Can elastometry be used for a better identification of cirrhosis?

Increased stiffness has been directly associated to fibrosis. Stage 4 fibrosis defines cirrhosis. Can elastometry (transient elastography, acoustic radiation force impulse imaging) be used for a better identification of cirrhosis? The answer is obviously yes, provided hepatologists, radiologists, pathologists, and biologists combine their expertise because severe chronic liver disease is a complex subject. Considering the pathogenesis of cirrhosis, it is likely that factors such as parenchymal extinction (leading to atrophy of the liver mass with approximation of portal tracts and hepatic veins) and exudation (congestion) play a role in liver stiffness not mentioning heart failure, liver necrosis and extrahepatic cholestasis. Neglecting these factors, elastometry will lead too often to a wrong appreciation of the degree and type of liver damage and eventually to wrong medical decision.

[Expression of the elastic fibers components during the fœtal liver development]

Elastic fibers are composed of microfibrils containing fibrillin-1 and an elastic component, elastin. Microfibrils may not be associated with elastin. In the adult liver, fibrillin-1 and elastin are coexpressed within the stroma and portal tracts vessel walls. Fibrillin-1 is expressed alone around the bile ducts and within the Disse space. There is little work that has studied the elastic fiber organization during the fœtal liver development. Here, we studied the expression of fibrillin-1 and elastin by immunohistochemistry on 20 cases of fœtal liver. During the development of the portal tract, the two components are coexpressed on interstitial elastic fibers and within vessel walls. Fibrillin-1 is expressed alone around the bile structures during their maturation. Unlike adult liver, fibrillin-1 is expressed on thin and very irregular microfibrils within the Disse space. Our study shows that the elastic matrix development in the portal tract follows the development of the different structures, notably biliary structures. In the Disse space, microfibrils are not continuous. Their maturation may be in relation with the change of the hepatic blood flow after birth.

Spontaneous regression of focal nodular hyperplasia: a pathological report

Focal nodular hyperplasia (FNH) is a benign tumour of the liver, occurring in 0.6-3% of the general population. Most lesions are diagnosed incidentally. With the increasing use and improvement of diagnostic imaging, FNH is being observed more often. It has been shown, using radiological approaches, that most FNH remain stable, or even regress, over a long follow-up period. In addition, it is extremely rare that FNH were discovered in elderly. However, to our knowledge, there is no pathological report illustrating the regression of FNH. We report here a case showing the pathological changes occurring during the regression phase of FNH, with dense fibrous tissue mixed with arteries replacing hepatocytes nodules.

Cirrhosis: what else?

When the term cirrhosis was coined two centuries ago by Laennec, it designates by definition an end stage irreversible liver disease. Nowadays this word encompasses a whole range of disorders including some degree of reversibility for the early stage. It is therefore of prime importance to define the stages of the fibrotic process, based on the integration of knowledge about liver structure and function. In addition to morphological data, modern imaging techniques coupled to non-invasive biomarkers will probably help to better define and denominate this heterogeneous entity.

Differential effects of inactivated Axin1 and activated beta-catenin mutations in human hepatocellular carcinomas

Perturbations to the Wnt signaling pathway have been implicated in a large proportion of human hepatocellular carcinomas (HCCs). Activating beta-catenin mutations and loss of function mutations in Axin1 are thought to be functionally equivalent. We examined the Wnt pathway in HCC by comparing the expression of beta-catenin target genes and the level of beta-catenin-dependent transcriptional activation, in 45 HCC tumors and four cell lines. Among these samples, beta-catenin and AXIN1 were mutated in 20 and seven cases, respectively. We found a significant correlation between activated beta-catenin mutations and overexpression of mRNA for the target genes glutamine synthetase (GS), G-protein-coupled receptor (GPR)49 and glutamate transporter (GLT)-1 (P=0.0001), but not for the genes ornithine aminotransferase, LECT2, c-myc and cyclin D1. We also showed that GS is a good immunohistochemical marker of beta-catenin activation in HCC. However, we observed no induction of GS, GPR49 or GLT-1 in the five inactivated Axin1 tumors. Beta-catenin-dependent transcriptional activation in two Axin1-mutated HCC cell lines was much weaker than in beta-catenin-mutated cell lines. Our results strongly suggest that in HCC, contrary to expectation, the loss of function of Axin1 is not equivalent to the gain of function of beta-catenin. Our results also suggest that the tumor suppressor function of Axin1 in HCC may be related to another, non-Wnt pathway.

Genomic alteration is not associated with fatty and clear cell change in hepatocellular carcinomas and its precursor nodular lesions in cirrhotic liver

BACKGROUND

The aim of this study was to investigate whether fatty and clear cell areas in large regenerative nodules (LRN), dysplastic nodules (DN), and hepatocellular carcinoma (HCC) show higher degree of genomic mutation compared to non-fatty/clear cell area in the same nodule or non-lesional tissue.

