients with other chronic liver conditions (58).An inverse relationship with fibrosis was also found (60).Adiponectin was not a predictor of either steatosis or fibrosis in a small pediatric NAFLD sample (25).
In a sample of 35 healthy controls and 85 ultrasonography-proven NAFLD patients (45 euglycemic and 40 hyperglycemic), serum adiponectin levels were significantly lower in euglycemic NAFLD patients than in controls, and in hyperglycemic compared with euglycemic NAFLD patients.A significant inverse relationship was further found between adiponectin levels and hyperglycemia in the NAFLD sample as a whole (62).Hypoadiponectinemia in NAFLD is part of the metabolic disturbances associated with MS, with a significant inverse relationship found between serum adiponectin levels and hepatic fat content (71).A significant inverse relationship was found between serum adiponectin levels and all of insulin resistance, insulin level, glucose level, hemoglobin A1c blood glucose level and triglycerides level in a large sample of obese male patients.Furthermore, adiponectin levels were higher in individuals with lower hepatosteatosis (61,65).
NASH was associated with an increase in the messenger RNA expression of the adiponectin receptor ApoR2, which was further associated with hepatocellular apoptosis (70).NAFLD progression in morbid obesity is associated with an increase in hepatic adiponectin receptors expression (72).
An inverse correlation was found between adiponectin and bile acids, while a trend was shown between bile acid levels and NASH.This suggests that adiponectin may be protective through its effect on bile acid metabolism (70).
Based on large variations in mean adiponectin levels found in different studies, this adipokine appears unlikely to be useful as a sole diagnostic biomarker in NAFLD/NASH.However, its regular monitoring can aid in assessing NAFLD progressing or regression over time in the same patient.For example, Wong et al (73) found that adiponectin levels increased after a three-year follow-up compared with baseline in NAFLD patients who received dietary counselling and were encouraged to increase physical activity.However, changes in adiponectin levels over this period were not correlated with increased, static or decreased NAFLD activity score (73).
Another prominent adipocytokine associated with NAFLD is leptin.Leptin is a 16 kDa proteohormone that is believed to be involved in the regulation of food intake, energy balance and body weight by increasing energy expenditure (74,75).Leptin is synthesized by differentiated adipocytes but has also been documented in other tissues including skeletal muscle, liver and the placenta (76)(77)(78)(79).Leptin concentrations in the blood reflect total body fat (80).
Leptin levels were higher in an obese pediatric population with concomitant NAFLD than in obese control patients without steatosis (68), and in patients with more advanced fibrosis in a small pediatric NAFLD sample (25).Serum leptin was further associated with insulin resistance (65,81).A model combining serum suboptimal adiponectin and elevated leptin was used to predicted NASH or borderline NASH (63).
Increases in leptin levels and decreases in adiponectin levels occurred similarly in patients with simple steatosis or NASH during obesity reversal after bariatric surgery, suggesting that these changes are due to morbid obesity and occur independent of liver disease (82).In a separate study, leptin and adiponectin levels remained independent predictors for NASH in obese patients (36).Similarly, progressive leptin increases and progressive adiponectin decreases occurred with increasing steatosis severity.Leptin increases were further correlated with fibrosis (83).Both leptin and adiponectin were associated with NASH in another study.However, the association between these adipokines and fibrosis assessed by FibroTest (LabCorp, USA) was no longer present (63).
Ghrelin is a 28 amino acid peptide, produced mainly by the stomach, which plays a major role in energy balance (84,85).However, small amounts of ghrelin are synthesized in the hypothalamus, pituitary, bowel, placenta and kidney (85,86).Ghrelin stimulates lactotroph and corticotroph secretion, and influences gastroenteropancreatic function and insulin secretion, as well as glucose and lipid metabolism.In addition, ghrelin has cardiovascular and antiproliferative effects (87,88).
Ghrelin has adipogenic properties (85).Ghrelin levels correlate negatively with insulin resistance and hyperinsulinemia, and contribute to the feedback mechanism by which body weight is regulated (89).Low levels of ghrelin are found in diabetic children, which may represent a defense mechanism against hyperglycemia (90,91).
Serum levels of retinol binding protein 4 (RBP4), another adipokine associated with insulin resistance, were also higher in NAFLD patients compared with controls (68).Adiponectin and resistin levels were negatively correlated with leptin and RBP4 levels (68).Based on these findings, all of adiponectin, leptin, resistin, ghrelin and retinol RBP4 may be appropriate biomarkers of NAFLD or MS, but are unlikely to serve as diagnostic markers on their own.
Metabolic profiles were further compared between individuals with simple steatosis and patients with NASH in NAFLD samples.These parameters were often comparable between simple steatosis and NASH, although similar trends as those observed between controls and NAFLD were maintained between simple steatosis and NASH (30,33,35,36,95,(113)(114)(115)(116)(117)(118)(119).Similarly, the incidence of MS, hypertension and diabetes mellitus was comparable between individuals with simple steatosis and patients with NASH in NAFLD samples (35).
Table 3 illustrates the sensitivity of different biomarkers to differentiate the possible progression of liver damage in NAFLD/NASH.
