Although stem cells and cancer cell lines are sources of liver cancer organoids, human and mouse tissues have been used to establish organoids (Fig. 2). Organoid models of liver cancer from different sources and their applications are summarized in Table 1. In 2009, Clevers et al. used Lgr5 intestinal stem cells to establish a culture system to maintain basic crypt-villus physiology in vitro. A single crypt can fission to produce multiple crypts, further forming a 3D structure with crypt-like and villous epithelial areas, that is, small intestinal organoids. This is the first truly self-renewing organoid that mimics the actual intestinal tissue structure. Since then, organoid research has entered a new era [64]. In 2013, Takebe et al. differentiated induced pluripotent stem cells (iPSCs) into hepatocyte-like cells and co-cultured them with umbilical vein endothelial cells and mesenchymal stem cells. These multiple type cells self-organized within 48 h to form macroscopically visible aggregates, known as “liver buds”, characterized by the formation of an endothelial network formation and expression of liver specific marker genes (AFP, ALB, retinol binding protein 4, transthyretin). After transplantation in mice, liver buds rapidly connected with recipient blood vessels to form a functional vascular network, resulting in a vascularized and functional mature “human liver”. This study provided a novel approach for the treatment of clinical transplantation of organs and has great prospects for regenerative medicine [65]. In the same year, Miho et al. used bioreactors to build a dense environment composed of fibrocytes and collagen fibers, and primary murine hepatocytes were encapsulated to form liver organoids comparable to endogenous liver. Morphological analysis revealed that hepatocytes in organoid tissues contained more organelles, such as mitochondria, Golgi apparatus, and endoplasmic reticulum, than those of hepatocytes in culture dishes. Moreover, structural liver organoid tissues exhibited a variety of liver specific functions, and microvascular networks could be formed in tissues after transplantation. Thus, organoids can be used to study the structural composition of an artificial liver [66]. Over time, advanced technologies, such as microfluidic platforms, biochips, and genetic engineering have been used for organoid establishment, and human as well as mouse liver cancer organoids have been used to provide diversity.

Fig. 2

Different organoid models of liver cancer

Table 1 Organoid models of liver cancer from different sources and their applicationsOrganoids derived from liver cancer cell lines

Human and animal cell lines are widely used for the in vitro modeling of liver cancer as they respond to specific genetic and epigenetic changes and are widely used for cell proliferation and metastasis studies. These lines include cells with hepatocyte-like (HepG2, HepRG, and Huh7) [67], cholangiocarcinoma cell-like (HuCCA-1 and MT-CHCO1) [68], and fibroblast-like (SKHep1 and SNU-475) phenotypes. Cancer cell lines are cost-effective, easily cultured, and infinitely transmissible. Single or multiple cancer cells can spontaneously aggregate in an ideal reactor and mix to form a spheroidal model with some of the characteristics of in vivo tumors, including a multilayered structure resembling tissues and the continuous multiplication dynamics of cancer cells [69]. Leite et al. generated a spheroidal structure of HepaRG through dimethyl sulfoxide-induced differentiation in a bioreactor culture for seven weeks. The core consisted of biliary cells, and the periphery consisted of hepatocytes. In addition to morphological and structural similarities, the organoids exhibited CYP450 enzyme activity and hepatotoxicity to drugs [70]. In another study, HepG2 hepatocellular carcinoma cells and HCT-116 human metastatic colon cancer cells were co-cultured, with hyaluronic acid and gelatin providing a customized spatial structure for these cells. The average size of the generated organoids steadily increased over time. The expression patterns of HCT-116 cell surface markers under different culture conditions were compared. In the 2D environment, cells were only observed with epithelial phenotype (zona occludens 1[ZO-1] and β-catenin), while mesenchymal markers (N-cadherin) were negative. When the cells transited to the organoid environment, they showed weak expression of ZO-1, E-cadherin, and vinculin. However, β-Catenin, N-cadherin, and matrix metalloproteinase-9 (MMP-9) were expressed. This suggests that the organoid environment supports a shift towards mesenchymal, mobile, and metastatic phenotypes. The authors of the study demonstrated that liver tumor hybrid organoids can provide a better metastatic tumor model compared to 2D cultures [71].