METHODS

We examined 22 nodular lesions (9 HCC, 5 DN and 8 LRN) from seven cirrhotic livers removed at transplantation. Frozen sections were used for manual microdissection of areas with fatty/clear cell change. DNA from microdissected tissue was amplified using arbitrarily primed polymerase chain reaction (AP-PCR), and PCR products were run on polyacrilamide gel generating a "fingerprint" band pattern. Autoradiographs were analysed using Adobe Photoshop version 6.0. Fingerprints from lesional tissue were compared to reference tissue and the total number of bands in excess or defect was calculated and divided by the total number of bands identified, obtaining the genomic damage fraction (GDF).

RESULTS

Increasing GDF average values were seen from cirrhotic liver (0.13+/-0.04), to LRN (0.16+/-0.1), DN (0.28+/-0.08) and HCC (0.30+/-0.07). A statistically significant difference in GDF values was documented between cirrhotic liver and DN (p=0.008) and HCC (p=0.005) and between HCC and LRN (p=0.02). No significant difference was documented between DN and HCC, and between LRN and cirrhotic liver. Eleven nodules containing fat/clear cell areas were compared to the other 11 nodules without fat/clear cell areas. The GDF was not different between the two groups: 0.29+/-0.11 versus 0.25+/-0.12; p=0.5. The average value of genomic damage fraction between fat/clear cell areas (0.29+/-0.11) and no fat/clear cell areas (0.25+/-0.1) within the same nodules were not significantly different (p=0.11).

CONCLUSION

Fatty and clear cell change in nodular lesions in cirrhotic liver may be an epigenetic phenotypic modification caused by microenvironmental factors such as ischaemia rather than indicating areas of increased malignant potential per se.

[Hepatitis C-autoimmune hepatitis overlap syndrome: five new cases]

PURPOSE

The hepatitis C-Autoimmune hepatitis overlap syndrome is an uncommon condition whose management can be difficult in both diagnosis and treatment.

PATIENTS AND METHODS

Five cases of hepatitis C and autoimmune hepatitis overlap syndrome, brought by a retrospective study, are reported. Hepatitis C was proven by a positive HCV viral load and the autoimmune hepatitis was proven by characteristic immunological and/or histological features.

RESULTS

All patients were female, the mean age at diagnosis was 54.2 years (+/-6.6), only one patient had a history of autoimmune disease (autoimmune thyroiditis). Four patients had significant autoantibodies levels (over 1/320) and 4 had a high serum gammaglobulin level (over 1.5 times the upper normal limit). Liver biopsy showed characteristic features of autoimmune hepatitis in all cases and characteristic features of chronic HCV infection in 3 cases. Corticosteroid and/or immunosuppressive treatment was given as a first line therapy in 4 cases while an antiviral treatment has been tried in 4 cases with a good viral response in 2 of them. Corticosteroid and/or immunosuppressive treatment enhanced HCV viral load in all cases, however a histological improvement was observed in every case in which a control biopsy has been performed (3 cases).

CONCLUSION

This study highlights the deciding contribution of the initial histological findings in the diagnosis of such a HCV/autoimmune hepatitis overlap syndrome; it also demonstrates that histological outcome is not necessarily compromised by corticosteroid and/or immunosuppressive treatment.

Classification tool for the systematic histological assessment of hepatocellular carcinoma, macroregenerative nodules, and dysplastic nodules in cirrhotic liver

AIM: To design a classification tool for the histological assessment of hepatocellular carcinoma (HCC), dysplastic nodules (DN), and macroregenerative nodules (MRN) in cirrhotic liver.

METHODS: Two hundred and twelve hepatocellular nodules (106 HCC; 74 MRN; 32 DN) were assessed systematically, quantitatively, and semiquantitatively as appropriate for 10 histological features that have been described as helpful in distinguishing small HCC, DN, and MRN in cirrhotic livers. The data were analyzed by multiple correspondence analysis (MCA).

RESULTS: MCA distributed HCC, DN, and MRN as defined by traditional histological evaluation as well as the individual histological variables, in a “malignancy scale”. Based on the MCA data representation, we created a classification tool, which categorizes an individual nodular lesion as MRN, DN, or HCC based on the balance of all histological features (i.e., vascular invasion, capsular invasion, tumor necrosis, tumor heterogeneity, reticulin loss, capillarization of sinusoids, trabecular thickness, nuclear atypia, and mitotic activity). The classification tool classified most (83%) of a validation set of 47 nodules in the same way as the routine histological assessment. No discrepancies were present for DN and MRN between the routine histological assignment and the classification tool. Of 25 HCC assigned by routine assessment in the validation set, 8 were assigned to the DN category by the classification tool.

CONCLUSION: We have designed a classification tool for the histological assessment of HCC and its putative precursors in cirrhotic liver. Application of this tool systematically records histological features of diagnostic importance in the evaluation of small HCC.

