The Role of HCC Cell Lines in Liver Cancer Research

Hepatocellular carcinoma (HCC) cell lines are the workhorses of liver cancer research, serving as the primary platform on which scientists test drugs, map genetic vulnerabilities, and study how tumors spread. Dozens of established lines exist, each carrying a distinct genetic fingerprint that reflects a particular slice of the disease. Yet the convenience of growing cancer cells in a dish comes with real trade-offs, from misidentified origins to drug responses that don’t always match what happens in patients. Understanding which lines researchers use, why they choose them, and where the models break down is essential for interpreting almost any laboratory finding in liver cancer biology.

The Most Common Lines and a Persistent Case of Mistaken Identity

Two cell lines dominate the published literature: HepG2 and Hep3B. They are easy to grow, widely available from cell banks, and have been used in thousands of studies over several decades. Because they share superficial similarities, many research teams have treated them as interchangeable stand-ins for liver cancer. In reality, the two lines differ in important ways, including their p53 status, hepatitis B virus (HBV) integration, and sensitivity to various drugs.¹

A more fundamental problem is that HepG2 is not, strictly speaking, an HCC line at all. Transcriptional profiling has shown that HepG2 cells correlate most strongly with hepatoblastoma, a childhood liver tumor, rather than with adult hepatocellular carcinoma. One recent analysis reported a correlation coefficient of 0.62 between HepG2 and hepatoblastoma gene expression, confirming the hepatoblastoma origin despite the line’s frequent misclassification as HCC in the literature.² This matters because conclusions drawn from HepG2 experiments may not generalize to adult liver cancer patients at all. Researchers who rely on HepG2 as their sole HCC model risk building on a foundation that doesn’t represent the disease they’re trying to treat.

How Viral Background Shapes Cell Line Choice

Chronic hepatitis B infection is one of the leading causes of HCC worldwide, so researchers need cell lines that carry integrated HBV DNA to study virus-related tumor biology. The available options vary widely. Among the commonly used lines, Hep3B carries integrated HBV sequences and expresses HBx transcripts, while HepG2 and SK-Hep1 do not.³ A set of Korean cell lines, including SNU-761, SNU-878, and SNU-886, were specifically established from HBV-positive tumors and confirmed to carry integrated HBV DNA with detectable HBx, HBc, and HBs transcripts.³

The virus’s relationship with tumor suppressor genes adds another dimension to cell line selection. A study of Japanese HCC cell lines found that lines carrying integrated HBV sequences did not have p53 point mutations, while the lines harboring p53 mutations were HBV-negative. That observation suggests two distinct routes to liver cancer: one driven by viral integration and another by p53 loss.⁴ For any experiment probing the intersection of viral hepatitis and tumor genetics, picking the right cell line isn’t a minor technical detail. It determines whether the results have any relevance to the patient population being studied.

Modeling Drug Resistance in a Dish

Sorafenib was the first systemic drug approved for advanced HCC, but resistance develops in most patients. To understand why, several groups have created sorafenib-resistant cell lines by exposing HCC cells to gradually increasing drug concentrations over months. One such effort, starting with PLC/PRF5 cells, produced two resistant clones whose half-maximal inhibitory concentrations climbed to roughly 1.8-fold and 4.6-fold above the parental line. The resistant clones showed increased expression of the drug efflux transporter MRP3, which may physically pump sorafenib out of the cell before it can act.⁵

Resistance isn’t just about drug transport, though. A separate study generated five sorafenib-resistant liver cancer lines and found that they had undergone a dramatic shift in identity. The cells lost epithelial markers like E-cadherin, gained the mesenchymal marker vimentin, became more invasive, and lost liver-specific gene expression altogether. When researchers removed sorafenib from the culture medium, these resistant cells showed rebound growth, a phenomenon that has also been observed clinically when patients stop therapy.⁶ These cell-line models of resistance have become standard tools for screening next-generation therapies and combination strategies.

A similar approach has been applied to lenvatinib, another front-line HCC drug. A genome-wide screen in lenvatinib-resistant Huh7 cells identified over 1,200 genes whose knockout promoted survival under drug pressure, with enrichment in pathways that lenvatinib is designed to block.⁷ Genes like NF1 and DUSP9, both negative regulators of pro-survival signaling, emerged as candidates whose loss could drive resistance. This kind of unbiased screening would be nearly impossible without well-characterized cell lines.

