Purpose: The current lack of tools for easy assessment of cancer stem cells (CSC) prevents the development of therapeutic strategies for hepatocellular carcinoma (HCC). We previously reported that keratin 19 (K19) is a novel HCC-CSC marker and that PET with 18F-fluorodeoxyglucose (18F-FDG) is an effective method for predicting postoperative outcome in hepatocellular carcinoma. Herein, we examined whether K19+ HCC-CSCs can be tracked using 18F-FDG-PET.

Experimental Design: K19 and glucose transporter-1 (GLUT1) expression was evaluated by IHC in 98 hepatocellular carcinoma patients who underwent 18F-FDG-PET scans before primary tumor resection. Standardized uptake values (SUV) for primary tumors and tumor-to-nontumor SUV ratios (TNR) were calculated using FDG accumulation levels, and values were compared among K19+/K19 patients. Using hepatocellular carcinoma cell lines encoding with a K19 promoter–driven enhanced GFP, 18F-FDG uptake and GLUT1 expression were examined in FACS-isolated K19+/K19 cells.

Results: In hepatocellular carcinoma patients, K19 expression was significantly correlated with GLUT1 expression and FDG accumulation. ROC analyses revealed that among preoperative clinical factors, TNR was the most sensitive indicator of K19 expression in hepatocellular carcinoma tumors. In hepatocellular carcinoma cells, FACS-isolated K19+ cells displayed significantly higher 18F-FDG uptake than K19 cells. Moreover, gain/loss-of-function experiments confirmed that K19 regulates 18F-FDG uptake through TGFβ/Smad signaling, including Sp1 and its downstream target GLUT1.

Conclusions:18F-FDG-PET can be used to predict K19 expression in hepatocellular carcinoma and should thereby aid in the development of novel therapeutic strategies targeting K19+ HCC-CSCs. Clin Cancer Res; 23(6); 1450–60. ©2016 AACR.

Translational Relevance

Non/less-invasive methods for assessing cancer stem cells (CSC) have yet to be developed, which has prevented the development of novel therapeutic strategies for hepatocellular carcinoma (HCC). We previously reported that keratin 19 (K19) is a novel HCC-CSC marker and that PET with 18F-fluorodeoxyglucose (18F-FDG) is useful for predicting postoperative outcomes in hepatocellular carcinoma. However, the relationship between K19+ HCC-CSCs and 18F-FDG-PET remains unclear. Therefore, we hypothesized that K19+ HCC-CSCs could be tracked in hepatocellular carcinoma via 18F-FDG-PET. Our findings indicate that, among preoperative clinical factors, the tumor-to-nontumor standardized uptake value ratio is the most sensitive indicator of K19 expression in human hepatocellular carcinoma and that K19 regulates 18F-FDG uptake through TGFβ/Smad signaling and its downstream target, glucose transporter 1 expression in hepatocellular carcinoma cells. 18F-FDG-PET will contribute to the development of new therapeutic strategies focusing on K19+ HCC-CSCs.

Hepatocellular carcinoma (HCC), which accounts for a majority of liver cancers, is the sixth most common cancer and the second leading cause of cancer-related deaths worldwide (1). Although various therapies have been developed or are currently being developed, the prognosis for patients with hepatocellular carcinoma is far from satisfactory. To develop the new phase of hepatocellular carcinoma treatment, the clinical application of cancer stem cells (CSC) is essential. CSCs possess the ability to self-renew and to generate heterogeneous populations of cancer cells that exhibit high motility and proliferation rates (2, 3). These stem cell–like features of CSCs contribute to rapid tumor growth, the resistance of tumors to chemotherapy/radiotherapy, and the epithelial–mesenchymal transition (EMT; refs. 4–6). In hepatocellular carcinoma, various molecules expressed during liver development, such as epithelial cell adhesion molecule (EpCAM), cluster of differentiation (CD) 90, CD133, and Sal-like protein 4 (SALL4), have been identified as CSC markers (7–11). Independently, we previously reported that keratin 19 (K19), a hepatic progenitor cell marker, is a novel HCC-CSC marker associated with EMT and TGFβ/Smad signaling. Furthermore, we proposed that K19+ HCC-CSCs could be new therapeutic targets of TGFβ receptor 1 inhibitor in hepatocellular carcinoma (12). However, the lack of a non- or less-invasive pretreatment evaluation method for identifying HCC-CSCs has hindered our ability to fully predict patients' outcomes and evaluate the therapeutic efficacy in patients with hepatocellular carcinoma.

On the other hand, PET with 18F-fluorodeoxyglucose (18F-FDG) analysis, which is based on the increased glucose uptake in cancer cells, has been shown to be an effective method for diagnosis or assessment of treatment efficacies in various malignancies (13–15). Indeed, we previously reported that the standardized uptake value (SUV) for the primary tumor and the tumor-to-nontumor SUV ratio (TNR), as calculated from the levels of 18F-FDG accumulation, in hepatocellular carcinoma is correlated with the expression of glucose transporter-1 (GLUT1), a key glucose transporter in both normal and cancer tissues (16). In addition, our previous study revealed that TNR is an independent predictor of postoperative recurrence and overall survival in hepatocellular carcinoma (17, 18). However, the relationship between K19+ HCC-CSCs and 18F-FDG-PET has yet to be elucidated. Therefore, we hypothesized that K19 expression can be predicted by 18F-FDG-PET in hepatocellular carcinoma.

