Division of Research, Central Texas Veterans Health Care System, Department of Medicine, Scott & White Digestive Disease Research Center, Scott & White and Texas A&M Health Science Center College of Medicine, Temple, TX
Potential conflict of interest: Nothing to report.
E.G. was supported by a research project grant from the University “Sapienza” of Rome (Rome, Italy) (Fondo per gli Investimenti della Ricerca di Base [FIRB] grant no. RBAP10Z7FS_001 and by Research Program for the Development of Research of National Interest [PRIN] grant no. 2009X84L84_001). D.A. was supported by FIRB grant no. RBAP10Z7FS_004 and by PRIN grant no. 2009X84L84_002).
Hepatic progenitor cells (HPCs) play a major role in liver repair and regeneration. We evaluated HPC involvement in pediatric nonalcoholic fatty liver disease (pNAFLD). Thirty biopsies of consecutive children and adolescents with untreated NAFLD (19 with nonalcoholic steatohepatitis [NASH] and 11 without NASH) were studied using immunohistochemistry and immunofluorescence. HPCs and HPC-expressing adipokines (e.g., adiponectin, resistin, and glucagon-like peptide 1 [GLP-1]) were counted and correlated with steatosis, inflammation, hepatocyte ballooning, fibrosis, and NAFLD activity score (NAS). The HPC compartment was expanded in pNAFLD, especially in children with NASH, and was independently associated with degree of fibrosis (r = 0.303; P = 0.033). NASH livers were also characterized by increased hepatocyte apoptosis, cell-cycle arrest, and an expanded pool of intermediate hepatocytes. Adiponectin expression in HPCs of pNAFLD patients was down-regulated with respect to the healthy liver, and this expression was inversely correlated with NAS score (r = −0.792; P < 0.001) and steatosis (r = −0.769; P < 0.001). Resistin expression in HPCs increased in pNAFLD and was related to degree of fibrosis (r = 0.432; P < 0.05). GLP-1 was overexpressed in HPCs of pNAFLD patients, and GLP-1 expression was related to degree of steatosis (r = 0.577; P < 0.05) and NAS (r = 0.594; P < 0.01). Conclusions: HPC activation is involved in the response of the liver to oxidative stress in pNAFLD and is correlated with fibrosis and the progression toward NASH. HPCs express adiponectin, resistin, and GLP-1, which become available to resident liver cells and are strongly associated with the severity of NAFLD. These results may have important pathophysiological implications in the modulation of hepatic insulin resistance and the progression of liver injury. (HEPATOLOGY 2012;56:2142–2153)
Nonalcoholic fatty liver disease (NAFLD) is one of the most important causes of liver-related1 morbidity and mortality in children.2 Its general prevalence is currently estimated at 3%-10% in children in westernized countries, but because of its close association with the epidemic of pediatric obesity, this estimate rises to 50%-70% in obese subjects.3 NAFLD includes a spectrum of diseases ranging from simple fatty liver to nonalcoholic steatohepatitis (NASH) and may progress to cirrhosis and hepatocellular carcinoma,1 although this outcome is rare in children.4
NASH development is characterized by intricate interactions between resident and recruited cells that enable liver-damage progression. The increasing general agreement is that the cross-talk between hepatocytes, hepatic stellate cells (HSCs), and macrophages in NAFLD has a main role in the derangement of lipid homeostasis, insulin resistance (IR), danger recognition, immune-tolerance response, and fibrogenesis.5 However, much remains unknown about the intra- and intercellular mechanisms underlying the progression toward NASH.
In NAFLD, hepatic stem/progenitor cells (HPCs) are activated when oxidative stress inhibits the regenerative capacity of mature hepatocytes, supporting the hypothesis that HPC expansion is a component of the adaptive response of the liver to oxidative stress.6 HPC activation and proliferation determines the appearance of a ductular reaction (DR) that is characterized by tortuous structures with no discernable lumen (i.e., reactive ductules), composed of progenitor cells with highly variable marker profiles.7-9 Several studies have suggested that DR drives hepatic fibrogenesis during liver injury.10 In NASH, DR is independently correlated with progressive portal fibrosis, raising the possibility of a second (i.e., periportal) pathway for fibrogenesis that is independent of the deposition of zone 3 subsinusoidal collagen by HSCs.11 In pediatric NASH, portal fibrosis may represent the predominant form of fibrosis.12, 13
A class of circulating factors named adipokines are involved in the development of NAFLD and its progression to fibrosis.14 The liver is a source of various adipokines, although the cell types responsible for their production15, 16 and their involvement in NAFLD progression remain to be elucidated.