Studying How Liver Cancer Spreads

Understanding metastasis requires cell lines that differ in their invasive potential. Chinese researchers developed the MHCC series, including the highly metastatic MHCC-97H and moderately metastatic MHCC-97L, specifically for this purpose. When compared with non-metastatic lines like Huh-7 and PLC, the metastatic lines showed significantly higher expression of Twist, a transcription factor that drives cells toward a more mobile, mesenchymal state. MHCC-97H, which is the more aggressive of the pair, expressed the highest Twist levels and the lowest E-cadherin, fitting neatly with the idea that losing cell-to-cell adhesion is a prerequisite for spreading.⁸

Small regulatory RNAs also play a role. Overexpression of miR-106b in PLC, Huh7, and Hep3B cells activated the epithelial-to-mesenchymal transition, boosting migration in culture and metastasis in animal models. Knocking miR-106b down reversed the shift, restoring epithelial markers.⁹ These paired gain-and-loss experiments across multiple cell lines are a bread-and-butter technique in HCC metastasis research.

MicroRNAs, Metabolism, and Apoptosis

HCC cell lines have been central to mapping how small non-coding RNAs regulate tumor behavior. MiR-122, the most abundant microRNA in healthy liver tissue, is typically downregulated in liver tumors. When researchers restored miR-122 levels in HepG2 and Hep3B cells, proliferation dropped and apoptosis increased. The mechanism traced to direct suppression of the Wnt signaling pathway, specifically the Wnt1 protein.¹⁰ Meanwhile, miR-122 also turned up in metabolic studies of cancer stem cells, where CD133-positive HCC cells, thought to represent a stem-like subpopulation, showed more active sugar metabolism (glycolysis) than their CD133-negative counterparts. MiR-122 appeared to regulate this metabolic shift.¹¹

Other microRNAs have distinct roles. MiR-34a inhibits HCC glycolysis and slows cell growth by targeting LDHA, an enzyme involved in converting pyruvate to lactate, a hallmark of cancer metabolism.¹² MiR-101, which is found at lower-than-normal levels in HCC, promotes apoptosis by targeting the anti-death protein Mcl-1. Restoring miR-101 in hepatoma cell lines sensitized them to both starvation and chemotherapy-induced death.¹³ Each of these findings started as a cell-line observation and offered a potential therapeutic angle worth testing further.

High-Throughput Screening for New Drugs

One of the most practical uses of HCC cell lines is as the foundation for large-scale drug screens. A screen built around SNU-387 and SNU-398 cells tested over 23,000 compounds, including small molecules and natural product extracts, looking for agents that selectively killed cells expressing high levels of the oncofetal protein SALL4. The platform was designed to compare SALL4-high and SALL4-low versions of the same cell line side by side, filtering out compounds that killed indiscriminately.¹⁴

A different approach used image-based phenotypic screening in co-cultures of HCC cells with normal hepatocytes. By observing which compounds killed cancer cells while sparing their healthy neighbors, the team identified anti-folate compounds with HCC-specific toxicity, with pyrimethamine showing the strongest selective effect.¹⁵ Co-culture screening is a clever workaround for a longstanding problem: many compounds that kill cancer cells in isolation also destroy healthy liver tissue, which makes them useless as therapies.

CRISPR Screens and Gene-Level Vulnerability Maps

Genome-wide knockout screens using CRISPR technology have turned HCC cell lines into tools for systematically mapping which genes a tumor needs to survive. One such screen identified 13 clinically relevant target genes enriched in cell cycle and survival pathways, with NCAPG standing out as an essential oncogene for HCC growth.¹⁶ This type of forward-genetic approach lets researchers cast a wide net across the entire genome rather than testing one gene at a time based on prior assumptions.

Cancer Stem Cells Hiding Within Cell Lines

Not all cells within an HCC line behave the same way. A subpopulation marked by the surface protein CD133 has been shown to possess stem-like properties, including higher proliferative potential, greater colony-forming ability, self-renewal capacity, and the ability to initiate tumors when implanted into mice. Another marker, CD90, has been correlated with tumorigenicity and metastatic potential across panels of HCC cell lines. CD90-positive, CD45-negative cells were found in all HCC lines tested.¹⁷ Single-cell multiomics studies have reinforced this picture, revealing extensive heterogeneity among cells within the same line, with subpopulations differing in their capacity for epithelial-to-mesenchymal transition, proliferation rate, and response to low-oxygen conditions.¹⁸ The practical implication is that even a “single” cell line is actually a mixture, and the balance of subpopulations can shift with culture conditions.