Investigation of liver fibrosis in clinical practice

The liver is composed of different hepatic fibrogenic cells: hepatic stellate cells, portal fibroblasts, fibroblasts of the Glisson capsule surrounding the liver and vascular smooth muscle cells and the second layer cells present around centrolobular veins. During liver disease, one or several populations of these cells are activated, transformed into myofibroblasts and secrete the extra-cellular matrix. There are markers to identify hepatic stellate cells either quiescent (CRBP-1) or activated (alpha-smooth muscle actin). Liver biopsy, the current "gold-standard" to estimate liver fibrosis cannot be used anymore as a "gold standard". Furthermore, it is a costly procedure with adverse effects feared by patients and clinicians. Alternative to liver biopsy using non-invasive-tests or technics include FibroTest-ActiTest, transient-elastography, hepatic vein transit time using contrast ultrasonography, magnetic resonance imaging. As a routine test, the FibroTest-ActiTest is a validated one for patients with chronic hepatitis C. The advantage of the non-invasive tests or technics is that they provide a rapid and quantitative estimation of fibrosis. With these new methods, it is possible to follow the progression of the disease and its regression either spontaneously or under treatment. In conclusion, clinicians have in their hands several painless tools to explore liver fibrosis that can be easily repeated.

Association of Helicobacter species with hepatitis C cirrhosis with or without hepatocellular carcinoma

BACKGROUND AND AIMS

Recent studies have suggested that bacterial coinfection with Helicobacter species in patients already infected with hepatitis C virus (HCV) could be involved in the development of cirrhosis and hepatocellular carcinoma (HCC). A retrospective cross sectional study was performed in order to explore the association between Helicobacter species and HCV associated liver diseases.

METHODS

The presence of Helicobacter species was tested by polymerase chain reaction on liver samples from four groups of patients.

RESULTS

Helicobacter 16S rDNA was found in only 4.2% of liver samples from control patients (n=24) and in 3.5% of liver samples from patients with non-cirrhotic chronic hepatitis C (n=29) while it was found in 68.0% of liver samples from patients with HCV positive cirrhosis without HCC (n=25) as well as in 61.3% of cirrhotic liver samples from patients with HCV positive cirrhosis and HCC (n=31). In addition, when the HCC tumour tissue was tested (n=21), 90.5% of samples were positive. DNA from Helicobacter pylori- and Helicobacter pullorum-like organisms was found.

CONCLUSIONS

There is an association between the presence of Helicobacter species DNA in the liver and hepatitis C cirrhosis, with or without HCC. Indeed, the presence of these bacteria could be the result of structural changes in the liver. Alternatively, Helicobacter species could be a co-risk factor in HCV chronic liver diseases. This result warrants prospective studies to determine the possible causal role of these bacteria in the progression of chronic hepatitis C.

Identification of homozygous deletions at chromosome 16q23 in aflatoxin B1 exposed hepatocellular carcinoma

Loss of heterozygosity (LOH) represents the most frequent genetic alteration observed in hepatocellular carcinoma (HCC). Chromosome 16q is of particular interest as it exhibits LOH in 29% of HCC tumors and is frequently lost in breast, prostate, ovarian and gastric carcinomas. We genotyped 157 HCC tumors for 17 microsatellite markers distributed on chromosome 16q and determined a common region of LOH localized between the markers D16S518 and D16S504. By refining the boundaries of two interstitial LOH and two homozygous deletions, the critical region was delimited to 180 kb between D16S3096 and D16S3029. This region is located in intron 8 of the WWOX/FOR gene, but a search for mutations in all coding exons of this gene in 27 HCC tumors and cell lines did not reveal any tumor somatic alterations. Furthermore, by RT-PCR, no abnormal transcripts of this WWOX/FOR gene was detected in nine HCC cell lines. Finally, analysis of the p53 gene mutations with the clinical parameters of all tumors revealed that the two homozygous deletions have occurred in tumors presenting a R249S mutation. Our data revealed a relationship between chromosome 16q homozygous deletions and R249S p53 mutations in tumors where the patient had been exposed to aflatoxin B1 (P=0.002). These results are consistent with a role of aflatoxin B1 in the instability of chromosome 16q at the fragile site FRA16D. However, the nature of the specific gene that is altered during hepatocarcinogenesis remains to be elucidated.

Immunohistochemical detection of hcv in cirrhosis, dysplastic nodules, and hepatocellular carcinomas with parallel-tissue quantitative RT-PCR

Hepatitis C virus is a major risk factor for hepatocarcinogenesis in humans. In situ detection of the virus in early sequential lesions of hepatocarcinogenesis could provide information about the role of the virus in the transformation and promotion process. Parallel in situ detection of HCV proteins and RNA in human tissues were performed in 55 posthepatitis C cirrhosis, 17 dysplastic nodules (DN), and 25 hepatocellular carcinomas (HCC), using immunohistochemistry and tissue quantitative RT-PCR. A consistent cytoplasmic hepatocellular staining was obtained in 73% of cirrhosis cases (with or without HCC) and in 55% DN cases. A few tumoral hepatocytes were unambiguously stained in 28% HCC. The percentage of positive cells and the intensity of immunostaining significantly decreased from cirrhosis to HCC through DN, whereas there was no difference in the prevalence of positivity or the number of viral copies between cirrhosis and HCC using tissue-quantitative RT-PCR. Finally, RT-PCR levels were found parallel with the immunostaining in cirrhosis but not in HCC. These results suggest that HCV protein synthesis may persist but be down-regulated during sequential hepatocarcinogenesis. A putative role of HCV proteins on cell proliferation and differentiation during the early steps of carcinogenesis cannot therefore be excluded.