This study aimed to evaluate the activation of the HPC compartment and the expression of adiponectin, resistin, and glucagon-like peptide-1 (GLP-1) by HPCs in patients with pediatric NAFLD (pNAFLD) and to determine how these parameters correlate with the histological aspects of liver damage.
Patients and Methods
Patients and Clinical Data.
This retrospective study included 30 consecutive untreated children and adolescents (boys, 13; girls, 17; median age, 10.9 years; range, 8-14) with biopsy-proven NAFLD, who were referred to Bambino Gesù Children's Hospital during January 2010-January 2011. Secondary causes of steatosis were excluded, including alcohol abuse (≥140 g/week), total parenteral nutrition, and drugs. Using standard clinical, laboratory, and histological evaluations, hepatitis A, B, and C, cytomegalovirus, Epstein-Barr virus infections, autoimmune liver disease, metabolic liver diseases, Wilson's disease, and alpha-1-antitrypsin deficiency were ruled out. The children included in our analyses showed clinical and pathological features resembling those observed in our general pediatric population with NAFLD.17 These patients were diagnosed with NAFLD through liver biopsy recommended because of persistently (i.e., 6 months) elevated alanine aminotransferase (ALT) levels and the presence of an echogenic texture of the liver on ultrasonography. Furthermore, the patients received no dietary or other therapeutic treatment regimens before diagnosis. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and the recommendations of the local ethics committee. Informed consent was obtained.
Weight, height, and body mass index (BMI) were measured.17 ALT, aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (GGT), total triglycerides (Tg), and total cholesterol were evaluated using standard laboratory methods. Insulin was measured using a radioimmunoassay (Myria Technogenetics, Milan, Italy). Glucose and insulin were measured at 0, 30, 60, 90, and 120 minutes of an oral glucose tolerance test (OGTT) performed with 1.75 g of glucose/kg of body weight (up to 75 g). Degree of IR and sensitivity was determined by the homeostatic model assessment of insulin resistance (HOMA-IR), and the insulin sensitivity index (ISI) was derived from an OGGT, respectively.17
Liver Biopsy and Histopathological Analysis.
The clinical indication for biopsy was either to assess the presence of NASH and the degree of fibrosis or to diagnose other likely independent or competing liver diseases. Liver biopsy was performed after an overnight fast by using an automatic core biopsy 18-gauge needle under general anesthesia and ultrasound guidance. Two biopsy passes within different liver segments were performed for each subject. Liver biopsy specimens were fixed in buffered formalin and embedded in paraffin. Standard histological stains were performed. Steatosis, inflammation, hepatocyte ballooning, and fibrosis were scored using the NAFLD Clinical Research Network (CRN) criteria.12 Features of steatosis, lobular inflammation, and hepatocyte ballooning were combined to obtain the NAFLD activity score (NAS). As recommended by a recent NASH CRN article,18 a microscopic diagnosis, based on overall injury pattern (i.e., steatosis, hepatocyte ballooning, and inflammation), as well as the presence of additional lesions (e.g., zonality of lesions, portal inflammation, and fibrosis), has been assigned to each case.19 Accordingly, biopsies were subdivided into not steatohepatitis (not-SH) and definite steatohepatitis (definite-SH) subcategories.18 Additional details are described in the Supporting Materials and Methods.
Liver specimens from 6 lean, nondiabetic children (boys, 4; girls, 2; median age, 13 years; range, 12-16) without liver disease were used as controls. Control liver fragments were obtained from patients who underwent laparotomy or laparoscopic procedures (for cholecystectomy), from liver donors (orthotopic liver transplantation), or incidental “normal” liver biopsies (i.e., children exhibiting persistent or intermittent elevations of liver enzymes for >6 months). Histological analysis was performed by a single pathologist blinded to clinical and laboratory data.
Immunohistochemistry and Immunofluorescence.
