The recognition that adherence to size and invasion based criteria such as the Milan criteria greatly reduces post-transplant tumor recurrence
 has made it is possible to improve survival outcomes for large numbers of cirrhotic patients with relatively small HCCs. The development of effective therapies, such as TACE, that shrink tumor size has allowed a larger number of patients to undergo transplantation
. However, with increasing understanding that factors other than tumor size may predict tumor phenotype, there is a need to revisit the optimal criteria for predicting post-transplant recurrence. The recent findings that certain progenitor cell markers portend poor prognosis of HCC
 promoted us to investigate the role of these in post-transplant tumor recurrence.
Hepatocellular carcinoma has a number of characteristics that are compatible with the existence of a circulating cancer stem cell (CSC) population
[22, 23]. The tumors are heterogeneous, frequently demonstrate intrahepatic recurrence after liver transplantation in spite of the absence of known metastatic disease, and different subpopulations of cells derived from HCCs have dramatically different abilities to form tumors in immunodeficient mice
[15, 16]. While phenotypic identification of these cells within tumors has been controversial, evidence exists supporting the importance of CD133, CD90, CD13, EpCAM and CD44 as potential markers to identify this cell population
[14–19, 24]. These observations have led to the widespread belief that CSCs are an important part of HCC biology. In spite of this consensus, identification of CSCs has been difficult due to lack of specificity of markers and limitations of model systems.
We reasoned that true HCC CSCs might be enriched in tumor cells that survived doxorubicin-based chemoembolization (TACE) and took advantage of the fact that some, but not all patients at our center undergoing liver transplantation for HCC underwent TACE prior to transplant. We thus stained HCC specimens and adjacent liver tissue for CD133, CD44, CD90, and EpCAM. Each of these has been previously reported to be associated with HCC CSCs. CD133 was used to identify CSCs in several types of tumors including HCC
. CD44, a stem cell biomarker, has been used as a target to attempt to eradicate CSCs in HCC
[17, 24] and has also been linked to invasion and metastatic potential
[25, 26]. EpCAM, a progenitor cell marker, was previously correlated with poor prognosis of HCC
[15, 27], and CD90 was reported as a key marker for the selection of CSCs
The results showed that all of these markers were elevated in liver tumor tissues compared to adjacent non-tumor liver, but there were considerable differences between individual tumors. This suggests that the expression patterns of various stem cell markers in tumor initiating cells with stem/progenitor cell features may be different in each HCC, possibly due to the heterogeneity of activated signaling pathways in normal stem/progenitor cells where these tumor initiating cells may originate. Expression of CD133 and EpCAM were significantly elevated in post-TACE tumors. This finding is similar to the recent observation that HCC tumors recurring post TACE frequently had a morphology suggestive of a mixed hepatocholangiocellular phenotype that stained positive for stem cell markers
. The greater presence of CD133 and EpCAM positive cells in the post-TACE samples could result either because this cell population selectively survived treatment, or because recurrent tumor growth resulted in a dedifferentiation process that generated positive cells. The present study is not able to distinguish between these possibilities. In addition, since this was a retrospective study, we did not have samples from these tumors before TACE, and the decision to perform TACE was made on clinical grounds. It is therefore possible that the TACE patients had larger or more aggressive tumors. The TACE-treated tumor did tend to be larger than the non-TACE tumors but we were not able to detect any differences in pre-op tumor growth rates or TMN score between the two groups (Table
Because of the possible difference between TACE-treated and non-TACE treated tumors we analyzed recurrence data separately for these two groups. We observed that the intensity of EpCAM, but not CD133 staining was a strong predictor of tumor recurrence in patients who were transplanted after TACE treatment. This was not true for the other markers studied. It supports prior studies that have identified EpCAM as a marker that best identified HCC-derived cells with tumor forming potential
The results obtained in this study show a significant relationship between tumor recurrence rates and pretransplantation EpCAM expression in patients who underwent pre-transplant TACE. Recurrence only occurred in 1 of 23 patients without TACE as compared to 4 of 16 patients who had had prior TACE and thus the infrequency of the event in the non-TACE group made it impossible to assess correlations.