The aims of this study were to demonstrate that K19+ CSCs can be tracked via 18F-FDG-PET in hepatocellular carcinoma tumors. For these analyses, the expression of K19 and GLUT1 was investigated in human hepatocellular carcinoma surgical specimens. Furthermore, we examined the efficacy of SUV, TNR, and various preoperative clinical factors for detecting K19 expression in hepatocellular carcinoma. Moreover, a transgene vector that expressed enhanced GFP (EGFP) under the control of the human K19 promoter was transfected into four hepatocellular carcinoma cell lines to visualize K19+ cells as HCC-CSCs (12). Subsequently, FACS-isolated K19+/K19 cells were used to explore the underlying mechanism governing the regulation of 18F-FDG uptake through TGFβ/Smad signaling in K19+ HCC-CSCs.

Patients

From January 2007 to December 2011, 316 consecutive patients diagnosed with hepatocellular carcinoma by CT and/or MRI were subjected to curative resection at Kyoto University Hospital (Kyoto, Japan). Of these, only 163 patients underwent 18F-FDG-PET during preoperative examination. The remaining 153 patients were excluded because they had previously undergone transarterial chemoembolization and/or radiofrequency ablation (RFA; 72 patients), did not consent to the 18F-FDG-PET procedure (53 patients), or exhibited high C-reactive protein (CRP > 5.0 mg/dL) and/or fasting blood sugar (FBS > 150 mg/dL; 28 patients) levels (Supplementary Fig. S1). In addition, of the 163 patients subjected to 18F-FDG-PET analysis, 62 were subsequently excluded because they did not provide consent for the use of surgical tissues, two were excluded after diagnosis with cholangiocellular carcinoma by pathologic examination, and one was excluded due to difficulties in pathologic assessment resulting from tumor necrosis. Thus, a total of 98 patients were included in this retrospective study (Supplementary Fig. S1). The clinical–pathologic characteristics of the patients are summarized in Supplementary Table S1. Tumor recurrence was investigated until the death of the patient or the end of the study (December 31, 2014). The follow-up period from surgery until death or the endpoint of the study ranged from 157 to 2,891 days (mean 1,408 days).

Written informed consent for the use of resected tissue samples was obtained from all 98 patients in accordance with the Declaration of Helsinki, and this study was approved by the Institutional Review Committee of our hospital.

18F-FDG-PET and image analyses

All PET imaging procedures and image analysis were performed as described previously (16–18). Briefly, maximum SUVs were calculated for quantitative analyses of the levels of tumor 18F-FDG uptake as follows: SUV = C (kBq/mL)×ID (kBq)−1×body weight (kg)−1, where C represents the tissue activity concentration measured by PET and ID represents the injected dose. TNR values were calculated as follows: TNR = tumor SUV/nontumor SUV, where the nontumor SUV is equal to the average of the SUVs at five points in nontumor liver tissues.

IHC

Immunohistochemical analyses were performed as reported previously (19, 20). Mouse anti-human K19 (Dako) and rabbit anti-human GLUT1 primary antibodies (sc-7903, Santa Cruz Biotechnology Inc.) were used at 1:100 and 1:200 dilution, respectively. K19 and GLUT1 expression levels were semiquantitatively assessed. Samples were considered positive for a particular marker when expression of that marker was observed in greater than 5% of the tumor cells examined. Each slide was evaluated by two independent investigators (T. Kawai and K. Yasuchika) who were blinded to the patient outcomes by anonymizing the samples prior to assessment.

Generation of transgenic hepatocellular carcinoma cell lines

The human hepatocellular carcinoma cell lines Huh7, HLF, and PLC/PRF/5 were obtained from the ATCC. All cells were authenticated by short tandem repeat profiling analysis before receipt and were propagated for less than 6 months after resuscitation. The cells were cultured at 37°C with 5% CO2 in RPMI1640 medium (Invitrogen) supplemented with 10% FBS (MP Biomedicals), and penicillin/streptomycin (Meiji Seika).

We generated a transgene plasmid vector that expressed EGFP under the control of the human K19 promoter as described previously (12). In summary, the plasmid vector pHCK-2952, which was constructed by inserting the human K19 promoter sequence into the pGL3-Basic (Promega) vector, was kindly provided by Professor Shuichi Kaneko (Kanazawa University, Kanazawa, Japan; ref. 21). The human K19 promoter region was isolated from pHCK-2952 by restriction digestion with the XhoI and HindIII enzymes (Takara Bio) and then ligated to the plasmid EGFP-1 (pEGFP1; BD Biosciences), which was linearized by digestion with XhoI and HindIII. The transgenic vector was then introduced into hepatocellular carcinoma cells by transfection using Lipofectamine LTX reagent (Invitrogen), according to the manufacturer's protocol. Stably transfected cells were selected for by cultivating in the presence of 200 μg/mL G418 (Sigma-Aldrich) over 30 days. We confirmed proper transgene insertion by PCR and immunocytochemistry analyses, as described previously (12).

Flow cytometry and single-cell culture analysis

We prepared the cultured cells as described previously (12, 22, 23). Dead cells were eliminated via 7-aminoactinomycin D (Beckman Coulter) staining. Single-cell culture analysis was then performed as described previously (12, 22, 23). Individual isolated cells were each sorted into 96-well culture plates using a FACSAria device (BD Biosciences), and wells were visualized by light microscopy at 10 to 16 hours after sorting to confirm that each well contained only one cell. After isolation of each clone, the cells were expanded and subjected to flow cytometry analysis.

