I was talking with a colleague the other day who’s done a lot of work on central nervous system disease over the years, and it reminded me of something that I said years ago on this blog (and was the first time I was quoted in the Wall Street Journal). Was that an opinion about some company’s stock or the biopharma climate? Nope: it was the statement that “we don’t know jack about how central nervous system drugs work”.
I’m still willing to stand by that one, although it’s certainly true that there are some drugs (and some categories of drugs) that we understand a bit better than others. But my point was directed to what people outside the field might imagine that we know about (for example) drugs for depression. Think about that one – it’s a very serious problem for those suffering from it, there are a lot of such people, and there are quite a few drugs for the condition, some of them famous and recognizable enough to have appeared in book titles. A lay observer could easily be forgiven for thinking that we understand how antidepressants work. But we don’t.
It’s a famous enough field that even the stated mechanism-of-action of many of the drugs in it is known to far more people than most drug mechanisms. Ask ten thousand random people how ibuprofen works and I don’t think you’ll get very far. But ask them how Prozac works, and I’ll bet someone pops up and says that it makes your whatchacallit, serotonin, go up. Serotonin is that stuff that makes you feel better, you know.
Well, that is actually what Prozac and the other SSRI drugs will do, up to a point – they will definitely affect the serotonergic signaling in the your brain. Past that, though, I would not want to put any money down, except against the proposition that serotonin levels = happiness. The unofficial motto of all CNS research is “It’s actually a lot more complicated than that. . .”, and so it is in this case, too. The long search for connections between serotonergic-related gene variants and depression has been frustrating in the extreme, for example, and if the story were a simple one you’d have hoped to find something there.
Here’s a review of the field from a drug discovery perspective, and the references therein make several points clear. The response rate to the SSRI drugs (and all antidepressants) is variable and unpredictable. Large numbers of patients show partial responses, at best, and many of them discontinue the course of treatment. The efficacy of the newer agents is not provably better than the older ones, although (since they have fewer side effects) people are willing to stick with them longer. In some cases that’s a good thing, because the onset of what beneficial effects there are can take quite a while to show up. And sometimes there are beneficial effects for a while, which then slowly disappear.
If we knew more about what was happening in the brain during depression, we might be able to do better. But that takes us back to the first paragraph: what happens, on a neuronal level, when a person has an onset of major depression? Despite a great deal of work on this question, we still have to throw our hands up in the air. The number of factors at work, the subtlety of their interplay, and the crudity of our tools with which to study them all conspire to keep us ignorant of the real story.
Let me say here that I’m not a nihilist: I believe that there is a real story, and that we will eventually know it. It could turn out to be something like “If you have more than X% of Type 473 neurons in subregion 79Q of the forebrain, you’re at risk for slipping into Alternate Neuronal Network Firing Patterns #2907 through 3043 if the balance of long-axon activity decreases, potentiating a feedback loop of protein synthesis that produces mixed-GPCR phenotypes #973 through 1028, spreading distally through axonal networks 384 through. . .”
But we have no way to tell anything about this level of detail in a living human brain, in neither the spatial nor temporal resolutions needed to start tracking patterns like this down. We’ve made vast amounts of progress, but it’s still like trying to map out how each snowflake fits together with its neighbors in a huge drifted pile. Actually, that would be easier: snowflakes are all roughly the same size compared to the variations between neurons, and they’re all made out of water, as opposed to having tens of thousands of proteins scattered through them. And, of course, they’re not constantly all talking to each other, or not that I know of.
The most progress has been made on brain function in its more mechanical aspects (such as the visual cortex). That is very much hard enough, but when you start moving into higher brain functions like cognition and emotion, you have most certainly slipped the surly bonds of earth as far as mechanistic explanations go. I’m in awe of the brain, and yes, it does indeed give me the creeps that the phenomenon of “being in awe” depends on my brain as well. Consciousness is one of the hard problems out there.
But we don’t need to understand consciousness (fortunately) to have better antidepressants than we do now. In fact, I think that useful agents are more likely to come empirically. Our progress is not going to be pretty, and it’s not going to be swift and sure, but as we investigate more combinations of neuroactive agents and look out for more effects of existing compounds (of all kinds), I actually think that the chances are good that we can come up with something better than we have now. Admittedly, that’s partly because what we have now isn’t that great. It’s going to be quite a while before we can take pride in our detailed cellular and molecular understanding of depression – but in the meantime, I think that we can still help people suffering from it, and if we can’t take pride in that, no matter how we stumble up on the treatments, what can we take pride in?