Markers of MS cannot be used alone to predict NAFLD/NASH, but their periodic monitoring offers an avenue to predict those patients that carry an increased risk of developing NAFLD/NASH, or those at increased risk for disease progression or extrahepatic manifestations.

METABOLOMICS AND PROTEOMICS
Metabolomics and proteomics are some of the techniques used to identify new markers of NAFLD/NASH.The compounds whose serum levels were observed to differ between patients with NASH and non-NASH controls could be associated with biochemical perturbations associated with liver dysfunction (eg, reduced creatinine) and inflammation (eg, eicosanoid signalling) in an exploratory metabolomics approach using ultra performance liquid chromatography-mass spectrometry (120).
Metabolomics analysis revealed that NAFLD was associated with higher plasma concentrations of homocysteine and total cysteine and lower plasma concentrations of total glutathione than controls in nondiabetic subjects with hepatic steatosis or NASH.Markedly higher levels of glycocholate, taurocholate and glyco-cheno-deoxycholate were further observed in subjects with NAFLD.Plasma concentrations of long-chain fatty acids were lower, while those of free carnitine, butyrylcarnitine and methylbutyrylcarnitine were higher in NASH.Several glutamyl dipeptides were also higher and cysteine-glutathione levels were lower in NASH and steatosis (96).Elsewhere, serum metabolomics revealed levels of gamma-Glu-Val, gamma-Glu-Thr, gamma-Glu-Leu, gamma-Glu-His, gamma-Glu-Phe and gamma-Glu-Arg were higher in simple steatosis than in NASH (121).Gammaglutamyl dipeptides, biosynthesized through a reaction with gamma-glutamylcysteine synthetase, indicated production of reduced glutathione (121).Using a metabolomics approach, Li et al (122) further identified serum glucose, glutamate, lactate and taurine levels as biomarkers of NAFLD and constructed a model to predict the different stages of NAFLD progression based on the different levels of these biomarkers in each individual.
A proteomics analysis revealed that the metabolic profile was different depending on an individual's level of obesity.This analysis was useful in distinguishing steatosis from NASH in a sample of patients with biopsy-proven NAFLD (123).Proteomics analysis was used to classify the serum protein fingerprint of NAFLD in 35 NAFLD patients and 35 healthy controls.Hemoglobin subunit α was upregulated in NAFLD compared with controls.High baseline hemoglobin levels predicted an increased risk of developing NAFLD during a three-year follow-up period in a large sample of previously healthy individuals (124).
Using two distinct shotgun proteomic techniques, Miller et al (125) found that several proteins are in NAFLD compared with healthy controls.Of these, afamin, apolipoprotein E, CD5 molecule-like, complement C3, insulin-like growth factorbinding protein 3, vitamin D-binding protein and lymphocyte cytosolic protein 1 were upregulated in serum sample of patients with NAFLD, compared with controls (125).Furthermore, apolipoprotein E, catalase, CD5 molecule-like, lymphocyte cytosolic protein 1 and vitamin D-binding protein were upregulated in serum sample of patients with NASH compared with those with simple steatosis (125).

MARKERS OF HEPATIC FIBROSIS
A strong correlation was found between NASH, diabetes mellitus and fibrosis in a large sample of overweight patients (127).Several noninvasive markers of fibrosis have been used in NAFLD patients, including NAFLD fibrosis score, AST/platelet ratio index, FIB-4 score and BARD score.These markers can help predict those patients who would be at highest risk of developing liver-related complications or death (128)(129)(130).Advanced fibrosis, inferred from higher scores in these assays, was associated with liver-related adverse events and shorter cumulative survival.Liver-related outcomes were not associated with the degree of steatosis or the presence of NASH, indicating the necessity to adequately assess liver fibrosis, using noninvasive markers, to predict outcomes in NAFLD patients.All four assays successfully differentiated between patients with low risk and patients with higher risk, for both liver-related adverse events and cumulative survival (129).These markers were also used elsewhere to characterize fibrosis (131).These noninvasive fibrosis scoring systems had good negative predictive value but poor positive predictive value, suggesting that they can be used to exclude fibrosis in NAFLD patients and, hence, prevent biopsy in patients without fibrosis (132).
De Lédinghen et al (133) describe a novel physical parameter based on the properties of ultrasonic signals acquired by the FibroScan (Echosens, France).This test, implemented together with FibroScan, can be used with a high degree of sensitivity to simultaneously assess fibrosis and steatosis.
FibroTest is a commercial algorithm based on different biochemical markers used to assess and monitor liver fibrosis progression (134,135).FibroTest further characterized fibrosis in samples of obese patients, identifying those patients with advanced fibrosis and, thus, in need of biopsy (136).
Neuman et al (12) showed that transforming growth factor-beta correlation with fibrosis stages F1 to F3 is strong.However, in patients with cirrhosis, this marker drops significantly because the stellate cells have been reduced (12).
