In the field of liver cancer research, a number of genetic and genome-wide profiling studies have detailed the genomic landscape of HCC and advanced the understanding of the underlying mechanisms involved in the multi-step hepatocarcinogenesis. Deregulated, non-resolving inflammation is thought to be central to the development and progression of HCC: emerging evidence has clarified that HCC-associated inflammation can be driven by genetic events promoting the secretion of factors guiding the build-up of an inflammatory microenvironment (intrinsic pathway) and by exogenous factors able to maintain a persistent inflammatory response (extrinsic pathway) (PMID 29872724). In this study, we have employed microarrays to profile the specific gene expression signature in circulating PBMCs of patients with HCC that could be associated with tumor progression and different response to treatment.
Results have been analyzed in relation to two main factors: the degree of response to treatment (complete or partial) and time of analysis of genes expression: before treatment (time 0) and three months after treatment (time 1).
From the analysis of the results, it emerged first that at time 0 there was no significant difference between the groups of subjects at RTqPCR level on circulating PBMCs, except for two genes: ENTPD1 and ICOS, respectively over-expressed and under-expressed at baseline in patients with partial response to ablative treatment.
The ENTPD1 gene (nucleoside triphosphate diphosphohydrolase-1) encodes an enzyme expressed on many cell types including leukocytes, endothelial cells, Kuppfer cells, regulatory T lymphocytes [23–25]. The enzyme catalyzes the degradation to AMP of extracellular ATP, which plays an anti-tumor role by activating the immune response against the tumor and directly in the death of neoplastic cells. Furthermore, the gene expression at the level of the regulatory T lymphocytes inhibits the antitumor activity of natural killer cells and promotes tumor growth, while the expression at the endothelial cell level facilitates angiogenesis [26].
Over-expression of ENTPD1 has been demonstrated in several cancers including melanoma [27], leukemia [28], pancreatic adenocarcinoma [29], ovarian carcinoma [30], while in colorectal adenocarcinoma lower expression levels of gene have been correlated to a lower metastatic tumor power and a longer survival [31].
In a study on 324 patients underwent to surgical resection for HCC, Cai et al have demonstrated that the expression of the ENTPD1 gene at the level of tumor cells and normal hepatocyte cell lines is an independent predictor of worse prognosis after resection of the tumor [32]. According to the data of the literature, in our study the gene is significantly upregulated at T0 in patients who reached the partial response at three months from treatment, thus delineating a prognostically unfavorable gene expression profile in terms of response to treatment. This novel data seems to confirm the observation of Bastid et al [33] that CD39/ENTPD1 inhibition may have anticancer activity and that ENTPD1 inhibitory compounds have a possible favorable safety profile.
The ICOS gene (Inducible T Cell Costimulator), is a CD28 family receptor, which modulates costimulatory signals necessary to fully activate lymphocyte T cells. It is expressed on T cells and, interacting with its ICOSL ligand, present on APCs (antigen presenting cells, such as B cells, macrophages and dendritic cells) and on some tumor cells, it stimulates T lymphocyte activity. The role of ICOS in oncogenesis is twofold, because, depending on the context, it can stimulate the CD8 + T lymphocytes and their antitumor action or activate the T-regulators, favoring the immune escape [34].
The ICOS + T-Regs are the predominant component in the microenvironment of HCC, in fact they are much more present, both in the liver and in the plasma, of affected subjects, compared to controls and, repressing the immune response against the tumor, correlate with a worse prognosis [35].
Studies on HepG2 cells show that the ICOS-ICOSL pathway is fundamental for neoplastic progression (if silenced, cell proliferation and invasion are inhibited) and its action is mediated by the PI3K pathway [36]. Similar results have also been confirmed on other tumors (breast, stomach, ovary) [37–39]. In some tumors, such as the colon, there are opposite results because CD4 + ICOS + promote the action of Th1 which inhibit tumor invasion and stimulate the CD8 + response against neoplastic cells [40].