Although every tumor cell line can generate organoid structures, such culture systems do not contain a complete tumor microenvironment (TME). Cancer cell survival is dependent on the TME, as it regulates cell survival, renewal, and differentiation, and influences the gradient distribution of nutrients, metabolic waste, and drugs [72]. The TME is a highly complex structural system composed of cellular and non-cellular components. Various cells in the TME can be tumor suppressive or tumor-promoting. Including tumor cells, immune cells (adaptive immune cells, myeloid immune cells, innate immune cells), stromal cells (cancer-related fibroblasts, adipocytes, neurons, and nerves), and vascular cells (vascular endothelial cells, lymphatic endothelial cells, pericytes) [73]. Non-cellular components include ECM, growth factors, cytokines, chemokines, enzymes, matrix proteins, and metabolic intermediates, which perform the function of information exchange [74, 75]. Wang et al. used Matrigel to grow HCC cell lines (HCCLM3 and Hep3B) along with non-parenchymal cells under 3D conditions and efficiently self-organized them, ultimately forming macroscopically visible organoids within 24 h post-inoculation. The addition of endothelial cells and fibroblasts enhanced the expression of angiogenic markers (vascular endothelial growth factor receptor 2, vascular endothelial growth factor, hypoxia-inducible factors-α), tumor-associated inflammatory factors (C-X-C-motif receptor 4, C-X-C motif chemokine ligand 12, tumor necrosis factor-α), and molecules related to epithelial-mesenchymal transition (transforming growth factor-β, vimentin, MMP9) compared with single HCC organoids. This model was used to verify that non-parenchymal cells are important components in the formation of liver cancer organoids. It can also be used as a stable and easy-to-operate organoid generation method for high-throughput assays [76]. Huh7 is a commonly used HCC cell line. Qiu et al. generated complete HCC organoids using luciferase-expressing Huh7 cells, endothelial cells, and mesenchymal stromal cells derived from human iPSCs. Organoids expressed specific immunofluorescence signals of these three cells (AFP, hCD31, and vimentin). The formed organoids secreted ALB and exhibited liver functional activity. In addition, the authors also introduced different types of controllable TME into organoid models to study the role of the TME in HCC tumor growth. For example, HCC organoids were orthotopically implanted into the liver of mice with fibrosis, and it was found that a fibrotic TME could promote tumor proliferation unidirectionally, while tumor proliferation could not promote the level of fibrosis. However, the development of HCC tumor is not significant in the model of nonalcoholic fatty liver disease. Finally, the organoids were implanted into humanized mice, which resulted in smaller tumors. This indicates that the immune response is important for slowing down tumor growth at early stage. The authors also found that in addition to Huh7, other liver cancer cell lines, such as HepG2, could also successfully form liver cancer organoids. This orthotopic liver cancer model established by using tumor organoids is beneficial to study the TME and tumor development process of HCC [77]. In addition to using the existing cancer cell lines, Song et al. obtained cell lines named AMC-H1 and AMC-H2 from the primary tissues of two HBV infected HCC patients. These two cancer cells can achieve cohesion in a short time, express AFP and have HBV DNA, and retain highly specific genetic changes. Co-culture of AMC-H1 and AMC-H2 with three types of stromal cells (LX2, WI38, and human umbilical vein endothelial cell) under 3D culture conditions formed multicellular tumor spheroids. The multicellular spheroids showed a more compact and clear morphology and a selective response to the standard therapies for HCC, such as sorafenib, 5-fluorouracil, and cisplatin. Thus, this model can be used for screening of selected drugs providing information for clinical drug application [78].

