Abstract
Background/Aim: Recently, vessels encapsulating tumor clusters (VETC) pattern and macrotrabecular massive (MTM) pattern of hepatocellular carcinoma (HCC) have been reported as aggressive histological types. These histological patterns showed an immunosuppressive tumor immune microenvironment (TIME). Since there have been no reports on the differences of these two subtypes simultaneously, this study examined the immunophenotypes and TIME of MTM-HCC and VETC-HCC immunohistochemically. Patients and Methods: Seventy-four cases of previously diagnosed HCC, including 32 MTM-HCCs, 21 VETC-HCCs, and 21 conventional HCCs, were enrolled in immunohistochemical analysis. We conducted immunohistochemical analysis. Results: We found that MTM-HCC showed less frequent expression of HepPar-1, which is one of the most common hepatocytic markers. In MTM-HCC, the frequency of high expression levels of Keratin19, carbonic anhydrase (CA) IX, and PD-L1 was higher compared to VETC-HCC and conventional HCC. PD-L1 expression was found in 34.4% of MTM-HCC, 0% of VETC-HCC, and 19.0% of conventional HCC. The rate of PD-L1 expression in MTM-HCC was significantly higher than the others (p=0.0015). PD-L1 expression was significantly associated with epithelial cell adhesion molecules and CA IX expression, which are representative markers of tumor stemness and hypoxic conditions, respectively. The CD8 infiltration in VETC-HCC was significantly lower than that in conventional HCC. Conclusion: MTM-HCC had different immunophenotypes and TIMEs compared to HCC with the VETC pattern. Although both had immunosuppressive TIME, the elements forming TIME were quite different. To enhance the immune checkpoint inhibitor efficacy, changing TIME from a suppressive to an active form is essential.
- Macrotrabecular hepatocellular carcinoma
- vessels encapsulating tumor clusters hepatocellular carcinoma
- immunophenotypes
- tumor immune microenvironment
Primary liver cancer was the sixth most common cancer and the third leading cause of cancer-related death worldwide in 2020 (1). Liver cancer remains a global health challenge and its incidence is growing worldwide. It is estimated that more than one million individuals will be affected by liver cancer annually by 2025 (2, 3).
In patients with unresectable hepatocellular carcinoma (HCC), the combination therapy with atezolizumab and bevacizumab has become the first-line treatment regime for unresectable HCC because it has resulted in better overall and progression-free survival outcomes than sorafenib (4). The combination of immune-checkpoint inhibitors with anti-angiogenic antibodies or tyrosine kinase inhibitors can drive immune cell infiltration into ‘cold’ tumors, thereby converting them to ‘hot’ tumors and increasing response rates (5). To obtain optimal efficacies for these treatments, assessment of the tumor immune microenvironment (TIME) is required. In some solid tumors, such as lung, uterine cervical, breast, and head and neck cancers, the assessment of TIME is routinely conducted by immunohistochemical staining for programed death-1 (PD-1) or PD-L1 (6-9). However, in HCC, no assessments have been conducted for b-catenin mutated/activated HCC and non-viral associated HCC is known to show resistance to immune check point inhibitors (9-11).
Recently, vessels encapsulating tumor clusters (VETC) pattern has been reported as a predictor of sorafenib benefit in patients with HCC (12) and is associated with unfavorable prognosis (12, 13). HCCs with the VETC pattern are often accompanied by the macrotrabecular massive (MTM) pattern, which is also known as an aggressive histological type (13, 14). We previously reported that HCCs with MTM pattern and VETC pattern had significant numbers of the VETC pattern and the MTM pattern, respectively (14). Moreover, these histological patterns showed high PD-L1 expression and immunosuppressive TIME (15, 16).
Due to their morphological characteristics, the VETC pattern and MTM pattern are easy to recognize. However, diagnosis of HCC is expected to be difficult with the limited samples obtained from biopsies. To obtain correct diagnoses, pathologists must be able to recognize the immunophenotype of these aggressive HCC subtypes.
Although MTM and VETC patterns were often observed within the same nodule (13, 14), there are no reports on the differences between these two subtypes simultaneously. We examined the immunophenotypes and TIME of HCC with the MTM pattern and VETC pattern immunohistochemically in this study.
Patients and Methods
Patients. Seventy-four cases of previously diagnosed HCC were enrolled in this study. All cases were resected at Kurume University Hospital between 2006 and 2015. We randomly selected 32 HCCs with the MTM pattern and 21 HCCs with the VETC pattern based on our previous study (14). Twenty-one conventional HCCs were randomly selected. Pathological assessment was conducted by two pathologists (JA and MN). Clinicopathological findings are shown in Table I. This study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of Kurume University (approval #19017).
Clinicopathological characteristics of 74 hepatocellular carcinomas (HCCs).