Expression and cellular localization of fibrillin-1 in normal and pathological human liver

BACKGROUND

The expression and the distribution of fibrillin-1 and elastin were studied in normal and pathological human liver samples.

METHODS

As controls, histologically normal/subnormal liver samples (n = 24) were used. Pathological samples corresponded to seven cirrhosis and eight hepatocellular carcinomas (HCC) developed on cirrhotic (four) or noncirrhotic (four) liver.

RESULTS

In normal liver, fibrillin-1 and elastin co-localized in vessel walls and portal tract connective tissue. Fibrillin-1 alone was detected along sinusoids and in portal spaces at the interface with the limiting hepatocytic plates and close to the basement membrane of bile ducts. By transmission electron microscopy, typical bundles of microfibrils were detected both in Disse space and in portal zones. Cirrhotic nodules were usually rich in fibrillin-1 along sinusoids; fibrillin-1 and elastin were co-localized in fibrotic septa surrounding nodules. In HCC, fibrillin-1 was present between tumoral hepatocytes; stromal reaction around the tumors contained both fibrillin-1 and elastin.

CONCLUSIONS

Fibrillin-1 was associated with elastin in portal mesenchyme and vessel walls of normal liver, in fibrotic septa around cirrhotic nodules and stromal reaction around HCC, but was expressed alone in the perisinusoidal space. The functional roles for fibrillin-1 in non-elastic tissues, such as the liver, remain to be elucidated.

Recurrence and accelerated progression of hepatitis C following liver transplantation

The patient described had recurrent hepatitis C following OLT. This hepatitis appeared early postOLT and progressed to fibrosing cholestatic hepatitis, a severe form of HCV recurrence. Factors such as genotype 1, high viral load and severe damage on the first postOLT biopsy may indicate a more severe outcome. We have hypothesized that, in parallel to what is known for hepatitis B, this rare form of recurrence was linked to a high expression of virus C proteins in the liver graft. Severe form of hepatitis C recurrence should be treated early with the best currently available treatment which is a combination of IFN and ribavirin. Large series of patients with comparable virological, histological and immunological inclusions criteria are necessary to evaluate the efficacy of this treatment.

[Liver transplantation and constrictive pericarditis]

We report the case of a patient with refractory ascitis due to a constrictive pericarditis who underwent a liver transplantation with the initial diagnosis of cryptogenic cirrhosis. The cardiac origin was suspected 5 months post surgery when a liver biopsy showed lesions in favor of a post sinusoidal shunt. The diagnosis was confirmed by the increased values of the right intra-ventricular pressures. We discuss the causes of the delay of the diagnosis and, in particular, the difficulty to interpret vascular liver lesions. Such vascular lesions were present on the needle biopsy performed prior to transplantation but wrongly interpreted as cirrhosis.

[Iron overload and cancer]

Most experimental and human data support the hypothesis that iron overload is a risk factor for cancer in general and liver cancer in particular. This oncogenic effect could be explained by an overproduction of reactive oxygen species and free radicals. In cirrhosis due to genetic haemochromatosis (homozygosity for the mutation C282Y in the HFE gene) there is an increase incidence of hepatocellular carcinoma. Few cases have been reported in genetic haemochromatosis with iron overload but without cirrhosis. In hepatocellular carcinoma developed in non cirrhotic liver there is a mild iron overload in more than 50% of cases. In these patients heterozygous and compound C282Y mutations are found in 36%. In black Africans, iron overload genetically determined but not linked to mutations in the HFE gene increases also the risk of hepatocellular carcinoma. Among the many factors (viral hepatitis, alcohol, tobacco etc.) which play a role in hepatic carcinogenesis, iron overload is probably an important one and therefore should be treated.

Expression of hepatocyte growth factor in human hepatocellular carcinoma

BACKGROUND/AIMS

We have shown that hepatocyte growth factor, secreted by human liver myofibroblasts, promoted in vitro invasion of human hepatocellular carcinoma cell lines. The aim of this work was to measure hepatocyte growth factor expression in 29 human hepatocellular carcinomas and the corresponding peri-tumoral livers.

METHODS

We used reverse transcription-polymerase chain reaction, in situ hybridization, ELISA and Western blot.

RESULTS

Sixty-two of tested hepatocellular carcinomas were positive by reverse transcription-polymerase chain reaction. With in situ hybridization, a signal was found in every sample. In many cases, the signal was localized in cells labeled with an anti-smooth muscle alpka-actin antibody, while hepatocytes were mostly non-labeled. ELISA, performed in 15 pairs of hepatocellular carcinomas and surrounding livers, detected hepatocyte growth factor in every sample with wide variations. Hepatocellular carcinomas that had developed in non-cirrhotic livers contained essentially the same amount of hepatocyte growth factor as the matching non-tumoral liver. In cirrhotic livers, the hepatocyte growth factor content of the tumors was significantly lower than that of the surrounding cirrhotic livers.

CONCLUSIONS

These data indicate that hepatocyte growth factor is expressed at significant levels in every hepatocellular carcinoma tested and that its expression takes place in the stromal myofibroblasts.

