Over the last decade, an improvement in diagnostic techniques has increased the detection of various benign liver tumors. These lesions pose difficulties in management. A critical question is whether they should be resected (to prevent complications such as hemorrhage or malignant transformation) or be conservatively managed. The prognosis and fate of various lesions, such as focal nodular hyperplasia (FNH), nodular regenerative hyperplasia (NRH) and hemangiomas are better understood and the management is mainly conservative unless there are associated symptoms or more importantly, if malignancy cannot be ruled out.1, 2 However, management of hepatic adenomas (HA), a rare benign hepatic tumor, is considerably more challenging. Adenomas have been classically associated with the use of estrogen-containing oral contraceptives or androgen-containing steroid anabolics.3, 4 Such tumors are usually observed as solitary masses and are asymptomatic. Additionally, glycogen storage diseases, especially Type I (von Gierke) and Type III, are known risk factors for the development of hepatic adenoma.2, 5 In this condition, the tumors often times occur in multiple with a higher propensity to undergo malignant transformation. Macroscopically, HA presents as a solitary, yellowish mass due to lipid accumulation and gives a pseudo-encapsulated appearance because of the compression of adjacent hepatic tissue.6 Some of the histological features include areas of fatty deposits and hemorrhage, cord- or plate-like arrangements of larger hepatocytes containing excessive glycogen and fat, sinusoidal dilatation (the result of the effects of arterial pressure as these tumors lack a portal venous supply), absent bile ductules, and presence of few and non-functioning Kupffer cells.2, 6 The extensive hypervascularity and a lack of a true capsule makes this tumor prone to hemorrhage.
Despite these well-characterized features, the prognosis of hepatic adenomas is not well established.2, 6 The tumors can regress in size following discontinuation of inciting factors such as oral contraceptives; at other times they continue to exhibit a stable or progressive disease. However, the risk of associated hemorrhage and neoplastic transformation is always present. In view of these risks, surgical resection is often recommended. Performance of surgery is influenced by the multiplicity or size of the tumors and associated symptoms on one hand and surgical risk to the patient on the other.1, 7–10 However, molecular predictors have not been available that would assist in identifying a subset of patients with greater propensity, for example, of developing malignancy. While this has been tried by screening for elevated α-fetoprotein levels as well as radiologically by serial examinations to identify first signs of such transformation, the success has been limited.6 The true prognostic indicators would come from understanding the molecular basis of this disease, which is what Zucman-Rossi et al. attempted in their report in the current issue of HEPATOLOGY.11
In the study, the authors examined 96 liver tumors with a diagnosis of hepatic adenoma. Further examination enabled them to categorize these tumors into three major subsets. The predominant subset of around 46% of all patients fell into the category of adenomas displaying HNF1α mutations. These tumors were predominantly a homogeneous group of firmly diagnosed adenomas in 84% of the cases with classical features, including marked steatosis and lack of cytological aberrations and inflammation. Out of the remaining adenomas in this group (7 tumors), 3 had some association with HCC while the remaining 4 were borderline between FNH and adenoma and thus would not change the management. Thus, it would be safe to state that around 93% of all hepatic adenomas displaying HNF1α mutations would have an overall benign course and a conservative approach would be sufficient. However, additional molecular aberrations were not explored in a minor subset, which might help identify this subgroup, which possibly could exhibit malignant transformation.
The second subset of hepatic adenomas accounting for around 12% of tumors was the one that exhibited β-catenin gene alterations. Most of these alterations were encompassing the “degradation domain” within the CTNNB1 gene altering the serine/threonine phosphorylation sites, resulting in its protein stabilization. While the percentage of these alterations was lower than what was reported previously, β-catenin nuclear translocation might be secondary to non-mutational causes, as suggested by immunohistochemical analysis.12 The current study did not perform immunohistochemistry for β-catenin, which could have addressed the significance of sustained nuclear β-catenin accumulation secondary to its gene alterations versus other causes. This information would have extended our understanding of the role of β-catenin in the pathogenesis of adenomas. It is quite possible that the discrepancy in β-catenin nuclear translocation and observed mutations might be secondary to tyrosine kinase activation, as suggested previously and could be secondary to TGFα or HGF/c-Met activation, interactions likely to be more benign than previously speculated.13, 14 This is especially true since some of our unpublished work has now depicted HGF-dependent β-catenin activation to be transient but functional, which can be successfully monitored by the hepatocyte ubiquitination machinery and by sequestration of β-catenin at the membrane by E-cadherin. Detection of β-catenin or axin mutations and its correlation with elevated expression of two of its known target genes, glutamine synthetase and GPR49, confirmed the sustained activation of β-catenin, although this could not rule out transient activation of β-catenin in other categories.