For immunohistochemistry (IHC), sections were incubated overnight at 4°C with primary antibodies (Abs) against cytokeratin-7 (CK-7), adiponectin, resistin, GLP-1, p21waf1, and cleaved caspase-3. A complete list of primary Abs is included in the Supporting Materials and Methods. Samples were then incubated for 20 minutes at room temperature (RT) with secondary biotinylated Ab and, successively, with streptavidin/horseradish peroxidase (LSAB+; code K0690; Dako Cytomation, Glostrup, Denmark). Diaminobenzidine (code K3468, Dako) was used as the substrate, and the sections were counterstained with hematoxylin.
For double immunofluorescence (IF) staining, nonspecific protein binding was blocked with 5% normal goat serum. To perform double immunostaining with two mouse primary Abs (CK-7/adiponectin, CK-7/resistin, and CK-7/GLP-1), we followed a three-step protocol: Sections were incubated with anti-CK-7, an antimouse secondary fluorescent Ab (Alexa Fluor 488; Invitrogen Ltd., Paisley, UK) was applied, and the second primary Ab was prelabeled with a fluorophore by using an APEX-594 labeling Kit (Invitrogen) and was applied to the section. All Abs were diluted (1:50) and incubated at RT for 1 hour. Slides were counterstained with 4',6-diamidino-2-phenylindole.
Negative controls (the primary Ab was replaced with preimmune serum) were included for all immunoreactions.
Sections were examined with a Leica Microsystems (DM 4500 B; Weltzlar, Germany) microscope equipped with a JenOptik Prog Res C10 Plus Videocam (JenOptik AG, Jena, Germany). Observations were processed with an image analysis system (IAS, Delta Sistemi, Rome, Italy) and were independently performed by two pathologists in a blinded fashion. Only biopsies containing at least five portal spaces were considered. The HPC compartment was evaluated by counting the number of CK-7-positive cells within the bile/reactive ductules by using a previously reported procedure.7, 20 Solitary CK-7-positive HPCs or those in small clumps that were localized in the parenchyma or at the portal interface were included in these counts because they should be considered as a histological sectioning of bile/reactive ductules through a transverse plane without any unique IHC markers to distinguish them from cells within bile/reactive ductules.8 Cholangiocytes lining the interlobular bile ducts were excluded from the counts. The number of HPCs within the ductular reaction was counted within the entire section and expressed as the number of CK-7-positive cells per high-power field (HPF; at ×20).7
Intermediate hepatocytes (IHs) were defined as cells with sizes between those of hepatocytes and HPCs (<40 but >6 μm in diameter), with faint CK-7 immunoreactivity in the cytoplasm and reinforcement at the plasma membrane.8 The presence of IHs was scored as reported elsewhere: 0 = no IH; 1 = single occasional IHs; and 2 = clusters of IHs.20
Apoptosis and cell-cycle arrest were assessed by counting the number of hepatocytes that stained strongly positive for cleaved caspase-3 (at the cytoplasm level) and p21waf1 (at the nuclear level). The apoptotic and p21 indices were calculated by dividing the average number of positive cells by the average number of hepatocytes and expressing the quotient as a percentage for each section. At least 30 lobular fields at ×40 magnification were analyzed (∼1,000 hepatocytes) for each section.
Adiponectin, resistin, and GLP-1 expression by CK-7-positive HPCs was evaluated in serial sections and confirmed with IF. Data were expressed both as number of positive cells per HPF and as a percentage of positive cells. Adipokine expression by hepatocytes was semiquantitatively evaluated and expressed as a percentage of positive cells.
Data are indicated as mean ± standard error of the mean (SEM). The Student's t test or Mann-Whitney's U test was used to determine differences between groups for normally or not normally distributed data, respectively. Similarly, Pearson's correlation coefficient or Spearman's nonparametric correlation was used. The association of independent factors with the degree of fibrosis was evaluated using stepwise multivariate logistic or linear regression analysis. A P value of <0.05 was considered statistically significant. Statistical analyses were performed using SPSS statistical software (SPSS, Inc., Chicago, IL).
The diagnosis of NAFLD was confirmed histopathologically. Eleven subjects (37%) had simple steatosis (i.e., isolated fat deposition in hepatocytes), whereas 19 subjects (63%) exhibited varying degrees of fibrosis and inflammation. NAS scores ranged from 0 to 8. According to the NASH CNR criteria,12, 18 patients were divided into two distinct groups: not-SH (N = 11) and definite-SH or NASH (N = 19). Fibrosis of some degree was observed in 25 biopsy samples (83%): stage 1c in 11 samples (37%), stage 2 in 11 (37%), and stage 3 in 3 (10%). Five biopsy samples (17%) showed no fibrosis. Control biopsy samples had healthy liver histological features.