Liver stiffness measurement estimated fibrosis in NAFLD patients.However, no differences were found between stages 1 and 2 of fibrosis (137).Both the pediatric NAFLD fibrosis index and transient elastography were higher in individuals with significant fibrosis in 67 children with biopsy-proven NAFLD (10 with fibrosis stage ≥2 and 57 with fibrosis stage <2) (138).Combining the two models yielded a cut-off of 8.2 for the pediatric NAFLD fibrosis index, below which clinically significant fibrosis can be ruled out.A pediatric NAFLD fibrosis index score of ≥8.2 led to transient elastography, and a transient elastography score of <8.6 kPa further revealed early liver fibrosis while a transient elastography score of ≥8.6 kPa predicted significant fibrosis with 100% accuracy and, hence, a need for liver biopsy (138).

OXIDATIVE STRESS
A strong correlation was found between oxidative stress markers and insulin resistance in obese adolescents with NAFLD (139).Oxidative stress, as indicated by higher plasma reactive carbonyl species levels, may be a direct risk factor for developing NAFLD (140).
Oxidative stress was assessed based on the balance of SOD 2, an enzyme with antioxidant activity, and cytochrome p450 2E1 (CYP2E1), an enzyme with pro-oxidant activity, based on data from 100 NASH patients, 31 simple steatosis patients and 90 healthy controls.SOD 2 47T>C and CYP2E1 1053C>T variants were genotyped using polymerase chain reaction.While the distribution of genetic variants was not different among groups for either enzyme, the presence of the higher activity SOD 2 C allele was higher in NASH patients compared with the other two groups (141).
CYP2E1 plays an important role in fatty acid metabolism, and it leads to the formation of toxic lipid peroxides.A recent study showed that lipid peroxidation was significantly greater in biopsies obtained from pediatric NAFLD patients than in patients with normal liver histology (P<0.001),without any significant differences in hepatic CYP2E1 expression.Furthermore, lipid peroxidation and CYP2E1 protein content were comparable between patients with simple steatosis and NASH patients (142).Thus, CYP2E1 may be responsible for reactive oxygen species overproduction, which plays an important role in the progression of NAFLD to NASH.While CYP2E1 expression profiles may not differ between simple steatosis and NASH, high CYP2E1 activity may be associated with progression to NASH.Furthermore, increased CYP2E1 expression may be an adaptive mechanism to prevent lipid overload.Recent evidence further shows a correlation between CYP2E1 activity, reactive oxygen species overproduction, mitochondrial assembly of reactive oxygen species and insulin resistance.These effects are detailed in recent reviews (143)(144)(145).
In addition, alcohol consumption was found to be a risk factor for NASH, because increased expression and activity of the pathways for alcohol catabolism, including CYP2E1, alcohol dehydrogenase and aldehyde dehydrogenase was found in liver tissues of NASH patients compared with normal controls, suggesting increased scavenging of alcohol from circulation in individuals with NASH (146).
Dietary long-chain fatty acids are an additional source of oxidative stress in NAFLD/NASH.Long-chain fatty acids represent normal intermediates in fat metabolism.Depending on the hepatocellular capacity to metabolize free fatty acids, the concentration of long-chain fatty acids and that of their activated form, acyl-coenzyme A, may be lipotoxic.Free fatty acids damage hepatic mitochondria by downregulating their beta oxidation and producing oxidative stress.Furthermore, accumulation of ceramides and diacylglicerol may decrease triglyceride to fatty acid ratio leading to lipotoxicity (147).

CONCLUSION
Unfortunately, to date, many critical gaps exist in the reference interval database of most of the biomarkers that have been identified for evaluation of hepatic steatosis.In addition, there are critical differences in our current knowledge of the normal levels of these biomarkers in healthy individuals depending on sex, ethnicity, age, nutritional status and comorbidities.The identification of the patients who are at risk for developing steatohepatitis that will advance to cirrhosis and complications of end-stage liver disease remains a challenge.Noninvasive biomarkers are being developed to replace liver biopsy.Hopefully, these biomarkers will provide accurate and reproducible predictive outcome data.A greater understanding of the evolution of the disease must also evaluate the role of lipid metabolism on all liver cells such as hepatocytes, stellate cells and Kupffer cells.The investigation of biomarkers may lead to therapeutic options that could be used to prevent inflammation and fibrosis in the individuals with fatty liver disease.These biomarkers would be a powerful evaluative tool for monitoring of patients with steatohepatitis.The key to a successful prevention program will depend on the early identification, treatment and monitoring of high-risk individuals by measuring a number of disease-specific biomarkers including the ones presented in the present review.Biomarkers are essential to screening and treatment strategies for patients with fatty liver disease, and diagnosing patients with lifethreatening NAFLD and NASH more quickly.This would enable classification and staging of disease using a simple blood test for biomarkers, thus avoiding a liver biopsy.Managing the underlying comorbidities generating NASH syndrome is achievable and should improve the natural history of this challenging disease.In addition, biomarker strategies can platform the biomedical research.
Moreover, we are convinced that the researchers studying MS, in which category NAFLD/NASH belongs, will have to evaluate not only the liver-related mechanisms, biomarkers and injuries, but also other tissues and organs cross-influencing the liver such as the gut, adipose tissue and the central nervous system.
A team in which the biotechnology industry will work with the researcher, hepatologist, cardiologist, endocrinologist and nutrition specialist may be beneficial for the patient.