Paradoxical pro-invasive effect of the serine proteinase inhibitor tissue factor pathway inhibitor-2 on human hepatocellular carcinoma cells

We have previously shown that human liver myofibroblasts promote in vitro invasion of human hepatocellular carcinoma (HCC) cells through a hepatocyte growth factor (HGF)/urokinase/plasmin-dependent mechanism. In this study, we demonstrate that myofibroblasts synthesize the serine proteinase inhibitor tissue factor pathway inhibitor-2 (TFPI-2). Despite the fact that recombinant TFPI-2 readily inhibits plasmin, we show that it potentiates HGF-induced invasion of HCC cells and is capable of inducing invasion on its own. Furthermore, HCC cells stably transfected with a TFPI-2 expression vector became spontaneously invasive. HCC cells express tissue factor and specifically factor VII. Addition of an antibody to factor VII abolished the pro-invasive effect of TFPI-2. We suggest that TFPI-2 induces invasion following binding to a tissue factor-factor VIIa complex preformed on HCC cells. Our data thus demonstrate an original mechanism of cell invasion that may be specific for liver tumor cells.

Multiple black hepatocellular adenomas in a male patient

A 65-year-old man presented with multiple liver tumours. Imaging techniques could not differentiate between adenomas and hepatocellular carcinomas. He had no relevant past medical history. Liver function tests were normal except for a 1.5-fold rise in GGT. AFP was normal. Viral markers were negative. During laparoscopy, numerous black tumours of different sizes were seen. These tumours were adenomas without malignant transformation. Tumoral hepatocytes contained a brown pigment in the canalicular area without evidence of cholestasis. This pigment was Fontana positive and looked like Dubin-Johnson pigment by electron microscopy. The expression of the canalicular multispecific organic anion transporter (cMOAT) was decreased in the tumours but normal in the non-tumoral liver ruling out the diagnosis of Dubin-Johnson syndrome. There was mild iron deposition possibly related to an homozygous H63D mutation in the HFE gene. Three years after their discovery, the size of the tumours remained stable. It is concluded that this male patient with multiple adenomas and mild iron overload is at risk of developing an hepatocellular carcinoma and that the black colour of adenomas is probably due to a partial defect in excretion of organic anions.

Detection of Helicobacter species in the liver of patients with and without primary liver carcinoma

BACKGROUND

Several studies have shown the presence of Helicobacter species in the human biliary tract and in the intestinal tract of animals. In this study, the presence of Helicobacter species in liver samples from patients with primary hepatic carcinomas was evaluated.

METHODS

Sixteen liver specimens were studied (8 from patients with primary liver carcinoma and 8 from patients without primary liver carcinoma). Histology with standard stains, culture, and polymerase chain reaction (PCR) amplification using two sets of primers located in the 16S ribosomal DNA (rDNA) were used to detect the presence of bacteria. Amplified products were sequenced to determine the genus and species of the bacteria. A search for other genes that were specific for Helicobacter pylori also was carried out by PCR.

RESULTS

PCR performed with the 16S rDNA primers revealed the presence of bacteria from the genus Helicobacter in all of the liver specimens from patients with primary liver carcinoma (eight of eight patients) and in one specimen from a patient without primary liver carcinoma (one of eight patients). When the nucleotide sequence of > 80% of the 16S rDNA was determined, the closest similarity was with the 16S rDNA from H. pylori in eight patients. In 1 patient sample from which only 398 nucleotides were sequenced, the closest match was Helicobacter felis.

CONCLUSIONS

The results presented in this study indicate that Helicobacter species can be present in the liver of patients with primary hepatic carcinoma, but their eventual role in the carcinogenesis process, although it is plausible, remains to be proven. Based on sequence similarity, it seems that Helicobacter species that are related closely to H. pylori but are distinct from it have been found.

Detection of Helicobacter species in the liver of patients with and without primary liver carcinoma

BACKGROUND

Several studies have shown the presence of Helicobacter species in the human biliary tract and in the intestinal tract of animals. In this study, the presence of Helicobacter species in liver samples from patients with primary hepatic carcinomas was evaluated.

METHODS

Sixteen liver specimens were studied (8 from patients with primary liver carcinoma and 8 from patients without primary liver carcinoma). Histology with standard stains, culture, and polymerase chain reaction (PCR) amplification using two sets of primers located in the 16S ribosomal DNA (rDNA) were used to detect the presence of bacteria. Amplified products were sequenced to determine the genus and species of the bacteria. A search for other genes that were specific for Helicobacter pylori also was carried out by PCR.

RESULTS

PCR performed with the 16S rDNA primers revealed the presence of bacteria from the genus Helicobacter in all of the liver specimens from patients with primary liver carcinoma (eight of eight patients) and in one specimen from a patient without primary liver carcinoma (one of eight patients). When the nucleotide sequence of > 80% of the 16S rDNA was determined, the closest similarity was with the 16S rDNA from H. pylori in eight patients. In 1 patient sample from which only 398 nucleotides were sequenced, the closest match was Helicobacter felis.

CONCLUSIONS

The results presented in this study indicate that Helicobacter species can be present in the liver of patients with primary hepatic carcinoma, but their eventual role in the carcinogenesis process, although it is plausible, remains to be proven. Based on sequence similarity, it seems that Helicobacter species that are related closely to H. pylori but are distinct from it have been found.

