Steatosis of indeterminate cause in a pediatric group: is it a primary mitochondrial hepatopathy?

ABSTRACT CONTEXT AND OBJECTIVE: In children, hepatic steatosis may be related to inborn errors of metabolism (IEMs) or to non-alcoholic fatty liver disease (NAFLD). The aim of this study was to assess and characterize steatosis of indeterminate cause through morphological and morphometric analysis of liver tissue. DESIGN AND SETTING: Cross-sectional study at the Departments of Pathology of Faculdade de Ciências Médicas, Universidade Estadual de Campinas (FCM-Unicamp) and Faculdade de Medicina de Botucatu, Universidade Estadual Paulista (FMB-Unesp). METHODS: Eighteen consecutive liver biopsies obtained from 16 patients of ages ranging from 3 months to 12 years and nine months that were inserted in a database in the study period were analyzed using optical microscopy and transmission electron microscopy. Through electron microscopy, the mitochondrial density and mean mitochondrial surface area were determined in hepatocytes. Ten patients ranging in age from 1 to 14 years were used as a control group. RESULTS: "Pure" steatosis was detected, unaccompanied by fibrosis or any other histological alteration. Microvesicular steatosis predominated, with a significant increase in mean mitochondrial surface area. CONCLUSION: Microvesicular steatosis may be related to primary mitochondrial hepatopathy, especially due to reduction of b-oxidation or partial stagnation of oxidative phosphorylation. For these reasons, this form of steatosis (which should not be called "pure") is likely to represent an initial stage in the broad spectrum of NAFLD. We have drawn attention to cases of steatosis in the pediatric group, in which the microvesicular form predominates, since this may be associated with mitochondrial disorders.


INTRODUCTION
In children, the presence of steatosis is generally related to inborn errors of metabolism (IEMs).The clinical presentation varies and includes changes to general health condition; neurological, digestive and respiratory disorders; rapid and progressive neurological deterioration; myopathy; cardiomyopathy; craniofacial dysmorphism etc.Some forms of the disease present essentially as liver disease with hepatomegaly, jaundice, vomiting, lethargy and abnormalities of liver function.
When these diseases are ruled out after clinical and laboratory investigation, non-alcoholic fatty liver disease (NAFLD) should be considered.This term, which was initially used for adults, has become more inclusive. 1It is currently used to describe all age groups, even children. 2,3In more recent years, several studies on childhood 4 and adolescent hepatic steatosis in cases of NAFLD have been conducted, especially in relation to obese children, [5][6][7][8][9][10][11][12][13] since the prevalence of NAFLD is higher in these types of children.[16] According to Mandel et al. 17 and Bioulac-Sage et al., 18 childhood microvesicular steatosis may be associated with mitochondrial abnormalities relating to increased numbers of organelles. 18,19ere is also a correlation with respiratory chain defects and consequently impaired oxidative phosphorylation. 20,21When this type of abnormality occurs, transfer of electrons to oxygen molecules generates reactive oxygen species, thus producing superoxide anions and hydrogen peroxide. 22Furthermore, with stagnation of mitochondrial oxidative capacity, fatty acids accumulate in the cytosol, thus reaffirming the close relationship between steatosis and mitochondriopathy.
The microvesicular form alone may also be associated with some clinical conditions, such as Reye's syndrome, Jamaican vomiting sickness and congenital deficiency of b-oxidation. 23In an experimental study on pigs, it was observed that the regenerative capacity of the hepatocytes in macrovesicular steatosis was more effective than in microvesicular steatosis, thereby implying that the latter has a less favorable prognosis. 24e cases that were studied did not have any clinical association with any previously cited factor and the findings were correlated with clinical and laboratory data.There is a continuing debate on the advisability of performing liver biopsy, because of the risk of complications, high costs and lack of effective therapy.
However, liver biopsy is the only precise diagnostic method, and it is essential for grading and staging the disease.Furthermore, it enables patient inclusion in clinical screening programs. 25,26

OBJECTIVE
The aim of our study was to evaluate and characterize "pure" steatosis in a pediatric group by means of morphological (optical and electron microscopy) and morphometric analysis.