RT-PCR and qRT-PCR

The total RNA was extracted with RNeasy Mini Kit (Qiagen) and RNase-free DNAse (Qiagen). Approximately 1 μg of total RNA was then reverse transcribed into cDNA using an Omniscript Reverse Transcription Kit (Qiagen), according to the manufacturer's protocol. Primers were generated for the following genes: K19, GLUT1, K19 open reading frame, Sp1, and β-actin (Supplementary Table S2). RT-PCR analysis was performed as described previously (24). qPCR and qRT-PCR assays were performed using SYBR Green PCR Master Mix (Applied Biosystems) on the ABI 7500 system (Applied Biosystems). Each target was run in triplicate, and expression levels were normalized to those of β-actin.

Western blot analysis

Western blot analysis was performed as reported previously (25). Primary antibodies specific to phospho-smad2 (pSmad2; Ser465/467, #3108; Cell Signaling Technology), Smad2 (#5339, Cell Signaling Technology), Sp1 (ab133596, Abcam), GLUT1 (sc-7903, Santa Cruz Biotechnology), and GAPDH (sc-25778, Santa Cruz Biotechnology) were used at 1:1,000 dilutions. HRP-conjugated bovine anti-rabbit IgG (1:2000; Molecular Probes) secondary antibody was used for detection of pSmad2, Smad2, Sp1, GLUT1, and GAPDH.

Knockdown and overexpression of K19

For K19 knockdown experiments, hepatocellular carcinoma cells were transfected with a 10 nmol/L concentration of the K19-siRNA (#4427037-s7998 or #4427037-s7999; Invitrogen) or control-siRNA (#4390843; Invitrogen) using Lipofectamine LTX reagent (Invitrogen), according to the manufacturer's protocol. K19 expression was significantly downregulated by both K19 siRNAs (Supplementary Fig. S2A). According to the same result acquired with both siRNAs, K19-siRNA (#4427037-s7999) was shown as representative data.

For K19 overexpression experiments, human K19 open reading frame was amplified by RT-PCR and ligated to the CMV6-AC plasmid (PS100020; OriGene) digested with Sgf1 and RsrII. Hepatocellular carcinoma cells were then transfected with this K19 expression vector or with mock vector (PS100020, OriGene) using Lipofectamine LTX reagent (Invitrogen), according to the manufacturer's protocol. K19 expression was significantly upregulated by the K19 expression vector (Supplementary Fig. S2B). For Western blot analysis, hepatocellular carcinoma cells were harvested 48 hours posttransfection.

Reagents

The TGFβR1 inhibitor LY2157299 was obtained from Axon Medchem. The compound was dissolved in 100% DMSO (Sigma) and diluted with RPMI1640 to the desired concentration with a final DMSO concentration of less than 0.5%.

18F-FDG accumulation in hepatocellular carcinoma cells

To determine FDG uptake in vitro, freshly sorted hepatocellular carcinoma cells were plated at a density of 2.0 × 105 cells per well in 12-well plates. In the experiments using TGFβR1 inhibitor LY2157299, 0.5 μmol/L LY2157299 or DMSO control was added. After 24 hours of incubation, the medium was replaced by 1 mL of medium containing 555 kBq of 18F-FDG and was added to each well. 18F-FDG was allowed to accumulate in cells in the incubator over times ranging from 30 to 120 minutes. After incubation for the respective times, the medium was removed and washed three times with ice-cold PBS; the cells were then dissociated with trypsin and collected into tubes. The 18F-FDG radioactivity was immediately determined using a gamma counter (Cobra II Auto-gamma; Packard), with nonspecific background assessed in identically treated cells that were not incubated.

Statistical analyses

Statistical analyses were performed using SPSS version 17.0 (SPSS Statistics, Inc.) and GraphPad Prism software Version 5.0 (GraphPad Software Inc.). Data are presented as the mean ± SD of three or more independent experiments. Student t tests, Mann–Whitney U tests, Fisher exact tests, or χ2 tests, repeated measures ANOVA, and log-rank tests were used for analyses of statistical significance.

Recurrence-free survival (RFS) and overall survival (OS) after the operation were calculated using the Kaplan–Meier method and analyzed with the log-rank test. Significant variables from univariate analyses were entered in the multivariate analysis using a Cox regression model with forward stepwise selection. We plotted ROC curve for SUV, TNR, and preoperative clinical factors and calculated the area under each ROC curve (AUC). The optimal cut-off values for SUV and TNR were calculated using the maximum sum of sensitivity and specificity, as well as the minimum distance to the top left corner of the ROC curve. The clinical cut-off value was used for the assessment of preoperative clinical factors tested. Statistical significance was defined as P < 0.05.

K19 and GLUT1 expressions in human hepatocellular carcinoma surgical specimens

18F-FDG accumulation is reflected by GLUT1 expression in various tumors, including hepatocellular carcinoma (16, 26–28). To detect the K19 and GLUT1 expression in human hepatocellular carcinoma clinical samples, 98 surgically resected primary hepatocellular carcinoma tumors were subjected to immunohistochemical analysis. K19 and GLUT1 expression was observed in 14 of 98 (14%) cases and in 15 of 98 (15%) cases, respectively (Fig. 1A). Notably, there was a significant correlation between K19 expression and GLUT1 expression (Fig. 1B; P < 0.01). In K19+ hepatocellular carcinoma, the majority of the K19+ and GLUT1+ cells were found in the invasive front of hepatocellular carcinoma and in the K19-expressing zone, respectively. In contrast, GLUT1+ cells were distributed in various areas in K19 hepatocellular carcinoma.

