Hepatocellular carcinoma risk-stratification based on ASGR1 in circulating epithelial cells for cancer interception

Hepatocellular carcinoma risk-stratification based on ASGR1 in circulating epithelial cells for cancer interception Amparo Roa-Colomo 1 , 2 , María Ángeles López Garrido 3 , Pilar Molina-Vallejo 4 , Angela Rojas 5 , 6 , Mercedes González Sanchez 3 , Violeta Aranda-García 7 , Javier Salmeron 2 , Manuel Romero-Gomez 5 , 6 , Jordi Muntane 6 , 8 , 9 , Javier Padillo 10 , Jose María Alamo 10 , Jose A. Lorente 4 , 11 , María José Serrano 4 , 12 , 13 * and M. Carmen Garrido-Navas 4 * 1 Clinical Medicine and Public Health Doctoral Program, University of Granada, Granada, Spain, 2 Gastroenterology and Hepatology Department, San Cecilio University Hospital, Granada, Spain, 3 Gastroenterology and Hepatology Department, Virgen De Las Nieves University Hospital, Granada, Spain, 4 Genyo-Centro Pfizer-Universidad De Granada-Junta De Andalucía De Genómica e Investigación Oncológica, Granada, Spain, 5 Seliver Group, Institute of Biomedicine of Seville (IBiS)/ Hospital Universitario Virgen Del Rocío/CSIC/Universidad De Sevilla, Seville, Spain, 6 Spanish Network for Biomedical Research in Hepatic and Digestive Diseases (CIBERehd), Carlos III Health Institute (ISCIII), Madrid, Spain, 7 Statistician at Fundación para la Investigación Biosanitaria Andalucía Oriental Alejandro Otero (FIBAO), Hospital Virgen de las Nieves, Granada, Spain, 8 Institute of Biomedicine of Seville (IBiS), Hospital University Virgen del Rocío/CSIC/University of Seville, Seville, Spain, 9 Department of Medical Physiology and Biophysics, University of Seville, Seville, Spain, 10 General and Gastrointestinal Surgery Division, Virgen del Rocío University Hospital, Sevilla, Spain, 11 Legal Medicine Department, Medicine School, University of Granada, Granada, Spain, 12 Comprehensive Oncology Division, Clinical University Hospital, Virgen de las Nieves-IBS, Granada, Spain, 13 Department of Pathological Anatomy, Faculty of Medicine, University of Granada, Granada, Spain Purpose: Lack of diagnostic and prognostic biomarkers in hepatocellular carcinoma impedes stratifying patients based on their risk of developing cancer. The aim of this study was to evaluate phenotypic and genetic heterogeneity of circulating epithelial cells (CECs) based on asialoglycoprotein receptor 1 (ASGR1) and miR-122-5p expression as potential diagnostic and prognostic tools in patients with hepatocellular carcinoma (HCC) and liver cirrhosis (LC). Methods: Peripheral blood samples were extracted from LC and HCC patients at different disease stages. CECs were isolated using positive immunomagnetic selection. Genetic and phenotypic characterization was validated by double immunocytochemistry for cytokeratin (CK) and ASGR1 or by in situ hybridization with miR-122-5p and CECs were visualized by confocal microscopy. Results: The presence of CECs increased HCC risk by 2.58-fold, however, this was only significant for patients with previous LC (p= 0.028) and not for those without prior LC (p= 0.23). Furthermore, the number of CECs lacking ASGR1 expression correlated significantly with HCC incidence and absence of miR-122-5p expression (p= 0.014; r= 0.23). Finally, overall survival was significantly greater for patients at earlier cancer stages (p= 0.018), but this OPEN ACCESS EDITED BY Umberto Malapelle, University of Naples Federico II, Italy REVIEWED BY Abilash Gangula, University of Missouri, United States Dirk Roggenbuck, Brandenburg University of Technology Cottbus-Senftenberg, Germany Francesco Pepe, University of Naples Federico II, Italy *CORRESPONDENCE María José Serrano, [email protected] M. Carmen Garrido-Navas, [email protected] SPECIALTY SECTION This article was submitted to Molecular Diagnostics and Therapeutics, a section of the journal Frontiers in Molecular Biosciences RECEIVED 19 October 2022 ACCEPTED 16 November 2022 PUBLISHED 28 November 2022 CITATION Roa-Colomo A, López Garrido MÁ, Molina-Vallejo P, Rojas A, Sanchez MG, Aranda-García V, Salmeron J, Romero-Gomez M, Muntane J, Padillo J, Alamo JM, Lorente JA, Serrano MJ and Garrido-Navas MC (2022), Hepatocellular carcinoma riskstratification based on ASGR1 in circulating epithelial cells for cancer interception. Front. Mol. Biosci. 