Thursday, February 27, 2014

Genome sequencing for rare diseases

Rare Disease Day
Tomorrow is Rare Disease Day 2014, an annual event intended to raise awareness of rare diseases and their impact on patients' lives.  In recognition of this day, we wish each year to remind readers of the importance of investing in research on rare diseases.  Rare diseases often strike early and hard, devastating lives that have yet just begun, with tragic consequences for the lives of those directly affected, and for the families who care for them.

Common, later onset diseases and problems are serious and can also be devastating.  They are worth understanding, but often they are common because they aren't due to severe or fatal gene mutations, and often aren't severe for years.  They accumulate their underlying pathology gradually over years or decades, often if not typically due to lifestyle factors that can be avoided or exposures reduced -- delaying or preventing the disease from ever striking.

But not so for devastating early onset diseases.  Their rarity means that not nearly the same effort or funds are spent on them, even though in reality they may be more understandable than common diseases if the effort were spent.

We have in the past suggested that in many ways there's no real advantage for healthy people to have their genome sequenced.  Today we are reposting a piece from a year or so ago (reworked a bit) in which we suggest that it's a different story for people with rare, unexplained diseases.  Rather than sequencing a genome for predictive reasons, which we believe will not often be all that useful, sequencing in the hopes of explaining a perplexing disease is another matter entirely.

We often criticize the spending of taxpayer money on what we see as fruitless gene searches, but there are traits and diseases that truly are genetic, or, if associated genes aren't yet known, physiologically act as if they are. That means that the trait seems closely related to gene function in ways that could indicate that genetic variants are responsible for the variation of the trait from 'normal'.  We think these are where the genetics money should be spent.  Cancer is one example, though usually of late onset, because it is about a lineage of cells behaving abnormally for their context, that arises during live and thus is amenable to genetic approaches.  Pediatric diseases and disorders are further examples, but there are certainly others.

Finding causal genes even for what look like 'single-gene disorders' isn't always easy, and even when it can be done, certainly doesn't always lead to therapy.  At the very least, when it is possible, it can be an important and valued piece in the puzzle of who one is.  And we think this is where heavy-duty research investment should be made.


Pink boulder, Shirehampton Road, Bristol
For no obvious reason, one of the boulders lining the north side of Shirehampton Road has been painted pink. By whom and why - unknown.
© Copyright Jaggery and licensed for reuse under this Creative Commons Licence.

Periodic paralysis -- a single gene disorder striking close to home 
The periodic paralyses are a rare set of ion channel disorders that are still not well-understood.  Partly of course it's because they are so rare (prevalence is on the order of 1 in 100,000 to 200,000), and partly because the normal functioning of ion channels isn't itself well-understood.  Channelopathies themselves are not rare -- epilepsy and cystic fibrosis are more well-known examples of ion channel dysfunction -- and now that ion channel-related diseases have been recognized, progress on understanding them is being made.

As the Periodic Paralysis Association website says,
Periodic Paralysis is a group of disorders whereby patients become weak due to triggers such as rest after exercise or certain foods.  These disorders are part of a broader class of disorders called ion channelopathies, in which a genetic defect in a muscle ion channel results in symptoms of episodic stiffness or weakness in response to certain triggers.
There are various periodic paralyses (hypo and hyperkalemic pp, and Anderson Tawil syndrome), and they are often difficult to diagnose.  Indeed, many people go for years without a diagnosis.  Most physicians may have heard of them once, long ago but very often it's not a diagnosis that immediately comes to mind when faced with someone even with classic symptoms.  Indeed, even now but especially in the past, people with these disorders could live a lifetime with neither diagnosis nor therapy -- an extensive bit of sleuthing has led us to think the famous pioneering Victorian poet, Elizabeth Barrett Browning, who was notoriously debilitated with a mysterious disease about which she wrote prolifically in her love letters to the poet (and her future husband) Robert Browning, had HKPP, as we surmised in detail here.  The disorder wasn't recognized when she was alive, so it's no surprise that EBB's doctors were completely at a loss as to what was causing her perpetual weakness. We'll talk more about this tomorrow.  It's more of a surprise when the diagnosis is missed today, as it needn't be; a computer search for diseases associated with abnormally low or high serum potassium should put a physician on the trail.  But it too often is.

As regular readers of MT know, we write a lot about complex diseases, and about how the idea of genes 'for' disease can be a naive one.  For many traits, perhaps most traits, in organisms, multiple genes contribute and most of the genetic aspect of variation of the trait is due to multiple, small contributions from many different genes.  Each individual with a given trait value (like, say blood pressure, height, glucose or cholesterol levels) has a unique genotype that contributes to that value (not to mention environmental contributors).  The hope that it will be possible to identify simple causation is manifest, and understandable, even if the reality is different.  That hope is what feeds the GWASification of everything, that is currently at such a fevered pitch.

