The cystic fibrosis gene 20 years later

There's a Genetics "News Focus" piece in today's Science magazine called "The Promise of a Cure: 20 Years and Counting", written by Jennifer Couzin-Frankel. We can't link to it, but we will describe it a bit because it relates to several of the issues we've written about ourselves this week.

The piece is about cystic fibrosis (CF), the gene for which was found 20 years ago, and announced with great excitement and hope for a cure. Gene therapy was the immediate focus, and many labs began the hunt for a method to deliver healthy copies of the gene into the lungs of people with CF. However, years into the search, it became apparent that gene therapy, at least as it was being tried, was not going to work, and research turned to better treatment and drug therapies. In the last 20 years, life expectancy for people with CF has risen by about 10 years, to 37. This isn't due to anything genetics has taught scientists, but to more aggressive and earlier treatment to keep lungs clear.

Even if the early promise of a cure hasn't yet been fulfilled,

in a funny way, "science has benefited more from the CF gene than CF has benefited from the science," says John Riordan, a biochemist

now at the University of North Carolina at Chapel Hill and one of the co-discoverers of the CF gene. Much has been learned about the genetics and physiology of CF, and new approaches to gene therapy may be on the horizon. Millions of dollars have been spent over the last decade or so on developing new drugs to treat the disease, and two are now in clinical trials. At least 1000 different mutations in the CF gene have been reported over 20 years; this isn't unusual, even for 'simple' single gene diseases. One of these mutations is quite common, while many have been seen only once; again, this is the usual case. As we wrote in our post on June 15, one of the speakers at the Bristol meeting described instances when it is useful to know a causative allele for determining therapy, and indeed both of the new drugs now in trials are targeted toward one of the known mutations for CF--one is meant to be helpful in patients with the most common mutation, while the other targets those with a mutation that explains the disease in only a few percent of patients.

Although we have never worked on cystic fibrosis, we well remember the excitement with which the finding of the CF gene was announced, and the belief that gene therapy was the next frontier. It was very sobering and discouraging to many when gene therapy turned out to be much trickier than anticipated (will stem cell therapy, so much hyped today, be equally recalcitrant?). Much has been learned about cystic fibrosis in the 20 years since the gene was identified, and it's still possible that genetics will contribute to treatment or even a cure. But, it's a cautionary tale.

The tale is also informative about the various strategies and approaches. The first step, at least as far as genetics goes, is to identify the causal gene(s). Once this is done, specific studies can be done focused on patients (and, where needed comparative control individuals), to identify the spectrum of mutations at the gene(s) and their clinical effects. This makes it possible to focus molecular and cell biology on the genes effects. Mapping or genomewide association studies contribute little more.

One difficulty in gene therapy is getting a good experimental animal model, and this was frustratingly true for CF; in future years, hopefully cell-culture rather than transgenic animal models will be developed (perhaps based on expression manipulation in stem cells of various kinds). This array of knowledge and tools can then lead to therapy that is 'genetic' in that it is directed against the genetic physiology, but is not necessarily genetic either in terms of germ-line alterations to protect future generations, nor not even necessarily related directly to the specific mutation, to DNA or mRNA, etc. It can 'simply' be an attack on the pathophysiology that might take any number of forms.

The motivation of GWAS to identify previously unknown pathways fall somewhere in the middle ground. Mapping (in families, not association studies) found the CF gene, which showed that ion channel biology was involved. Once that happened, mapping studies were no longer needed, and new variation in the CFTR gene could be found by sequencing affected persons. Once a pathway has been found by mapping or any other method, the pathway can be studied directly.

A cogent contemporary question is when can we assume that most pathways have been identified by mapping? Pathways often involve many genes, any one of which could be mutated to have a major effect. If it is highly penetrant and viable, it should be findable in families, in which case 'linkage' analysis is the statistically best approach. Or, well-designed but not massively large genomewide association studies in appropriate samples could identify the offending pathway member.

So long as any gene in the network has mutations with detectable effect, the network is found and its remaining members can then be studied molecularly and in affected people. One might expect that this would be the case for most important gene networks and hence most diseases. In this sense, the new wave of very large genomewide association studies probably is not going to be all that useful, even if it certainly will make important findings now and then. It is here that the discussion of the relative importance of the investment in large GWAS should be.

The Science article also points out both the idea that once found, study of causal genes probably can lead to effective treatments, as well as that it takes time. Media hyperbole doesn't help, or put another way, isn't necessary for support to be given to research on genes that really are important. Likewise, not all avenues need to be followed up if they are very costly, once it becomes possible to focus on known causal genes. Existing resources can be more focused. And currently, there are hundreds of disease-related genes where this could be the case.