On March 4, scientists reported that people who ate a diet rich in protein during their middle-aged years (ages 50-65) were four times more likely to die from cancer than those who ate a low-protein diet, and had a 75 percent increase in overall mortality.
Valter Longo of the University of Southern California, who led the study, likened the risk of a high-protein diet to that of smoking. The key to these startling findings—underreported in many media accounts—is the role of a molecule called insulin-like growth factor-1 (IGF-1).
This same molecule was the prime player in a highly publicized study published three years ago by Longo, along with Jaime Guevara-Aguirre of the Institute of Endocrinology, Metabolism and Reproduction in Quito, Ecuador, and their colleagues, that looked at patients in a remote area of southern Ecuador. The subjects of the study had a rare mutation that interfered with the normal functioning of IGF-1, blocking the body's normal growth pathway. As a result, adults with the disease (known as Laron syndrome) rarely grew beyond 3 1/2 feet tall.
But the scientists observed an unusually beneficial side effect of the disease: The "Larons" did not get cancer, according to the researchers, and did not get diabetes, even though they were often obese. (Read "On Beyond 100" in National Geographic magazine.)
For the recent study, published in the journal Cell Metabolism, the scientists analyzed data from the Center for Disease Control and Prevention's National Health and Nutrition Examination Survey (NHANES).
In an unexpected twist, the researchers also reported evidence that after age 65, a high-protein diet seemed to have the opposite effect: It protected against mortality.
In other words, a low-protein diet during middle age appears to be protective, but the same low-protein diet in older adults has deleterious effects.
To explore these puzzling results and the role of IGF-1, National Geographic spoke with Longo earlier this week.
For a number of years you've been studying a population in Ecuador with a growth disorder, and in 2011 you reported the relative absence of cancer and diabetes mortality in people with Laron syndrome. How does this new report tie in to that earlier research?
The Laron study was a continuation of our work that we started 20 years earlier. When we put it all together, the obvious thing was, in both in mice and humans, when they eat [a low-protein diet], we predicted that they would have low levels of IGF-1. So I started bugging everybody in the world that had any databases. It turned out that here at USC, downstairs from our lab, we had some people who had been studying NHANES for years. It was a perfect database to look at, because it was nationally representative; it's done by the CDC, and the sample and the measurements are considered to be very reliable.
There are several surprising claims in this new study, but let's talk about perhaps the biggest: A high-protein diet during middle age was associated with a much greater risk of dying of any cause, especially cancer, than a low-protein diet. In fact, you liken the mortality risk of the high-protein diet to that of smoking. How would you explain this association?
The more protein you eat—and this is very clearly shown by clinical work by others—the more IGF-1 activity you have. And we see the same in mice. There was a paper that came out in the same issue of Cell Metabolism, looking at all kinds of combinations of macronutrients, and in fact the best combination for longevity was low-protein, high carbs.
And just to be clear: When you say high-protein or medium-protein diet versus low protein, what exactly do you mean?
High protein means 20 percent or more of the calories is coming from protein; moderate is 10 to 20 percent; low is less than 10. We looked at animal-based and plant-based proteins.
One of the other surprises in the study is what you refer to as a "dramatic switch"—a point when a low-protein diet seems to stop being protective and starts having negative effects. How did you identify that shift, at what age does it occur, and why do you think it happens?
So when I first sat down with the epidemiologist, the analysis was done for the whole survey, from ages 50 on up. And we saw nothing. But we were seeing in mice that if you starve a young mouse or a middle-aged mouse, it does very well. But we started seeing that when we starved old mice in the same way, they were struggling. That's when I started thinking, "Hmm, I think that there are two major phases, if not more."
I went back to the epidemiologist and said, "Can you reanalyze the data and break it down into two groups? Up to 65 years old, and after 65. And sure enough, she came back and said to me, "Valter, it's incredible." Now we know why we didn't see it before. In the original analysis the effects in the old were offsetting the effects in the middle-aged.
You also conducted complementary studies of high-protein and low-protein diets in mice and yeast. Those studies suggested a role for IGF-1. Can you explain your working hypothesis about the possible role of IGF-1 in cancer, and in mortality in general?
We know that in either animals or humans, if the diet is low in protein, the level of IGF-1 goes down—or way down, depending how severely you restrict protein. If you look at humans, it's a pretty large change [in IGF-1 levels] between the high-protein group and the lowest-protein group, with the moderate-protein group in between.
So then how does it lead to cancer? Well, that was basically our paper three years ago on the Larons, which showed that if you take the blood from the Larons [which essentially has no IGF-1] and you expose human cells to it, and then you try to damage those cells to make them precancer cells, the absence of IGF-1 has a protective effect in two ways. First, as we have known for years from yeast, the lack of IGF-1 protects DNA from damage in the first place. Second, the lower IGF-1 now makes the mammalian cell more likely to go into what is called "programmed cell death," or apoptosis.
And the cell dying in that instance is a good thing?
Yes—because it's a cell that's about to become a cancer cell. So there's a dual protective effect. You try to prevent DNA damage from happening in the first place, but when the damage does happen, the [precancerous] cell is much more likely to die.
We showed this very clearly in this new study, where a cancer [in mice] progresses more slowly on the low-protein diet.
This new paper suggests that animal proteins, as in red meat, have a different effect on this biochemical pathway than plant proteins, as in vegetables. At the biochemical level, why would cells see a difference between animal proteins and plant proteins?
When we did the mouse studies, we did see a trend for decreasing IGF-1 with soy protein versus animal protein. We are doing more studies now, looking at different protein sources and IGF-1 levels. We suspect that the amino acid profile of plant proteins is going to control IGF-1 differently, but we need to test that more carefully.
Just to make sure the proper caveats are introduced here: This study finds an association between diet and mortality, but doesn't prove that a high-protein diet in middle age causes an increased risk of dying. And one of the limitations with the NHANES study is that it is based on a single, one-time recall of a person's diet over a 24-hour period, and then you're associating mortality up to 18 years later from that single dietary report.
Well, there's not many things you can do that are not going to have limitations. I think this needs to be viewed in the context of a number of other studies done, including the Laron study.
Frank Hu at Harvard looked at low carbohydrate diets, and even if he didn't break it down into age segments, he still saw that people eating low carbohydrates in fact also had high levels of animal protein intake, and they were the ones with the highest overall mortality risk, the highest cancer risk, the highest cardiovascular risk.
And certainly if you look at long-lived populations around the world, this is really consistent with that. So whether it's the California Adventists, the Okinawans, the southern Italians—the common denominator seems to be a low-protein, high-plant-food-based diet.
Every experimental result points to the next experiment. What's the next step here?
We're doing a clinical trial that we started a year ago, and in that clinical trial we're basically saying to people, "Don't change what you do in terms of diet." But we do ask them to switch once a month for five days to a diet that mimics fasting. We want to see if this periodic exposure to these more extreme diets can actually reprogram the system so the IGF-1 pathway is switched to a more beneficial mode, similar to what you would see in somebody having the perfect diet.