The precise role of P53 in early-stage liver cancer remains unclear, and the physiological similarity between murine and human models is limited. In this study, we developed a CRISPR/Cas9-mediated TP53KO human iPSC-derived hepatic organoid model (eHEPO) to replicate the early stage of liver cancer within a pc-ME.
Genome-wide transcriptome and comparative GSEA studies in TP53KO-eHEPOs identified enrichment in gene sets related to fibrosis, ECM modification, and inflammation. Furthermore, TP53KO-eHEPOs demonstrated correlations with liver cancer subsets, specifically in CHIANG and HOSHIDA, particularly associated with interferon signaling and altered Wnt/β-catenin signaling subclasses. In line with the established role of TP53 in cancer, TP53KO-eHEPOs exhibited increased correlations with other cancer types. Notably, a cluster of genes associated with early cancer initiation and progression, including MYC, CD44, ALDH4A1, MUC1, GPC3 and GPC4 exhibited significantly elevated expression levels in TP53KO-eHEPOs. Due to the wide acceptance of these markers in the initiation of liver cancer and their high expression in various other cancers, they hold potential as therapeutic targets [45–48]. The implications of these findings are further substantiated by previous research. For instance, Dhar et al. (2018) demonstrated a link between TP53 inactivation and CD44-mediated MDM2 translocation in hepatic progenitor cell [48]. This specific subset of progenitor cells, marked by heightened danger signals, exhibited AKT-mediated proliferation, thus contributing to the initiation of cancer. To provide insights into the clinical significance of the genes enriched in our DEG list, we conducted a comprehensive analysis of their implications for overall survival among HCC patients. Leveraging the Kaplan-Meier plotter and utilizing liver cancer RNA-seq data, we explored the survival outcomes of individuals with varying expression levels of specific genes. Critically, individuals with elevated expression levels of COL24A1, S100A9, and COL11A1 (Supplementary Fig. 6A), as well as CXCL3, CXCL5, and CCL20 (Supplementary Fig. 6B), and ADAM12, SPP1, and ITGA2 (Supplementary Fig. 6C), along with KRT80, MUC15, and MUC1 (Supplementary Fig. 6D), exhibited lower overall survival rates. These findings underscore the potential prognostic significance of these gene sets in liver cancer patients.
To validate hepatic progenitor mediated fibrotic liver tissue alterations, Masson's trichrome staining and associated histological changes serve as significant markers [44, 59]. In our histological staining, we observed that TP53KO-eHEPOs exhibit heightened atypical features, pseudo-glandular-tubular rosettes, steatohepatitis-like inflammatory areas, and ballooning-like hepatocytes, accompanied by increased collagen deposition. Similar histological changes and an elevated collagen signal have also been reported in various studies. For instance, Guan et al.'s liver fibrosis model is defined in hepatic organoids generated from a genetic disease model of Autosomal Recessive Polycystic Kidney Disease (ARPKD). In this study, ARPKD mutation triggered myofibroblasts to form collagen fibers by increasing TGFβ in cholangiocytes, and that PDGFb expression increased and the STAT3 signaling pathway was activated in collagen-producing fibroblasts [16]. Another noteworthy study demonstrated that Thioacetamide (TAA) and Free fatty acids (FFA) induced hepatic organoids exhibit increased COL1A1, COL3A1, PDGFRB, ECM modification, and pro-inflammatory cytokines, resembling fibrosis and steatosis-like histological changes. Importantly, similar to our results, collagen staining became thick around the perisinusoidal space after treatment [60].
The functional loss of TP53 results in aberrant regulation of cell death and cell proliferation, a characteristic observed in the majority of cancers [61, 62]. Correlated with this, we observed that our TP53KO-eHEPOs exhibit increased proliferation in the EM medium. Interestingly, this proliferation of TP53KO-eHEPOs is attenuated in DM and iDM compared to WT. We hypothesize that this increase in TP53KO-eHEPOs is likely attributed to the role of TP53 as the "guardian of the genome" and its critical function in preserving genomic integrity and the homeostasis of cell differentiation [63]. In this regard, we conducted multiple analyses to further assess the differentiation state of our TP53KO-eHEPOs. The results indicated well-differentiated hepatic cell characteristics, including increased ALB secretion, positive PAS staining (glycogen storage), and enhanced expression of A1AT, ALB, and CK18 in the IF staining. Interestingly, our model shares certain similarities with findings in a TP53 knockout mouse model. In the TP53 knockout mouse model, knockout cells exhibited a blast-like shape and increased albumin secretion, characterized by hyperproliferative behavior and the retention of stemness properties [64]. Likewise, TP53KO organoids display hyperproliferative properties in the EM medium. However, in the DM and iDM medium, in which cells are induced to undergo a well-differentiated stage, TP53KO-eHEPOs demonstrate superior differentiation compared to WT, which is due to the lack of the genomic barrier function of TP53.