The K19+ patients exhibited significantly lower RFS and OS rates, with the median RFS being 183 days for K19+ patients and 1,070 days for K19 patients, and the median OS being 464 days for K19+ patients and 2,112 days for K19 patients, respectively (Fig. 1C and D; Table 1). In contrast, no significant differences were detected between the RFS and OS rates of GLUT1+ and GLUT1 patients (RFS, P = 0.42; OS, P = 0.55; Fig. 1C and D; Table 1). The significance of K19 expression in predicting postoperative RFS and OS was confirmed by log-rank test and multivariate analyses (Table 1 and Supplementary Table S3). Specifically, a log-rank test revealed that K19 expression, high preoperative total bilirubin concentrations (>1.0 mg/dL), tumor size (≥5 cm), poor tumor differentiation, and microvascular invasion were associated with worse RFS rates (Supplementary Table S3). Regarding OS rates, K19 expression, high preoperative aspartate aminotransferase concentrations (≥40 IU/L), tumor size (≥5 cm), the presence of multiple tumors, poor tumor differentiation, and microvascular invasion were detected by log-rank test (Supplementary Table S3). Furthermore, multivariate analysis demonstrated that K19 expression was an independent predictor of both postoperative recurrence and lower OS rates (Table 1).

Efficacy of SUV, TNR, and preoperative factors for the evaluation of K19 expression in hepatocellular carcinoma

Our finding that K19 expression is significantly correlated with GLUT1 expression in hepatocellular carcinoma surgical specimens prompted us to examine whether SUV and TNR could be utilized to identify K19 expression in human hepatocellular carcinomas. First, consistent with the results of our previous study, (16) GLUT1+ patients exhibited significantly higher SUV and TNR than GLUT1 patients (data not shown). Subsequently, as shown in Fig. 2A, we found that SUV and TNR were significantly higher in K19+ (SUV, 7.6 ± 3.5; TNR, 2.8 ± 0.8) than in K19 patients (SUV, mean = 4.2 ± 2.2; TNR, mean = 1.5 ± 0.8; P < 0.01). However, no significant differences were observed between K19+ and K19 patients in preoperative α-fetoprotein (AFP) levels and the amounts of protein induced by vitamin K antagonists-II (PIVKA-II; Fig. 2B). Notably, ROC analysis revealed that among the preoperative clinical factors, including AFP and PIVKA-II, TNR was the most sensitive indicator of K19 expression in hepatocellular carcinoma tumors (Fig. 2C; Table 2).

EGFP marking of K19+ cell populations in heterogeneous hepatocellular carcinoma cell lines

To investigate the underlying mechanism of high FDG accumulation in K19+ hepatocellular carcinoma, we generated transgenic hepatocellular carcinoma cell lines to visualize K19+ cells, as described previously (12). Initial, our RT-PCR analyses demonstrated that the Huh7 and PLC/PRF/5 cell lines expressed K19, whereas HLF cells did not (Supplementary Fig. S3A). Histologic analyses demonstrated that Huh7 and PLC/PRF/5 cell lines contained both K19-expressing and nonexpressing cells. Subsequently, each of these three cell lines was transfected with the K19-EGFP reporter vector (Supplementary Fig. S3B), and double staining of K19 and GFP confirmed that our selective GFP-labeling method resulted in successful (efficiency of >95%) marking of K19-expressing cells (Supplementary Fig. S3C and S3D). As indicated by our FACS results, the ratio of K19+ cells differed among the cell lines. Specifically, 20.6% ± 3.9% of the Huh7 cells and 14.8% ± 2.7% of the PLC/PRF/5 cells (n = 3) exhibited EGFP expression (Supplementary Fig. S3E). PCR analyses confirmed that the sorted EGFP and EGFP+ cells contained equal copy numbers of the reporter gene (Supplementary Fig. S3F).

As we previously reported (12), FACS-isolated single K19+ cells generated both K19+ and K19 cell fractions during single-cell culture analyses. In contrast, single K19 cells produced only a K19 cell fraction (Supplementary Fig. S3G). Therefore, we further analyzed the K19+ and K19 populations derived from a single K19+ cell.

18F-FDG accumulation in K19+ and K19 hepatocellular carcinoma cells

To test the correlation between K19 expression and SUV/TNR in hepatocellular carcinoma patients, we first examined 18F-FDG uptake in K19+ and K19 hepatocellular carcinoma cells. The kinetics of 18F-FDG uptake in K19+ and K19 Huh7 cells indicated that 18F-FDG uptake was significantly higher in K19+ Huh7 cells than in K19 Huh7 cells (Fig. 3A). In addition, in K19+ Huh7 cells, siRNA-based K19 knockdown significantly decreased the 18F-FDG uptake (Fig. 3B). On the other hand, K19 overexpression in K19 Huh7 cells consistently and significantly increased the 18F-FDG uptake (Fig. 3C). Meanwhile, treatment with the TGFβR1 inhibitor LY2157299 resulted in significant suppression of FDG accumulation in K19+ Huh7 cells (Fig. 3D). Similar results were obtained in PLC/PRF/5 cells (Fig. 3A–D). These findings strongly suggest that K19 functions as a regulator of 18F-FDG accumulation in hepatocellular carcinoma cells and that evaluation of 18F-FDG accumulation might therefore be useful for monitoring the response of K19+ HCCs to TGFβR1 inhibitors.

