Jian-Gang Zhang ,Hua-Yu Yang ,Yi-Lei Mao

Department of Liver Surgery, Peking Union Medical College (PUMC) Hospital, PUMC & Chinese Academy of Medical Sciences (CAMS), Beijing 100730, China

Liver plays a central role in various physiological functions,including metabolism,biliary secretion,production of plasma proteins,regulation of hormones as well as detoxication.Because of its multidimensional functions,liver diseases such as viral hepatitis,non-alcoholic fatty liver disease,non-alcoholic steatohepatitis,fibrosis and liver cancer may lead to serious consequences.However,the therapeutic options for these diseases are limited [1].Preclinical models are important to explore the managements of patients with liver diseases.However,the traditional cell line and xenografted mice are not satisfying in preclinical studies [2].We need to create patient-derived preclinical models with high robustness of cell fate maintenance,high-level tissue self-assembly accuracy and ability to exhibit dynamics of physiological processes.This article reviewed recent progresses,applications and future prospects of novel preclinical model systems with primary human liver cells,including spheroid systems,patient-derived organoid (PDO),microfluidic systems,bioprinting and patient-derived xenograft (PDX).

Cultivation of primary human hepatocytes (PHHs) in spheroid with or without additional non-parenchymal cells (NPCs) has been utilized to perform liver toxicity evaluations and establishment of models of liver diseases [3–5].Patient-derived liver spheroid system usually consists of PHHs 3D aggregates with stable hepatic phenotypes and functionality in a long period of time.It can closely mimic human liver in various environment to achieve the modeling of chronic liver-related diseases such as steatosis and insulin resistance.Since spheroids derived from different donors show different responses to steatosis induction [5],patient-derived liver spheroid system demonstrated interpatient heterogeneity in terms of susceptibility.Furthermore,the ability to integrate NPCs into the spheroid system makes it possible to investigate how NPCs influence drug-induced liver injury [4],providing a more comprehensive perspective.Compared to other model systems,spheroid exhibits cost-effectiveness and high throughput capacity.However,due to the relatively simple architecture,it is not possible to mimic sophisticated localization of different cell types.Furthermore,the addition of NPCs further complicates culture conditions to meet the needs of different types of cells [6].

Similar to spheroid systems,PDO also mimics the donor organ.However,PDO exhibits a higher level of self-assembly accuracy and lineage-specific differentiation compared to spheroid [7].Generation of PDO could be achieved from various cells including primary hepatocytes,induced pluripotent stem cells and tumor cells,which could further combine with gene editing technologies to serve as novel platforms for studies of gene function and regulation.Moreover,a recent study [8]has shown the possibilities to promote the expansion of different adult liver cell types,which are diffi-cult to culture,under the fine control of growth factors and other small molecules.Therefore,cell sources of PDO may be greatly enriched.Organoid derived from tumors (tumoroids) also mimics numerous characteristics of the original tumors.Hepatocellular carcinoma (HCC) and cholangiocarcinoma tumoroids have been derived from needle biopsies [9].Because of their similarity to the original tumors in both genetic signature and phenotypic features,tumoroids have been used to evaluate drug responseinvitro,showing both intratumor and interpatient heterogeneity [10]and a promising correlation with clinical outcomes.However,the practice of organoid is limited by the low success rate of establishment as well as the lack of tumor microenvironment.Its relatively complex culture process and the requirement for growth factors and small molecules also make it labor-intensive.There is another concern that genetic and epigenetic alternations might happen during the delicate process of establishment and reprogramming [7],which will introduce uncertainties into the model.

Microfluidic system offers a novel platform to mimic basic liver architectures and functionsinvitro.Within the hepatic acinus,liver tissue can be divided into periportal zone,transition zone and pericentral zone.Different zones tend to get different level of oxygen and nutrients and carry out distinct metabolic functions.It has been reported that microfluidic systems that consist of PHHs and NPCs have been established with oxygen zonation and at least partial immune response [11].Moreover,a microfluidic system using PHHs and Kupffer cells has been established to model hepatitis B virus infection [12].Building upon the physiological environments provided by microfluidic systems,it is promising that more infectious diseases can be modeled in this way.Apart from the applications mentioned above,microfluidic systems have been utilized in many aspects such as drug-induced liver injury studies,pharmacokinetic studies,drug discovery and development,modeling of various liver diseases as well as testing hepatotoxicity.Furthermore,microfluidic systems have great potential to integrate with other modeling systems.For instance,bioprinting technology may be useful in achieving the accurate localization of different cell types in microfluidic systems.Combining different modeling systems or even constructing an organs-on-chip system will certainly simulate the functions of human liver under pathological and physiological conditions.However,it is difficult to achieve high-throughput screening in microfluidic systems to date,which holds back the practice of microfluidic systems.Microfluidic systems are also difficult to maintain long-term culture without contamination.On the macroscale,machine perfusion technology provides organ level platforms for pharmacology and toxicology studies as well as organ-oriented treatment exploration [13]similar to microfluidic system.When it comes to orchestrating spatial structures,bioprinting may be of great significance.

