Physiologically relevant media are associated with overlapping metabolic responses in primary human hepatocytes and Huh7 cells

This study demonstrates that Huh7 cells cultured in physiologically relevant media exhibit metabolic responses and lipid accumulation patterns comparable to primary human hepatocytes, supporting their utility as a scalable in vitro model for studying metabolic dysfunction-associated steatotic liver disease (MASLD).

The Gist

The Big Picture: The Liver's "Factory" Problem

Imagine your liver is a massive, busy factory that processes food and fat. When this factory gets overwhelmed with too much fat, it starts storing it in giant warehouses inside the building. This condition is called MASLD (Metabolic Dysfunction-Associated Steatotic Liver Disease), which is basically a fancy way of saying "fatty liver."

Scientists want to study how to fix this factory, but they can't just cut open a person's liver to watch it work in real-time. So, they use models (simulations) to study it in a lab.

For a long time, scientists have had two main types of "factory simulations":

1. The Gold Standard (Primary Human Hepatocytes or PHHs): These are real liver cells taken directly from human donors. They are like master craftsmen. They know exactly how a real human liver works, but they are fragile, expensive, hard to find, and they die quickly in the lab.
2. The Workhorse (Huh7 Cells): These are immortalized cancer cells that can grow forever. They are like durable robots. They are cheap, easy to get, and never die, but scientists often worry they are "dumb" and don't act like real human cells.

The Question: Can we teach the "durable robot" (Huh7) to act like the "master craftsman" (PHH) if we give it the right environment?

The Experiment: Changing the Diet

In the past, scientists fed these cells "junk food" in the lab—huge doses of sugar or single types of fat that don't exist in real human bodies. This is like feeding a human a diet of only pure sugar water; it breaks the system and gives fake results.

In this study, the researchers decided to feed both cell types a physiologically relevant diet.

• The Menu: Instead of junk, they gave the cells a mix of fats found in a normal human diet (a mix of saturated and unsaturated fats) and human serum (like blood plasma).
• The Goal: To see if the "robot" (Huh7) could mimic the "craftsman" (PHH) when both are eating a realistic diet.

What They Found: The Robot Learned to Act Human

Here is the breakdown of the results, using our factory analogy:

1. Survival of the Fittest

• The Craftsmen (PHH): They were fragile. About 30% of them died before the experiment even started. They were also very sensitive; different batches of craftsmen acted differently (because they came from different people).
• The Robots (Huh7): They were tough. Almost all of them survived, and they all acted the same way.
• Takeaway: The robots are much easier to work with, but the craftsmen are more "real."

2. The Fat Storage (The Warehouses)

• Both cell types took in fat from the diet and stored it in lipid droplets (tiny fat bubbles inside the cell).
• The Result: When fed the realistic diet, the robots (Huh7) started building these fat warehouses in almost the exact same way as the real human cells. They looked the same size and had the same chemical makeup.
• Analogy: It's like giving a toy car and a real car the same gas; both engines sputtered and ran in a very similar way.

3. The Energy Output (Exhaust Fumes)

• Lactate (Sugar Exhaust): The robots (Huh7) were burning sugar very fast and spitting out a lot of lactate (like a car revving its engine). The real human cells (PHH) were more efficient and didn't produce as much.
• Ketones (Fat Exhaust): The real human cells were better at burning fat for energy, producing more ketones. The robots were a bit slower at this.
• Takeaway: The robots are great at handling fat storage, but they still burn sugar a bit differently than a real human liver.

4. The "Brain" (Genetics)

• When the researchers looked at the cells' "instruction manuals" (RNA/Genes):
  • The Robots: Their genes changed in a predictable way based on the diet. They adapted to the new food.
  • The Craftsmen: Their genes were chaotic. The biggest difference wasn't the food; it was who the donor was. One donor's cells reacted totally differently from another's.
• Analogy: The robots all followed the same script. The craftsmen were improvising based on their own personal history.

The Verdict: A New Tool for Science

The study concludes that Huh7 cells (the robots) are actually very good at mimicking human liver cells (the craftsmen) if you feed them a realistic diet instead of lab junk.

• Why this is good news: Scientists can now use the cheap, durable, and easy-to-grow Huh7 cells for most of their experiments. They can test drugs and study fat storage without needing expensive, fragile human cells for every single test.
• The Catch: Because the real human cells (PHH) still show unique differences based on the person they came from, we still need them for the final, critical checks—especially when studying inflammation or specific genetic diseases.

In short: The "durable robot" has been upgraded with a realistic diet and is now a reliable stand-in for the "fragile master craftsman" for studying fatty liver disease. This means we can study liver disease faster, cheaper, and more effectively.

Technical Summary

1. Problem Statement

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a global health burden affecting ~30% of adults. Studying its early pathophysiology (specifically intrahepatocellular triglyceride, IHCTG, accumulation) in humans is difficult due to the invasiveness of liver biopsies and the lack of high-fidelity in vivo models.

• The Gap: While Primary Human Hepatocytes (PHH) are the "gold standard" in vitro model, they suffer from high costs, low viability post-thaw, donor-to-donor variability, and limited scalability. Conversely, immortalized cell lines (e.g., Huh7, HepG2) are scalable but often fail to recapitulate human metabolic phenotypes, largely because they are typically cultured in non-physiological media (e.g., high glucose, single fatty acids, fetal bovine serum) for short durations.
• Objective: To determine if an optimized Huh7-based model, cultured in physiologically relevant media (human serum, chronic exposure, mixed fatty acid ratios), can replicate the metabolic responses of PHHs, thereby offering a scalable alternative for mechanistic MASLD studies.