K19 regulates GLUT1 expression through TGFβ/Smad signaling

To elucidate the mechanism underlying K19 and 18F-FDG uptake, we focused on TGFβ/Smad signaling, a signaling pathway involved in CSC maintenance. As we previously showed, TGFβ/Smad signaling is activated at the steady state and suppressed by K19 knockdown in K19+ HCC-CSCs (12). Meanwhile, previous studies demonstrated that the transcription factor Sp1, a downstream target of TGFβ/Smad signaling, binds and transactivates the GLUT1 promoter (29, 30). A separate study reported that the activation of TGFβ/Smad signaling plays a central role in glucose-induced cell hypertrophy in fibroblasts and epithelial cells (31). These notions prompted us to explore the mechanism governing the regulation of GLUT1 expression through TGFβ/Smad signaling in K19+ HCC-CSCs.

For these analyses, the Huh7 and PLC/PRF/5 cell lines were utilized due to their expression of K19, GLUT1, and Sp1 (Fig. 4A). Compared with the K19 population, Sp1 and GLUT1 were detected at significantly higher levels in K19+ Huh7 and PLC/PRF/5 cells (Fig. 4B). To assess whether K19 regulates GLUT1 expression through TGFβ/Smad signaling, we performed gain/loss of K19 function experiments. In K19+ Huh7 and PLC/PRF/5 cells, siRNA-mediated K19 knockdown resulted in decreased pSmad2, Sp1, and GLUT1 expression (Fig. 4C and D). Furthermore, K19 overexpression resulted in increased pSmad2, Sp1, and GLUT1 expression in K19 Huh7 and PLC/PRF/5 cells (Fig. 4C and D). On the basis of these results, we concluded that K19 regulates 18F-FDG accumulation through TGFβ/Smad signaling via a mechanism that includes Sp1 and GLUT1.

Despite advances in treatments, such as surgical resection, liver transplantation, RFA, and regional/systemic chemotherapy, hepatocellular carcinoma is associated with one of the poorest prognoses among carcinomas due to the difficulty in controlling tumor recurrence and metastasis. Therefore, the identification and clinical application of CSCs, which are involved in tumor recurrence/metastasis, contribute to the improvement of hepatocellular carcinoma treatment outcomes. We previously reported that K19 is a new HCC-CSC marker associated with TGFβ/Smad signaling and EMT and that K19+ HCC-CSCs could be new therapeutic targets for TGFβ receptor 1 inhibitors. Consistent with previous studies (32–34), K19 expression was found to comprise an independent prognosticator of poor hepatocellular carcinoma outcomes in both the current cohort, as well as those evaluated in our prior study. However, patients with K19+ hepatocellular carcinomas in this study exhibited poorer outcomes than those in our previous study, which is likely due to the correlation between K19 expression and tumor size/differentiation observed in the current, but not the prior analysis. Likewise, a separate study reported that K19 expression is a prognosticator of poor outcomes, is associated with tumor size/differentiation, and plays a key role in hepatocellular carcinoma invasion (35).For the clinical application of these highly malignant K19+ HCC-CSCs as a new therapeutic target, the development of tools for non/less-invasive assessment of K19 expression in hepatocellular carcinoma is essential.

In the clinical field of cancer treatment, 18F-FDG-PET has been established as a relatively noninvasive diagnostic tool for detecting/staging of cancers and for assessing the therapeutic efficacy of chemotherapies for many cancers. This image inspection technique can be used to evaluate glucose metabolism in vivo by measuring the uptake of the glucose analogue FDG. Besides, GLUT1 expression is known to be a major factor that affects FDG uptake (28). As for the role of 18F-FDG-PET in hepatocellular carcinoma, we previously found that although the frequency of GLUT1 expression is low, GLUT1 expression is correlated with elevated 18F-FDG accumulation in hepatocellular carcinoma and that hepatocellular carcinoma with high TNR values (≥2.0) exhibits more aggressive malignancy potential (16–18).

Notably, our analyses using human samples showed the relevance of K19 expression to 18F-FDG accumulation. By analyzing 98 hepatocellular carcinoma surgical specimens, we demonstrated that patients with K19+ tumors exhibited significantly higher SUV and TNR values and a higher incidence of GLUT1 expression than those with K19 tumors. Moreover, it should be noted that among the preoperative clinical factors, TNR is the most sensitive predictor of K19 expression in hepatocellular carcinoma. In the current study, TNR was more sensitive than SUV for the prediction of K19 expression. Differentiation between tumor tissue and physiologic noise in the liver by 18F-FDG-PET can sometimes be difficult, as normal liver tissue exhibits heterogeneity in 18F-FDG uptake, and physiologic noise is therefore apparent. Background nontumor liver uptake varied to some extent, and the contrast between the liver tumor and background liver noise may be dependent on the extent of tumor uptake itself as well as on the general levels of liver background. For these reasons, TNR would be expected to provide a better predictive value than SUV in this study. Besides, in this study, we included only patients with blood glucose levels <150 mg/dL and CRP levels <5.0 mg/dL for 18F-FDG-PET analysis. However, no significant differences in clinical–pathologic backgrounds, RFS, or OS were observed between patients of the current study and those of our previous study (data not shown), which indicates that there was no apparent selection bias in this study. In clinical setting, CT and/or MRI are routinely used for the diagnosis and monitoring of hepatocellular carcinoma. By combining 18F-FDG-PET with CT and/or MRI, a more precise prediction of K19 expression in hepatocellular carcinoma might be achieved.