Bioprinting can fabricate functional tissue constructs layer by layer using cell-laden bioinks.Liver tissues have been bioprinted from PHHs and NPCs.That the ability to express liver-specific proteins and maintain long-term functionality expand the applications of bioprinting in pharmaceutical studies and precision medicine [14].Bioprinted intrahepatic cholangiocarcinoma tumor model with a high viability has been reported to have different gene expression profiles and drug responses compared to 2D culture [15],suggesting that bioprinting has an essential potential in drug testing and personalized tumor model.Besides,hepatotoxicity and liver-related disease models have been established using bioprinting.3D bioprinted hepatorganoids were shown to prolong the survival of liver failure in a mouse model [16],serving as a proofof-concept that functional 3D bioprinted human liver tissues may be used as transplantation donors.Therefore,bioprinting shows great potential in regenerative medicine to bridge to gap between the number of patients waiting for liver transplantation and organ supply.From another perspective,3D bioprinted HCC suitable for anti-HCC drug screening also shows great potential in clinical applications [17].It can be established swiftly with high success rate.At the same time,typical features as well as tumorigenic potential of the original tumor can be maintained in 3D bioprinted HCC.Apart from progresses made in the evaluation of biological functions and drug response using bioprinting technology,more recent studies focus on optimizing bioprinting process as well as bioinks [18].For instance,new methodologies have been introduced to expand bioink library,increase bioprinting capabilities,realize complex self-assembly in a large scale [19],expand cell source and materials,realize precise control on the microscale as well as standardize bioprinting process [20].These advances will greatly push the boundaries of biofabrication window and more studies using primary human liver cells are needed.

PDX models are widely used in preclinical studies such as modeling pediatric liver cancers and other rare cancers [21]for its ability to maintain disease heterogeneity and potential in personalized medicine.Compared to other model systems,PDX can mimic tumor microenvironment more realistically in immunodeficient mice.Therefore,PDX model of liver cancer may be possible to predict drug responses as well as treatment outcomes.Banks of liver PDX models were established to identify predictive markers and perform genetic analysis [22],which are valuable resources in oncology research.Furthermore,PDX in humanized mice is an important step towards the evaluation of immunotherapeutic approaches.However,PDX models have several drawbacks.The engraftment efficiency is relatively low and PDX which is successfully transplanted tends to have a higher degree of malignancy.Besides,complex process of establishing PDX makes the model labor-intensive.Thus further studies are needed to optimize the methodologies.

There are other patient-derived preclinical model systems such as circulating tumor cell-derived models [23]and precision-cut liver slices as well as studies focusing on the optimization of microenvironment of 3D primary human liver cell cultures [24],which are beyond the scope of our discussion.

In conclusion,different 3D preclinical model systems with primary human liver cells show different applicable scenarios as well as advantages and disadvantages.Constructing multiple models is the ideal situation for preclinical modeling to adapt to different stages of drug development and studies with different purpose.Future directions include establishing biobanks of preclinical models,maintaining intratumor heterogeneity,raising the efficiency of establishment,expanding cell sources and integrating functional immune component and other NPCs to construct liver-related disease models with more sophisticated microenvironment.As for drug development and the implementation of precision medicine,balancing the feasibility of applying high-throughput methods and the intactness of microenvironment is the key to create an effi-cient and reasonable workflow.In the future,comprehensive biological function evaluation and independent clinical studies of various models mentioned above are needed to expand the application and promote standardization.

Acknowledgments

None.

CRediT authorship contribution statement

Jian-Gang Zhang:Data curation,Investigation,Writing -original draft.Hua-Yu Yang:Data curation,Investigation,Writing -original draft.Yi-Lei Mao:Conceptualization,Resources,Supervision,Writing -review &editing.

Funding

None.

Ethical approval

Not needed.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.