9:1074277. doi: 10.3389/fmolb.2022.1074277 COPYRIGHT © 2022 Roa-Colomo, López Garrido, Molina-Vallejo, Rojas, Sanchez, Aranda- García, Salmeron, Romero-Gomez, Muntane, Padillo, Alamo, Lorente, Serrano and Garrido-Navas. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Molecular Biosciences frontiersin.org01 TYPE Original Research PUBLISHED 28 November 2022 DOI 10.3389/fmolb.2022.1074277 difference was only maintained in the group with the presence of CECs (p= 0.021) whereas progression-free survival was influenced by the absence of ASGR1 expression. Conclusion: Identification and characterization of CECs by ASGR1 and/or miR- 122-5p expression may be used as a risk-stratification tool in LC patients, as it was shown to be an independent prognostic and risk-stratification marker in LC and early disease stage HCC patients. KEYWORDS hepatocellular carcinoma, liver cirrhosis, circulating tumor cells, precision medicine, cancer interception 1 Introduction Liver cancer affected more than 900,000 individuals worldwide in 2020 and the estimated incidence is expected to rise to 1.4 million individuals in 2040 (Cancer Tomorrow, 2022). Hepatocellular carcinoma (HCC) is the most common liver cancer, accounting for 75%–85% of all primary liver tumors (Sung et al., 2021). The risk of HCC is known to be increased by external factors (such as excessive alcohol consumption, viral infections, aflatoxin exposure) inducing liver inflammation and hepatic fibrosis progression; however, the impact of internal factors (beyond fatty liver disease) increasing the risk of developing HCC is still poorly studied (Forner et al., 2018). The known HCC risk factors promote liver cirrhosis (LC), which is usually a prior finding in 80% HCC patients (Llovet et al., 2021). In fact, 2%–5% of cirrhotic patients will develop HCC annually (Fateen and Ryder, 2017), so surveillance by liver ultrasound screening is performed in LC for early diagnosis of HCC (Forner et al., 2018;Reig et al., 2021). Unfortunately, curative treatments (based on surgery or liver transplant) are available in less than 60% of cases due to late diagnosis (Cadier et al., 2017). There is an absence of tissue biopsies because diagnosis of HCC in LC patients is mainly based on dynamic imaging tests, what reduces the ability to molecularly characterize the tumor (Reig et al., 2021). In addition to imaging tests, serum markers such as alpha-fetoprotein (AFP) that is linked to the evolutionary stage of the tumor, have been used for diagnosis and prognosis of HCC (Fateen and Ryder, 2017). However, the fact that 1) AFP may be elevated in chronic hepatitis without HCC (Force et al., 2022)2) many small-sized HCCs have normal AFP levels (Carr et al., 2018), 3) 30% of HCC patients have normal AFP levels at diagnosis (Lee et al., 2019) and 4) high AFP levels were found in HCC-free individuals (Kobeisy et al., 2012), highlights the lack of sensitivity and specificity of this marker (Bialecki and di Bisceglie, 2005). Although AFP levels greater than 400 ng/ml are considered diagnostic of HCC as well as a marker of bad prognosis marker (Bialecki and di Bisceglie, 2005), it is not recommended to use AFP as a sole marker for HCC surveillance (Hanif et al., 2022) and the combination with other blood-based biomarkers is suggested to improve HCC diagnosis (Wang and Zhang, 2020). Liquid biopsies have the potential to improve sensitivity/ specificity as they are non-invasive, represent better the tumor heterogeneity and can be used to monitor disease evolution. The most widely studied type of liquid biopsy with demonstrated clinical utility are Circulating Tumor Cells or CTCs. The evaluation of CTCs in terms of occurrence, cell count and phenotypic characterization has demonstrated its prognostic value in several solid tumors, including lung (Bayarri-Lara et al., 2016), (de Miguel-Pérez et al., 2019), breast (Nadal et al., 2013), (Ye et al., 2019) and colorectal (Delgado-Ureña et al., 2018) cancers. With respect to HCC, several meta-analyses