Tissue preparation and pathological assessment. Liver specimens were fixed in 10% formalin and embedded in paraffin. We cut 4-μm consecutive sections and stained them with hematoxylin and eosin. When a tumor nodule consisted of plural histological grades, the highest grade was adopted. For pathological diagnosis, we followed the WHO classification (17). The MTM pattern and the VETC pattern were defined following previous reports (13, 14). We conducted IHC using TMA sections, following the procedure of our previous study. Each slide contained 22 cores. We punctured paraffin-embedded samples in two cores from each case. Briefly, in HCC with the MTM pattern, tumor cells proliferated with thick trabecula consisting of more than six cells. In HCC with the VETC pattern, the tumor was composed of small tumor clusters covered with endothelial cells that are easily detectable morphologically.
Immunohistochemical staining. IHC was performed using the following antibodies: HepPar-1 (OCH1E5, Agilent Technologies, Inc., Santa Clara, CA, USA), arginase-1 (Sigma-Aldrich, St. Louis, MO, USA), glypican-3 (GPC3) (clone 1G12, Nichirei Bioscience Inc., Tokyo, Japan), Keratin (K) 19 (clone RCK108, Dako; Agilent Technologies, Inc.), epithelial cell adhesion molecule (EpCAM, clone VU1D9, Cell Signaling Technology, Danvers, MA, USA), carbonic anhydrase (CA) IX (H-11, Santa Cruz Biotechnology, Dallas, TX, USA), PD-L1 (clone E1L3N, Cell Signaling Technology,), and CD8 (clone 4B11, Leica Biosystems, Newcastle, UK).
The expression of PD-L1 was evaluated in conjunction with tumoral and stromal expression. Based on previous reports, the cut-off value for PD-L1 expression was set as 1% (18, 19). CD8-positive tumor infiltrating lymphocytes were also evaluated. These cells were manually counted in five high-power fields of view (magnification: ×400) and the average number of cells per mm2 was determined. Cut-off values were determined as >1% for K19 and EpCAM and were >50% for HepPar-1, arginase-1, GPC3, and CA IX. IHC staining for HepPar-1, K19, CD8, and PD-L1 was performed using BOND-III (Leica Microsystems, Newcastle, UK). IHC for the others was conducted using BenchMark ULTRA (Roche Diagnostics, Tucson, AZ, USA).
Statistical analysis. Intergroup differences in categorical variables and IHC results were examined using the Chi-square test or Fisher’s exact test. Differences of p<0.05 were considered significant. All analyses were conducted using JMP version 16 software (SAS Institute Inc., Cary, NC, USA).
Results
Clinicopathological findings. Patients with HCC with the MTM pattern were significantly younger those of the others (p=0.0431). Tumor markers, such as α-fetoprotein (AFP) and des-γ-carboxy prothrombin (DCP), of the MTM pattern and VETC pattern were significantly higher than those of conventional HCC (p<0.0001 and p=0.0392, respectively). Tumor size of the MTM pattern and VETC pattern HCC was significantly larger than that of conventional HCC (p<0.001). HCC with the MTM pattern and VETC pattern showed significantly high histological grades (p<0.001) and high frequency of portal vein invasion (p=0.0054).
Immunohistochemical staining of HCCs. The results of IHC are summarized in Table II. The number of cases with HepPar-1 high expression in the MTM pattern was significantly lower than that in the VETC pattern and conventional HCC (p<0.0001). There were no significant differences among the three groups in arginase-1 expression although that expression in the MTM pattern tended to be lower compared to the others. K19 expression was found only in the MTM pattern and not found in the VETC pattern and conventional HCC (p=0.0301). The expression of CA IX and PD-L1 was found in the MTM pattern and conventional HCC and was not found in the VETC pattern (p=0.0064 and 0.0015, respectively). There were no significant differences in Glypican-3 and EpCAM expression among the groups. The VETC pattern of HCC exhibited a significantly lower CD8 infiltration compared to conventional HCC (Figure 1). No significant differences were found between the others.
The association between hepatocellular carcinoma (HCC) subtypes and hepatobiliary markers.
Representative microphotographs of the hepatocellular carcinoma (HCC)-macrotrabecular massive pattern, HCC-vessels encapsulating tumor clusters pattern, and conventional HCC are shown in the upper, middle, and lower arrows, respectively. The microphotographs in the same row originate from the same case. Scale bar indicates 100 mm.
Correlation between PD-L1 expression and several hepatobiliary markers. The results of correlation between PD-L1 expression in the tumor and stroma and hepatobiliary markers are summarized in Table III. PD-L1 expression was significantly correlated with low HepPar-1, high EpCAM, and high CA IX expression (p=0.0305, p=0.0224, and p=0.0033, respectively).
Correlation between PD-L1 expression and several hepatobiliary markers.
Discussion
In this study, we found that HCC with the MTM pattern showed less frequent expression of HepPar-1, which is one of the most common hepatocytic markers. Moreover, in MTM-HCC, the frequency of high expressions of K19, CA IX, and PD-L1 was higher compared to VETC-HCC and conventional HCC. PD-L1 expression was associated with tumor stemness and hypoxic conditions. In addition, HCC with the MTM pattern and VETC pattern had different immunophenotypes and TIME.
Clinicopathological findings of HCC with the MTM pattern and VETC pattern are mostly comparable with our previous study (14). Briefly, HCC with the MTM pattern and VETC pattern demonstrated aggressive features, such as high levels of tumor markers, large tumor size, less differentiation, and frequent portal vein invasion. Although this cohort was composed of a small number of cases, the accompanying biological characteristics of these subtypes were similar with previous reports (12-14).