On the other hand, given the pivotal functions of various signaling pathways in the maintenance and proliferation of embryonic stem/progenitor cells, including the Notch, Wnt/β-catenin, and TGFβ/Smad signaling pathways, it is feasible that such pathways also function in CSCs (36–38). Indeed, we previously demonstrated that K19+ HCC-CSCs, in which TGFβ/Smad signaling is active under steady-state conditions, exhibit high proliferative activity and a tendency to cause EMT (12). These findings highlight the necessity for further investigation into the mechanism by which TGFβ/Smad signaling contributes to the clinical application of K19+ CSCs in hepatocellular carcinoma. In the current study, we used K19 promoter–driven EGFP-marked cells to isolate K19+ populations of human hepatocellular carcinoma cell lines. Our findings demonstrate that K19+ cells exhibited significantly higher 18F-FDG uptake and significantly higher expression levels of Sp1, a downstream transcription factor of TGFβ/Smad signaling, and GLUT1. Moreover, our gain/loss of K19 function experiments clearly showed that K19 regulates 18F-FDG uptake through TGFβ/Smad signaling pathway and that this regulation occurred in an Sp1/GLUT1–dependent manner. We also observed that FDG accumulation was significantly suppressed in K19+ cells treated with the TGFβR1 inhibitor LY2157299, suggesting the potential clinical application for monitoring the response of this compound. Besides, among the human hepatocellular carcinoma specimens examined in this study, less than half of the GLUT1+ tumors exhibited K19 expression. GLUT1 expression was previously reported to be regulated by Notch signaling in breast cancer (39, 40). Given the significance of Notch signaling in hepatocellular carcinoma (41), it is conceivable that this signaling pathway is also involved in the regulation of GLUT1 expression in these cells.

Various studies regarding the role of K19 in hepatocellular carcinoma show the aggressive properties of K19+ hepatocellular carcinoma and prompt us to develop new therapeutic strategies for it. Accompanied with the efficacy of TGFβ receptor 1 inhibitor against K19+ HCC-CSCs, detecting K19 expression using 18F-FDG-PET study before hepatocellular carcinoma treatment aids to enable individualized medicine for K19+ hepatocellular carcinoma patients. Further prospective clinical trials examining whether 18F-FDG-PET comprises an effective method for identifying K19+ hepatocellular carcinoma patients, and for assessing the therapeutic efficacy of TGFβ receptor 1 inhibitors in K19+ hepatocellular carcinoma patients, will advance the clinical application of K19+ HCC-CSCs.

In conclusion, the results of our study indicate that K19 regulates 18F-FDG uptake in an Sp1/GLUT1–dependent manner through the TGFβ/Smad signaling pathway and that 18F-FDG-PET is an effective method for identifying K19 expression in hepatocellular carcinoma tissues. We believe that further study of K19+ HCC-CSCs and 18F-FDG-PET will contribute to the development of novel approaches for the treatment of hepatocellular carcinoma.

No potential conflicts of interest were disclosed.

Conception and design: T. Kawai, K. Yasuchika, T. Higashi, Y. Miyauchi, S. Ogiso

Development of methodology: T. Kawai, K. Yasuchika, T. Higashi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Kawai, S. Seo, K. Yasuda, E. Hatano

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T. Kawai, K. Yasuchika, T. Ishii, S. Ogiso, K. Fukumitsu

Writing, review, and/or revision of the manuscript: T. Kawai, K. Yasuchika, S. Ogiso, E. Hatano

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Higashi, Y. Miyauchi, H. Kojima, R. Yamaoka, H. Katayama, E.Y. Yoshitoshi, S. Ogiso, S. Kita, K. Fukumitsu, Y. Nakamoto

Study supervision: K. Yasuchika, T. Ishii, Y. Miyauchi, H. Kojima, R. Yamaoka, H. Katayama, E.Y. Yoshitoshi, S. Ogiso, K. Fukumitsu, E. Hatano, S. Uemoto

The authors wish to thank Dr. Makiko Kagaya and Professor Shuichi Kaneko (Kanazawa University, Kanazawa, Japan) for providing the plasmid vector pHCK-2952.