demonstrated the diagnostic and prognostic utility of detecting CTCs (Fan et al., 2015), (Sun et al., 2017), (Cui et al., 2020), although different isolation and detection technologies may undermine their clinical utility. Isolation methodologies based on EpCAM such as CellSearch ® are widely used, although elevated EpCAM expression levels were only found in metastatic HCC lesions compared to primary and vascular invaded tumor (Tsuchiya et al., 2019). Furthermore, elevated EpCAM expression levels were also linked to poor prognosis (Shimada et al., 2019). Thus, using alternative methodologies for CTC isolation, such as those based on size (Wang et al., 2018), (Qi et al., 2018) or including other biomarkers such as glypican-3 (Court et al., 2018) among others could overcome CellSearch ® limitations. However, Wang et al. (2018) reported lack of correlation between CTCs and recurrence after liver transplantation even though CTCs were detected in more than two-thirds of HCC patients. This suggests that not only improvements for CTC isolation are needed, but also that CTC characterization might be useful for diagnostic and prognostic purposes. Furthermore, presence of Circulating Epithelial Cells (CECs) in patients with pre-malignant diseases such as chronic obstructive pulmonary disease (COPD) had prognostic and diagnostic value (Romero-Palacios et al., 2019) in the context of Cancer Interception (Serrano et al., 2020), although it has been poorly studied in the context of liver cirrhosis (Chen et al., 2020). In fact, in most studies including LC patients, CTCs/CECs were not detected (Vona et al., 2004) or reported percentages were very low (Takahashi et al., 2021), possibly due to either patient selection or isolation methodology. Frontiers in Molecular Biosciences frontiersin.org02 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 One liver-specific biomarker used to isolate CTCs from HCC patients is the asialoglycoprotein receptor 1 (ASGR1) (Li et al., 2014; Courtetal.,2018), that represents the human lectin subunit 1 of the asialoglycoprotein receptor. ASGR1 heterodimerizes with ASGR2 (human lectin subunit 2) to produce a transmembrane protein primarily present in sinusoidal and basolateral hepatocellular membranes. The main role of ASGR1 is to bind galactosyl residues, facilitating glycoproteins turnover (Muramatsu, 2007). In fact, it was recently studied in the context of cardiovascular diseases (Xie et al., 2021;Wang et al., 2022a) demonstrating that its inhibition reduces hypercholesterolemia and atherosclerosis (Xie et al., 2021) by increasing cholesterol excretion (Wang et al., 2022a). In the context of liver cancer, the role of ASGR1 as tumor suppressor was first suggested via its interaction with LASS2 (longevity assurance homolog 2 of yeast LAG1) (Gu et al., 2016)andmore recently by its association with DNA methylation (Zhu et al., 2022), confirming the prior findings of loss of ASGR1 expression in HCC tissue (Shi et al., 2013;Witzigmann et al., 2016). Another potential biomarker of HCC is miR-122-5p, a microRNA involved in multiple physiological processes in the liver that was found to suppress cell proliferation and malignant transformation of hepatocytes (Hu et al., 2012). Downregulation of miR-122-5p in HCC patients as well as HCC-derived cell lines was demonstrated together with an inverse correlation with cyclin G1 expression (Gramantieri et al., 2007) and upregulation of miR-122-5p was shown to repress the epithelial to mesenchymal transition (EMT) through the WNT/βcadherin signaling pathway via Snail 1/2 (Jin et al., 2017). In plasma, miR-122-5p together with other four miRNAs were described as a diagnostic tool (Jin et al., 2019) and recently, its prognostic role was demonstrated when survival of HCC patients was greater in those with higher miR-122-5p expression (Wang et al., 2022b). In this proof-of-concept study, we evaluated the potential diagnostic and prognostic role of circulating epithelial cells (CECs) using an isolation methodology based on immunomagnetic selection with cytokeratins 7/8 (CK) followed by ASGR1 and miR-122-5p characterization in a cohort