So, it is a bit ironic that we have a daughter with HKPP, a disorder that is generally considered to be a monogenic condition (caused by a single mutation). To date, causal mutations have been identified in three ion channel genes, but this doesn't explain the disease in all those who have it.  Some of the known mutations disrupt the structure of the channel so that it malfunctions in response to specific environmental triggers.  One is a sodium channel gene, and one is a calcium channel gene, which is interesting because calcium channels don't seem to even be used by skeletal muscles as sodium channels are, so it's difficult to understand why disrupted calcium channels can shut down these muscles, but it seems to be.  Insulin is also related to the process, but the periodic paralyses don't seem to be related to diabetes.  Is the trait due to a channel mutation, or mutations of more general sorts that affect the ion concentrations that normal channels respond to?  It seems to be a mix.  But it doesn't seem to be a hopeless polygenic sea of contributing variants, because the symptomatology is so specific and localized to specific tissue.

The problem exemplifies the importance of partial sequestration and modularity, and others of the basic principles of life that we often write about.  An ion channel is used by a cell to sense and relate to its environment: to shove excess negative or positive molecules out or import them in, to keep the ionic or pH (chemical) balance suitable for the reactions that must occur inside the cell, and an appropriate difference from the outside world of, say, the blood stream.  In simplified terms, if the cell is too salty relative to the blood stream, or too unsalty, the cell can burst, or be drained of water, or be unable to import needed ingredients or export waste, etc.  It's a fundamental way that cells relate to their environment.  And many different genes are involved in the ion channels, or chemical pores, through which these molecules shuffle in and out.

Nonetheless, as we've blogged about before (here, e.g.), even these 'simple' processes are complex.  Many genes may be involved, at least among different cases, but it is not always the case that multiple minor contributions from different genes are required to add up to trouble.  In some cases, and HKPP may be one, there is what is called multiple unilocus causation:  In a given case, only one variant gene may be responsible, but in different cases different genes -- but only one gene per case.

Some people can trace a specific form of periodic paralysis through generations in their family, and others are the only known family member to be affected.  And, the same mutation in a single family can have very different symptoms, from very infrequent, or even no attacks of weakness, to waking daily with paralysis.  And, essentially the same phenotype, or at least spectrum, is due in different individuals to mutations in different genes.  Or different people with the same variant can have different symptoms. Other examples of similar multiple unilocus causation include retinitis pigmentosa, an inherited disease that leads to blindness in middle age, and another is congenital deafness.

Some individuals, including our daughter, have none of the known mutations.  We know this because a physician in Germany, Dr Frank Lehmann-Horns, generously donates genotyping and sequencing services to anyone who has been diagnosed with one of these disorders.  Affected individuals naturally would very much like to know the cause of their disorder, however, and when the cost of whole genome sequencing really is $1000 per genome, they will likely have their genomes sequenced so that a systematic hunt for causation may be undertaken by interested researchers.

Of course, finding the causative mutation in such situations, with hundreds of ion-channel genes, and their regulation, to search through, won't be easy when, as in our daughter's case, there aren't other affected family members to compare.  We all differ from each other at millions of loci in our genomes, and determining which one causes a given case, even focusing in on ion channel genes alone, is a challenge

Affected individuals don't need to know what causes their disorder in order to treat it, certainly, because it is the ion concentration that's the trait, regardless of its origin -- at least as is understood today.  Indeed knowing the gene that causes a monogenic disease is rarely useful in treatment: hundreds of such 'Mendelian' traits are known but few really treatable based on the gene in question. But, patients often worry, and indeed it's often the case that some doctor won't believe their diagnosis unless they have an identified mutation, so the identification can be important for that reason.  And, identifying as completely as possible the suite of mutations that cause this, and any multiple unilocus disorder could be useful in understanding how things go awry, and could in principle lead to better treatment.

Of course, we study and write about aspects of genetic causation and generally see complexity when others yearn for simplicity, but there is the danger that when the story strikes close to home, we might naturally drift towards a search for simple causation -- making the very 'gene for' mistake we criticize when others do it.

Still, while we do think that complex traits should not be treated as though they were simple, traits that really are relatively simple are a different matter.  The search to understand the genetic basis of complex multilocus disease is challenging.  The search to understand multiple unilocus traits, and to know whether they are only the clearest subset of multilocus versions in the population is somewhat different -- single gene changes might be easier to track and confirm when they are inherited.  The unexplained cases, like unexplained heritability that we've written about, may be those due to multiple, individually minor, genetic variants.  As we have often said, and said even before our daughter's diagnosis, the truly genetic disorders are where the money should go, at least to show that understanding causation at the gene level is an important way to approach life.

Similar issues apply to evolution.  A multiple unilocus trait favored by natural selection could arise in different individuals in a population because of mutations in different genes with similar effect.  Over time, the population could come to be made of individuals who had the favored trait.  But this doesn't mean that they share the same genotype or that there would be detectable evidence for natural selection in any specific part of the genome -- because many different genes could each have experienced only weak selection in the population as a whole.  If there are many roads to Toledo, none of them need to be superhighways.
------------------

Update: Our daughter outed us last year here on MT, noting in a comment to a post of Holly's that we'd all been enrolled in a sequencing study of rare Mendelian diseases.  The study is ongoing, but our exomes have been completely sequenced now and are currently being analyzed, but nothing yet found.  We certainly hope they identify the genetic basis of her disease, though it's probably a long shot.  But it would mean a lot to her to know the cause of the disease that too often rules her life.

1 comment:

Jennifer said...

And then there is lipedema (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3309375/) where people for years and years are told they are fat by their doctors and that it is all their fault. There are ways to manage this disease, and simply going on a diet isn't the answer.