Nonetheless, 6/12 cases in this class were associated with HCC, being either borderline lesions or having concomitant HCC. Histologically, there were frequent cytological abnormalities and pseudo-glandular formation. Thus, a subgroup of hepatic adenomas with atypical histological features displaying activation of β-catenin, occurring at abnormally higher frequency in males than what is known for traditional adenomas, would indicate a higher propensity for malignant transformation. Additional prospective studies would need to confirm the requirement for an aggressive approach in such cases.
Finally, the last group of hepatic adenomas displayed no mutations in HNF1α or CTNNB1. While this report fails to identify the molecular basis in this group, they do identify a smaller subgroup of tumors displaying inflammation and pseudo-glandular formation. This subgroup was placed as being closer to telangiectatic FNH with dystrophic rather than pre-neoplastic features and hence not a high risk for malignant transformation. Again, additional prospective studies would be essential. In addition, the authors fail to elaborate on 3/12 cases of hepatic adenomas displaying simultaneous inflammation and mutations in the CTNNB1. This comprised a fourth of hepatic adenomas that displayed mutations in β-catenin gene, suggesting that ongoing inflammation might have additional significance and hence would need to be investigated in a larger sample size. On the other hand, the larger mutation-free and non-inflammatory subgroup in this category did not display any specific clinical or morphological features and were presumed innocent requiring a conservative follow-up. Again, additional prospective studies would be needed to strengthen this claim as well.
This study introduces a molecular classification of hepatic adenomas, detecting patients at a higher risk for malignancy, who might be candidates for early surgical resection or perhaps chemoprophylaxis. This classification is based on identifying a subgroup of adenomas with no genetic aberrations (Fig. 1A) or mutations in β-catenin gene, the latter being associated with HCC (Fig. 1B-C) and implied as a tendency for malignant transformation. β-Catenin also plays an important role in hepatic physiology. It regulates hepatocyte proliferation, survival and differentiation during normal liver growth, development and regeneration.15–22 However, as in other tumors, aberrant activation of β-catenin is implicated in liver tumors, as well.23 Irrespective of the basic mechanism of HCC development (dedifferentiation or maturation arrest hypothesis), β-catenin activation has been reported in various “stages” of tumorigenesis, including the dysplastic foci, hepatic adenoma, hepatoblastoma and HCC.12, 24–26 Multiple mechanisms of β-catenin activation or stabilization have been reported in these cases including mutations in the β-catenin gene or CTNNB1, or components of its degradation machinery such as AXIN1&2 and GSKβ inactivation.13, 24, 26–30 In mice, deletion of APC in liver, another degradation component of β-catenin, leads to HCC.31 Recently, upregulation of a member of Wnt receptors, Frizzled-7, was shown as another possible mechanism of β-catenin activation in HCC.32 In our recent unpublished observations, we have also identified β-catenin activation in oval cells (liver stem cells), which might be precursors of a subset of HCC. Thus, there appears to be an array of mechanisms of β-catenin activation in an array of liver pathologies leading to liver neoplasia.
Whatever the mechanism, the convergence at β-catenin makes it an attractive therapeutic target.33, 34 Current investigations are being aimed at identifying inhibitors of β-catenin for preclinical and clinical applications. It would be of essence to see the benefit of such agents on hepatic adenomas that display elevated β-catenin and its target genes. These investigations could also be extended to chemoprophylaxis if there is evidence of presence of multiple risk factors. In cases of hepatic adenomas, this could be the presence of β-catenin mutations, lesser than usual steatosis, presence of cytological abnormalities, presence of simultaneous inflammation and pseudo-glandular structures and possible occurrence in males. The verdict for such an adenoma would most likely be guilty!