This work was supported by grants from the Scientific Research Fund of the Japan Science and Technology Agency (research project number: 24659606).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Llovet
JM
,
Zucman-Rossi
J
,
Pikarsky
E
,
Sangro
B
,
Schwartz
M
,
Sherman
M
, et al
Hepatocellular carcinoma
.
Nat Rev Dis Primers
2016
;
2
:
16018
.
2.
Clarke
MF
,
Dick
JE
,
Dirks
PB
,
Eaves
CJ
,
Jamieson
CH
,
Jones
DL
, et al
Cancer stem cells–perspectives on current status and future directions: AACR Workshop on cancer stem cells
.
Cancer Res
2006
;
66
:
9339
44
.
3.
Pardal
R
,
Clarke
MF
,
Morrison
SJ
. 
Applying the principles of stem-cell biology to cancer
.
Nat Rev Cancer
2003
;
3
:
895
902
.
4.
Dean
M
,
Fojo
T
,
Bates
S
. 
Tumour stem cells and drug resistance
.
Nat Rev Cancer
2005
;
5
:
275
84
.
5.
Phillips
TM
,
McBride
WH
,
Pajonk
F
. 
The response of CD24(-/low)/CD44+ breast cancer-initiating cells to radiation
.
J Natl Cancer Inst
2006
;
98
:
1777
85
.
6.
Pattabiraman
DR
,
Weinberg
RA
. 
Tackling the cancer stem cells - what challenges do they pose?
Nat Rev Drug Discov
2014
;
13
:
497
512
.
7.
Yamashita
T
,
Ji
J
,
Budhu
A
,
Forgues
M
,
Yang
W
,
Wang
HY
, et al
EpCAM-positive hepatocellular carcinoma cells are tumor-initiating cells with stem/progenitor cell features
.
Gastroenterology
2009
;
136
:
1012
24
.
8.
Ma
S
,
Chan
KW
,
Hu
L
,
Lee
TK
,
Wo
JY
,
Ng
IO
, et al
Identification and characterization of tumorigenic liver cancer stem/progenitor cells
.
Gastroenterology
2007
;
132
:
2542
56
.
9.
Ma
S
,
Tang
KH
,
Chan
YP
,
Lee
TK
,
Kwan
PS
,
Castilho
A
, et al
miR-130b Promotes CD133(+) liver tumor-initiating cell growth and self-renewal via tumor protein 53-induced nuclear protein 1
.
Cell Stem Cell
2010
;
7
:
694
707
.
10.
Yang
ZF
,
Ho
DW
,
Ng
MN
,
Lau
CK
,
Yu
WC
,
Ngai
P
, et al
Significance of CD90+ cancer stem cells in human liver cancer
.
Cancer Cell
2008
;
13
:
153
66
.
11.
Oikawa
T
,
Kamiya
A
,
Zeniya
M
,
Chikada
H
,
Hyuck
AD
,
Yamazaki
Y
, et al
Sal-like protein 4 (SALL4), a stem cell biomarker in liver cancers
.
Hepatology
2013
;
57
:
1469
83
.
12.
Kawai
T
,
Yasuchika
K
,
Ishii
T
,
Katayama
H
,
Yoshitoshi
EY
,
Ogiso
S
, et al
Keratin 19, a cancer stem cell marker in human hepatocellular carcinoma
.
Clin Cancer Res
2015
;
21
:
3081
91
.
13.
Rigo
P
,
Paulus
P
,
Kaschten
BJ
,
Hustinx
R
,
Bury
T
,
Jerusalem
G
, et al
Oncological applications of positron emission tomography with fluorine-18 fluorodeoxyglucose
.
Eur J Nucl Med
1996
;
23
:
1641
74
.
14.
Stokkel
MP
,
ten Broek
FW
,
van Rijk
PP
. 
The role of FDG PET in the clinical management of head and neck cancer
.
Oral Oncol
1998
;
34
:
466
71
.
15.
Eary
JF
,
Conrad
EU
. 
Positron emission tomography in grading soft tissue sarcomas
.
Semin Musculoskelet Radiol
1999
;
3
:
135
8
.
16.
Seo
S
,
Hatano
E
,
Higashi
T
,
Nakajima
A
,
Nakamoto
Y
,
Tada
M
, et al
P-glycoprotein expression affects 18F-fluorodeoxyglucose accumulation in hepatocellular carcinoma in vivo and in vitro
.
Int J Oncol
2009
;
34
:
1303
12
.
17.
Seo
S
,
Hatano
E
,
Higashi
T
,
Hara
T
,
Tada
M
,
Tamaki
N
, et al
Fluorine-18 fluorodeoxyglucose positron emission tomography predicts tumor differentiation, P-glycoprotein expression, and outcome after resection in hepatocellular carcinoma
.
Clin Cancer Res
2007
;
13
:
427
33
.
18.
Kitamura
K
,
Hatano
E
,
Higashi
T
,
Seo
S
,
Nakamoto
Y
,
Yamanaka
K
, et al
Preoperative FDG-PET predicts recurrence patterns in hepatocellular carcinoma
.
Ann Surg Oncol
2012
;
19
:
156
62
.
19.
Ishii
T
,
Yasuchika
K
,
Fujii
H
,
Hoppo
T
,
Baba
S
,
Naito
M
, et al
In vitro differentiation and maturation of mouse embryonic stem cells into hepatocytes
.
Exp Cell Res
2005
;
309
:
68
77
.
20.
Ishii
T
,
Yasuchika
K
,
Machimoto
T
,
Kamo
N
,
Komori
J
,
Konishi
S
, et al
Transplantation of embryonic stem cell-derived endodermal cells into mice with induced lethal liver damage
.
Stem Cells
2007
;
25
:
3252
60
.
21.
Kagaya
M
,
Kaneko
S
,
Ohno
H
,
Inamura
K
,
Kobayashi
K
. 
Cloning and characterization of the 5′-flanking region of human cytokeratin 19 gene in human cholangiocarcinoma cell line
.
J Hepatol
2001
;
35
:
504
11
.
22.
Ishii
T
,
Yasuchika
K
,
Suemori
H
,
Nakatsuji
