Saturday, January 21, 2017

Headbangers' Ball

I am so excited for today's post.  Before we get there, though, I have a favor to ask.  I love writing this blog, I'm having so much fun doing it, and I, personally, think the quality is pretty good.  If it just stays between me and some Facebook friends, great, but as with most things in life, there's always that dream of things you love taking off.  And so I'm trying to build my reader base.  So please, if you read this post and find it interesting, share it! Facebook, Twitter, a gchat to a friend who's into this sort of thing, whatever.  I am eternally grateful and love y'all for indulging in my passion.  And  on to the show.

Anyone who knows me will tell you that I love sports.  Love.  It's been at the top of every dating profile I've ever had, it's on my about me page here, it's a defining feature of my life.  There's a huge problem with my love though: my two favorite sports are football and hockey, and it's difficult for me to reconcile my enjoyment watching with my knowledge of what they do to the brain.  This has been a major conflict in my life, and I'm a major proponent for improved concussion protocols.  This post isn't specifically about traumatic brain injury in sports, but some new potential advances in the diagnosis of concussions, and I'll probably talk about how that can be integrated into concussion protocols, because I just can't help myself.

Concussions occur following some sort of trauma to the brain; while we normally think of it as someone getting hit in the head, they can also occur following a blast or sudden acceleration.  They're a mild form of traumatic brain injury, although the classifications of concussions and TBI are fuzzy and not agreed upon.  The effects of concussions include headaches, emotional reactivity, memory problems, visual problems, attention problems, sleep problems, and other cognitive problems.1  These effects seem to be due to a confluence of mechanisms.  The first is the most obvious: when a person gets hit in the head or experiences a rapid acceleration, it can be too much for the cerebral spinal fluid, the fluid that surrounds the brain and protects it from minor disruptions.  Think of it like a ping pong ball floating in a glass of water.  You could walk around holding the glass, and the ping pong ball would bob around, never actually making contact with the side of the glass.  If someone pushed you, though, that force would be enough to make the ball hit the wall of the glass.  You end up with damage to the cortex wherever the brain hits the skull, and often times in areas directly across from each other, as a hard enough blow can make the brain hit the skull on one side, then rebound back and hit on the other.  The second mechanism of concussions is a little more complicated.  Unlike our ping pong ball, the brain isn't hard; it's more the consistency of jello.  That means that when the brain makes contact with the skull, a wave of force travels through it, disrupting a lot of things that aren't just the cerebral cortex, that outside gray matter.  That wave of force often does damage to the white matter that's just inside the gray matter.  White matter is made up of bundles of axons, the communicating part of neurons, and connects different parts of the brain.  Damaging these connections can have a whole host of results, including slower thinking, emotional effects, memory problems, and movement problems, and those effects can persist for at least a year following the injury.  These white matter injuries have been suggested to be the major underlying mechanism of traumatic brain injury symptoms, more so than cortical damage.2  For a fantastic demonstration of how concussions cause damage, check out this video that was part of a NY Times article by Sam Borden, Mika Grondahl, and Joe Ward.

Currently, concussion diagnosis primarily occurs by neurocognitive testing and where someone falls on a measure called the Glasgow Coma Scale.  The GCS measures verbal response, pupillary response, and motor response.  If a score is sufficiently low, brain imaging is often ordered to rule out more critical issues like bleeding.  Neurocognitive testing may include tests of memory, word fluency, information processing speed, and attention.  TV shows and movies have led to the idea that the first thing a doctor will do is put someone in an MRI, CT, or PET scanner, but the truth is that most concussions don't have any sort of marker that you can see in imaging.3  EEGs are often also inconclusive when it comes to diagnosis4, and overall, 88% of people who suffer from symptoms of a concussion do not recognize that fact.5

In the past month or so, a few news stories have come out championing new diagnostic screening tools for concussions.  One of the side effects that often accompanies a concussion is difficulty sorting and processing sounds.  Basically, all sounds are on the same level, so picking out the person talking to you from background noise in a restaurant or a singers' voice over screaming at a concert is nearly impossible.  My sister suffered from a concussion when she was sixteen after hitting her head on the kitchen floor, and this is one of the things about the experience that is most salient to me.  She was so hypersensitive to sounds that leaving the radio on a low level, mostly out of everyone else's hearing range, would drive her absolutely crazy.  She couldn't be around large groups of people, couldn't go to the mall, in part because she couldn't deal with the noise.

