CRAIG THOMPSON HOPED to bring Warburg’s science into the era of molecular biology, to connect the old knowledge that cancer cells take up lots of glucose and ferment it to the new knowledge of how oncogenes signal cancer cells to divide and multiply. But before moving ahead, Thompson faced a considerable obstacle: working with the metabolic enzymes would require an expertise in the biochemistry of Warburg’s era, an expertise that Thompson and most members of his lab still lacked.
Thompson called a lab meeting, and the half dozen postdocs and PhD students came to an agreement: each member of the team would be assigned a different metabolic enzyme and be required to learn how to study it in a modern lab. That meant digging up the classic papers on enzymes by Warburg and his contemporaries. It also meant spending a lot of time with Britton Chance, a legendary biochemist of Warburg’s era, then in his mid-80s, who was still at the University of Pennsylvania. “We would go over to Britton’s lab, and he would roll out old machines that hadn’t been used in 20 years,” Thompson recalled. “We would dust them off and start using them again.”1
That Chance, who died in 2010, helped facilitate Thompson’s return to Warburg’s science was somewhat ironic. Chance had exchanged a number of letters with Warburg in the 1950s and 1960s and visited Warburg in Berlin in the summer of 1966. Chance’s notes from the visit suggest that Warburg spent most of the time lecturing him on photosynthesis and cancer. It’s not clear if Chance argued with Warburg, but it’s unlikely that he found Warburg’s arguments persuasive. Chance had investigated Warburg’s claim that cancer cells from abdominal fluid rely almost exclusively on fermentation and arrived at the opposite conclusion. As far as Chance could tell, the mitochondrial power stations of the cancer cells were in good working order.2
Chance proved more helpful than Thompson could have hoped. “He knew all the tricks,” Thompson said. “He became our reference source for all the things that aren’t written in scientific articles.” Thompson’s lab had stepped into a scientific time machine. Though it was no longer in use, Chance had even held on to one of the original Warburg manometers. One afternoon the members of the lab gathered around the device to examine it, fully aware that their new investigations could be traced back to the simple U-shaped tubes Warburg had spent his life gazing at.
Even with Chance’s help, Thompson had no idea if his hunch about metabolism was correct. When it was time for Vander Heiden to decide if he would continue his work in the lab or finish the medical training portion of his MD-PhD program, Thompson told him to return to medicine. If they were truly onto something important, the field would still be there waiting for him in a few years. In the meantime, metabolism was still considered so archaic that it would be hard to even publish their research.
By 2004, seven years after he and his lab had turned to metabolism, Thompson’s doubts were fading. He had found precisely the type of gene he had been looking for. It wasn’t an entirely new discovery. The gene, AKT, had already been identified as a cancer-causing gene that drove cells to divide and grow. But Thompson found that AKT had another role, a role that appeared even more fundamental: it drove cells to swallow glucose and ferment, to switch on the Warburg effect. Thompson discovered that he could turn on “the full Warburg effect” simply by inserting a mutated AKT gene into a healthy cell, causing the cell to build overactive AKT proteins.3
AKT does not act alone. Like a commander in a castle directing the guards to lower the drawbridge so that a shipment of wood can be hauled in, the protein made by AKT directs other proteins to open passageways through a cell’s membranes so that glucose can come inside. AKT also relays the news that food has arrived to mTOR, a protein that plays a key role in governing what a cell will do next based on the supplies at hand. Depending on the type of tissue and the food available, the glucose can be broken down for spare parts, burned to heat and power the cell, or stored for later use.
In a healthy cell, AKT waits for the proper signals to arrive from growth factors before letting in glucose. These messages inform the cell that food is arriving. But when AKT is hyperactivated, as it often is in cancer, the commander of the castle no longer waits for updates. The drawbridge opens, and the cell scavenges for more and more glucose. The Warburg effect and fermentation soon follow. A once satisfied cell is now insatiable.