2. Methodology

The study compared two models under their respective optimal chronic culture conditions:

• Cell Models:
  • PHH: Cryopreserved primary human hepatocytes from multiple donors (mixed gender).
  • Huh7: Immortalized hepatoma cell line.
• Culture Conditions:
  • Media: Both were cultured with 2% Human Serum (HS) instead of Fetal Bovine Serum (FBS).
  • Substrates: Cells were exposed to a mixture of four fatty acids (Oleic, Palmitic, Linoleic, $\alpha$-Linoleic) in two physiological ratios:
    • OPLA: Predominantly unsaturated (45:30:24:1).
    • POLA: Predominantly saturated (44:45:10:1).
    • Control: No-fat media.
  • Duration: Huh7 cells were cultured for 7 days; PHH cells were cultured for 4 days (due to lower viability constraints).
• Assays Performed:
  • Viability: Trypan Blue exclusion and ATP luminescence.
  • Metabolism: Media glucose/lactate/NEFA uptake, TG secretion, intracellular glycogen, and ketone bodies (3-hydroxybutyrate).
  • Lipidomics: Intracellular TG composition and De Novo Lipogenesis (DNL) quantification using stable isotope tracers (deuterium).
  • Morphology: Confocal microscopy for lipid droplet (LD) size and number.
  • Transcriptomics: Bulk RNA sequencing (5 PHH donors) with Principal Component Analysis (PCA) and Gene Set Enrichment Analysis (GSEA).

3. Key Contributions

• Physiological Media Optimization: Demonstrated that culturing Huh7 cells in human serum with chronic (7-day) exposure to mixed fatty acids shifts their phenotype to closely resemble human hepatocytes.
• Direct Comparative Analysis: Provided a head-to-head comparison of metabolic fluxes, lipid storage, and gene expression between the "gold standard" (PHH) and the optimized cell line (Huh7) under conditions mimicking early MASLD.
• Validation of Scalability: Established that Huh7 cells can serve as a robust, scalable "workhorse" model for studying early-stage lipid accumulation, provided the culture conditions are physiologically relevant.

4. Key Results

A. Viability and Morphology

• Viability: Huh7 cells showed significantly higher viability (96%) compared to PHH (68%) prior to seeding. PHH exhibited lower ATP levels and higher donor-dependent variability.
• Lipid Droplets (LDs): Both models accumulated larger LDs (>2µm²) in response to fatty acid treatment.
  • Huh7: Responded to both OPLA and POLA with increased large LDs.
  • PHH: Responded significantly to OPLA, but POLA did not induce the same magnitude of large LD formation.

B. Lipid Metabolism and Uptake

• Fatty Acid Uptake: Both models took up ~70–80% of media NEFAs. PHH showed higher variability between donors.
• IHCTG Accumulation:
  • Baseline: PHH started with significantly higher baseline IHCTG (20x higher than Huh7), suggesting they retain donor-specific steatotic characteristics.
  • Response: Both models increased IHCTG with OPLA treatment. POLA treatment tended to increase IHCTG in Huh7 but showed a weaker effect in PHH.
• Lipid Composition: The fatty acid composition of accumulated IHCTG in both models reflected the media composition (OPLA vs. POLA), indicating similar handling of exogenous lipids.
• De Novo Lipogenesis (DNL): Both models showed similar percentages of DNL-derived palmitate contribution to total TG (~12–13% in FA-treated cells), comparable to fasting DNL in healthy humans.

C. Glucose and Oxidative Metabolism

• Glucose Utilization: Huh7 cells consumed nearly all available glucose (~95% uptake), whereas PHH left ~40% of glucose in the media.
• Lactate vs. Ketogenesis:
  • Lactate: Huh7 media had significantly higher lactate levels (indicating glycolytic reliance), while PHH had lower lactate.
  • 3-Hydroxybutyrate (3-OHB): PHH media had ~3-fold higher 3-OHB levels than Huh7, suggesting PHH are more efficient at fatty acid oxidation (ketogenesis).
• Glycogen: Intracellular glycogen levels were similar across both models and treatments.

D. Transcriptomics

• Huh7: The primary source of variance was the treatment condition (OPLA/POLA vs. Control). Pathways related to proliferation and oxidative phosphorylation were upregulated; inflammatory pathways were suppressed.
• PHH: The primary source of variance was donor identity (47.3% of variance), followed by treatment. PHH showed upregulation of inflammatory and epithelial-mesenchymal transition (EMT) pathways, reflecting donor-specific baseline differences and stress responses.

5. Significance and Conclusion

• Overlap: Despite differences in baseline lipid content and viability, Huh7 cells cultured in physiologically relevant media recapitulate key metabolic features of PHHs, including fatty acid uptake, lipid droplet morphology, TG composition, and DNL responses.
• Divergence: PHHs retain unique characteristics essential for modeling MASLD progression, specifically donor-dependent variability, inflammatory stress responses, and enhanced fatty acid oxidation (ketogenesis). Huh7 cells, being transformed, show muted inflammatory responses.
• Implication: The study validates the use of Huh7 cells as a scalable, cost-effective model for investigating the initiation of MASLD (early lipid accumulation) when cultured under physiologically relevant conditions. However, for studying MASLD progression involving inflammation, fibrosis, or donor-specific genetic susceptibility, PHHs remain indispensable.

Limitations Noted: Differences in culture duration (4 vs. 7 days) and basal media composition (glucose levels) may influence metabolic fluxes. Future work should aim for 3D culture systems to better mimic liver zonation.