Mouse liver tumor-derived organoids

Cao et al. used dimethylammonium nitrite to induce PLC in mice and established a liver cancer organoid model. A variable number of tumors appeared in the liver of each mouse, such that 129 liver tissue/tumors were obtained from 53 mice. Organoid culture of PLC mouse tumors resulted in the successful establishment of 91 models, with a success rate of 70.5%. These tumor organoids were morphologically diverse and could be cultured, frozen, or revived for long periods of time. After homografting in immunodeficient mice, approximately 20% of the derived organoids could rapidly re-initiate tumors, suggesting that the tumor-derived organoids in these mice were self-renewing and highly tumorigenic. Moreover, the tumor organoids expressed specific markers for hepatocytes (AFP/hepatocyte nuclear factor 4α) and cholangiocytes (epithelial cell adhesion molecule [EpCAM]/cytokeratin 19). The xenograft tumors had the same histological features and expression profiles as the primary tissues. In addition, the evaluation of response to anticancer drugs could be carried out in tumor organoid. It shows that organoids can promote the development of liver cancer (stem cell) biology, drug development and personalized medicine [79].

Jeon et al. injected plasmid vectors containing tumor suppressor genes into the livers of adult C57BL/6 mice via hydrodynamic tail vein injection. Genetic modifications that produce double mutations in tumor protein 53 (Tp53) and phosphatase and tensin homolog (Pten) induced the formation of liver tumors within approximately four months. By correlating human target gene data, it was found that the survival rate of neurofibromin2 (Nf2) or tuberous sclerosis complex 2 (Tsc2) low expression group was lower. Furthermore, Nf2 or Tsc2 gene mutations accelerated tumor formation in a double mutant background of Tp53 and Pten. Organoids were prepared from liver tumors generated in mice. Pten + Tp53 mutant organoids showed cystic structures, while Tp53 + Pten + Nf2 mutant organoids showed dense morphology. After the organoids were irradiated with the specified dose of radiation, the organoids with mutations showed different sensitivity to radiation. This study helped clarify the underlying mechanisms of individual differences in radiosensitivity by testing the radiosensitivity of liver tumors in vitro [80].

Patient-derived organoids

Liver cancer-patient-derived organoids are mainly derived from punch biopsy and surgically collected tumor specimens [40]. In addition, body fluids from cancer patients (pleural fluid or ascites) have been used to establish pancreatic, breast, and lung cancer organoids [81].

Broutier et al. were the first to construct patient-derived PLC organoids by establishing liver and pancreatic organoid protocols with self-renewal capacity in humans and adult mice in 2016 [26]. They extended the culture system for healthy hepatocytes to pathology one year later by designing a liver cancer isolation medium based on normal liver organoid medium, which induces liver tumor organoid growth. First, specimens of the three main subtypes were obtained by surgical resection. Three HCC, three ICC, and two cHCC-ICC organoids with different degrees of differentiation were established. Organoids created using this method retained the specific morphology, marker analyses, transcriptomic features, and gene mutations of the tumor tissue of origin. For example, HCC organoids highly expressed HCC (AFP and glypican-3) and hepatocyte markers (ALB, apolipoprotein E, and transthyretin), but downregulated typical ICC markers, which are some ductal-related markers. In contrast, the ICC organoids highly expressed ICC markers (EPCAM and S100 calcium binding protein A11). Injection of patient-derived organoids into the subcutis of immunodeficient mice revealed tumor growth, and the resulting tumors strongly resembled organoids. This finding suggests that PLC-derived organoids have the potential for tumor formation and metastasis after long-term in vitro expansion and in vivo transplantation. In addition, liver cancer organoids can identify patient-specific drug sensitivities. A correlation has also been shown between the sensitivity of some drugs and the mutational spectrum. Interestingly, the extracellular regulated protein kinases (ERK) inhibitor SCH772984 can inhibit the growth of some HCC and ICC tumor organoids. In vitro experiments also proved that SCH772984 could significantly reduce tumor growth, indicating that SCH772984 may represent a potential therapeutic agent for PLC. Tumor organoids can replicate the histological, transcriptomic, and genomic profiles of PLC in vitro, helping researchers in their deeper understanding of liver cancer biology and the development of personalized treatments [19].