MATERIAL AND METHODS
All pediatric patients who were diagnosed as presenting morpho-   the percentage of hepatocytes affected by microvesicular steatosis, all cells present in the fragments analyzed were counted (cells with small-droplet steatosis and those with large-droplet steatosis) and then the percentages of each cell type were calculated.
Specimens processed for TEM were fixed in a 2.5% glutaraldehyde solution, postfixed in a 1% osmium tetroxide solution and embedded in araldite resin.Ultrathin sections (70-80 nm) were observed under a LEO 906 (Zeiss) electron microscope.For each patient, a liver tissue block was analyzed to determine the mitochondrial surface area in 10 microscope fields at a magnification of 6,000 x and the mean mitochondrial density, by counting the number of mitochondria present in 10 hepatocytes.Since this was a retrospective study, identification of the liver acinus zone at the time of inclusion was not a cause for concern.The acinus zone could not be characterized in this study because of the small dimensions of the TEM blocks.However, hepatocytes were randomly selected in all cases, thereby minimizing the error relating to sampling.
The results obtained were compared with the mean values found in the control group of patients.The inclusion criteria for the control group were that these individuals should be under 14 years of age and present a liver biopsy that was diagnosed as normal (n = 10).In both analyses, image capture was analyzed with the aid of the TPS Dig 1.30 software.The difference in steatosis between the patients and controls was evaluated using analysis of variance (ANOVA) with log transformation.P values < 0.05 were arbitrarily taken to be significant ("acceptable error limit").

RESULTS
All the samples were described as having fatty liver, without fibrosis and/or deposition in hepatocytes or Kupffer cells.For each sample evaluated, Analysis on the mitochondrial ultrastructure showed that the most frequent morphological abnormalities were megamitochondria, with or without crystalline inclusions, and a highly irregular mitochondrial shape, with degenerative changes (decreased electron density and loss of cristae).
Morphometric analysis on mitochondrial density showed that there were no significant differences between the control group, microvesicular steatosis group and macrovesicular Analysis on the mean mitochondrial surface area showed that there was a significant difference between the microvesicular steatosis group and the control and macrovesicular steatosis groups.These results are presented in Figure 3 and images are shown in Figure 4.

DISCUSSION
Since mitochondria are the sites for many metabolic processes (such as tricarboxylic acid production, fatty acid β-oxidation, synthesis of urea and other substances, and use of glucose and fatty acids as fuels for ATP synthesis), a disorder in any of these mechanisms may cause severe damage to the cell and thus to the tissue.Respiratory chain defects are not rare and are known to cause early liver failure in infancy, 21 thereby leading to a change in fatty acid oxidation. 27Changes to mitochondrial β-oxidation may result in accumulation of fatty acids in the cytosol, and this is typically represented by microvesicular steatosis. 28,29 our study, a significant elevation in mean mitochondrial surface area was observed, and this proved to be substantial in the patients with predominantly microvesicular steatosis.In contrast, the mitochondria in the patients with predominantly macrovesicular steatosis were considered to be normal in size.These     In our case study, 64% of the patients with microvesicular steatosis and 57% of the patients with macrovesicular steatosis showed an elevation in at least one hepatic serum aminotransferase.Some authors such as Rashid and Roberts 15 and Molleston et al. 16 have reported higher rises in AST than in ALT.However, this relationship is quite typical of alcoholic hepatitis, in which the AST/ALT ratio is greater than one. 2In our case study, only Another indication of compromised mitochondria is the frequent observation of megamitochondria, often showing crystalline inclusions.Such findings have been correlated with Wilson's disease 31 and nonalcoholic steatohepatitis (NASH). 32More recently, it has been indirectly correlated with oxidative stress in NAFLD cases. 33,34e mechanisms that induce mitochondrial "hypertrophy" remain obscure.However, changes to the respiratory chain or to other mitochondrial metabolic enzymes, whether inborn or acquired, may lead to an increase in b-oxidation and partial stagnation of oxidative phosphorylation (OxP).We hypothesize that an increase in mitochondrial volume occurs as a consequence.
As a result, the volumes of the inner membrane (and cristae) and inner membrane space (sites relating to b-oxidation and OxP) increase in an attempt to compensate for metabolic processes that were partially or totally interrupted.Taking into account the "twohit" hypothesis proposed by Day and James 35 to explain the pathogenesis of NAFLD, it is important to note that steatosis by itself does not cause the development of liver disease.However, it is a factor that can sensitize the liver to the damaging effects of a "second hit." Therefore, factors causing innocuous stress in healthy livers may lead to the development of NASH in steatotic livers. 36e results from this study have contributed substantially towards evaluations on childhood steatosis of indeterminate cause, drawing close attention to cases with predominantly microvesicular steatosis.No similar studies on pediatric groups have been published in the literature, and we would particularly like to highlight that the data on the control group in this study, which has also never previously been published, may be useful as a reference for conducting other electron microscopy studies on pediatric livers.