Thompson, Dang, and other researchers at the forefront of the metabolism revival were coming to see that metabolic enzymes that govern how cells eat and use fuel are not merely a by-product of other more fundamental cellular processes—not merely an “epiphenomenon,” as one prominent cancer gene researcher described the Warburg effect in 2004. They are woven into cancer signaling networks and sit at the origin of the disease. Cells might develop mutations that spur growth before they develop mutations that allow them to overeat, but such cells aren’t likely to survive. “If you don’t have enough cement, and you try to put a lot of bricks together,” said Dang, “you’re going to collapse.”4
According to Thompson, overeating glucose doesn’t pose an equal threat to all tissues. Cancers rarely originate in fat and muscle cells, Thompson suspects, because those tissues are better equipped to manage an influx of energy. Like a castle that has a huge cellar for excess wood, the cells are able to store much of the glucose rather than burning it. The epithelial cells that line our organs are in a more precarious situation. When AKT won’t stop telling an epithelial cell to eat, it will soon have more glucose than it can handle. It searches for a way to adjust. Thompson believes that the Warburg effect can be understood as one such adjustment. As he explained, “Getting more glucose than you want might be what the Warburg effect is.”
A cell that is overeating due to overactive AKT proteins won’t necessarily turn cancerous. Genes known as tumor suppressors are designed to put a stop to a cell’s self-destructive impulses. The cell might still manage to register internal damage and turn to apoptosis—or be induced to do so by an immune response. And even if growing cells can’t be eliminated, the emerging colony will still have to develop new tricks before turning into a deadly cancer. The cells will need to evolve to outwit immune attackers and grow new blood vessels to support their new appetites. Eventually, some of the cells in the new colony will have to break free and venture off on their own in search of new places where they can grow.
But if a cell that can eat all it wants isn’t yet cancerous, it is an emerging threat to its neighbors. If a cell manages to survive in this gluttonous state, Thompson said, “it really doesn’t ever think about dying. Now it thinks, ‘I’ve got a lot of fuel; I could do a lot of other things.’ ”5
One specific thing an overeating cell can do is build. With fermentation now supplying plenty of energy, the mitochondria no longer have to worry about doing the same and will set aside many of its nutrients for construction projects, just as soldiers in a castle, presented with a huge amount of wood, might come to see an opportunity to build new castles. In this analogy, the soldiers had not been longing for wood. They had never even asked for it. But now that the wood is arriving, their imaginations are running wild. Never mind one castle. There is enough material to support two.
The cell’s mitochondria now function like workshops within the castle, taking the wood coming in (the nutrients) and, with the help of scores of craftspeople (the enzymes), turning it into the various parts needed for new castles (the daughter cells). Biologists suspect that it is this special ability of the mitochondria to create the building blocks for new cellular components—fats, proteins, and DNA—that explains how the symbiotic relationship between the mitochondria and the host cell evolved. The host cell learned to take advantage of the energy the mitochondria could provide. But the most valuable thing it gained wasn’t more power. It was the ability to reinvent itself.6
OVEREATING CANCER CELLS use their excess glucose in other creative ways as well. In 2009, Georgia Hatzivassiliou and Kathryn Wellen, of the University of Pennsylvania, were postdocs in Craig Thompson’s lab when they made a critical finding. It was known by then that cancer isn’t merely a disease of genetic mistakes, or mutations. Every cell carries the entire genome, the blueprint for the body, in its DNA. A lung cell differs from a liver cell because different parts of the DNA code are expressed. When an individual cell alters the state it’s in by expressing different genes, it is known as an epigenetic change.
By 2009, it was also known that nutrients can influence which parts of the blueprint are read. The most famous example of the phenomenon is found in bees. There is no genetic difference between the larvae that grow into worker bees and the larvae that grow into queens. The queen develops ovaries and a larger abdomen because she is fed royal jelly. Scientists have now identified the specific gene that the royal jelly silences to make the developmental shift from worker to queen possible.
What was not clear, until Wellen and Hatzivassiliou’s breakthrough, was whether glucose plays a direct role in changing which genes are expressed. Before new genes can be expressed, the tightly wound DNA inside of the cell needs to open up, just as a paper blueprint in the form of a rolled-up scroll will need to be spread out for reading by a construction worker. Wellen and Hatzivassiliou demonstrated that excess glucose can trigger the unwinding process. And the glucose isn’t merely a signal. The glucose molecule itself will be transformed into a component of the proteins that open up portions of our DNA for gene expression. Wellen described the process as a cycle that pushes the cell in a new direction. The arrival of nutrients change which genes are expressed, and the newly expressed genes, in turn, change the metabolic state of the cell, allowing still more food to be consumed.