In 2022, Xian et al. established 52 organoids from 153 patients with PLC, with a success rate of 29% for HCC organoids, 52.9% for ICC organoids, and 100% for cHCC-ICC organoids. Compared with the PDX model, the success rate of tumor organoid establishment was higher and took less time. The factors influencing organoid establishment were determined by comparatively analyzing the clinical data from patients: samples with larger tumor volumes, microvascular invasion, terminal PLC stage, and high levels of AFP expression had a greater likelihood of establishing successful organoids. The occurrence of sorafenib resistance represents a difficult obstacle in the treatment of HCC. Acquired sorafenib-resistant organoids were generated form four HCC patients with acquired resistance to the drug. It was found that the resistant organoids were enriched in genes related to cancer stemness-related gene (myelocytomatosis viral oncogene homolog [c-myc] and epidermal growth factor receptor-related gene) and epithelial–mesenchymal transition-related gene (transforming growth factor-β1 and elongation factor 2-related gene), but the heterogeneity between organoids was still obvious. Further testing revealed that targeting the mechanistic target of the rapamycin (mTOR) signaling pathway could effectively treat acquired sorafenib-resistant HCC, which may be caused by inducing phosphorylated S6 kinase. This study not only generated a wide range of PLC organoids but also acquired sorafenib-resistant HCC organoids. It will help clarify the mechanism of acquired drug resistance in HCC and screen possible targeted therapies for HCC [82].

Fibrolamellar carcinoma (FLC) of the liver is a rare subtype of HCC, comprising approximately 1% of HCC in all patients. FLC has no gender preference, and its distinctive feature is the unimodal age distribution between young teens and thirties. About 80% of FLC occurs in adolescents and young people without previous liver disease. Patients harbor a 400 kb deletion on chromosome 19, producing a fusion between exon 1 of DnaJ heat shock protein family (Hsp40) member B1 (DNAJB1), a heat shock protein, and exons 2–10 of protein kinase cAMP-activated catalytic subunit alpha (PRKACA), the catalytic subunit of protein kinase A [83]. In situ hybridization and/or reverse transcription polymerase chain reaction to detect DNAJB1-PRKACA fusion transcripts are very useful for the confirmatory of FLC [84]. Narayan et al. used tumor and adjacent non-tumor liver tissues from nine patients (mean age 22.6 years) to derive human liver organoids for FLC disease modeling. These organoids were divided into six normal organoids, three primary tumor organoids, and 12 metastatic organoids. Metastatic organoids originated from different anatomical sites including the liver, lungs, and omentum. DNAJB1-PRKACA fusion transcripts were detected by polymerase chain reaction in both the FLC tumor samples obtained and patient-derived organoids, and DNAJB1-PRKACA fusion proteins were detected using probes. H&E staining showed that the organoids were similar to their tumor of origin and had the polygonal morphology of FLC cells, as well as granular eosinophilic stroma and prominent nucleoli. Transcriptome analysis further validated the tight phenotype of the tissue of origin and the organoid; thus, suggesting that the organoid recapitulated the features of the primary tumor of the patient [85].

Moreover, hepatoblastoma, known as “childhood liver cancer”, organoids were also established. They can be used for modelling and drug testing of pediatric solid hepatocellular carcinoma [86].

In addition to resected liver cancer tissues, Nuciforo et al. discovered that organoids can originate from tumor needle biopsies. They collected samples from diagnostic puncture biopsies to obtain both tumor and non-tumor samples. HCC organoids showed dense globules without a lumen, and non-tumor tissues grew as unicellular layers of vesicles. Biopsy-based organoids retained the same pathological features, somatic genetic alterations, and mutational features as the original tumor. Transplantation into immunodeficient mice produced xenograft tumors [40]. In such situations, tumor needle biopsy allows for the simultaneous acquisition of multiple samples with little or no damage to the patient, thereby overcoming the limitations of surgically resected specimens. At the same time, this protocol enables the establishment of organoids before extended surgery, providing a reference basis for disease diagnosis and clinical medication [87].

Genetically engineered organoids

Specific modifications of genomic sequences can cause changes in target proteins that control cell differentiation. Organoids can be designed to generate human liver cancer organoids by modifying genomic sequences, endowing the organoids with specific functions, and effectively modeling the disease [88]. Modifications are usually performed using both gene editing methods and targeted transport.