N
,
Ikai
I
,
Uemoto
S
. 
Alpha-fetoprotein producing cells act as cancer progenitor cells in human cholangiocarcinoma
.
Cancer Lett
2010
;
294
:
25
34
.
23.
Sasaki
N
,
Ishii
T
,
Kamimura
R
,
Kajiwara
M
,
Machimoto
T
,
Nakatsuji
N
, et al
Alpha-fetoprotein-producing pancreatic cancer cells possess cancer stem cell characteristics
.
Cancer Lett
2011
;
308
:
152
61
.
24.
Ishii
T
,
Fukumitsu
K
,
Yasuchika
K
,
Adachi
K
,
Kawase
E
,
Suemori
H
, et al
Effects of extracellular matrixes and growth factors on the hepatic differentiation of human embryonic stem cells
.
Am J Physiol Gastrointest Liver Physiol
2008
;
295
:
G313
21
.
25.
Yamanaka
K
,
Hatano
E
,
Narita
M
,
Kitamura
K
,
Yanagida
A
,
Asechi
H
, et al
Olprinone attenuates excessive shear stress through up-regulation of endothelial nitric oxide synthase in a rat excessive hepatectomy model
.
Liver Transpl
2011
;
17
:
60
9
.
26.
Pedersen
MW
,
Holm
S
,
Lund
EL
,
Højgaard
L
,
Kristjansen
PE
. 
Coregulation of glucose uptake and vascular endothelial growth factor (VEGF) in two small-cell lung cancer (SCLC) sublines in vivo and in vitro
.
Neoplasia
2001
;
3
:
80
7
.
27.
Brown
RS
,
Goodman
TM
,
Zasadny
KR
,
Greenson
JK
,
Wahl
RL
. 
Expression of hexokinase II and Glut-1 in untreated human breast cancer
.
Nucl Med Biol
2002
;
29
:
443
53
.
28.
Higashi
T
,
Saga
T
,
Nakamoto
Y
,
Ishimori
T
,
Mamede
MH
,
Wada
M
, et al
Relationship between retention index in dual-phase (18)F-FDG PET, and hexokinase-II and glucose transporter-1 expression in pancreatic cancer
.
J Nucl Med
2002
;
43
:
173
80
.
29.
Santalucía
T
,
Boheler
KR
,
Brand
NJ
,
Sahye
U
,
Fandos
C
,
Viñals
F
, et al
Factors involved in GLUT-1 glucose transporter gene transcription in cardiac muscle
.
J Biol Chem
1999
;
274
:
17626
34
.
30.
Viñals
F
,
Fandos
C
,
Santalucia
T
,
Ferré
J
,
Testar
X
,
Palacín
M
, et al
Myogenesis and MyoD down-regulate Sp1. A mechanism for the repression of GLUT1 during muscle cell differentiation
.
J Biol Chem
1997
;
272
:
12913
21
.
31.
Wu
L
,
Derynck
R
. 
Essential role of TGF-beta signaling in glucose-induced cell hypertrophy
.
Dev Cell
2009
;
17
:
35
48
.
32.
Uenishi
T
,
Kubo
S
,
Yamamoto
T
,
Shuto
T
,
Ogawa
M
,
Tanaka
H
, et al
Cytokeratin 19 expression in hepatocellular carcinoma predicts early postoperative recurrence
.
Cancer Sci
2003
;
94
:
851
7
.
33.
Durnez
A
,
Verslype
C
,
Nevens
F
,
Fevery
J
,
Aerts
R
,
Pirenne
J
, et al
The clinicopathological and prognostic relevance of cytokeratin 7 and 19 expression in hepatocellular carcinoma. A possible progenitor cell origin
.
Histopathology
2006
;
49
:
138
51
.
34.
Kim
H
,
Choi
GH
,
Na
DC
,
Ahn
EY
,
Kim
GI
,
Lee
JE
, et al
Human hepatocellular carcinomas with "Stemness"-related marker expression: keratin 19 expression and a poor prognosis
.
Hepatology
2011
;
54
:
1707
17
.
35.
Govaere
O
,
Komuta
M
,
Berkers
J
,
Spee
B
,
Janssen
C
,
de Luca
F
, et al
Keratin 19: a key role player in the invasion of human hepatocellular carcinomas
.
Gut
2014
;
63
:
674
85
.
36.
Rizzo
P
,
Osipo
C
,
Foreman
K
,
Golde
T
,
Osborne
B
,
Miele
L
. 
Rational targeting of Notch signaling in cancer
.
Oncogene
2008
;
27
:
5124
31
.
37.
Klaus
A
,
Birchmeier
W
. 
Wnt signalling and its impact on development and cancer
.
Nat Rev Cancer
2008
;
8
:
387
98
.
38.
Mishra
L
,
Shetty
K
,
Tang
Y
,
Stuart
A
,
Byers
SW
. 
The role of TGF-beta and Wnt signaling in gastrointestinal stem cells and cancer
.
Oncogene
2005
;
24
:
5775
89
.
39.
Efferson
CL
,
Winkelmann
CT
,
Ware
C
,
Sullivan
T
,
Giampaoli
S
,
Tammam
J
, et al
Downregulation of Notch pathway by a gamma-secretase inhibitor attenuates AKT/mammalian target of rapamycin signaling and glucose uptake in an ERBB2 transgenic breast cancer model
.
Cancer Res
2010
;
70
:
2476
84
.
40.
Landor
SK
,
Mutvei
AP
,
Mamaeva
V
,
Jin
S
,
Busk
M
,
Borra
R
, et al
Hypo- and hyperactivated Notch signaling induce a glycolytic switch through distinct mechanisms
.
Proc Natl Acad Sci U S A
2011
;
108
:
18814
9
.
41.
Villanueva
A
,
Alsinet
C
,
Yanger
K
,
Hoshida
Y
,
Zong
Y
,
Toffanin
S
, et al
Notch signaling is activated in human hepatocellular carcinoma and induces tumor formation in mice
.
Gastroenterology
2012
;
143
:
1660
9.e7
.