Moreira, Roger K.

AbstractObjectivesSteatohepatitic hepatocellular carcinoma is a distinct variant of hepatocellular carcinoma strongly associated with underlying nonalcoholic steatohepatitis. The molecular biology of steatohepatitic hepatocellular carcinoma is not fully elucidated, and thus we aimed to investigate the molecular underpinnings of this entity.MethodsTranscriptomic analysis using RNAseq was performed on eight tumor-nonneoplastic pairs of steatohepatitic hepatocellular carcinoma with comparison to conventional hepatocellular carcinoma transcriptomes curated in The Cancer Genome Atlas. Immunohistochemistry was used to validate key RNA-level findings.ResultsSteatohepatitic hepatocellular carcinoma demonstrated a distinctive differential gene expression profile compared with The Cancer Genome Atlas curated conventional hepatocellular carcinomas (n = 360 cases), indicating the distinctive steatohepatitic hepatocellular carcinoma morphology is associated with a unique gene expression profile. Pathway analysis comparing tumor-nonneoplastic pairs revealed significant upregulation of the hedgehog pathway based on GLI1 overexpression and significant downregulation of carnitine palmitoyltransferase 2 transcript. Glutamine synthetase transcript was significantly upregulated, and fatty acid binding protein 1 transcript was significantly downregulated and immunohistochemically confirmed, indicating steatohepatitic hepatocellular carcinoma tumor cells display a zone 3 phenotype.ConclusionsSteatohepatitic hepatocellular carcinoma demonstrates a distinctive morphology and gene expression profile, phenotype of zone 3 hepatocytes, and activation of the hedgehog pathway and repression of carnitine palmitoyltransferase 2, which may be important in tumorigenesis.

Biliary complications (strictures and leaks) represent major limitations in living donor liver transplantation. Mesenchymal stem cells (MSCs) are a promising modality to prevent biliary complications because of immunosuppressive and angiogenic properties. Our goal was to evaluate the safety of adipose‐derived MSC delivery to biliary anastomoses in a porcine model. Secondary objectives were defining the optimal method of delivery (intraluminal versus extraluminal) and to investigate MSC engraftment, angiogenesis, and fibrosis. Pigs were divided into 3 groups. Animals underwent adipose collection, MSC isolation, and expansion. Two weeks later, animals underwent bile duct transection, reanastomosis, and stent insertion. Group 1 received plastic stents wrapped in unseeded Vicryl mesh. Group 2 received stents wrapped in MSC‐seeded mesh. Group 3 received unwrapped stents with the anastomosis immersed in an MSC suspension. Animals were killed 1 month after stent insertion when cholangiograms and biliary tissue were obtained. Serum was collected for liver biochemistries. Tissue was used for hematoxylin‐eosin and trichrome staining and immunohistochemistry for MSC markers (CD44 and CD34) and for a marker of neoangiogenesis (CD31). There were no intraoperative complications. One pig died on postoperative day 3 due to acute cholangitis. All others recovered without complications. Cholangiography demonstrated no biliary leaks and minimal luminal narrowing. Surviving animals exhibited no symptoms, abnormal liver biochemistries, or clinically significant biliary stricturing. Group 3 showed significantly greater CD44 and CD34 staining, indicating MSC engraftment. Fibrosis was reduced at the anastomotic site in group 3 based on trichrome stain. CD31 staining of group 3 was more pronounced, supporting enhanced neoangiogenesis. In conclusion, adipose‐derived MSCs were safely applied to biliary anastomoses. MSCs were locally engrafted within the bile duct and may have beneficial effects in terms of fibrosis and angiogenesis.

Hepatocellular adenomas (HCAs) often pursue an innocuous clinical course. Recent work has elucidated important subtypes of HCA and biomarkers to identify them, including HCA at an increased risk for malignant transformation. Another key complication of HCAs is the risk of spontaneous tumoral hemorrhage, which may be life-threatening. Identification of a predictive biomarker for this clinical complication would therefore be of clinical value. It has been suggested that Argininosuccinate Synthase 1 (ASS1) immunohistochemistry (IHC) identifies HCA with a high propensity for hemorrhage. The aim of our study was to validate ASS1 IHC as a predictive marker of hemorrhage. Eighty-nine HCAs were collected for ASS1 IHC and subtyped according to published criteria. Clinical records were examined for evidence of tumoral hemorrhage. Twenty-one (23.6%) HCAs were complicated by clinically detected hemorrhage and were more likely to be resected (P=0.0002). Hemorrhage complicated all WHO subtypes of HCA. There was no association between hemorrhage and HCA subtype (P=0.92). Neither the distribution of ASS1 expression nor the intensity of ASS1 expression compared to normal liver showed a significant association with hemorrhage (P=0.051 and 0.34). Interlaboratory comparison of 8 cases showed good agreement regarding the intensity (6/8 and 7/8) and distribution of staining (7/8 and 7/8) across 3 laboratories performing ASS1 IHC. In conclusion, all subtypes of HCA may be complicated by hemorrhage. ASS1 IHC expression did not correlate with hemorrhagic complications. Caution is prudent before routine implementation of ASS1 IHC in clinical practice.