Anthropometric and laboratory data are described in Table 1. These parameters were similar for the two groups.
Table 1. Anthropometric and Laboratory Parameters of the Study Population
pNAFLD Whole Sample
Mann-Whitney's U Test
(N = 30)
(N = 11)
(N = 19)
Data are reported as median value.
Abbreviation: WC, waist circumference.
HPCs in pNAFLD Biopsy Samples.
Overall, pNAFLD biopsy samples showed an expanded HPC population (42.17 ± 2.33 CK-7-positive cells per HPF), compared to those of healthy livers (23.17 ± 1.72; P < 0.01), and were characterized by CK-7-positive reactive ductules at the portal-parenchymal interface (Fig. 1; Table 2). Definite-SH biopsies showed a more extensive expansion of the HPC population (46.41 ± 2.74), compared to not-SH biopsies (34.89 ± 3.24; P < 0.05). Values in the latter, however, were higher than those in healthy livers (P < 0.05). In pNAFLD, HPC number was correlated with steatosis (r = 0.417; P < 0.05) and NAS scores (r = 0.403; P < 0.05), but not with hepatocyte ballooning or lobular inflammation.
Table 2. Number of HPCs in pNAFLD and Absolute Number and Percentage of Adipokine-Positive HPCs
pNAFLD (All Patients)
pNAFLD No Fibrosis
(N = 6)
(N = 30)
(N = 5)
(N = 25)
(N = 11)
(N = 19)
Data are expressed as mean ± SEM.
P < 0.05 versus healthy livers.
P < 0.01 versus healthy livers and pNAFLD with no fibrosis.
P < 0.05 versus healthy livers and pNAFLD with not-SH.
Taking in account the degree of fibrosis (Table 2), we found no difference in the number of HPCs between pNAFLD biopsies without fibrosis (28.29 ± 3.56) and healthy livers (P = 0.087). pNAFLD biopsies with fibrosis displayed a significant HPC compartment expansion (45.29 ± 2.27), compared to biopsies from healthy controls (P < 0.01) and pNAFLD biopsies without fibrosis (P < 0.01). In pNAFLD patients, HPC number was correlated with degree of fibrosis (r = 0.582; P < 0.05; Fig. 2A).
The number of HPCs in pNAFLD showed no significant correlation with insulin serum levels, HOMA-IR, or ISI.
Logistic regression analysis revealed that the number of HPCs in pNAFLD was significantly associated (t = 2.320; P = 0.033) with degree of fibrosis, but not inflammation or NAS scores.
No intermediate hepatocytes were found in not-SH biopsies. In contrast, single occasional or clusters of intermediate hepatocytes were present in definite-SH biopsies (Fig. 1C,D). The presence of intermediate hepatocytes in pNAFLD was correlated with the number of HPCs (r = 0.685; P < 0.001), hepatocyte ballooning (r = 0.534; P < 0.01), and NAS scores (r = 0.428; P < 0.05), but not with fibrosis, steatosis, or inflammation.
Evaluation of Hepatocyte Cell-Cycle Arrest and Apoptosis in pNAFLD Biopsies.
The p21waf1 and apoptotic indices in hepatocytes (Fig. 3) were higher in definite-SH biopsies (9.64 ± 1.64 and 11.57 ± 1.93, respectively) than in not-SH biopsies (3.00 ± 2.41 and 4.33 ± 1.94, respectively; P < 0.05 for both). Moreover, the p21waf1 and apoptotic indices in hepatocytes were correlated with HPC number (r = 0.698, P < 0.01; and r = 0.758, P < 0.01, respectively) and NAS score (r = 0.476, P < 0.05; and r = 0.470, P < 0.05, respectively), but not with other variables.
Adiponectin, Resistin, and GLP-1 Expression in pNAFLD Biopsies.