Let me detour briefly into sound physics: everyday sounds aren't made up of just a single frequency, or pitch, like the ones used in a hearing test.  Different frequencies of sound waves are essentially overlapping each other.  The lowest of these frequencies is called the fundamental frequency, or F0.  That's the frequency that carries information about pitch, prosody, vocal stress, all the things that are necessary for identifying speech and who's speaking.  The better you are at discerning F0, the better you are at being able to pull critical sounds out of environmental noise.6  The other, higher frequencies, are called harmonic frequencies and are less important to our purposes.  Given this, a group of researchers at Northwestern University in Chicago measured a brain response called the frequency following response (FFR).  The FFR is measured by placing three electrodes on the scalp and measuring electrical activity in the auditory brainstem, an area where signals from different parts of the brain and different parts of the processing stream converge, following the presentation of periodic sounds.

Researchers found that children diagnosed with a concussion following a sports injury tended to have smaller responses to the F0 than did healthy kids, and the size of the response was negatively correlated with the number of symptoms the child was experiencing.  The kids also tended to have slightly, but significantly, slower FFR responses, suggesting that both the magnitude and the speed of of this processing can be diminished in children with concussions.7  Basically, there seems to be a difference in both the magnitude and the timing of neural signals when processing speech following a concussion, and the timing difference may be reflective of the previously mentioned white matter damage in concussions.  What this boils down to is that children experiencing a concussion have less ability to identify, group, and sort sounds, all of those symptoms my sister experienced, but, more importantly, that these symptoms are correlated with an observable biological measure.  The researchers also found that as the kids' symptoms reduced, the FFR to F0 started to recover, which suggests that this is correlated with injury state.

So is this quick, non-invasive brain recording really a new diagnostic measure?  The follow up analysis showed that using this brain response to try to split the children into concussed and non-concussed groups had a 90% sensitivity rate (said 90% of the concussed children had a concussion) and a 95% specificity rate (said 5% of the control kids had a concussion when they didn't).  That's pretty good  for a diagnostic test.  Take it with a grain of salt, because they only looked at this phenomenon in kids and the sample size was small, only 20 subjects.  In my opinion, that would be concerning for the diagnosis of some things, but if you get a good bonk on the head without a concussion and happen to be in that 5% that get a false diagnosis, any concussion treatment isn't going to hurt you.  It's definitely going to take a while for this to start making its way into hospitals, and just because of the way science works, I actually doubt that it ever will, but the hopeful part of me would love to see some more research into the F0 FFR.

My second new diagnostic tool has been showing up in a lot more news articles than the auditory stuff, probably because it's a lot easier to conceptualize as a diagnostic tool: the tried and true blood test.  When the brain is damaged, it increases the production of a protein called tau.  This happens in Alzheimer's  disease, dementias, and some forms of Parkison's.8  It also happens following a head injury.9  Tau is produced in the brain, crosses the blood brain barrier into the cerebral spinal fluid, and ends up floating around in the peripheral blood stream.  In theory, you could measure the level of tau in the blood to diagnose whether damage to the brain has occurred.  Initially there were a few problems with this, revolving around the fact that even in people suffering from brain damage, levels of tau in the blood are extremely low.  This meant that a) a blood assay needed to be developed that was sensitive enough to detect tau, and b) the tau detected had to be significantly different from the amount in healthy controls.  Whenever you have something that occurs in low levels, any sort of test has to be SUPER precise in order to differentiate from even lower low levels.

