The article is pretty good, actually, in that it doesn't try very hard to claim that they're curing the world of its ills. There's a little in there, but mostly it deals with Fragile X.

Randi Hagerman (the researcher quoted extensively in the article) is one of the leading lights in Fragile X research. She and her husband, Paul, described the gene, developed the RFLP that we now use to diagnose the illness, and did much of the fundamental work to explain the genetic-expression behavior of the gene. It is not a simple inheritance model, and the expression of the gene is quite confusing. She's a superstar.

As far as the broader issue of autism (and even more confusingly, autism spectrum), Fragile X has always seemed to me to be a blind alley. People with Fragile X (I've worked in that community as a physician) have a very specific affect and behavior pattern that doesn't look a lot like the behavior of people with autism (a community I know all too well as a physician and a parent of an autistic young man). Most of the early research in autism was tainted by the inclusion of Fragile X patients, and most of the combined research is just confusing.

I hope that the drug proves useful in Fragile X, although pharmacotherapy for these kinds of disorders has frustrated us over and over again. These are simply very hard diseases to affect very much. At the least, though, it'll be another step toward understanding a serious disease. And I'll continue to wait and watch for anything that will help in autism, but I REALLY don't expect much from this specific drug.

The FDA has had a table of valid genetic biomarkers for medications for several years now. While many of these are cancer drugs looking at specific metatabolic or receptor issues, our old friend warfarin (a "blood thinner" with a narrow therapeutic index, a reputation for causing a lot of trouble and a genomic profile that accounts for about half of the known variation in the drug) and the pain drug codeine are on that list as well. There's even a research website devoted to genetic calculation of warfarin dosing.

Carbamazepine (Tegretol) can cause a rare life-threatening reaction called Stevens-Johnson Syndrome (Toxic Epidermal Necrolysis), but it's mostly limited to individuals with a specific Human Leukocyte Antigen (HLA-B*1502). Again, known for quite a while and a part of the basic biology of the drug.

It's a fairly well-written article, but it's kind of breathless about stuff that I was really excited about back in the '90's when my medical school teachers were really excited about it too. The best news is that the FDA has really stepped up in the past few years to make this actionable data that a practicing clinician can use.

Okay, so let's give credit - this is a legitimate researcher doing interesting (if highly preliminary) work. From his bio, accessible from TFA, you find that:

We have been assessing CSF and plasma samples from human subjects at the Washington University ADRC and have found that decreased CSF [alpha-beta] and increased tau are harbingers of cognitive decline in cognitively normal elderly.

Which suggests that the increase in CSF beta protein seen in sleep deprivation might actually be a harbinger of protection from Alzheimer Disease (AD). Or not, and it's not possible to know right now. Your speculation is just as valid as mine.

The problem is that we don't know if the protein causes the neuronal damage in AD or is a side effect of the damage, like a clot or a scar. Dr. Holzman's research bio makes that clear, and it also makes it clear that the damage, whatever it is, starts decades before the protein levels become abnormal. So if you want to avoid sleep deprivation, that's cool, and the fact that most people reading this site are hopelessly sleep deprived most of the time is probably cool too. Either way, our other lifestyle issues will likely collect us long before our brains start to rot.

On the plus side, we now know how to make mice demented. It's not much, but it's something.

Amen. The vaccine has showed animal immunogenicity, which is not a bad thing, but since the animals in question don't get AIDS from HIV, their immune systems don't react the same way that human ones do. Which means you need to proceed to human testing, and that takes a long time.

Phase I trials are important, and announcing them is not a bad thing. And nobody particularly expects cures in the HIV-positive population, although circulating HIV may be interesting (if the virus can cause a practical immune response in subjects with HIV but who have fairly normal T4 counts and you can show reduced circulating viral load, you have an interesting data point for efficacy).

My biggest problem with this kind of press release is that they don't include the details. I'd be interested in knowing why this vaccine is likely to work better than the last two hundred that have been tried, what the actual animal studies showed, and so on. Oh well. I'm not going to be waiting up this weekend to hear more. It will be a couple of years before we know whether this one works.

Unfortunately (OK, it's not unfortunate at all, actually), option #1 isn't just taboo, it's impossible in any human society. Even if we had a completely accurate test (which we certainly do NOT have), and even if you could somehow prevent all positive contacts from continuing their infectious behavior (and I'm not sure anything short of summary execution would be reliable), you'd still have leakers, avoiders, corruption, and resistance. Not even North Korea has managed to avoid HIV, although they're close, at least by report. Largely, I suspect, because summary execution is a routine thing for them.

Option #2 is science fiction for now. The genetic resistance to HIV is conferred by the lack of a particular cell-surface receptor, so you'd have to find a way to effectively eliminate that piece of genetic material from every genome in the body. And since T cells are quite long-lived, you'd have to mess with a lot of quiescent DNA to do it. Maybe some day, but not soon.

Which leaves vaccination. I'm in agreement with your skepticism on this one, not because it's impossible but because HIV, due to its unique targeting system, has been very intractable. You do have to target relatively stable regions of its proteins or its DNA, but this isn't unique to HIV, and we've solved it with polyvalent vaccines elsewhere (think Menactra, or the recent HPV vaccine, or even the flu shot). HIV is a pretty wimpy virus from a spread perspective, so a good polyvalent vaccine would seem possible. The problem is practical immunogenicity, and that's the issue that has torpedoed previous vaccines.