[Premalignant lesions and hepatocellular carcinoma on non cirrhotic liver overloaded with iron]

A 68-year old mildly obese Caucasian man underwent hepatic resection for multinodular hepatocellular carcinoma which had developed in the left lobe of a non-cirrhotic liver. The only risk factors found were heavy drinking, smoking, and serum markers of hepatitis B virus without virus genome in hepatocytes. The non tumoral liver was mildly fibrotic and iron overloaded (hepatic iron index: 1.6) with three types of iron-free lesions: (i) periportal clear hepatocyte foci, (ii) hyperplastic nodules and (iii) dysplastic or neoplastic nodules with well to moderately-differentiated hepatocellular carcinoma. The genetic investigation was negative for the C282Y and the H63D mutations of the HFE gene. This observation illustrates the multistep process of carcinogenesis in the non-cirrhotic liver and raises the question of i) the origin of this iron overload possibly linked to insulin resistance syndrome and ii) the role of iron as a co-carcinogen.

Hepatocarcinoma cells stimulate hepatocyte growth factor secretion in human liver myofibroblasts

Hepatocellular carcinoma (HCC), the main type of primary liver cancer, is characterized by a high rate of intra-hepatic invasion. The stroma of HCC is infiltrated by myofibroblasts. We have previously shown that hepatocyte growth factor (HGF) secreted by human liver myofibroblasts greatly increased the in vitro invasiveness of 3 human HCC cell lines. In this study we show that the conditioned medium (CM) from the same HCC cell lines dose-dependently stimulates HGF secretion by myofibroblasts. This effect was post-transcriptional as no increase in HGF mRNA was observed. We show that the effect of CM is not due to IL-1, IL-6, IGF-1, bFGF or PDGF, previously shown to stimulate HGF synthesis in other models. Our data demonstrate that HCC cells increase HGF secretion by liver myofibroblasts in a paracrine way that could act to enhance invasion.

Increased incidence of HFE C282Y mutations in patients with iron overload and hepatocellular carcinoma developed in non-cirrhotic liver

BACKGROUND/AIMS

Histological and biochemical iron overload has been reported in non-tumoral liver of most patients presenting an hepatocellular carcinoma (HCC) developed in non-cirrhotic liver (NCL). The aim of our study was to investigate HFE mutations in patients with HCC in NCL.

METHODS

Thirty-five patients with HCC in NCL were included either retrospectively or prospectively. Clinical data, iron and viral status, and HFE gene mutations were compared between groups with (I+, n = 19) or without histological iron overload (I-, n = 16).

RESULTS

Twenty per cent of patients were HBV or HCV positive. Fifty-four per cent had hepatocytic iron overload at histology. Mean hepatic iron concentration was 100.2 +/- 14.6 micromol/g in I+ versus 23.2 +/- 2.1 micromol/g in I- (p<0.001). Among the 19 I+ patients, eight mutations were found: two C282Y/C282Y, three C282Y/WT, two C282Y/H63D and one H63D/H63D. None of these mutations was found in the I- group. There was no significant difference concerning the H63D heterozygous mutation between I+ or I- patients.

CONCLUSIONS

In patients with HCC in NCL, HBV and HCV markers are rare (20%), and mild iron overload is frequent (54%). In patients with HCC in NCL and iron overload, C282Y mutations are frequent (36.8% of cases) and significantly increased (p<0.009) compared to HCC in NCL without iron overload; these mutations are mostly heterozygous. H63D heterozygosity is not associated with liver iron overload. Because of the small size of the series, HFE C282Y mutation should be investigated on a larger scale in patients with HCC in NCL with iron overload in order to confirm this association.

Osteonectin/SPARC is overexpressed in human hepatocellular carcinoma

Osteonectin (ON)/SPARC is a glycoprotein involved in extracellular matrix remodelling. ON expression by myofibroblasts has been reported in fibrotic human liver. As ON also plays a role in cell adhesion, differentiation, and proliferation, this study was designed to document its expression in human hepatocellular carcinoma (HCC). Tissues from 26 HCCs of various histological grades and architecture and from surrounding non-tumour liver (23 cirrhotic or fibrotic, three non-fibrotic) were tested by in situ hybridization and immunohistochemistry. Immunohistochemical detection of alpha-smooth muscle actin (alpha-SMA) was performed on serial sections or in combination with hybridization. Large amounts of ON mRNA and protein were detected in the tumour capsule, in the fibrous bands, and along capillaries within HCCs. The signal was located in cells suggestive of myofibroblasts, as confirmed by positive staining for alpha-SMA. In HCC, ON protein was always detectable, with strong staining in high-grade tumours, whereas it was mostly undetectable in non-tumour tissues. A clear difference was also shown for ON transcripts, except in a few cases with chronic active hepatitis, where ON transcripts were also expressed at a high level. Overexpression of ON transcripts in HCC vs. non-tumour liver was confirmed by RNA blot in 20/22 patients tested. In conclusion, ON is strongly expressed by the stromal myofibroblasts of human HCC, especially of high grade. This expression could play a role in tumour progression.