of patients suffering from LC, with and without subsequent HCC. 2 Material and methods 2.1 Study design and sample collection This prospective cohort study included 113 patients aged between 32 and 86 years suffering from hepatocellular carcinoma (HCC, N= 71) or cancer-free liver cirrhosis (LC; N= 42). Inclusion criterion for cirrhotic patients was diagnosis of LC by dynamic imaging tests (computerized axial tomography and/ or magnetic resonance imaging). Tumor staging for HCC patients followed the BCLC (Barcelona Clinic Liver Cancer) classification (Forner et al., 2018). All patients were aged over 18 and had no other liver disease beyond liver cirrhosis and/or HCC (exclusion criteria). For analytical purposes, the HCC cohort was divided into those diagnosed at an earlier stage (eHCC = BCLC 0-A; N= 30) and those diagnosed at a later stage (aHCC = BCLC B-C-D; N= 41). Of the 30 eHCC patients, only 11 were subjected to liver transplant. Patients were recruited between 2017 and 2020 at the Gastroenterology and Hepatology Units of three Andalusian University hospitals: San Cecilio (Granada), Virgen de las Nieves (Granada) and Virgen del Rocio (Seville). Informed consent was obtained from all patients before blood extraction and the study protocol was approved by the Hospital´s Ethics Committee following the ethical guidelines of the 1975 Declaration of Helsinki. Not signing the informed consent and not fulfilling the inclusion criteria were exclusion criteria. Samples were anonymized upon blood collection to ensure patients’privacy and clinicians involved in the project updated the database for clinical information. Patients’clinical and pathological characteristics are shown in Table 1. 2.2 Circulating epithelial cell isolation Peripheral blood samples (15 ml) were collected in K2-EDTA Vacutainer tubes at the time of diagnosis. Blood samples were enriched in peripheral mononuclear blood cells using gradient centrifugation with Ficoll-Histopaque ® -1119 and CECs were isolated using the Carcinoma Cell Enrichment Kit (130-060- 301, Miltenyi Biotec) based on pan-anti-cytokeratin, as previously described (Nadal et al., 2013). Isolated CECs were then spun down onto Poly-L-lysine-coated glass slides using a cytospin (Hettich) for subsequent phenotypic and genetic characterization (one glass slide was prepared per 7.5 ml peripheral blood). 2.3 CECs enumeration and phenotypic characterization by ASGR1 detection Before isolating and characterizing CECs from patients, antibody specificity for ASGR1 was assessed on the hepatocellular tumor cell line HEPG2 (reference 85011430, lot. 2440) obtained by the Centre of Scientific Instrumentation at the University of Granada. The primary alveolar epithelial cell line hAELVi (InSCREENeX GmbH) was selected as a negative control for ASGR1 expression as stated in the human protein atlas and peripheral blood mononuclear cells (PBMCs) from a healthy donor were used as negative control for both antibodies (Supplementary Figure S1). Both cell lines were tested with STR assay for cell authenticity and for mycoplasma contamination. All experiments were done in duplicates. Subsequently, isolated CECs from patients were enumerated and characterized using double immunocytochemistry with chromogenic staining for cytokeratin (CK) followed by fluorescent detection (for ASGR1). CK detection was done using the Carcinoma Cell Frontiers in Molecular Biosciences frontiersin.org03 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 TABLE 1 Risk factors for HCC. Abbreviations are: HCC, hepatocellular carcinoma; n, number of individuals; SD, standard deviation; Dx, diagnosis; INR, international normalized ratio; AFP, alpha-fetoprotein; HCV, hepatitis C virus; CEC, circulating epithelial cells; IQ, interquartile; CK, cytokeratin; ASGR1, asialoglycoprotein receptor 1. For quantitative analyses t-student and Mann-Whitney tests were used depending on data normality. For qualitative data, Chi-Square was used except for those cases with less than 20% of the data with lower than 5 expected frequencies, in which case, Fisher test was used. pvalues are: ***p<0.001 and **p<0.05. Bold is the