Immune Evasion Studies

One reason HCC is difficult to treat with immunotherapy is that tumor cells can hide from the immune system by downregulating the surface molecules that T cells use to recognize foreign tissue. Studies in HCC cell lines have shown that defects in the machinery responsible for loading and displaying these molecules can lead to escape from immune killing.¹⁹

The phenomenon can be extreme. In two of four sarcomatoid HCC cell lines tested, surface HLA class I expression was completely undetectable. The underlying cause was a large deletion spanning the gene encoding β2-microglobulin, a protein required for assembling the antigen-display complex. Interferon-gamma treatment partially rescued some of the downstream processing components but could not fix the gene deletion itself. The same β2-microglobulin loss was confirmed in the original patient tumors, indicating the cell-line findings weren’t an artifact of culture.²⁰ These observations have direct implications for designing T-cell-based immunotherapies, since treatments that rely on T-cell recognition won’t work if the tumor has deleted the very genes those T cells need to see.

Where Cell Lines Fall Short

Growing cancer cells on flat plastic surfaces for decades introduces problems. A large-scale investigation of 278 widely used tumor cell lines found that repeated passaging can cause genetic drift, with sublines sharing only about 80 to 94 percent identity with their originals at standard identity-testing loci.²¹ That drift means the HepG2 cells in one lab may not behave identically to HepG2 cells in another, and the longer a line has been in culture, the wider the gap can become.

Flat culture also strips away the three-dimensional architecture and cellular neighborhood that tumors inhabit in the body. When HCC cells are grown in 3D scaffolds instead of on flat plastic, they upregulate adhesion molecules like E-cadherin and fibronectin and form spheroids with more tissue-like characteristics.²² The gene expression profile in 3D culture looks meaningfully different from what the same cells produce in 2D, suggesting that flat-culture results may misrepresent how tumor cells actually behave inside a patient.

Patient-derived organoids have started to expose the gap more precisely. A comparison of standard cell lines, patient-derived organoids, and patient-derived cell cultures found that conventional cell lines from commercial banks showed the highest sensitivity to most drugs, while patient-derived organoids displayed more heterogeneous, patient-specific responses. The biggest divergence appeared with targeted therapies and drugs with narrow therapeutic windows, though no significant difference was seen for standard chemotherapies.²³ In other words, conventional cell lines may overstate how well a targeted drug works because they lack the resistance mechanisms present in real patient tumors.

Moving Into Animal Models and Engineered Platforms

Xenograft models, where human HCC cells are implanted into immunodeficient mice, remain important for bridging the gap between culture dish and clinic. When HepG2-derived tumors were grown in mice and examined after sacrifice, they showed the hallmarks of aggressive liver cancer: dense proliferating cells, high Ki-67 staining (indicating rapid division), and active new blood vessel formation.²⁴ Not every cell line forms tumors easily, however. Among seven novel lines established from patient-derived xenografts, only two showed reliable tumor-forming ability in mice, with one requiring a supporting matrix gel and a higher cell dose to take hold at all.²⁵ The unpredictability is a reminder that behavior in a dish does not automatically predict behavior in a living organism.

On the engineering side, organ-on-a-chip technology is beginning to incorporate HCC cell lines into more physiologically relevant settings. One liver-on-a-chip platform combined HepG2 cells with endothelial cells and hepatic stellate cells inside a microfluidic device, forming three-dimensional spheroids that experience fluid flow and cell-to-cell communication absent from static culture.²⁶ These multi-cell-type platforms aim to recreate the tumor microenvironment, including the stromal cells, blood vessel lining, and physical forces that influence how cancer cells grow and respond to treatment. They are still relatively new and not yet standardized, but they represent a trajectory away from the simplistic single-cell-type monolayer that has dominated for decades.

What Single-Cell Technologies Are Revealing

Perhaps the most striking recent development is the application of single-cell sequencing to cell lines themselves. Rather than treating a flask of Huh7 or MHCC-97H cells as a uniform population, researchers are now profiling individual cells and discovering that each line harbors subpopulations with strikingly different molecular profiles. A multiomics study of five liver cancer cell lines with varying metastatic potential found that molecular heterogeneity within a single line was associated with differences in the capacity for epithelial-to-mesenchymal transition, proliferation speed, and response to low oxygen.¹⁸ This is both good news and a cautionary tale. Good news because it means even a single cell line can model some of the diversity found in real tumors. A cautionary tale because it means bulk measurements from a cell line are averages that may obscure the behavior of the subpopulation most relevant to clinical outcomes. As single-cell tools become more accessible, they are likely to change how researchers interpret and report cell-line experiments, pushing the field toward more granular, subpopulation-aware analyses rather than treating each line as a single entity.