Using both viral and non-viral approaches to organoid modification, a target gene is specifically encapsulated and aggregated at the target site in response to the stable delivery of the gene into the organoid. When selecting vectors, the time of expression, cellular characteristics, and length of the genetic fragments must be considered. Viral approaches that stably deliver genetic information to the offspring by transfection often use adenoviruses [89], lentiviruses [90], or retroviruses [91]. Sun et al. pre-generated reprogrammed human hepatocytes (hiHeps) from fibroblasts and expanded them using simian virus 40 large T antigen. HiHep cells develop an eosinophilic cytoplasm and elliptical nucleus and are enriched in mitochondria, similar to mature hepatocytes. The formed organoids have the same structure and function as the liver. Later, lentiviral vectors for oncogene transfection were used to overexpress the c-myc oncogene and RAS profiles in hiHep organoids; thus, establishing oncogenic transformation models of HCC and ICC. The organoids formed after these gene modifications had the typical features of HCC and ICC cells and could even undergo hepatocyte-to-cholangiocyte lineage conversion [92].

The organoid model of the disease was constructed in vitro, and the disease-causing genes were precisely manipulated by gene editing technology. The two technologies complement each other and represent an approach that will significantly advance the clinical translation of organoids towards the development of personalized medicine [93]. Frequently used genetic methods include RNA interference, transposons, and CRISPR-Cas [94]. The CRISPR-Cas nuclease system achieves multiple purposes, including gene repair, mutation, knockout, knock-in, and fusion, based on its broad applicability, stability, and ease of use [95]. As mentioned earlier, FLC, as a special class of HCC phenotypes, can be reproduced using CRISPR technology in different genetic backgrounds, including protein kinase A-associated mutations and BRCA1 associated protein 1 (BAP1)-driven mutations. In this study, after the cloning products were generated, they were selected to establish clonogenic organoids. Although different mutational backgrounds have been shown to cause hepatocyte dedifferentiation, only the combined deletion of BAP1 and protein kinase cAMP-dependent type II regulatory subunit alpha led to hepatocyte trans-differentiation into hepatic ductal/progenitor-like cells. The engineered organoids showed a heterogeneous polycystic appearance with the loss of highly polarized tissue. In addition, FLC mutant organoid models and tumor tissues are generally similar, with altered gene and protein expression present in the original tumor being consistently altered in the organoid. Knockdown techniques can be applied to normal liver organoids, triggering the onset of HCC [96]. Genetic engineering can also be used to simulate the function of human cancer genes in normal liver organoids. For example, Lam et al. simulated the early stage of liver cancer by manipulating the TP53 status of knockout and overexpression of R249S in normal liver organoids. First, CRISPR knockout of TP53 in liver organoids produced in tumor-like morphological alterations, enhanced stemness, and unrestrained proliferation in vitro. Further, overexpression of mutant R249S in TP53 knockout organoids observes a spontaneous increase in tumorigenic potential and true HCC histology in xenografts. Mutation of TP53, rather than simple loss, confers early clonal advantages and pro-survival functions in hepatocarcinogenesis. It shows that the organoids after gene editing can facilitate the analysis of cancer mutations and provide an opportunity to monitor early stage tumorigenic changes [97]. Artegiani’s team used CRISPR/Cas9 technology to knockout BAP1 gene function in normal human cholangiocyte organoids. Compared with wild-type liver ductal organoids, BAP1 mutant organoids exhibited complete loss of cell organization and polarity, cells presented a very irregular shape, and increased motility. Transcriptomic analysis and quantitative mass spectrometry revealed changes in organoid junctions and cytoskeletal components after BAP1 mutation, such as claudin1, claudin2, and periplakin. Surprisingly, after restoring BAP1 expression in the nucleus, the morphology of organoids can be rapidly reversed and normal morphology can be reconstructed. It shows that organoids combined with genetic engineering can effectively reproduce the carcinogenic process of human tissues [98].