In an apparently contrasting way to these data, in our study ICOS was statistically significantly downregulated at T0 in patients who then reported RP to treatment. However, it has been shown that there is a different distribution of Treg lymphocytes, which are significantly more concentrated at the tumor level than in the peritumoral tissue and in the peripheral blood. Therefore our result, which deserves to be reconsidered on a larger sample of subjects, can be justified, at least in part, by the fact that this cellular population is more represented at the tumor level, thus constituting a smaller proportion of the cells that represent the PBMCs whose gene expression has been assessed [41].
Apart from the significant exceptions of ENTPD1 and ICOS genes, the gene expression at RTqPCR of patients' PBMCs is not modified at time 0 and therefore we can assume that any difference in gene expression is due exclusively to the effect of the treatment that is the local consequences on the tumor microenvironment of the ablation of the tumor.
To the best of our knowledge, there is no study in literature in which the effect of treatment of HCC on the genomic expression in circulating PBMCs was evaluated.
A first consideration is that the treatment of the tumor can determine significant changes in peripheral gene expression. This result is even more surprising considering that it has also been detected in patients in whom the volume of the neoplasm is relatively small (average volume of 11.6 ml in patients who have obtained a complete response).
We have documented variation of expression subsequent to treatment of six genes.
The CCDC88C gene, coding for the Daple protein, which belongs to the family of non-receptor guanine nucleotide exchange factors (GEF), acts on G1α by increasing the non-canonical way of Wnt, implicated in the modulation of cell motility [42]. It acts as an oncosuppressor in normal epithelial cells, while it is upregulated in tumor cells, which acquire particular invasive capacity. Gene expression levels have been shown to be increased in colorectal cancer and, to a greater extent, in hepatic and pulmonary metastases, emphasizing the role of GEF in tumor progression [43].
In our study the gene is downregulated at T1 in patients with complete response to treatment, confirming the fact that tumor ablation is accompanied by the down regulation of pathways that contribute to carcinogenesis, such as that of GEF family.
The CORO 7 gene codes for coronin 7 protein, which belongs to the family of coronins, membrane proteins that have different roles in maintenance of cellular structure, in cell-cell and matrix-cell interactions, in cytoskeletal organization, in chemotaxis and in cellular migration. The composition and characteristics of cell membrane proteins, including coronins, undergo rearrangement during malignant transformation. For these reasons these proteins can be considered tumor biomarkers. Wu et al. analyzed the expression profile of coronin-1C in two HCC cell lines in mouse: a higher level of protein expression correlates with a more invasive neoplastic phenotype and with a higher frequency of pulmonary metastases. In humans, immunohistochemical studies of biopsy samples of HCC have shown that patients with higher expression levels of coronin-1 exhibited a more advanced state of disease [44]. Wang et al. demonstrated the overexpression of coronin-1C protein in HCC samples in comparison to the surrounding healthy parenchyma; moreover expression levels correlates with the proliferation and metastatic potential of HCC cells, through activation of Rac-1 [45]. Gao et al. have shown overexpression of coronin 3 in hepatocellular cells [46]. In our study, confirming the data reported in the literature, gene expression levels are downregulated in patients with complete response to treatment, while they do not vary in patients with stable disease.
The LTB4R gene codes for the Leukotriene B4 receptor, a pro-inflammatory eicosanoid derived from arachidonic acid, which acts as a chemoattractant for phagocytes. It is precociously up-regulated in many tumors (ovary, breast, liver, esophagus, colon), where it acts as an oncogene through various mechanisms: for example, it negatively regulates the anti-proliferative action of TGF-β1 in breast carcinoma [47] and mediates proliferation induced by the oncogenic protein X of hepatitis B virus (HBx) in HBV-mediated HCC [48]. Because its upregulation correlates with a worse outcome, it has been shown to be a valid prognostic marker and a possible therapeutic target. The combined use of Celecoxib (COX2-inhibitor) and MK886 (5-lipoxygenase inhibitor), in pancreatic and colorectal carcinomas, inhibits tumor growth through various mechanisms, including the suppression of overexpression of LTB4R by MK886 [49]. As expected on the data of the literature, in our study the gene is downregulated in patients who have achieved complete response to treatment, while it does not vary in patients with stable disease.