CONCLUSION
It may be concluded from our results that changes to serum transaminases and the presence of macrovesicular steatosis in this pediatric group did not correlate with changes to mitochondrial density or size.In addition, increased mitochondrial size was probably a morphological expression relating to deranged beta-oxidation and partial stagnation of oxidative phosphorylation.This therefore supports the hypothesis that there may be a mild form of primary mitochondrial hepatopathy, which may then progress in a more worrisome manner, especially as a second hit.
The control group was composed of liver biopsies obtained from ten patients aged from one to 14 years who presented normal physical development.In this group, the indications for liver biopsy were similar to those in the study group, i.e. vague abdominal pain and/or minimal hepatomegaly associated with slightly and persistently elevated hepatic enzyme levels.The diagnoses for the control individuals were normal, confirmed either from optical or from electron microscopy findings.In all of these patients, the blood glucose levels were always normal.All the clinical data were obtained by consulting patient medical records.The present study was approved by the Research Ethics Committee of FCM-Unicamp (process no.523/2004).The liver specimens were obtained by means of percutaneous biopsy and were processed using optical microscopy and transmission electron microscopy (TEM).For optical microscopy, the specimens were fixed in 10% formalin and processed for embedding in Paraplast Plus.The 5 mm sections were stained with hematoxylin-eosin to semi-quantitatively analyze two parameters: a) intensity of total steatosis (macro and microvesicular); and b) estimated percentage of hepatocytes affected by microvesicular steatosis, compared with hepatocytes affected by macrovesicular steatosis.The intensity of total steatosis was classified from (1) to (4), as follows: (1) 5% to 25% of tissue affected by steatosis; (2) 25 to 50% of tissue affected; (3) 50 to 75% of tissue affected; and (4) more than 75% of tissue affected.To determine

Figure 1 .Figure 2 . 1 .
Figure 1.Mean mitochondrial density in the hepatocytes from patients with steatosis, compared with patients in the control group.Analysis of variance (ANOVA) with log transformation.
results corroborate the reports by Sherlock and Dooley 30 indicating that microvesicular steatosis is of greater clinical importance and may progress in a more troubling manner.Many other reports on the relationship between steatosis and mitochondrial alterations have been published in the literature.Studies by Sokol and Treem 27 have reported that steatosis in neonatal liver failure is related to increased mitochondrial density, with an occasional increase in mitochondrial size.Mandel et al.17 observed mitochondrial abnormalities in the matrix and cristae, and also an increase in the total number of mitochondria, in cases of microvesicular steatosis.Bioulac-Sage et al.18 reported mitochondrial proliferation ("oncocytic" changes) in patients with microvesicular steatosis.Despite an evident increase in mitochondrial size in patients with microvesicular steatosis, changes to mitochondrial density were not clearly expressed in this morphometric model.To explain this result, our suggestion is that the normal distribution of mitochondria in the cytosol becomes altered because of the existence of lipid vacuoles in hepatocytes.Changes to mitochondrial distribution then disrupt the regularity of the outcome, especially when comparing the results from patients with steatosis to those from the control group (in which there are no lipid vacuoles in the cell).

Figure 4 .
Figure 4. Electron micrographs showing differences in relation to mean mitochondrial surface area in the hepatocytes.A: Mitochondria in the hepatocytes from a subject in the control group; B: Mitochondria in the hepatocytes from a patient with macrovesicular steatosis; C: Mitochondria in the hepatocytes from a patient with microvesicular steatosis showing considerable dimensions (megamitochondria) and crystalline inclusions.

Figure 3 .
Figure 3. Mean mitochondrial surface area in the hepatocytes from patients with steatosis compared with patients in the control group.(*) represents a significant difference between the group with microvesicular steatosis, the control group and the group with macrovesicular steatosis.Analysis of variance (ANOVA) with log transformation (P < 0.05).
39% of the patients with steatosis (of both types) presented elevated serum transaminase levels with an AST/ALT ratio greater than one.Elevated serum aminotransferase levels seem to show no direct relationship with either type of steatosis.Thus, no relationship with increased mitochondrial volume is found, since the enzyme levels vary widely, both in patients with predominantly microvesicular steatosis and in those with predominantly macrovesicular steatosis.

Table 1 .
Description of some parameters evaluated from the 16 patients studied(18 samples)

Table 1
shows the total steatosis values obtained by means of semiquantitative analysis and the microvesicular steatosis values obtained by means of estimated analysis.
U/I; and gamma-glutamyl transpeptidase (GGT) from 15 to 585 U/I.In Table1, the biochemical parameters, weight and height of each patient are shown.