The percentages of adiponectin-, resistin-, and GLP-1-positive HPCs have been evaluated in serial sections using CK-7 as a progenitor cell marker11; the data were further confirmed through double IF. CK-7 immunostaining clearly and specifically recognized HPCs in reactive ductules, cholangiocytes of interlobular bile ducts, and intermediate hepatocytes, but was consistently negative in mature hepatocytes and stromal and endothelial cells (ECs). In healthy livers and pNAFLD biopsies, adiponectin, resistin, and GLP-1 were significantly expressed by hepatocytes, cholangiocytes, and HPCs, as observed both in serial sections and through double IF (Figs. 4-6). Stromal cells were negative for adiponectin, resistin, and GLP-1, whereas ECs of portal vessels and sinusoids were negative for resistin and GLP-1, but positive for adiponectin (Supporting Fig. 1), as described elsewhere.16 The separate channels of IF are shown in Supporting Figs. 2- 4, showing the specific expression of adipokines by HPCs within reactive ductules.
Evaluation of adiponectin expression showed that the percentage of HPCs expressing adiponectin and the absolute number of adiponectin-positive HPCs per HPF did not differ between healthy liver biopsies and pNAFLD biopsies (Table 2). Taking the diagnostic category into account (Table 2; Fig. 4), we found that the percentage of adiponectin-positive HPCs in definite-SH biopsies (6.94 ± 1.27) was significantly lower than that in not-SH biopsies (19.20 ± 1.72; P < 0.01) and healthy controls (18.20 ± 2.15; P < 0.05). In addition, in pNAFLD patients, the percentage of adiponectin-positive HPCs per HPF was inversely correlated with NAS score (r = −0.792; P < 0.001; Fig. 2B), steatosis (r = −0.769; P < 0.001), hepatocyte ballooning (r = −0.408; P < 0.05), and lobular inflammation (r = −0.487; P < 0.01), but not with fibrosis. Adiponectin expression in the HPCs of NASH patients appeared to be down-regulated in relation to the NAS score and steatosis, but not fibrosis.
Evaluation of resistin expression (Fig. 5) revealed that the percentage of HPCs expressing resistin was similar between healthy biopsies (22.00 ± 1.46) and pNAFLD biopsies (21.85 ± 1.85) and between definite-SH biopsies (23.94 ± 2.54) and not-SH biopsies (17.89 ± 2.04). However, the absolute number of resistin-positive HPCs per HPF in children with NAFLD (9.90 ± 1.25) was significantly higher than in healthy livers (5.10 ± 0.51; P < 0.05). Moreover, biopsies of livers with fibrosis displayed a higher number of positive cells (11.34 ± 1.35) than those of livers without fibrosis (3.90 ± 1.27; P < 0.05) and healthy controls (P < 0.05). In definite-SH biopsies, the number of resistin-positive HPCs per HPF (11.80 ± 1.69) was higher than that in not-SH biopsies (6.34 ± 1.03; P < 0.05) and controls (P < 0.05; Table 2). In pNAFLD patients, the absolute number of resistin-positive HPCs was directly correlated with fibrosis score (r = 0.432; P < 0.05; Fig. 2C), but not with other histological parameters. The more expanded the HPC pool, the greater was the resistin expression, which correlated with the degree of fibrosis.
The percentage of resistin-positive hepatocytes (Fig. 5C) in pNAFLD (20.06 ± 3.33) was higher than that in controls (6.4 ± 1.02; P < 0.05) and was correlated with the number of HPCs (r = 0.617; P < 0.05), steatosis (r = 0.591; P < 0.05), and NAS scores (r = 0.546; P < 0.05). No correlations were observed with other histological parameters, such as ballooning, inflammation, or fibrosis.
The percentage of GLP-1-positive HPCs (Fig. 6) in pNAFLD biopsies (10.77 ± 1.64) was higher than that in controls (4.6 ± 0.81; P < 0.05). The absolute number of GLP-1-positive HPCs per HPF was significantly (P < 0.05) higher in pNAFLD biopsies than in healthy livers, in definite-SH biopsies than in not-SH biopsies, and in biopsies with fibrosis than in biopsies without fibrosis (Table 2). Furthermore, GLP-1-positive HPCs correlated with steatosis (r = 0.577; P < 0.01), hepatocyte ballooning (r = 0.496; P < 0.05), and NAS score (r = 0.594; P < 0.01; Fig. 2D), but not with inflammation or fibrosis. HPC expansion in NAFLD patients was associated with enhanced GLP-1 expression, which correlated with steatosis, but not with lobular inflammation or fibrosis.