The KLF9 gene encodes for a factor belonging to the KLF (Kruppel-like factors) family, involved in cell proliferation and differentiation, both in physiological conditions and malignant transformation, through the link with the promoter sequences of target genes that can be activated or suppressed [50]. Fu DZ and coll. reported that KLF9 mRNA and protein levels were decreased in HCC tissues compared with normal liver tissues and the upregulation of KLF9 expression has antiproliferation and pro-apoptosis properties in HepG2 cells [51]. Sun J et al. found that by binding to the p53 promoter, KLF9 upregulates p53 levels and KLF9 overexpression significantly promotes tumor regression in the xenograft model [52]. These studies demonstrate the oncosuppressive role of KLF9 in HCC.
PREX1, phosphatidylinositol-3,4,5-trisphosphate dependent Rac exchange factor 2, encodes for a protein belonging to the RacGEFs family (Rac guanine nucleotide exchange factors) and is regulated with a positive feedback mechanism, by PI3K. RacGEFs promote Rac activation, involved in cytoskeletal rearrangement, cell migration and adhesion, and ROS production [53]. P-Rex1 acts as a oncogene and its mRNA is upregulated in breast, thyroid, prostate, lymphoid, ovarian, kidney, adrenal and melanoma tumors. At the protein level, P-Rex1 overexpression has only been confirmed in breast cancer, prostate cancer, and melanoma [54–56]. High P-Rex1 protein expression has been linked to decreased disease-free breast cancer patient survival [54].
In our study the gene is significantly down regulated in patients who have achieved complete response to treatment, thus emphasizing its possible role in hepatocarcinogenesis, even if any clear evidence is still present in the literature.
EAF2, ELL-associated factor 2, interacts with RNA polymerase II and promotes the transcription elongation phase. It is a tumor suppressor gene and its expression is downregulated in human prostate cancer specimens and cell lines [57]. Furthermore, EAF2 knockout mice developed lung adenocarcinoma, hepatocellular carcinoma, B cell lymphoma and high-grade murine prostatic intraepithelial neoplasia (mPIN) [58].
The altered pathways due to its down-regulation are multiple: it is implicated in the non-homologous mechanism of DNA repair (NHEJ); interacts with p53, synergizing its oncosuppressive effects; antagonizes the effects of the WNT / beta-catenin pathway, constitutively activated in many tumors. Furthermore, in knock-out mice for EAF2 there is a deficiency of the anti-angiogenic trombospondin 1 protein and a consequent increase in hepatic angiogenesis [59–61].
In our study, the gene is upregulated (p < 0.05) in patients who have achieved complete response to treatment. This result is apparently in contrast to the few data in the literature on the relationship between its expression and the development of hepatocellular carcinoma and deserves further studies to understand if the pathways involved in its upregulation after tumor ablation coincide with those detected until now only on an experimental basis.
In conclusion, the analysis of the genomic profile of circulating PBMCs in patients with hepatocellular carcinoma treated with locoregional procedure showed significant variation in the expression of some genes.
The amplification by RT-qPCR confirmed the data obtained from the analysis of microarrays on the different expression of some genes both in the comparison of patients with CR and PR before treatment, and in the comparison between T0 and T1 in patients who reported complete response to therapy.
The ablation of hepatocellular nodules with radiofrequency or chemoembolization can determine a significant modification of gene expression in circulating PBMCs, which is therefore conditioned by the extent of tumor necrosis and by the characteristics of the tumor microenvironment.
The most intriguing results were the upregulation at baseline of ENTPD1 and the down-regulation of ICOS in patients with partial response to treatment. To the best of our knowledge, there are no study relating ENTPD1 and ICOS gene signatures to outcome of treatment in HCC patients. Our findings highlight a new role for ENTPD1 and ICOS as useful biomarkers for prognostic classification and therapeutic stratification of patients with liver cancer.
This assumes an important biological and prognostic value in order to better understand the pathogenetic mechanisms underlying the carcinogenesis and in the perspective of the development of a target therapy for patients affected by this tumor.