Organoids in liver cancer research

The main problems in the treatment of patients with cancer are the limited number of therapeutic modalities and wide variation in efficacy [99]. Organoids are promising in vitro models that preserve the genetic characteristics and molecular heterogeneity observed in patients. Compared with 2D cell culture models, the 3D structure of organoids provides a model that is closer to the physiological state, which can completely reproduce the system complexity and has patient-like heterogeneity. Compared with in vivo mouse models, organoids can be constructed at a lower cost, with a higher modeling success rate, which is conducive to short-term and large-scale preclinical screening of new drugs [100, 101]. Currently, organoids are widely used in tumor development, drug screening, regenerative medicine, and precision medicine, which can help elucidate the underlying mechanisms of PLC and promote its research development. Current uses of organoids in liver cancer research are illustrated in Fig. 3.

Fig. 3

Organoid uses in liver cancer research

Basic liver cancer researchResearch on the mechanisms of tumor development

Simulating diseases at the tissue level is relatively difficult. However, organoids have a multicellular structure that corresponds to the state of cell development, proliferation, or regeneration under growth conditions. Several studies have reported the use of organoid models to explore tumor development, metastasis [102], and the TME [103]. Cancer-derived organoids retain tumor heterogeneity and recapitulate the parent tumor morphology, histopathology, hormone receptor status, and gene expression profiles [104,105,106]. At the same time, in more advanced biomedical research, the transformation and application of organoid technology can be accelerated by controlling the narrative engineering (morphogenesis, pattern, assembly, growth), biological environment (ECM, cell type, nutrients, stiffness), and comprehensive environment (gas, agitation, contraction, electrochemistry) in the process of organoid culture [27]. Furthermore, organoids allow the analysis of histological changes and mutated genes at the onset of cancer and the monitoring of liver cancer progression in vitro. For example, using organoids, activation of follistatin-like protein 1 (FSTL1) was shown to be correlated with the fibrotic state, whereas high FSTL1 expression was significantly associated with advanced liver tumors. By collecting conditioned media of hepatic stellate cell FSTL1 for the treatment of HCC cell lines and organoids, Loh et al. discovered that it promoted the growth and metastasis of HCC. If FSTL1 is blocked with the monoclonal antibody, the malignancy of HCC will be weakened. Similar results were found in HCC cell lines and organoids. In addition, western blotting confirmed that the key molecules in the AKT/mTOR/4EBP1/c-myc pathway (pAKT, p-mTOR, p-4EBP1, and c-myc) were enhanced in organoids treated with FSTL1; however, this pathway was inhibited after the addition of AKT inhibitor. These results indicate that FSTL1 drives hepatocarcinogenesis through the dysregulated AKT/mTOR/4EBP1/c-myc signaling pathway. It suggests that organoids can not only mimic the growth of HCC but they also allow to investigate molecular mechanisms leading to hepatocarcinogenesis [107].

Liver organoids are 3D cultures of bipotent liver precursors that can be manipulated to express cancer driver genes, which when are transplanted into mice can give rise to tumors. Cristinziano et al. used fibroblast growth factor receptor 2 (FGFR2) fusion proteins to design liver organoids from TP53 null mice. This liver organoid showed the characteristics of ICC after implantation in NOD-SCID mice. The tumors showed downregulation of genes linked to hepatocellular lineage, while genes of the transcriptional signature associated with embryonic cholangiocellular specification were upregulated. This suggests that FGFR2 fusion drives the transformation of bipotent liver organoids to cholangiocarcinomas. Furthermore, ERK1/2 was a necessary signal downstream of FGFR2 fusion in ICC. In this study, organoids and FGFR2 fusion proteins were used to understand biliary tumorigenesis and solve the problem of lack of available gene signature defined ICC models in current preclinical research [108].