A decade earlier, Vander Heiden had found that metabolism governed a cell’s decision to live or die. Wellen and Hatzivassiliou’s research on epigenetics was still more evidence that metabolism influences virtually every decision a cell makes. A 2019 paper in Nature revealed that even the lactate produced via the Warburg effect can directly change which genes are expressed in a cell. As Wellen once explained it, the idea that a cell could carry out its functions without sensing its metabolic state is like trying to take trips in a car with no awareness of the gas gauge. “Your cells need to be able to assess what their nutritional resources are,” Wellen said, “and then make decisions about what they’re doing based on that information.”7
It’s not hard to appreciate why natural selection favored cells that make decisions according to how much food they have access to. All life-forms evolved from single-celled organisms. When food is scarce, an organism’s best chance at multiplying (and passing along its genes) is to conserve resources and to survive the famine for as long as possible, so that it can reproduce when food finally arrives. Researchers believe this is why eating very few calories has been shown to extend life in so many different species. Robert Koch discovered the same phenomenon when he realized that the bacteria that cause anthrax would turn into resilient spores when no food was available. When food is abundant, the opposite strategy makes the most sense. The most successful single-celled organisms will be the ones that eat and multiply as soon as they come across food.
As Thompson explained, a cancer cell behaves like a growing single-celled organism that is binging on glucose. “When they see food in their environment, they move to capture as much as they can,” he said. “And when that food exceeds their need to survive, they begin to make copies of themselves, as many as they can.”
To drive this point home, Thompson once showed students a slide with images of a speck of mold growing across a slice of white bread. The speck gets bigger and bigger until it has turned into an ugly dark splotch. The title of the slide, “Everyone’s First Cancer Experiment,” is a starkly simple and vivid explanation of how cancer works. But the lesson did not end with this fundamental comparison between how cancer cells and microorganisms eat. As the mold grows across the bread, Thompson told the students, its supply of nutrients or water will run low. When that happens, some of the mold cells will break away and migrate to another piece of the bread. In the next slide, the students saw that the slice of white bread now had not only a big dark splotch, but also another smaller splotch on the other side of the bread.
“This is exactly what cancer does,” Thompson said. “It starts to grow in one place, it accumulates until it runs out of food, and then it finds a way to get to the new home.” This search for a new living space is metastasis, the spread of the disease from one place in the body to another, and it is typically why cancer kills us.8
BY THE END OF the first decade of the twenty-first century, metabolism researchers were talking about cancer in a way that felt not only new but startlingly so. And yet, they had also brought cancer science full circle. As Warburg put it upon discovering fermentation in cancer cells, “The most important fact that we have discovered in regard to the metabolism of carcinoma tissue is, we believe, that carcinoma tissue” behaves “like yeast.”9
Warburg had discovered something of immense importance in the 1920s: the overeating and fermenting of glucose is as fundamental to cancer as he had always argued. But when explaining the fermentation, Warburg, until the very end, found his vision obscured by the long shadow of Pasteur, who had assumed that fermentation arose only when oxygen was in short supply and respiration wasn’t possible. Warburg never grasped that Pasteur had himself missed something crucial: If yeast have all the glucose and nitrogen they want, they will ferment and grow regardless of how much oxygen is available. They ferment the glucose not out of necessity but because it allows them to process more cellular building supplies and grow more rapidly.