mean for each value as a whole. HCC-free N= 42 HCC-affected N=71 pvalues Sex [n(%)] <0.001*** Female 16 (38.1%) 7 (9.9%) Male 26 (61.9%) 64 (90.1%) Age (mean ± SD, years) 66 ± 9 66 ± 10 0.784 Female 67 ± 9 63 ± 11 Male 64 ± 9 66 ± 10 Cirrhosis Dx (mean ± SD, years) 61 ± 10 62 ± 11 0.529 Female 63 ± 10 63 ± 13 Male 59 ± 10 62 ± 11 Bilirubin levels (mean mg/dL) 0.96 ± 0.38 1.70 ± 2.82 0.034** Female 0.90 ± 0.44 1.16 ± 1.89 Male 1.0 ± 0.34 1.76 ± 2.91 Albumin levels (mean g/dL) 4.10 ± 0.40 3.70 ± 0.60 <0.001*** Female 3.90 ± 0.40 3.70 ± 0.70 Male 4.20 ± 0.30 3.70 ± 0.60 Platelets number (mean 10 6 /L) 146,524 ± 78,641 152,052 ± 103,980 0.767 Female 151,125 ± 69,214 173,957 ± 172,546 Male 143,692 ± 85,125 149,656 ± 95,491 INR (mean ± SD) 1.13 ± 0.35 1.23 ± 0.32 0.116 Female 1.26 ± 0.54 1.12 ± 0.18 Male 1.05 ± 0.12 1.25 ± 0.33 Prothrombin activity (%) 90 ± 21 77 ± 18 <0.001*** Female 84 ± 27 83 ± 77 Male 94 ± 16 77 ± 18 AFP baseline (mean ng/mL) 6.13 ± 16.65 1,111.62 ± 4,685.08 <0.001*** Female 11.15 ± 27.64 5,333.91 ± 13,330.00 Male 3.32 ± 1.38 619.03 ± 2,030.34 Alcohol, n (%) 0.17 Female 4 (33.3%) 4 (14.33%) Male 8 (66.7%) 24 (85.7%) HCV, n(%) 0.045** Female 7 (35%) 1 (6.3%) Male 13 (65%) 15 (93.8%) Child Pugh-Turcotte, n(%) <0.0001*** A(5–6) 40 (100) 48 (71.6) B(7–9) 0 15 (22.4) C (10–15) 0 4 (6.0) CEC phenotype, n(median; IQ range) 0.044** CK+/ASGR1+ 17 (2; 4) 29 (2; 2) CK+/ASGR1−7 (1; 4) 25 (2; 7) Negative 18 (0) 17 (0) Total sum of CECs 0.0002*** CK+/ASGR1+ 36 48 CK+/ASGR1−15 74 Frontiers in Molecular Biosciences frontiersin.org04 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 Detection Kit (130-060-301, Miltenyi Biotec) as previously described (Nadal et al., 2013). CECs from HCC and LC patients were then blocked with 10% goat serum and 1.5% FcR blocking in cell stain solution for 45 min and incubated with 1/100 diluted anti-ASGR1 rabbit antibody (Atlas Antibodies Cat# HPA012852, Merck) in 1% goat serum overnight. For visualization of ASGR1, 10 μg of goat anti-rabbit Alexa Fluor 633 antibody (A-21070, Thermo Fisher Scientific) were added for 20 min. Finally, slides were mounted using 4 μg of Hoechst 33342 (Thermo Fisher Scientific) and SlowFade™Gold Antifade Mounting Media (S36936, Thermo Fisher Scientific). Cell enumeration and characterization was performed in a laser confocal microscope (Zeiss LSM 710) and pictures were taken using a ×60 oil objective. CECs were reported as either CEC CK+/ASGR1+ (for positive ASGR1 staining) or CEC CK+/ASGR1− (for negative ASGR1 staining). 2.4 CECs genetic characterization by immunoFISH As a proof-of-concept, we selected 5 individuals positive for CECs belonging to each patient cohort (5 LC, 5 eHCC and 5 aHCC) and a second glass slide for each patient allowed genetic characterization by immunoFISH. This procedure was not performed in all patients due to: availability of enough sample, difficulties of the combination of chromogenic staining with fluorescence detection as well as financial reasons. In patients positive for CECs, fluorescence in situ hybridization (FISH) for the miR-122-5p in combination with immunofluorescence with cytokeratin was performed to determine liver-origin of CECs. Samples were treated with filtered 0.1% pepsin (Merck) in 10 mM HCL during 1 min at 37°C, washed twice with PBS and then hybridized for 1 h at 57°C with miRCURY LNA miRNA Detection Probe for miR-122-5p (Qiagen, Cat#339453) according to the manufacturer’s protocol. After increasing stringency washes with SSC samples were blocked (PBS 1x, 0.1% Tween-20, 2% goat serum, 1% BSA) and incubated with anti-digoxigenin antibody (Roche, Cat#11093274910) for 1 h at room temperature. After several PBS-Tween washes, samples were stained using the pre-filtered substrate FastRed for 1.5 h at 30°C. Staining was terminated by addition of KTBT buffer and subsequently, immunofluorescence for cytokeratin-FITC was performed. Supplementary Figure S2 shows internal controls for specificity of probes and methodology. 