Mimicking vascular secretion and the TME by co-culturing liver cancer organoids to identify the factors that drive tumor progression, including microbes, antigens, and inflammatory mediators, is possible. For example, the role of vascular secretory factors in regulating HCC progression was previously unclear. In addition to forming vascular structures, endothelial cells also coordinate cancer progression and metastasis through vascular factors. At present, many soluble and membrane-bound vascular secreted factors have been found, which can be roughly divided into endothelial cell adhesion protein (intercellular adhesion molecule 1, vascular cell adhesion molecule 1, E-selectin) and chemokines (interleukin-8 [IL-8], monocyte chemoattractant protein 1 [MCP1]). The patient derived organoids and endothelial cells was co-cultured in thiohyaluronic acid hydrogel, and the angiogenesis related proteins were analyzed using proteome. The secretion of MCP-1 and C-X-C motif chemokine ligand 16 was higher than that of PDX only or endothelial monocultures. The interaction between HCC and endothelial vessels also affects the immune microenvironment of HCC organoids, leading to the upregulation of vascular secreted factors (such as MCP-1 and IL-8) in co-cultures, promoting angiogenesis and macrophage polarization. Using organoids to investigate the effects of vascular secretory factors on the HCC phenotype and microenvironment could therefore serve as a platform for targeting vascular and microenvironmental interactions [109]. In another example, Liu et al. established a transwell co-culture system of mouse and human PLC-derived organoids with cancer-associated fibroblasts (CAFs) in vitro. CAFs promote organoid growth through paracrine signaling, and cancer cells in turn influence the production of secretory factors by CAFs. Furthermore, they studied the effects of CAFs on organoids derived from in vivo tumors and found that compared with transplanting organoids alone, the co-transplantation of organoids and CAFs had a higher success rate of tumor formation and heavier tumor weight. Even in the group of mice transplanted with organoids alone, the generated tumors could effectively recruit CAFs. The study indicates that the co-transplantation of CAFs and liver tumor organoids promotes tumor formation and growth in xenograft models. This model investigated the interactions between these two cell types and demonstrated the clinical significance of CAFs as important components of the TME in liver cancer development and drug resistance [110].

Novel markers

Numerous studies indicate that tumor lineages are linked to biomarkers to some extent. Tumor markers, which include proteins, carbohydrates, and nucleic acids, can be detected in patients’ blood, urine, and tissue samples. Furthermore, biomarkers have been validated as relatively simple and intuitive methods for the diagnosis, prognosis, and therapeutic assessment of PLC in conjunction with imaging [111]. AFP [112] and glycoconjugate antigen 199 (CA199) [113] are important indicators for the diagnosis of HCC and ICC. Elevated levels of these tumor markers are associated with high tumor incidence, low survival, and poor late prognosis in patients. These markers are also associated with microvascular tumor infiltration and invasiveness of tumor dedifferentiation [114].

However, the limitations of marker specificity and monitoring sensitivity prevent effective screening for PLC, precluding the use of these markers in some cases. For example, a retrospective analysis of liver transplant recipients showed that approximately 30% of the patients with HCC and significant symptoms, maintained AFP levels within the normal range [115]. In another report on patients with chronic liver disease, AFP detection alone was not recommended as a diagnostic marker. If the detection threshold is set low, it may lead to a high proportion of false positives, increasing the psychological pressure of patients and medical costs. If the optimal cutoff value of AFP is increased to 200 ng/mL, although the patient specificity can reach more than 99%, about 80% of HCC patients will be missed, resulting in false negative results [116]. Therefore, in addition to AFP and CA199, a variety of biomarkers have been proposed for individual or combined assessment. These include molecular and biochemical cellular markers (bone bridging proteins, de-γ-carboxy prothrombinogen, and alpha-fetoprotein-L3) [117], cancer stem cell markers (CD44, CD133, CD90, and EpCAM) [118], and non-cellular components (transforming growth factor beta, FGF, and vascular endothelial growth factor) [119]. These potential biomarkers should be further investigated in a tumor-organoid environment.

Presently, other notable markers have been identified. For example, Broutier et al. used RNAseq to analyze differentially expressed genes. They identified markers associated with HCC (chromosome 19 open reading frame 48, ubiquitin conjugating enzyme E2 S, and deoxythymidylate kinase) and ICC (complement C1q binding protein and stathmin 1). Kong et al. analyzed microarrays of six liver tumor tissues and discovered that the expression of the transmembrane tight junction protein claudin 6 (CLDN6) progressively increased from normal to paraneoplastic to HCC tumors. Furthermore, CLDN6 expression was significantly correlated with the poor survival of HCC patients, which was clinically significant. The high expression of CLDN6 promoted the proliferation of HCC cell lines; this phenomenon was also verified in organoid cultures, wherein CLDN6 increased the ability of organoid sphere formation and the formation of organoids of greater size and number [120]. Finally, Zou