Cholangiocarcinoma arising in Von Meyenburg complex associated with hepatocellular carcinoma in genetic haemochromatosis

A 61-year-old man had a liver resection for a bilobar mass thought to be, by imaging techniques, an hepatocellular carcinoma. He had been treated for the last 12 years by venesections for genetic haemochromatosis complicated by well-compensated cirrhosis. At surgery, prothrombin time and platelet count were normal, as was alpha-fetoprotein. On the resected specimen, the non-tumoral liver was not cirrhotic; septal fibrosis was present as well as mild iron overload and numerous Von Meyenburg complexes. The bilobar tumour was composed of two different parts: one was a cholangiocarcinoma arising from Von Meyenburg complexes, the other was a moderately differentiated hepatocellular carcinoma with a partially invaded capsule. The two tumours, in close proximity, did not communicate. This observation raises three questions: the relative risk of primary liver cancer including both hepatocellular carcinoma and cholangiocarcinoma in haemochromatosis without cirrhosis; the development of cholangiocarcinoma from Von Meyenburg complexes; the reversibility of cirrhosis in treated patients.

Osteonectin/SPARC is overexpressed in human hepatocellular carcinoma

Osteonectin (ON)/SPARC is a glycoprotein involved in extracellular matrix remodelling. ON expression by myofibroblasts has been reported in fibrotic human liver. As ON also plays a role in cell adhesion, differentiation, and proliferation, this study was designed to document its expression in human hepatocellular carcinoma (HCC). Tissues from 26 HCCs of various histological grades and architecture and from surrounding non-tumour liver (23 cirrhotic or fibrotic, three non-fibrotic) were tested by in situ hybridization and immunohistochemistry. Immunohistochemical detection of alpha-smooth muscle actin (alpha-SMA) was performed on serial sections or in combination with hybridization. Large amounts of ON mRNA and protein were detected in the tumour capsule, in the fibrous bands, and along capillaries within HCCs. The signal was located in cells suggestive of myofibroblasts, as confirmed by positive staining for alpha-SMA. In HCC, ON protein was always detectable, with strong staining in high-grade tumours, whereas it was mostly undetectable in non-tumour tissues. A clear difference was also shown for ON transcripts, except in a few cases with chronic active hepatitis, where ON transcripts were also expressed at a high level. Overexpression of ON transcripts in HCC vs. non-tumour liver was confirmed by RNA blot in 20/22 patients tested. In conclusion, ON is strongly expressed by the stromal myofibroblasts of human HCC, especially of high grade. This expression could play a role in tumour progression.

Beta-catenin mutations in hepatocellular carcinoma correlate with a low rate of loss of heterozygosity

To determine the frequency of Wnt/Wingless beta catenin pathway alteration in human hepatocellular carcinoma, a beta catenin and APC gene mutation screening was performed in a series of 119 tumors. An activating beta catenin mutation in exon 3 was found in 18% of the cases. Among tumors lacking beta catenin mutation, no APC mutation has been evidenced in a subset of 30 cases tested. The correlation between beta catenin mutation status and chromosome segment deletions was studied on a set of 48 hyperploid tumors. Chromosome 1p, 4q and 16p deletions were significantly associated with the absence of beta catenin mutation (P<0.05). Furthermore the Fractional Allelic Loss was significantly smaller in the beta catenin mutated tumors than in the non-mutated tumors (0.12 versus 022). Taken together, these results suggest, the existence of two carcinogenesis mechanisms. The first mechanism implies a beta catenin activating mutation associated with a low rate of loss of heterozygosity. The second mechanism, operating in a context of chromosomal instability, would involve tumor suppressor genes.

Evidence for the presence of a functional pregnane X receptor response element in the CYP3A7 promoter gene

Pregnane X Receptor (PXR) has been recently shown to regulate the inducible expression of CYP3A genes in response to xenobiotics and steroids. PXR forms a heterodimer with the retinoic acid receptor (RXR) and this complex binds to and transactivates an 18bp region containing two everted repeats TGA(A/C)CT separated by 6 nucleotides (ER6) and located at approximately -150 in the CYP3A4 promoter. In this work we have isolated and sequenced the proximal 5'-flanking region of CYP3A7 from two different human genomic libraries. In contrast to a previously reported sequence (Itoh et al., 1992), we did not observe any mutation in the 3'-half of the CYP3A7 ER6 element. Using electrophoretic mobility shift assays and cotransfection experiments we show that this element is able to bind the PXR:RXR complex and transactivates the expression of a down stream promoter in response to rifampicin, clotrimazole, and RU-486, three compounds known to specifically activate the human PXR. This is consistent with the fact that CYP3A7 mRNA is inducible in several primary cultures of human hepatocytes from different patients, as well as in two hepatocarcinoma cell lines HuH7 and HepG2, in response to these compounds. In contrast to a previous report (Blumberg et al., 1998), based on the sequence published by Itoh et al., we conclude that CYP3A7, like CYP3A4, is inducible in response to xenobiotics and presumably in a large proportion of the population.