Even so, Warburg’s explanations were not as inaccurate as his critics have sometimes argued. Though most cancers are no longer thought to have damaged mitochondrial power stations in the way Warburg described, there is a growing appreciation among cancer scientists that mitochondria play a critical role in the transformation of a normal cell into a cancer cell. According to Celeste Simon, of the University of Pennsylvania, one of the leading authorities on the respiration of cancer cells, a shrinking supply of oxygen does play a role in every cancer as the disease progresses. And as cancers grow, they will invariably end up with a limited supply of oxygen from the blood and ferment more glucose as a result. But, Simon explained, that a lack of oxygen contributes to the progression of cancer does not necessarily mean it is the underlying cause of the disease. “Warburg was a real visionary,” Simon said, but “some of his principles are probably not quite going to hold up the way he thought they would.”10
In the 1960s, Warburg continued to search for ways to treat and prevent cancer by providing cells with the vitamins, or coenzymes, needed to keep cellular breathing running smoothly. If the power stations could maintain their function, he reasoned, the shift to the backup generators (fermentation) might be slowed or stopped. But in his last years, Warburg also renewed his long-standing interest in another approach: shutting down the backup generators by starving cancer cells of the glucose they rely on. The logic was straightforward. It might be impossible to target only the cancer cells, but such cells would be especially vulnerable to glucose deprivation. As Warburg wrote in 1926, “An overpopulated city is more sensitive to stoppage of food supply than a normally populated city, even when the inhabitants can all endure hunger alike.”11
Toward the end of his life, seemingly aware that he was now too old to conduct experimental cancer research on his own, Warburg found a new scientific partner: Manfred von Ardenne, a brilliant German physicist who had worked on the Soviet atomic bomb project after the war and who arranged special permission for Warburg to travel across the Berlin Wall. In 1965, von Ardenne announced a new cancer therapy that involved heating the body to as high as 110°F for over half an hour while giving patients DL-glyceraldehyde, a compound Warburg was then testing as a means to disrupting a cell’s ability to ferment glucose.
Some 45 years later, Chi Van Dang, then at Johns Hopkins, decided to investigate this old idea of disrupting the fermentation of cancer cells. He began with a simple experiment designed to understand how cancerous and noncancerous growing cells respond to a lack of food. In the presence of glucose, Dang found, both healthy growing cells and cancerous cells took up the glucose and used it to power their growth. But when Dang removed the glucose from both groups of cells, the differences immediately became apparent. The healthy cells, sensing the lack of nutrients, slowed down their metabolism and shifted into a resting state. The cancer cells couldn’t stop themselves. In becoming cancerous, they had lost the internal checkpoints and feedback loops that tell a cell there’s nothing left to eat. They were like addicts, and when they couldn’t get their glucose fix, they would die. “Contrary to popular belief,” the University of Southern California biologist Valter Longo said, “cancer cells are dumb.”
Dang’s preliminary research led him and others to revisit the idea of starving cancer with modern therapies. But if the new molecular biologists turned metabolism experts were returning to ideas from generations past, they were also returning to a hard lesson that Warburg and others had learned long before: cancers are remarkably adept at finding new ways to fuel their growth. The molecular pathways, as MIT’s David Sabatini explained, “can go in many, many different directions and change very, very quickly.”
“You block glucose, they use glutamine,” as Dang put it, referring to another primary fuel used by cancers. “You block glucose and glutamine, they might be able to use fatty acids.”12
These challenges haven’t stopped the Warburg revival from leading to new cancer treatments. A biotechnology company started by Thompson and several other prominent metabolism researchers recently created a drug that treats one type of leukemia by inhibiting a mutated form of the metabolic enzyme IDH-2. The treatment is said to be the most significant advance for this specific leukemia in decades.13
At the same time, the metabolism revival is prompting researchers to rethink traditional cancer therapies. Matthew Vander Heiden is now exploring why a given chemotherapy will work against one cancer but not against another, even when the two cancers have identical mutations. Vander Heiden suspects the answer has a lot to do with the types of nutrients that were available in the particular tissue where the cancer formed. “It has nothing to do with the genetic mutations,” Vander Heiden said.14
FOR ALL THE EXCITEMENT surrounding new metabolism-focused therapies, and for all the money investors are now pouring into such drugs, the most important findings to emerge from the return to metabolism may ultimately be about cancer prevention rather than treatment. If Warburg was correct, in the broadest sense, about cancer originating in changes to how cells eat, but wrong that the process always begins with a struggle to breathe, then the most important question, when it comes to prevention, is simple: What does cause the Warburg effect?
One possibility is that the cancer cell’s huge appetite for glucose is strictly a product of unlucky mutations in genes that control cellular eating. If that’s the case, the Warburg revival will still have fundamentally changed our understanding of cancer, and yet it will not have told us anything new about cancer prevention.