2.5 Statistical analysis Descriptive analysis of variables was performed using SPSS, calculating measures of central trend and dispersion for the numerical variables; absolute frequencies and percentages for qualitative variables were also calculated. Normality of the data was studied with the Shapiro-Wilks test. A bivariate analysis was carried out to analyze possible factors related to the main variables. For numerical variables, the Student’s t test was applied for independent samples or Mann-Whitney in nonparametric cases. For qualitative variables, Pearson’s Chisquare test or Fisher’s exact test were applied. In addition, odds ratio and its 95% confidence interval were calculated for each variable. With those that were statistically significant, a multivariate logistic regression model was proposed to jointly predict which factors influenced tolerance to treatment. The variable selection method was performed by successive steps backwards, eliminating in each step those variables that did not significantly influence the model, applying the likelihood ratio test. To evaluate the goodness of fit of the model, the Hosmer- Lemeshow statistic was calculated and to calculate survival rates, overall survival (OS) and progression-free survival (PFS) were plotted as Kaplan Meier curves. Statistical significance was considered for p<0.05. Graphs were created using GraphPad. 3 Results 3.1 Clinical and pathological characteristics of the study cohort Our study cohort included 113 individuals, 42 cancer-free LC patients and 71 HCC patients (of which only 9 did not have prior LC). Univariate analysis of factors associated with HCC included sex, albumin levels and prothrombin activity (p<0.001), bilirubin (p= 0.034), etiology including hepatitis virus C (p= 0.045) and Child Pugh-Turcotte stage (Table 1). No relevant information was obtained using multivariate analysis (data not shown). 3.2 Circulating epithelial cells characterization in LC and HCC patients CECs were detected in 79 patients (69.9%) with significantly greater frequencies in HCC (54/71; 76.1%) than in HCC-free patients (24/42; 57.1%) (p= 0.023). Phenotypic heterogeneity of CECs was identified both intra and inter individual based on ASGR1. Particularly, two phenotypes were identified: ASGR1 positive (CK+/ASGR1+) and negative (CK+/ASGR1−), with varying intensities and sizes (Figure 1). No significant differences were observed between the ASGR1 positive (46/78; 58.9%) and the ASGR1 negative (32/ 78; 41.1%) phenotype (p= 0.33), being ASGR1 expression more prevalent in cirrhotic (17/24; 70.8%) than in HCC (29/54; 53.7%) patients. Importantly, all CECs with the ASGR1 positive phenotype also showed positivity for the miR-122-5p liverspecific marker (Figure 2, rows 1,3 and 5), demonstrating Frontiers in Molecular Biosciences frontiersin.org05 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 FIGURE 1 CECs heterogeneity in patients with liver cirrhosis or HCC. Heterogeneity is shown with respect to Cytokeratin (CK) and ASGR1 expression in two LC (liver cirrhosis) patients (top), two eHCC (early HCC) patients (middle) and two aHCC (advanced HCC) patients (bottom). In each case, one CEC positive for the two markers (CK and ASGR1) and one CEC negative for ASGR1 are shown. Hoechst is used as nuclear staining. Frontiers in Molecular Biosciences frontiersin.org06 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 FIGURE 2 Confocal microscopy of immunoFISH for miR-122-5p in patients with liver cirrhosis or HCC. Cytokeratin (CK) is shown in green (FITC) and miR122-5p probe is shown in red. Data from two patients for each group: liver cirrhosis (LC; top), early HCC (eHCC; middle) and advanced HCC (aHCC; bottom) are shown. For each patient group, a positive miR-122-5p signal (coinciding with positive ASGR1 staining) and a negative miR-122- 5p signal (coinciding with negative ASGR1 staining) are shown. Frontiers in Molecular Biosciences frontiersin.org07 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 their liver origin. In contrast, absence of miR-122-5p was observed in all patients with CECs showing ASGR1 negative staining (Figure 2, rows 2,4 and 6). 