The percentage of GLP-1-positive hepatocytes (Fig. 6C) in pNAFLD biopsies (12.94 ± 2.06) was higher than that in healthy livers (3.80 ± 0.49; P < 0.05) and was correlated with the number of HPCs (r = 0.704; P < 0.05). The percentage of GLP-1-positive hepatocytes was higher in definite-SH biopsies (16.50 ± 2.90) than in not-SH biopsies (7.86 ± 1.49; P < 0.05) and was correlated with ballooning (r = 0.549; P < 0.05) and NAS scores (r = 0.506; P < 0.05), but not with steatosis, fibrosis, or inflammation.
Intermediate hepatocytes were positive (approximately 10%) for adiponectin, resistin, and GLP-1 (data not shown).
The main findings of our study indicate the following: (1) The HPC compartment is markedly expanded in pNAFLD, especially in children with definite-SH (i.e., NASH), and this expansion is associated with degree of fibrosis and indices of hepatocyte damage (i.e., apoptotic and p21 indices), and NASH livers are also characterized by the expansion of the intermediate hepatocyte pool; (2) HPCs and hepatocytes express adiponectin, resistin, and GLP-1; (3) adiponectin expression in the HPCs of NAFLD patients appears to be lower than that in healthy livers and this expression is inversely correlated with NAS and steatosis; (4) resistin expression increases in the HPCs of pNAFLD patients and is correlated with degree of fibrosis; and (5) GLP-1 expression in HPCs is enhanced in pNAFLD in relation to steatosis, hepatocyte ballooning, and NAS.
In healthy conditions, HPCs are mostly quiescent cells that take no or minimal part in liver parenchyma renewal, because the replacement of necrotic and apoptotic hepatocytes occurs through the replication of adjacent hepatocytes within the lobules.7-9, 21 In chronic liver diseases, however, this primary pathway is impaired by various insults, thereby activating a secondary proliferative pathway of HPCs and the appearance of DR, which represents the morphological expression of HPC activation.7, 22 In NAFLD, oxidative stress plays a major role in the impairment of hepatocyte replication through the direct induction of p21 by hydrogen peroxide or by the triggering of the apoptotic cascade.11 Our previous studies on pediatric NAFLD have shown that oxidative stress-marker levels are increased in patients with NASH and correlate with the severity of SH.23 In the present study, we showed that massive HPC expansion, especially in children with NASH, is associated with degree of liver injury, hepatocyte apoptosis, and cell-cycle arrest. A population of intermediate hepatocytes was observed in NASH liver biopsies, but not in biopsies of not-SH patients. The number of intermediate hepatocytes in pNAFLD biopsies was correlated with both the number of HPCs and hepatocyte damage (i.e., ballooning), suggesting that the activation of the progenitor cell compartment and its hepatocyte differentiation are the result of the combination of impaired hepatocyte regeneration and increased hepatocyte injury caused by long-standing oxidative stress.