In HCC with the MTM pattern, the representative hepatocytic marker HepPar-1 showed significantly low expression and arginase-1 expression tended to be lower compared to HCC with the VETC pattern and conventional HCC. To date, there have been no reports on hepatocytic markers in HCC with the MTM pattern. However, these results are associated with the fact that HepPar-1 and arginase-1 expression are decreased with less differentiation (20) as most HCCs with the MTM pattern show poorly differentiated HCC. In surgical specimens, pathological diagnosis of HCC with the MTM pattern is easy. However, HCCs with the MTM pattern often showed different imaging patterns compared to conventional HCCs (21, 22). Therefore, these cases may undergo liver biopsy. When a characteristic trabecular component is not observed in a tiny sample, pathologists may be unsure of the diagnosis for HCCs and conduct IHC for hepatocytic markers as well as biliary markers. In this study, significant expression of K19, a representative biliary marker, was found only in HCC with the MTM pattern. In this situation, pathologists should not misinterpret HCC with the MTM pattern as poorly differentiated cholangiocarcinoma. It is well known that a small subset of HCC has K19 expression and shows poor prognosis (23). In addition, the prevalence of CA IX expression was notably higher in HCC exhibiting the MTM pattern when compared to other subtypes, in line with findings from a prior study (24). CA IX is not specific to HCCs and its expression is observed in various carcinomas, inducing hypoxic conditions (25-28). In pathological diagnosis, morphological and immunophenotypic findings as well as high levels of tumor markers can be useful.
In this study, to examine TIME of the MTM pattern and the VETC pattern, we conducted IHC using TMA sections. The MTM pattern showed significantly high expression of PD-L1 in both tumor and stromal cells. PD-L1 expression was found in 34.4% of HCC with the MTM pattern. A previous study by Liu et al. reported that PD-L1 expression was found in 43% of HCC with the MTM pattern and this rate was significantly higher than that of the non-MTM pattern HCC (16). In our study, no HCCs with the VETC pattern expressed tumoral PD-L1. Itoh et al. reported that the VETC pattern was associated with a significantly high rate of tumoral PD-L1 positivity in HCC tissues (38.1%) compared with non-VETC (8.3%) (15). On the other hand, Kurebayashi et al. demonstrated that most HCCs-MTM and VETC patterns belonged to immune-low/angiogenic subtypes and the expression of PDCD1LG1 and PDCD1LG2, which encode interferon-g-inducible immune checkpoint molecules, PD-L1 and PD-L2, respectively, was also down-regulated in immune-low/angiogenic subtypes (29). This difference may be attributed to the different cohort and different analytical methods. However, the underlying causes are unknown. HCC with the MTM pattern and the VETC pattern could have different tumoral microenvironments. Regarding the correlation between PD-L1 expression and other hepatic markers, low HepPar-1 expression, high EpCAM expression, and high CA IX expression are significantly associated. Higher histological grade HCC is known to frequently demonstrate lower HepPar-1 expression and showed significantly higher PD-L1 expression (30). EpCAM is a representative biliary marker as well as a cancer stem cell (CSC)/hepatic progenitor marker in HCC. Also, CA IX is a sensitive biomarker of hypoxia. Hypoxia can enhance HCC stemness (31). Liver CSCs contribute to the evasion of immunosurveillance by altering the phenotype of dendritic cells and impairing their recruitment (32, 33). Subsequently, in cases with high expression of EpCAM and CA IX, it is presumed that elevated PD-L1 expression was induced to evade immune surveillance.
Study limitations. First, the sample size used in this study was small. Second, we conducted the immunohistochemical study using TMA sections. Although we revealed that the MTM pattern and VETC pattern had different immunophenotypes and TIME, we were unable to verify the heterogeneity of TIME within the same nodule. To solve these problems, research with a large-scale sample and whole slide sections should be conducted.
Conclusion
We demonstrated that HCC with the MTM pattern had different immunophenotypes and TIMEs compared to HCC with the VETC pattern. Although both of them had immunosuppressive TIME as shown in previous reports, the elements forming TIME were quite different. To enhance immune checkpoint inhibitor efficacy, changing TIME from suppressive type to active type is essential. Therefore, conducting combination therapy using atezolizumab and bevacizumab is suitable for HCC with the MTM pattern and VETC pattern.
Acknowledgements
The Authors would like to thank Ms. Akiko Tanaka for her excellent preparation of sections.
Footnotes
Authors’ Contributions
JA designed this study and drafted the manuscript. JA, MN, RK, HK, and SO conducted immunohistochemical staining and assessed the results. JA, MN, YM, MT, KT, YY, and DK participated in pathological diagnosis. JA, ST, TM, and HN constructed tissue microarray sections. HS and TH acquired and collected the samples. SS and HY edited the manuscript.
Conflicts of Interest
The have no conflicts of interest to declare in relation to this study.
- Received November 15, 2023.
- Revision received December 13, 2023.
- Accepted December 18, 2023.
- Copyright © 2024 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