3.3 Correlation of CEC phenotypes with clinical and pathological data The amount of CEC CK+/ASGR1- was significantly greater in HCC than in HCC-free patients (p= 0.0096), significantly increasing the total CEC count in HCC compared to LC patients (p= 0.035) (Figure 3A). Furthermore, when accounting for disease stage: early (BCLC 0-A) or advanced (BCLC B-C-D), no differences were observed for the ASGR1 positive population (Figure 3B). However, Kruskal- Wallis tests revealed that the number of CECs with absence of ASGR1 expression was significantly different among disease stages (p= 0.017). There was a significant increase of CEC CK+/ASGR1− between LC and eHCC (p= 0.0051) but not between LC and aHCC (p= 0.08) or eHCC and aHCC (p= 0.23) (Figure 3C). Our results suggest that both presence of CECs and absence of ASGR1 expression in CECs are risk biomarkers of HCC. In fact, presence of CECs increased HCC risk by 2.58 -fold (p= FIGURE 3 CECs enumeration in blood in patients with liver cirrhosis or HCC. CECs are characterized by either double staining (CK+/ASGR1+) or single staining (CK+/ASGR1-) in liver cirrhosis (LC) and hepatocellular carcinoma (HCC) patients (A). The latter are divided into early (eHCC) and advanced (aHCC) disease for both, double (B) and single staining (C).pvalues for Kruskal Wallis test and multiple comparisons are:**p= 0.0096 and *p= 0.035 (A) and **p= 0.0051 (C). Frontiers in Molecular Biosciences frontiersin.org08 Roa-Colomo et al. 10.3389/fmolb.2022.1074277 0.023) (Figure 4A) and the number of CECs lacking ASGR1 expression significantly (p= 0.001) increased risk of developing HCC by 2.66-fold (Figure 4B). A significantly greater proportion of aHCC and eHCC patients had CECs (including ASGR+ and ASGR-) compared with LC (p= 0.049) (Figure 4C) and this increase came with a significant reduction on the number of CECs expressing ASGR1 in HCC compared to LC (p= 0.003) (Figure 4D). When pathological AFP levels were considered, 13/ 13 patients (100%) with AFP greater than 400 ng/ml showed HCC (11 of which had prior LC). However, there were a significant (p= 0.0117) proportion of individuals with lower AFP levels (54/93; 58.1%) that also developed HCC, highlighting the poor performance of AFP for identifying high-risk individuals (Figure 5A). Interestingly, there was a significant difference between ASGR1 expression when analyzing liver cirrhosis and cancer occurrence together (p= 0.033). Thus, of the total of individuals with CECs negative for ASGR1 expression, the majority were HCC patients with prior LC (65.6%; 21/32), while there was no difference between cancerfree LC (18.8%; 6/32) and HCC without prior LC (5/32; 15.6%). Contrarily, positive expression of ASGR in CECs was more frequent in LC (42.9%; 18/42) than in HCC without prior LC (22.2%; 2/9), suggesting that LC patients (independently on their cancer status) were more likely to have ASGR1 positive CECs (Figure 5B). In fact, HCC patients without previous LC had the greatest percentage of CECs without ASGR1 expression (median 100%) compared with HCC with LC (median 71.0%) and HCC- free LC (median 51%) patients although this difference was not significantly different (p= 0.21) (Figure 5C). Furthermore, we found that the percentage of CECs without ASGR1 expression increased upon liver dysfunction (Child Pugh-Turcotte stage) (Figure 5D), although this increase was not significant (p= 0.74). Presence of CECs without ASGR1 expression significantly correlated with cancer occurrence (p= 0.012). Indeed, ASGR1 expression in CECs negatively correlated with INR (p= 0.025). Finally, as expected, early HCC patients showed significantly greater overall survival (OS) than advanced HCC (p= 0.018) (Figure 6A). However, this difference was observed only in patients showing CECs (Figure 6B;p= 0.021) and not in patients without CECs (Figure 6C;p= 0.250). Furthermore, a FIGURE 4 Risk factors for HCC. (A) Shows presence/absence of CECs and (B) shows number of CECs expressing (or not) ASGR1 in liver cirrhosis (LC) or hepatocellular carcinoma patients (HCC). (C,D) show same information as in (A,B) respectively, but dividing HCC individuals by either early (eHCC) or advanced (aHCC) disease stages. Frontiers in Molecular Biosciences frontiersin.org09 Roa-Colomo et al. 10.3389/fmolb.2022.1074277