One of the primary goals of medicine is to prevent disease. What would a disease free world look like? No cancer, no stroke, no kidney failure, no Alzheimer’s, etc, etc. For those of us that are lucky enough to be young, not much would change. In youth, the risk of getting the diseases that kill most people is minute. As we age, the structure and function of our tissues deteriorate, and our risk of disease climbs accordingly. In order to effectively prevent disease, medicine must first learn to prevent this deterioration.

What can science tell us about interventions that effectively prevent disease? Sensationalistic science journalism would have you believe that everything from supplemental vitamins to flossing will prevent disease and extend lifespan. These reports do a great disservice. By exaggerating the extent to which interventions delay disease, they obscure the few interventions that do have a large impact. So what can you do? If you smoke, quit. If you’re sedentary or overweight, get moving and lose weight. If you’ve already done all that, you’re next best option is probably to restrict your caloric intake.

In healthy model organisms, caloric restriction is just about the only intervention that substantially delays disease and increases longevity (figure below). Although there are genetic manipulations with comparable benefits, these manipulations also cause voluntary caloric restriction. It’s unclear, and in my opinion unlikely, that they would extend lifespan if they didn’t also cause caloric restriction.

Caloric restriction extends lifespan and delays disease incidence in rodents. The extension of mouse lifespan is proportional to the severity of caloric restriction (left, source). Caloric restriction also delays the onset of various degenerative processes in rats (right, source)

Caloric restriction is really hard. Yesterday, my plans to fast yielded to a potato chip and Reese’s peanut butter cup fueled binger. I’ll keep trying to eat less, but I am also going to try and figure out how caloric restriction slows the aging process and whether or not something more palatable can produce the same benefits.

In large part, figuring out how caloric restriction slows aging means figuring out how gene expression is changed in response to caloric restriction. Doing this was the primary reason why I created the Bioconductor package crossmeta. As I have previously demonstrated, the reliability of gene expression signatures are greatly improved through meta-analysis with crossmeta. In this post, I present some of my results achieved using crossmeta to perform meta-analyses of microarray studies on caloric restriction and fasting.

Although it would be useful to understand the mechanisms through which caloric restriction slows aging, this knowledge might not be necessary in order to mimic caloric restriction. In order to mimic CR, it might be possible to just measure how gene expression is changed in response to caloric restriction and then find a drug or drug combination that causes similar gene expression changes. This approach, called Connectivity Mapping, was initiated by the Broad Institute when they assayed how gene expression is changed after treating cells with 1309 different compounds. These drug signatures, as well as predicted signatures for all two-drug combinations, can be queried using the ccmap Bioconductor package.

In order to find candidate CR mimetics, I used crossmeta to perform a meta-analysis of microarray studies on caloric restriction and then used the resulting gene expression signature to query the ccmap drug signatures (figure below). Some of the highest ranking drugs have extended lifespan to some extent (e.g. melatonin and staurosporine). If you have the resources to test the other top drugs and drug combinations, feel free to do so.

In addition to finding drugs and drug combinations to mimic or reverse a gene expression signature, ccmap can also be used to determine how similar different treatments are. As an example, I performed separate meta-analyses for short- and long-term CR in mice, CR in humans, and fasting in mice. These meta-analysis signatures were queried along with the ccmap drug signatures. The results demonstrate that short-term CR in mice causes gene expression changes that are similar to those caused by long-term CR, fasting, and to a lesser extent CR in humans (figure below). These data support other evidence in suggesting that CR will increase the length and health of human life.

A meta-analysis signature of microarray data from short term CR in mice overlaps strongly with meta-analysis signatures from long term CR and fasting (orange points). The ccmap drugs (green points) and predicted drug combinations (purple points) that most strongly overlap with short term CR are also highlighted.


In rodents, caloric restriction extends lifespan and delays the onset of a variety of diseases. In order to understand the mechanisms through which caloric restriction delays aging, it will be necessary to delineate how caloric restriction affects gene expression. Even without an understanding of CR’s mechanisms, it may still be possible to mimic CR with drugs that cause gene expression changes similar to those caused by CR.

In the above post, I present the results of my efforts to obtain a reliable expression signature for CR through meta-analysis and to find drugs and drug combinations that cause similar gene expression changes. I additionally demonstrate that fasting and CR in humans cause similar gene expression changes as those caused by CR in mice. These results support the hypothesis that CR can delay the onset of disease and extend lifespan in humans.