Such an assay was developed in the early 2010's by the Quanterix corporation, which yes, sounds kind of evil, but as far as I know, isn't.10  This assay allowed researchers to look at multiple potential concussion markers in the blood, and the only one that consistently arose as significant was tau.  For example, a study in 2014 looked at levels of tau in blood in hockey players and were able to discern a difference between  a baseline level before the season and after they had sustained a concussion.  Elevated tau was measured in the blood up to six days after the concussion, when players were allowed to return to practice.11  Two important takeaways here.  The first is that it does seem that tau increases following a concussion, and the second, equally important point, is that perhaps concussion protocols allow players back into the game too quickly.  That's all well and good, but it says nothing about our ability to differentiate between concussions and not, simply that we know tau increases.  That's where the recent news stories come in.

My first point that I want to clear up is that there have actually been news stories about two different concussion blood tests in the past month.  The first study was performed in athletes at the National Institute of Health.  It showed that increases in tau were evident in blood in both concussed athletes and control athletes up to seven days after playing in a game.12  This isn't the first study to show little difference between athletes regardless of concussion status13, so this raises some huge questions about the damage that's being done by small, non-concussive hits over time.  What this study did find that could be thrown under the "diagnostic" category is that concussed athletes with long recovery times following their injuries had significantly higher tau for up to three days than did concussed athletes with shorter recovery times.  This could be hugely important when it comes to evaluating athletes for timing on returning to play when it's right for them as opposed to a standard protocol, but there's one major caveat:  this test only had an 81% accuracy rate.  As critical as it could be to determine markers of recovery time for personalized protocols, this one definitely needs some more work before it's ready to be used on the sidelines.

The second blood test story is from a study done at Western University in hockey players because, well, it's Canada.  This study looked at preadolescent males roughly two days after suffering their concussion.  The blood test they used actually didn't end up using tau at all; rather, they found 17 metabolites out of 174 that, when looked at together, were 91% accurate at classifying kids that had sustained a concussion.14  Again, that's pretty good diagnostic accuracy, and I really like that this study looked at the way biomarkers together can indicate a diagnosis.  I am a little suspicious of something, given that previous studies hadn't found a lot other than tau to be related, but also encouraged that if the NIH study, for example, were to look at tau in combination with other things, accuracy may rise.  We also don't know if any of these metabolites appear immediately after an injury, or if more time is needed in order to judge.  The time period immediately following a concussion is critical in terms of treatment, so a diagnostic tool that was capable of leading to a determination within hours would be ideal.  In terms of sports, something that could be done on a sideline to determine whether a player was safe to go back into the game is everyone's dream, and two days just isn't practical for that.  The other problem with this study is a crazy small sample size: only 12 concussed kids and 17 controls were included.  While it was a good start, this is no where near ready for news outlets to slap a "Concussion Diagnostic Blood Test" headline on it.  There are a few other blood tests being developed that look at some novel markers, but this isn't quite yet the "game changer" the media has labeled it.

Saturday, January 7, 2017

Original Recipe, Extra CRISPR-y

Today we're not exactly talking about breaking news; CRISPR gene editing has been talked about extensively for a couple years now.  It was named the American Academy for the Advancement of Science's (AAAS) breakthrough of the year in 2015, and articles are coming out constantly with new applications and successes.  I, like most science journalists, find it super cool but pretty complicated, so I wanted to talk about what exactly it is, what has developed recently, and what it might be able to do.