The resident stem/progenitor cell pool participates in the repair of liver damage either through the replacement of dead cells or by driving fundamental repair processes, including fibrosis and angiogenesis.9, 24, 25 In this context, our results in pediatric subjects seem to confirm data from adult patients, in which DR is independently correlated with progressive portal fibrosis.11 In our study, HPC compartment expansion was independently correlated through multivariate logistic regression analysis with degree of fibrosis, indicating that in pediatric NASH, DR is also a main driver of fibrosis. DR might modulate hepatic fibrogenesis during liver injury through two possible mechanisms: (1) DR cells produce agents that are chemotactic for inflammatory cells and may activate HSCs,10 and (2) DR cells undergo epithelial-mesenchymal transition, contributing to the portal myofibroblast pool.10, 26
In this study, we found no correlation between the HPC number and IR markers. This finding agrees with those of previous studies, in which these markers (e.g., insulin, HOMA-IR, and ISI) were not associated with liver fibrosis in pediatric populations.13, 27, 28 Controversy persists in the literature, however, because (1) hyperinsulinemia is a common feature of pNAFLD, and some evidence exists that it may be an independent predictor of liver fibrosis,27 and (2) even if ISI is a better marker of insulin sensitivity than fasting insulin and HOMA-IR, it had no practical application for the prediction of liver fibrosis.27
We have previously shown how the activated and expanded HPC compartment expresses many neuroendocrine features.24, 25 Resistin, GLP-1, and adiponectin expression in HPCs adds new insights into this scenario. The specificity of the immunostaining of these adipokines in HPCs was confirmed using double IF for CK-7 and by the negative reactivity of stromal cells. Although adipokine expression is prominent in adipose tissue, it has also been shown to occur within the liver. Leptin, for example, is expressed by activated HSCs and correlates with fibrogenesis,29 adiponectin expression has been shown in liver tissue and is decreased in patients with NASH,16 and resistin is expressed in the human liver, where it influences IR and the development of liver inflammation and fibrosis.15 In pediatric livers, we showed that adiponectin, resistin, and GLP-1 were significantly expressed by HPCs and hepatocytes and showed marked changes throughout the course of pNAFLD. Parallel with their expansion, HPCs down-regulated their expression of adiponectin and up-regulated that of resistin and GLP-1. The inverse correlation we found between adiponectin and steatosis and NAS agrees with the current understanding16 that this adipokine (1) is inversely correlated with fat mass and down-regulated in obesity and type 2 diabetes, (2) suppresses hepatic glucose production, improves insulin signaling, and exerts insulin-sensitizing effects in the liver, skeletal muscle, and adipose tissue, (3) is down-regulated in steatotic liver disease through various mechanisms, including fetuin-A hepatic secretion, and (4) ameliorates necroinflammation and steatosis when administered in experimental NASH.
Several lines of evidence link the biology of resistin with hepatic inflammation. In rats, resistin administration significantly worsens inflammation after lipopolysaccharide injection,30 and activated human HSCs respond to resistin with increased expression of proinflammatory chemokines and nuclear factor-kappa B activation.15, 31 Hepatic resistin expression increases in alcoholic SH and NASH and is correlated with inflammatory cell infiltration.15 Resistin has been particularly associated with macrophage recruitment within the liver; this relationship could be related to the release of monocyte chemotactic protein-1 (Mcp-1), which contributes to macrophage infiltration. Treatment with human resistin can induce Mcp-1 in ECs and HSCs in vitro.14, 15 Resistin could play a role in modulating the response of heterogenic macrophages in NAFLD.32 NAFLD is classically associated with an increase in M1 proinflammatory cytokines (e.g., tumor necrosis factor-alpha and interlukin-12)32; however, the M2 macrophages could play a role in hepatic fibrosis because they typically express higher levels of transforming growth factor-beta, which tends to promote collagen expression by HSCs.33 The capability of resistin to induce “alternatively activated” M2 macrophages, rather than the “classically activated” M1 macrophages, could represent an interesting topic for future investigation. The correlation between resistin in HPCs and degree of fibrosis in pNAFLD is consistent with the current understanding of resistin expression.
GLP-1 has insulin-independent effects on glucose disposal in extrapancreatic tissues, including the liver.34 In hepatocytes, GLP-1 activates glycogen synthesis and has been implicated in the regulation of glucose homeostasis and IR in animal models of NAFLD.35, 36 A role for GLP-1 in the pathophysiology of cholangiocytes has also been shown. In particular, GLP-1 can sustain the growth of cholangiocytes in a biliary damage model, in which it is synthesized and locally released by proliferating cholangiocytes to sustain their proliferation in an autocrine/paracrine manner.37 Herein, we showed that GLP-1 was also produced by proliferating hepatic stem/progenitor cells. Its expression was correlated with HPC expansion and was higher in definite-SH biopsies than in not-SH biopsies, suggesting that GLP-1 plays a role in sustaining HPC proliferation in an autocrine/paracrine manner during NAFLD progression. Our findings open new perspectives on the pathophysiology of NAFLD and suggest several hypotheses to be tested. First, adipokines and, especially, GLP-1 and resistin may play a major role in promoting the resistance of HPCs to oxidative stress, favoring their differentiation during the course of NASH.37 A second hypothesis is that these adipokines could be secreted into the peribiliary plexus, and easily reach hepatocytes, in which they may participate in modulating hepatic IR. Third, these adipokines may represent a new tool in the cross-talk between HPCs and HSCs, a mechanism considered to be the basis of periportal fibrosis.