ABSTRACT Objectives Albumin messenger RNA (mRNA) expression is a marker of hepatocellular differentiation. Most published data are from review of tissue microarrays, and albumin in situ hybridization (ISH) expression across several tumor types is incompletely characterized. Methods Sections from 221 tumors were evaluated for albumin mRNA. Immunohistochemistry was used to confirm diagnoses. Albumin ISH was performed according to manufacturer-provided instructions. Fifty-nine cases were evaluated with both commercial ISH assays. Results Albumin mRNA was detected in all hepatocellular carcinomas (HCCs) and 81% of intrahepatic cholangiocarcinomas. Lung (20%), gallbladder (39%), hepatoid pancreatic (n = 1 of 1) adenocarcinoma, breast invasive ductal carcinoma (18%), yolk sac tumor (25%), and acinar cell carcinoma (29%) showed expression. Both assays were concordant in 93% of cases. Conclusions Albumin ISH was expressed in all HCCs studied. It was also positive in intrahepatic cholangiocarcinoma and patchy positive in gallbladder adenocarcinoma and a subset of other neoplasms, which can be a potential pitfall.

Steatohepatitis-like change has not been described in focal nodular hyperplasia (FNH). Steatohepatitis-like change in FNH may show overlapping features with steatohepatitic variant of hepatocellular carcinoma (HCC). This problem can be compounded if seen in FNH with widened cell plates or hepatocyte rosettes, other features that can also be seen in HCC. This study examined steatotic FNHs for the frequency of steatohepatitis-like change, especially in the setting of FNH with rosettes and/or widened cell plates. Thirty-three resection specimens of steatotic FNH from 3 institutions were evaluated for degree of steatosis, background liver steatosis, ductular reaction, and lymphocytic infiltrate, as well as presence of thick fibrous bands, thick-walled vessels, ballooned hepatocytes, Mallory-Denk bodies, dilated sinusoids, hepatocyte rosettes, and thick hepatic plates. Steatosis was distributed along fibrous septa as well as diffusely throughout the FNH. Steatohepatitis-like changes were focally present in 54% (18 cases). Thick plates>3 cells were focally found in 14 cases (42%); rosettes were common (70%). All cases showed at least 2 of the histologic features highly suggestive for the diagnosis of FNH such as thick bands of fibrosis, thick-walled vessels and/or ductular reaction and the typical map-like pattern of glutamine synthetase immunostaining. More than half of fatty FNH examined for this study had features of at least focal steatohepatitis-like changes. This finding should not be confused with steatohepatitic variant of HCC. Common typical features of FNH including thick-walled vessels, ductular reaction and thick fibrous bands are helpful for discrimination of FNH from HCC.

Context HepPar-1 and Arginase-1 are urea cycle enzymes used to distinguish hepatocellular carcinoma from other carcinomas. HepPar-1, but not Arginase-1, is known to be immunoreactive with normal human small intestine. Objectives To better define and compare the immunohistochemical staining patterns of HepPar-1 and Arginase-1 in adenocarcinomas arising in the small intestine, including the ampullary region. Design Staining for HepPar-1 and Arginase-1 was performed on 20 nonampullary small intestinal adenocarcinomas and 32 adenocarcinomas from the ampullary region. Ampullary adenocarcinomas were divided into intestinal morphology (15), pancreatobiliary morphology (14), and unclassifiable (3). Nonneoplastic small intestinal mucosa and colorectal adenocarcinomas were used as control groups. Results HepPar-1 stained 12 of 20 nonampullary small intestinal adenocarcinomas, with a median of 63% of cells staining in positive cases. It also stained 11 of 15 ampullary carcinomas with intestinal morphology, with a median of 75% of cells staining in positive cases. Two of 14 ampullary carcinomas with pancreatobiliary morphology were positive for HepPar-1. Arginase-1 showed positivity in 2 ampullary region carcinomas and diffuse positivity in 1 duodenal adenocarcinoma. Two of 22 colorectal carcinomas stained for HepPar-1 with none positive for Arginase-1. Conclusions HepPar-1, but not Arginase-1, usually shows positivity in small intestinal adenocarcinomas and ampullary adenocarcinomas with intestinal morphology, but only rarely shows positivity in ampullary adenocarcinomas with pancreatobiliary morphology. HepPar-1 positivity in metastatic adenocarcinoma with intestinal morphology is suggestive of an upper gastrointestinal primary site.

AimsBile salt export pump (BSEP) is a transporter expressed exclusively at hepatic canaliculi and drives bile‐salt efflux. Minimal data exist about BSEP expression in tumours. We hypothesized that BSEP immunohistochemistry would be specific for hepatocellular carcinoma (HCC).Methods and resultsTissue microarrays, including 48 HCC, 41 cholangiocarcinomas and 24 metastatic tumours in liver, were immunostained for BSEP. Expression was compared with common markers of hepatocytic differentiation including CD10, hepatocyte paraffin‐1 antigen (HepPar‐1), carcinoembryonic antigen, arginase‐1 (ARG) and glypican‐3 (GPC‐3). BSEP expression was assessed in normal tissues. Special attention was given to adrenal gland (normal and neoplasia). BSEP was easy to interpret and showed no background staining. Canalicular expression was seen in all normal livers, but not in other normal tissue. BSEP was 90% sensitive and 100% specific for HCC (canalicular in 33 of 43 positive cases). The sensitivity of ARG was slightly higher, but specificity was slightly lower (94% for both). HepPar‐1 was 90% sensitive and 97% specific. CD10, polyclonal carcinoembryonic antigen (pCEA) and GPC‐3 all had lower sensitivity (74, 81 and 54%, respectively).ConclusionsIn malignant tumours in the liver, BSEP marking was 100% specific and 90% sensitive for HCC. The specificity of BSEP for HCC obviates the need to identify a ‘canalicular’ pattern, which can limit the utility of other canalicular markers.