Additionally, we observed an enrichment of HIF1a, IFNA and STAT3 signaling pathways in TP53KO-eHEPOs. Interestingly a recent study showed that IL10 mediated senescence of hepatic stellate cells (HSC), increases STAT3 mediated TP53 activation and reduces fibrosis. However, lack of TP53, probably brought about by aberrant activation of STAT3 mediated downstream signals, is generally related with poor prognosis in several cancer types [65, 66]. In addition, we observed an augmented TGFB signal pathway in TP53KO-eHEPOs, accompanied by an attenuated Wnt signal pathway or an increased expression of Wnt inhibitory signals, such as DKK1 and CDKN2A. A recent study demonstrated that fibrous nest-type HCCs exhibit a positive correlation with Hoshida's liver cancer subclass S1, which is associated with Wnt/TGFB-mediated changes. The correlation between cancer samples and this particular subset, characterized by altered inflammation and ECM modifications, leading to a poor prognosis in patients [67]. Excitingly, we also identified a similar enrichment of Hoshida's S1 in our TP53KO-eHEPOs, which correlated with altered inflammation, ECM modification, and an increased fibrosis signal. The observed correlation underscores the effectiveness of our eHEPO model as a highly applicable tool for systematically investigating fibrosis stages. This investigation is contingent upon variations in mutational burden and changes in the tumor microenvironment. The model facilitates a granular examination, allowing for a step-by-step analysis of these intricate aspects.
Additionally, we examined alterations in stem cell markers in our eHEPO model due to their direct or indirect regulation by TP53. EpCAM is generally highly expressed in the early developmental stages of the liver, including embryonic stem cells, endoderm, liver progenitors, and immature hepatocytes. Its expression is attenuated during hepatocyte differentiation; however, EpCAM expression level increases once more when cells enter progenitor or cancer-like stages [68]. We observed a drastic reduction in EpCAM and CD24 protein levels in TP53KO-eHEPOs cultured in DM and iDM conditions in comparison to WT-eHEPOs. Consistent with the literature, CD24 expression decreases in well-differentiated stages (DM) compared to hepatic progenitors (EM) [69]. Additionally, CD24 is a well-known stem/progenitor cell marker, and its role in the regulation of the proliferation and differentiation balance suggests a potential association with the well-differentiated stage of TP53KO-eHEPOs [70]. CD133 appears to be another progenitor molecule and a cholangiocyte marker, which should be reduced in well-differentiated hepatocytes [71]. However, we observed an increase in CD133 in TP53KO-eHEPOs DM and IDM conditions, which may facilitate an aggressive inflammatory and fibrotic response [72, 73]. In our study, we observed increase in TGFB signal activation suggests a role in promoting fibrosis. Interestingly, a recent study using patient-derived TP53KO and TP53R249S organoids revealed an increased level of CD44 and CD133, indicating tumorigenic properties. However, a single loss of TP53 is generally associated with altered histology, aneuploidy, and chromosome segregation errors some of them also observed in our model. Tumorigenic transformation typically requires additional driver mutations or enforced microenvironmental changes [35]. Since this model was derived from adult human paired HCC tumor and adjacent non-tumoral liver tissues, compared to our iPSC-derived eHEPO model, the cells may exhibit slightly different responses due to differences in stemness repository (adult-fetal stem cells, respectively) and cell composition [74].
In conclusion, our in vitro eHEPO model demonstrated a more realistic representation of human physiology and gene expression patterns, aligning closely with the current knowledge in the literature regarding liver cancer initiation, fibrosis, and disease progression. It offers a valuable tool for studying the effects of TP53-like first-hit genes on hepatic progenitor cell transformation in various environments. We anticipate that this in vitro model will serve as a promising tool for investigating the combined effects of different microenvironments and gene knockouts. Moreover, it holds potential as a model for identifying early-stage drug candidates.