Activation of cultured rat hepatic stellate cells by tumoral hepatocytes

Hepatocellular carcinoma (HCC) is the main type of primary liver cancer, and it develops from hepatocytes. The stroma of HCC is infiltrated by myofibroblasts. In other settings, such as liver fibrosis, myofibroblasts are derived mainly from the activation of hepatic stellate cells (HSC). In this study, we investigated whether tumoral hepatocytes were able to activate HSC. HSC were isolated from normal rats and were plated in dishes coated with Matrigel, to prevent their spontaneous activation. HSC were exposed to conditioned medium (CM) from the rat HCC lines Fao and H5. Tumor cell CM elicited major morphologic changes, such as spreading and generation of cytoplasmic processes. Fao and H5 CM increased HSC proliferation to 1.60 and 1.76 times control values, respectively. The expression of alpha-smooth muscle actin was low or undetectable in control cells and was markedly increased by both tumor cell CM but not by normal rat hepatocyte CM. Desmin expression was also enhanced. Gelatinase A secretion was significantly increased 1.20-fold by Fao CM and 1.55-fold by H5 CM. Expression of beta-type platelet-derived growth factor receptor mRNA was increased 5.8-fold by H5 CM but was decreased to 13% of control levels by Fao CM. HSC activation by tumor cell CM was not prevented by urokinase or matrix metalloproteinase inhibitors, suggesting that Matrigel degradation was not central to the activation process. Finally, a blocking antibody to transforming growth factor-beta1 did not impede Fao CM-induced activation but significantly blocked the increase in matrix metalloproteinase-2 expression induced by H5 CM. Our results show that tumoral rat hepatocyte CM is able to induce the activation of rat HSC in culture. The lack of induction of beta-type platelet-derived growth factor receptor mRNA by Fao CM indicates that, in some cases, tumor-induced activation differs from classic fibrosis-type activation. Our data thus suggest that HSC recruitment and activation in HCC could be under the control of tumor cells.

Direct evidence that hepatocyte growth factor-induced invasion of hepatocellular carcinoma cells is mediated by urokinase

BACKGROUND/AIMS

We have shown that hepatocyte growth factor secreted by human hepatic myofibroblasts increased the in vitro invasion of the hepatocarcinoma cell line HepG2 through Matrigel. Our aim in this study was to evaluate the role of urokinase in this process.

METHODS

Expression of urokinase in HepG2 cells was measured by Northern blot and zymography, and plasminogen activation was shown by a chromogenic substrate assay. Cell invasion was assayed on Matrigel-coated filters. Urokinase and urokinase receptor transcripts in hepatocarcinoma were detected by reverse transcription-polymerase chain reaction. Activated hepatocyte growth factor was detected by Western blot with a hepatocyte growth factor-beta chain-specific antibody.

RESULTS

HepG2 cells expressed urokinase mRNA and secreted active urokinase. Urokinase expression was enhanced by hepatocyte growth factor at the protein and mRNA level. Notably, cell-surface-associated urokinase was increased 22-fold by hepatocyte growth factor. Hepatocyte growth factor also increased urokinase receptor mRNA expression. B428, a urokinase inhibitor, decreased by up to 70% HepG2 invasion induced by myofibroblasts and by 90% that induced by recombinant hepatocyte growth factor. This was not due to a decrease in the generation of activated hepatocyte growth factor by myofibroblasts. Finally, all 17 hepatocarcinoma samples tested expressed urokinase and urokinase receptor transcripts.

CONCLUSION

Hepatocyte growth factor-dependent, myofibroblasts-induced invasion of HepG2 cells is secondary to the induction of urokinase expression on tumor cells.

Myofibroblasts are responsible for collagen synthesis in the stroma of human hepatocellular carcinoma: an in vivo and in vitro study

BACKGROUND/AIMS

Marked changes in extracellular matrix occur in the stroma of hepatocellular carcinoma, as compared to normal or cirrhotic liver. The cell types responsible for extracellular matrix synthesis within hepatocellular carcinoma have not been clearly identified.

METHODS

In vivo collagen synthesis was studied by in situ hybridization and immunohistochemistry for types I, IV, V and VI collagen, together with immunolabeling of alpha-smooth muscle actin, a myofibroblast marker, and CD34, an endothelial cell marker. In vitro, extracellular matrix deposition by cultured myofibroblasts was studied by reticulin staining, immunocytochemistry and RNase protection.

RESULTS

All collagens studied were expressed in the stroma of the tumor, with a higher level of type VI and IV collagens than of type I and V. The majority of the cells expressing collagen transcripts in human hepatocellular carcinoma stroma were alpha-actin positive and CD 34 negative. In vitro experiments demonstrated that the hepatocellular carcinoma cell lines HepG2, HuH7 and Hep3B markedly increased extracellular matrix deposition by human liver myofibroblasts. This increase was mediated by a soluble mediator present in tumor cell conditioned medium. It was not explained by an increase in mRNA levels of extracellular matrix components, nor by a decrease in the secretion of matrix-degrading proteinases by myofibroblasts.

CONCLUSIONS

Myofibroblasts are the main source of collagens in the stroma of hepatocellular carcinoma. Our data also indicate that tumoral hepatocytes increase extracellular matrix deposition by cultured myofibroblasts, probably by post-transcriptional mechanisms. The generation of hepatocellular carcinoma stroma by myofibroblasts could thus be under control of tumoral cells.