One of the major confusions with CRISPR is that the name is used to describe two different things.  You'll often hear it used to refer to a gene editing technique, but the name for that technique was taken from a phenomenon in bacterial DNA that it exploits.  CRISPR the phenomenon stands for Clustered Regularly Interspaced Short Palindromic Repeats.  It is a complicated name, likely mostly based off of the discoverer's desire to make a pronounceable word, but is actually pretty descriptive of what they are.  CRISPRs are sections of bacterial DNA that contain repetitions of the same sequence of A-T and C-G base pairs over and over again.  These repetitions are separated by segments of foreign "spacer" DNA from viruses and other infectious agents that have attacked the bacteria at some point.  There's a three step process involving CRISPRs: first, the bacterial genome and the viral genome are both cut, and a section of the viral genome becomes a new spacer portion of a CRISPR.  Second, the bacterial DNA with the new spacer undergoes transcription- the process where DNA is made into an RNA template, used for producing more copies of the DNA.  Most importantly, the RNA allows for the third step in the process: it recognizes matching DNA that's introduced to the bacteria, an invader that the bacteria has seen before, and guides enzymes known as CRISPR associated systems, or cas proteins, to cut up and destroy the invader.1  In this way, the CRISPR/cas system is basically a bacterial immune system making antibodies new antibodies when introduced to a new foreign invader, and deploying antibodies when it recognizes a foreign invader.

That's the bacterial phenomenon part.  Scientists have discovered that this system can be altered and utilized to edit genes.  If, instead of invading viral DNA, they use specially designed DNA sequences, they can essentially tell the CRISPR/cas system to cut DNA at any point that they want, allowing them to manipulate an organism's genome by turning on or off very precise genes, removing mutations that may be due to repeats, or potentially allowing other sequences to be inserted.2

This is super cool, but gene editing isn't new- remember having to make glowing bacteria in a high school biology lab?  Even before CRISPR, gene editing has gotten a lot more advanced than just transfecting jellyfish DNA into bacteria, but people are still talking about CRISPR as if it is going to change everything.  It has several advantages over other gene editing technologies.  For one thing, it's cheap.  Compared with having to pay somewhere between $3,000 and $5,000 dollars for specialized DNA segments with enzymes that are kind of a craps shoot, labs can purchase CRISPR kits for around $75 dollars, and they are far more likely to work.  Because CRISPR RNA can recognize sequences of around 20 base pairs, you get more specificity than with some traditional methods that only recognize four base pairs.  Think of it as doing a find and replace in a document.  If you just look to replace "ed" with "ing", you're going to end up accidentally changing the word "seeds" to "seings", when that's really not what you what.  If you look to replace "crawled" with "crawling", you're going to get much more specific results.  CRISPR allows for the same sort of specificity, just with genes.  There are other methods of gene editing that work with even longer sequences of base pairs, and therefore allow for even more precision, but those have the problem listed above- that a special protein has to be created for everything you want to use it on, and there's a lot more room for error.  This is because the CRISPR/cas system comes with its own enzyme for cutting DNA.  Instead of having to have a cleaving enzyme, guiding RNA, and a system for healing the spliced DNA, all of which have to be correct, CRISPR eliminates those difficulties by essentially being a self contained package, making it cheaper, faster, and more efficient.  It also has the advantage of allowing you to target more than one sequence of DNA, allowing you to manipulate multiple genes at one time.3

With most new scientific techniques, people might talk a big game about the implications, but the actual applications are probably a least a dozen years off.  Not quite so with CRISPR.  It first started being looked at for gene editing in 2012, and in the four years since, over 1,000 papers have been published.  Research is moving so quickly that it is already coming up with solutions for some of the early limitations.  For example, scientists have already found an alternative enzyme with CRISPR sequences that cuts DNA without needing a specific DNA sequence nearby.  By tethering cas to an enzyme that doesn't just cut DNA but actually converts one base to another, an A to a C for example, we're moving forward on not just cutting DNA and removing segments, but actually changing DNA.  Originally, the CRISPR/cas complexes were too large to put into a lot of the vehicles that are commonly used to introduce them to a genome, but a possible solution for that has recently come along as well.4

The really amazing question with CRISPR is what being able to precisely edit genes means practically.  Here we get into the really cool stuff.

