High-fat diets may lead to liver cells becoming cancerous

Long before a liver tumor appears, a high-fat diet can push liver cells into a risky survival mode. That is the central finding of a new study led by researchers at the Massachusetts Institute of Technology, including Alex K. Shalek of MIT’s Institute for Medical Engineering and Sciences and the Koch Institute for Integrative Cancer Research.

The senior team also includes Ömer Yilmaz, an MIT biology associate professor and Koch Institute member, and Wolfram Goessling, co-director of the Harvard-MIT Program in Health Sciences and Technology. Their work, published in Cell, traces how chronic metabolic stress can make hepatocytes, the liver’s main workhorse cells, behave more like immature cells that are easier to turn cancerous.

The study focuses on a major public health problem: metabolic dysfunction-associated steatotic liver disease, or MASLD. It affects more than one-third of people worldwide, according to the paper. In its more severe form, metabolic dysfunction-associated steatohepatitis, or MASH, the liver can become inflamed and scarred. Over time, that damage can progress to cirrhosis, organ failure, and hepatocellular carcinoma, the most common form of liver cancer.

In the new work, the researchers argue that cancer risk rises even before the liver reaches cirrhosis. They point to a process that is not just about mutations. Instead, they show liver cells shifting into progressively dysfunctional states under long-term dietary stress, setting up conditions that can favor cancer later.

To understand what a high-fat diet does inside the liver, the team used mice and fed them a diet designed to trigger disease through food alone. They then tracked changes over months as the liver moved from fat buildup to inflammation, then scarring, and finally spontaneous tumors.

At key time points, the researchers used single-cell and single-nucleus sequencing methods to watch which genes switched on and off in individual liver cells. They paid special attention to hepatocytes, since these cells run much of the liver’s daily business, from processing nutrients to producing proteins and supporting detoxification.

The results showed a steady trade-off. Under chronic metabolic stress, hepatocytes turned on genes tied to survival and growth. At the same time, they dialed down genes linked to normal liver identity and function.

“This really looks like a trade-off, prioritizing what’s good for the individual cell to stay alive in a stressful environment, at the expense of what the collective tissue should be doing,” said Constantine Tzouanas, an MIT graduate student and co-first author.

Across the mouse timeline, nearly all animals on the high-fat diet developed liver cancer by the end of the study. The researchers say the shift toward an immature state helps explain why. Once cells lose parts of their mature identity, they can become more willing to divide and less constrained by the usual rules of liver function.

A central observation was that hepatocytes can revert toward an immature, development-like program. In simpler terms, the cells start to resemble a more stem-cell-like state. That change appears to offer protection in a harsh environment. Yet it also leaves cells poised to take the next step into cancer if the wrong mutation arrives later.

“These cells have already turned on the same genes that they’re going to need to become cancerous,” Tzouanas said. “They’ve already shifted away from the mature identity that would otherwise drag down their ability to proliferate. Once a cell picks up the wrong mutation, then it’s really off to the races and they’ve already gotten a head start on some of those hallmarks of cancer.”

The team also identified factors that may help control this shift. Those include transcription factors, which act like genetic switches. One highlighted candidate is SOX4, which usually appears during fetal development and in only a small number of adult tissues, not the liver.

"Our study did not rely on gene readouts alone. We also measured protein changes and tracked how liver fats shifted over time, Shalek told The Brighter Side of News.

"We found declines in proteins tied to normal liver function, including HNF4A, a key regulator of hepatocyte identity. We also observed reduced levels of secreted proteins made by hepatocytes, including albumin, a common clinical marker'.," he continued.

Meanwhile, they found signs of rising pathways linked to survival, regeneration, and signaling linked to cancer. Lipid measurements supported the same story. Early on, triglycerides climbed as the liver stored more fat. Later, cholesterol- and sphingolipid-related species rose more strongly, matching gene signals that suggested the liver was redirecting its chemistry under chronic stress.

One detail stood out. Genes involved in cholesterol synthesis rose, while a gene tied to ketone production fell. The study frames this as a resource shift inside hepatocytes, as they reroute shared chemical building blocks toward pathways that may help survival, but also reshape the liver’s internal balance.

To probe cause and effect, the researchers zeroed in on HMGCS2, a key enzyme for ketogenesis. Its expression dropped during metabolic stress in the mouse model, and lower HMGCS2 levels also linked to worse outcomes in human liver cancer datasets.

The team then created a liver-specific mouse knockout that lacked HMGCS2 in hepatocytes. Under a high-fat diet, these mice did not simply gain more weight than others. Instead, they showed signs of worse liver damage and higher circulating cholesterol, paired with lower ketone production after fasting.

Sequencing results suggested that losing HMGCS2 pushed hepatocytes further along the stress path, making their gene patterns look more “advanced” than diet alone. When the researchers used a tumor-triggering system during chronic dietary stress, the HMGCS2 knockout mice developed greater tumor outgrowth and larger nodules.

The paper also reports that, in multiple human cohorts, reduced HMGCS2 expression in non-cancerous liver tissue predicted higher liver cancer risk years later, in some cases up to 15 years. In the study’s framing, this supports the idea that the liver can carry a kind of biological “stress memory” that shapes cancer risk long before a diagnosis.

After mapping the shift in mice, the researchers tested whether human disease showed a similar arc. They analyzed liver tissue data from patients at different stages of MASLD and from people who had liver disease but no cancer yet.

They found the same basic pattern: genes needed for normal liver function fell over time, while genes linked to immature programs rose. The study also reports that gene-expression patterns helped predict survival outcomes after tumors developed.

“Patients who had higher expression of these pro-cell-survival genes that are turned on with high-fat diet survived for less time after tumors developed,” Tzouanas said. “And if a patient has lower expression of genes that support the functions that the liver normally performs, they also survive for less time.”

In mice, the timeline to cancer can be about a year. In humans, the researchers estimate the process may unfold over a much longer period, possibly around 20 years, though it can vary with alcohol use, viral infections, and other risk factors.

This work reframes liver cancer risk as more than a late-stage surprise. If hepatocytes gradually shift into a survival-first, immature-like state during years of metabolic stress, then clinicians and researchers may be able to spot danger earlier by looking for these gene and protein signatures in diseased liver tissue. That could improve how doctors identify which patients with MASLD or MASH face the greatest long-term cancer risk, even before cirrhosis.

The findings also point to new prevention strategies. The study highlights molecular “switches” that appear to steer hepatocytes toward risky cell states. Some targets already connect to active drug development. A drug targeting the thyroid hormone receptor has been approved for MASH fibrosis, and an approach that activates HMGCS2 is in clinical trials for steatotic liver disease, according to the study text.

Over time, that line of work could lead to treatments aimed not only at reducing fat and inflammation, but also at preventing the cellular reprogramming that primes tumors. The researchers also plan to test whether shifting back to a normal diet, or using weight-loss drugs such as GLP-1 agonists, can reverse parts of the stress program.

“We now have all these new molecular targets and a better understanding of what is underlying the biology, which could give us new angles to improve outcomes for patients,” Shalek said.

Research findings are available online in the journal Cell.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.