Richard David Feinman

It was in July of 2012 that I suddenly realized that we had won, at least scientifically. It was now clear that we had a consistent set of scientific ideas that supported the importance of insulin signaling in basic biochemistry and cell biology and that there was a continuum with the role of dietary carbohydrate restriction in obesity, diabetes or for general health.  The practical considerations, how much to eat of this, how much to eat of that, were still problematical but now we had the kernel of a scientific principle. In fact, it was not so much that we had the answer as that we had the right question.  In science, the question is frequently more important than the answer.  Of course, winning wasn’t the original idea. When my colleagues and I got into this, about ten years ago, coming from basic biochemistry, we hadn’t anticipated that it would be such a battle, that there would be so much resistance to what we thought was normal scientific practice.

Surprisingly, it was cancer studies that showed us how it all fit together.  The locale was the conference in Washington, D.C. called “Metabolism, Diet and Disease.”  The first day of the conference would probably be more accurately described as Metabolism, Drugs and Cancers.  A surprising thread in the various talks — surprising to me — was the interaction between obesity and cancer.  Less surprising, because of the numerous studies, was the effect of calorie restriction, the apparent basis for the connection between obesity and cancer: reducing calories has been shown to reliably increase longevity and to control cancer in animals. An additional thread was that the hormone insulin popped up as a major player in various experiments in cancer. The identification of downstream signaling elements — the compounds and proteins that transmit the information about cell stimulation to the interior of the cell — were important results and, again, they frequently pointed to the components of insulin pathways and even to an association between cancer and diabetes. This was not new to me — outstanding experiments had pointed the way to the critical role for insulin and I always wondered why the connection to dietary carbohydrate was not made.  In any case, the second day of the conference included presentations on dietary carbohydrate restriction. Although not listed on the organizing committee, Gary Taubes had been one of the organizers and deserves credit for a program with both cancer people and low-carbohydrate people.

My colleague Dr. Eugene Fine presented a poster at the Washington conference.  Posters don’t always have a big impact and we are grateful to Gary Taubes for making Gene’s poster known to the main speakers.  The work, now published in a journal article, describes a small study conducted with ten seriously ill cancer patients.  The study had the modest goal of showing that a ketogenic diet was a safe and feasible regimen and, in fact, the patients did well and six of ten had stable disease or partial remission. By itself, this study would usually been considered only a small step forward,  but, in fact, it was a key link in tying together the field of carbohydrate restriction and the field of cell signaling in normal and cancer cells.  A study such as this is difficult to do — patients need to have refused or failed chemotherapy; that is, for patients to agree to experimental nutrition approaches, they usually have exhausted the traditional regimens.  One way to look at the significance is that, given what we know about insulin and what we know about low carbohydrate diets, the experiment should have been done twenty years ago. There were simply two lines of thought that had to be brought together.

Now in hindsight, it seems that workers in carbohydrate restriction should have paid more attention to downstream cell signaling; we thought that the role of insulin in system biochemistry made it clear that carbohydrate restriction was built on a solid foundation. The resistance to the idea seemed incomprehensible and parochial but this conference made it clear that there was more that was needed to make a consistent biological story.  It became clear that there was a conceptual barrier to acceptance of carbohydrate restriction beyond the traditional resistance to anything associated with the Atkins diet. There was a mindset that prevented adequate synthesis of all the information.  Much of the valuable work on cell biology was informed by an approach to disease that tended to downplay the biochemistry at the upstream-stimulus level, namely, what you eat.  The major goal was to characterize the individual components in the inner working of the cell and to search for those components that were specifically malfunctioning in pathological states.  These agents, primarily proteins, could then be targeted with drugs, either directly or indirectly, sometimes through their synthesis at the genomic level. With the important observation that calorie restriction could ameliorate or prevent cancer, a link with obesity was established and, in some way, excess calories became interchangeable with weight gain. The obesity-cancer link became a serial link and it was assumed, as it is in nutrition, that preventing obesity was part of preventing associated pathologies.  (In this way, successes of carbohydrate restriction in improving cardiovascular risk factors, for example, have frequently been dismissed as due to the attendant weight loss despite evidence that the improvements in risk factors or other outcomes persist even in the absence of weight loss or even when benefits were demonstrated in eucaloric trials). The rationale, then, was that calorie restriction, the recognized approach to obesity, would point to those intracellular components that could be targeted for drug development. In many cases, this was a conceptual error.  In nutrition, it is likely that the doctrine of “a calorie is a calorie” is the single greatest impediment to understanding. The identification of obesity with excess calories, and the failure to look beyond this effect, that is, the failure to ask how the separate nutrients, carbohydrate, fat and protein individually affected cellular metabolism and how these effects interacted — there is, after all, no calorie receptor — compromised otherwise sophisticated and informative experiments. Similarly, we and other workers in carbohydrate restriction failed to see how important it was to look at downstream cellular signaling.