But at the turn of the twenty-first century, something unexpected happened: like two lost legs of an expedition suddenly encountering each other after years apart, two fields of cancer science that had long been separated began to converge. With each new advance showing that cancer is tied to how our cells eat, researchers studying cancer trends were arriving at advances of their own showing that cancer is tied to being overweight. Not even Doll and Peto had seen it coming. In “The Causes of Cancer,” they highlighted the importance of Tannenbaum’s studies of overfed mice, but with the exception of cancers of the endometrium and gallbladder in women, they remained unconvinced that overweight people were more likely to have cancer. The evidence for a connection between obesity in humans and cancer, they wrote in 1981, was “not particularly impressive.”15
Doll and Peto could not have anticipated how soon those words would be obsolete. In 1982, researchers at the American Cancer Society selected a population of more than 900,000 Americans and asked them to fill out surveys that included basic personal information, such as their weight, height, and smoking habits. By 1998, almost 60,000 of the participants had died of cancer, and the American Cancer Society was anxious to figure out why.
Among those digging through the data was Eugenia Calle, an American Cancer Society epidemiologist. An exercise enthusiast known to drag her colleagues to the gym, Calle wondered whether the participants in the study who had succumbed to cancer were more likely to have been overweight. The question had been asked before, but no one had ever looked for the relationship in such a large data set.
After Calle finished her number crunching, many researchers would never think of obesity and cancer in the same way again. Her study, published in the New England Journal of Medicine in 2003, found that being overweight or obese increased the risk of nearly every cancer she looked at. Compared with a woman of normal weight, the women in the highest-weight category were 62 percent more likely to die from cancer. The most obese men, in turn, were 52 percent more likely to die from cancer. Calle estimated that every year some 90,000 Americans were losing their lives to cancers linked to being overweight or obese.
Even Calle was surprised by the strength of the evidence. The connection between excess weight and cancer, she said, “was the rule more than the exception.”16 Tragically, in 2009, Calle was murdered during a robbery at her Atlanta condominium. In the following years, the importance of her contribution to our understanding of cancer would grow increasingly clear. As we have grown heavier—nearly three-quarters of American adults are now overweight or obese—the association with cancer has grown still more convincing. A 2017 Center for Disease Control analysis concluded that more than 600,000 Americans had been diagnosed with body-fat-related cancers in 2014 alone.
This harrowing number does not include deaths from prostate cancer and other cancers that may also be linked to obesity. (The evidence was not quite as strong for these other cancers.) As Rebecca Siegel, scientific director of Surveillance Research at the American Cancer Society, put it, when it comes to how obesity will influence cancer rates, we currently might be seeing only “the tip of the iceberg.”17
IT IS OFTEN SAID that we are losing the war on cancer. The numbers from the obesity studies suggest that we are losing it, in large part, to what we eat and drink each day. And yet, though it is now commonly said that obesity will soon surpass smoking as the leading cause of preventable cancers, these studies do not show that obesity causes cancer. It is entirely possible that excess weight leads to cancer, but it is also possible that another underlying phenomenon is at work that is driving both obesity and cancer.
Understanding the obesity-cancer connection has been challenging, in part, because getting to the bottom of the obesity epidemic itself has been challenging. The simple explanation—that people started to eat more food—leaves too many questions unanswered. Many societies have had an abundance of food without developing high rates of obesity or cancer. Tellingly, animals in the wild do not grow fat and sick when they are fortunate enough to get all the food they want. For animals, coming into an abundance of food is akin to winning natural selection’s lottery: a well-fed population can be expected to grow in number and thrive. Likewise, prior to the nineteenth century, obesity and cancer were rare regardless of how much food a human population might have. As one British specialist noted in 1908, it seemed conceivable that eating too much led to cancer, but it was not at all clear why overeating caused pathological rather than normal growths.18
There are other complicating factors as well. The Warburg effect involves cells overeating glucose, but our bodies are designed to keep blood glucose levels within a tight range regardless of how much food we consume. And Thompson’s own research had demonstrated that healthy cells will starve unless told to eat.
Obesity had been definitively linked to cancer at the very moment that scientists were rediscovering Warburg’s metabolism research. It seemed clear that the two phenomena fit together, that the way we eat must somehow be connected to the way that cancers eat, but the precise nature of the connection was not at all obvious. Some seventy years after Warburg first observed the ravenous appetite of the cancer cell, the mystery of the cancer-diet connection remained to be solved.