Disease research: The most common way of studying human diseases in animals is to use knockout mouse models.  These are lines of mice that have been bred to not have a certain gene that we think plays a role in a disease- obesity, cardiovascular disease, depression, etc.  We can use knockout mouse models to study behaviors caused by a gene, potential therapies, and the interactions of genes with environment, among other things.  CRISPR allows the possibility of creating new knockout mice by allowing us to a) target new genes, b) develop a knockout mouse line in a period of weeks instead of months, c) saving money, time, and several generations of mice's lives.5  These knockouts could also potentially have alterations in several genes, allowing for more complex and accurate models.6  CRISPR is being looked at to create primate models of disease that even more closely mimic human diseases, a path that hasn't really previously been open given the ethical concerns of creating a line of genetically engineered primates.7

Disease treatment: CRISPR has already made its way into clinical trials for disease treatment.  China has used it in non-viable human embryos to alter the genes that code for β-thalassaemia, a potentially fatal blood disorder.  Only 33% of the embryos survived and were successfully spliced, at which point they stopped the experiment until CRISPR research matures.8   A second group in China has also used CRISPR in non-viable embryos, this time attempting to splice in a gene that codes for HIV immunity.  This experiment was only successful in 15% of embryos, but these two experiments serve as a proof of concept for the potential application of treating genetic disorders or granting immunity from disease before embryo implantation.9  Obviously there are huge ethical controversies involving this research, which is another post for another time.  On the more immediate front, in October of this year the first trial of CRISPR in a living, adult human began.  Doctors in China delivered CRISPR modified cells into a patient with aggressive lung cancer; they have not released any followup reports on the success of the trial.  A similar trial is approved to begin in the US in 2017 to treat various cancers.  The idea of these studies is to use CRISPR to target mutation in genes that are causing cancer, attempting to stop the spread and development.  From what I can tell, this isn't a treatment for any current tumors, more that it will keep cancer from growing, spreading, and returning.10  Another very interesting disease treatment application is using CRISPR to treat naturally occurring viruses in pigs.  Although pigs are some of the most similar animals to humans biologically, these viruses have kept them from being widely used as hosts of transplant organs.  By addressing the virus issue, porcine organs have a much lower chance of rejection and may become more viable for human transplant.11  In mouse models, scientists have also used CRISPR to treat retinitisa pigmentosa, a cause of blindness,12 to correct tyrosinemia, a genetic liver disorder,13 and to improve cholesterol.14

Agriculture:  There are a couple big areas utilizing CRISPR in agriculture.  The first is the obvious: genetically modified plants that increase yield, nutritional content, and photosynthetic efficiency, faster and with more precision and efficiency than traditional GMO crops.15   Basically, more crops that provide more nutrients for us.  All good things, but you still have a lot of the GMO arguments currently going on (another post for another time).  Plants can also be modified with CRISPR to give us more options for plant based drugs and vaccines.  We can use plants to do the work of creating proteins and metabolites for us, and CRISPR can both increase our knowledge of these systems and, again, increase specificity.16 

Petrochemicals: I love this one because it goes beyond the obvious applications of "genetic engineering".  Fossil fuels are a problem.  Burning cleanly, having enough, the damage on the environment.  We really need to come up with efficient alternate energy sources, and CRISPR offers us a chance at just that.  By engineering bacteria, yeast, and fungi to control the hydrocarbons they produce, there's the opportunity for a lot of innovation in producing biofuels, plastic polymers, and adhesives.17

These applications don't even begin to get into controlling reproductive and feeding drives of animals, genetically modifying mosquitoes so they don't carry malaria, selective breeding of livestock, creating antibacterials and antibiotics that are more targeted and don't carry the risks of making immune bacteria, and a host of other applications.  What is truly amazing about CRISPR, besides the wide range of impacts, is the speed at which these developments are happening.  Basically all of the aforementioned breakthroughs have happened in the last three to four years,outpacing just about every other technology by miles.  From the first paper coming out in 201218 saying the CRISPR/cas system could potentially be manipulated to clinical trials in humans in a four year span is insane when it comes to science.  Keep an eye on CRISPR in the next couple years.  This one really does have a the promise to change a lot of things, and change them sooner rather than eventually.