The Washington conference represented two parallel approaches to the problem that continually brushed up against each other.  Gary Taubes and the other organizers made it happen and, if it wasn’t at the level of the Solvay conferences, it may prove to be a landmark. I take credit for standing up during the final panel discussion (at 22:40 of the video) and making my Emperor’s New Clothes speech which explicitly said that both the cell biology of cancer and the upstream biochemistry of obesity were all about insulin and insulin was controlled primarily by carbohydrate.

“Targeting insulin inhibition as a metabolic therapy in advanced cancer”

It is not really the personal involvement that makes me describe Gene Fine’s feasibility study of low-carbohydrate diets in cancer as having such impact; we, in fact, didn’t see it that way when he started. The study treated ten seriously-ill cancer patients with low-carbohydrate ketogenic diets and showed that it was a safe and feasible regimen for such patients. The rationale followed from the fact that rapidly-growing tumors have an increased requirement for glucose (cannot metabolize fat or ketone bodies well). However, in designing therapy, simply reducing carbohydrates to give the host an advantage was unlikely to be effective since blood glucose is regulated to stay fairly constant and, in addition, the tumors may over-express GLUT1 receptors (the insulin-independent transporter for glucose); that is, tumor cells are good at getting whatever glucose is made available in the bloodstream . Ketosis, however — the state associated with very low-energy or very low-carbohydrate intake — held some promise. Fine’s hypothesis was that if we think of cancer in terms of evolution, we could imagine cancer cells as having evolved through the life of the individual, an individual whose systemic environment, in a modern setting, would be unlikely to have any significant level of ketosis. In other words, there was little selective pressure for adaptation to the use of ketone bodies as a fuel source. The host, on the other hand, was well-adapted to this substrate — it is unlikely that our ancestors regularly had “four squares a day.” (We know from biochemistry that normal tissues are well adapted to ketosis).  Some fraction of cancer cell lines, then, might be expected to deal poorly with a ketotic environment.  Some might actually be inhibited.

Although the study had a small n, several of these advanced cancer patients became stable or showed partial remission. It turned out that those who became stable or showed partial remission had the highest level of ketone bodies.  The figure below shows that the predictor of this outcome was ketonemia (ketone bodies in the blood).   Patients who demonstrated stable disease or partial remission had a three-fold higher average ketotic response compared to those with continued progressive disease. In distinction, both groups had similar calorie deficits or, as shown below, degree of weight loss.

The figure below shows individual time points with different symbols for each patient. The figure shows, as expected, an inverse correlation between ketone bodies and insulin levels.

Ketonemia versus insulinemia: the lowest insulinemia correlated with the highest ketonemia levels. Uniquely colored symbols represent values for each patient.

The impact of such a small study was that it could be tied to the vast area of research on the downstream signaling in both cancer and normal cells — that is, the changes in the cell following stimulation by an external food or hormones. Of particular interest is the effect of calorie restriction.  It was widely understood that dietary calorie restriction would have a therapeutic effect on animals with cancer and, in addition, reducing calories was the only way to prolong life in animals.  Studies following this idea showed that it was an insulin pathway that was involved. Gene’s study nailed the link for us. It is an encouraging situation, because if this is the worst of times in nutrition, it is also the golden age of biology.  Being in a position to bring all of the science to bear on what kind of dietary recommendations you can make is heady indeed.  In the next post, I will try to present an introduction to the field of cell signaling. I will discuss a cancer study .  The starting point for the study is the observation that caloric restriction is capable of causing a decrease in the growth rate of some cancers. The experimental model will be a collection of mice with human tumors implanted, so-called exo-grafts.  It turned out that some of these tumors were sensitive to dietary restriction while some were not.  What’s the difference?  Could insulin be involved?