Will we ever witness the perfect drug in our lifetimes?

February 17, 2013 9:20 AM

From a medical/science standpoint, what are the limiting factors that make it difficult to produce pharmaceutical drugs that do not have negative side effects?
posted by 6spd to health & fitness (22 answers total)

I can't imagine any potential wonder drug being completely free of any side-effects, given the variability of human physiology. There are just too many variables.
posted by Thorzdad at 9:29 AM on February 17 [2 favorites]

Basically, every target in the body does lots of different things. It's not like there's a single protein switch you can flip to go "Ok, anxiety off now". Even if you could fit a drug to hit one and only one target, which would be very difficult in itself, that target would likely do a half-dozen things, only one of which is your actual therapeutic effect.
posted by tau_ceti at 9:46 AM on February 17 [2 favorites]

I think the thing is, there is always one person somewhere, who will have a bad reaction. I mean, if you went out and invented air, odds are there is someone who is allergic to it.

That’s why every pharm ad you see spends 15 secs telling you how this is the best drug in the world for your particular ,malady, and 45 secs telling you all the ways it will either make you drop dead, or want to kill yourself.

You’ll never have a perfect drug, until you have a perfect human.
posted by timsteil at 9:50 AM on February 17

Assuming by "drug" you mean a chemical compound, then no, it isn't going to happen.

The real problem with the question is that we have silly beliefs about what diseases are and we lie to ourselves about our track record. Things like anxiety or depression have been simplified to a false story of 'chemical imbalances", we give pat, simple answers to the nature of huge mysteries (alzheimers), and we are pretty much deeply exaggerating our understanding of many diseases.

The limiting factor is not money, the limiting factors are that many of the mechanisms of disease are also used in the normal day to day operation of our bodies. Not only will there never be perfect drugs, there are many diseases that we live with today that will very likely never be cured or even treated substantially better than they are today. If you want to understand the hopelessness of the situation, you really need look no further than what cancer really is (millions of decedents of yourself, just better) or alzheimers (which, despite having been the target of billions of dollars of research since the 80s, is still not even properly understood - the amyloid hypothesis is still dominant but hardly justified by trials of drugs targeting that as an actual mechanism).

The real medicinal stories of our lifetimes are going to be how deeply we misled ourselves about the value of the human genome project and how we took an incredible array of drugs (antibiotics) and through crappy politics ended up rendering them useless, making us worse off by the end of this century than we were at the end of the last.
posted by rr at 9:52 AM on February 17 [8 favorites]

The limiting factor is that we don't understand biology well enough to design drugs. The way drugs work is that a drug is a molecule that sticks to other molecules in your body. Depending on what the drug sticks to, it alters different processes in your body. For example, a good pain drug might stick to part of the cells in your nervous system that process pain signals so that the cells can't communicate with each other - this would mean that the drug stops the pain message from being sent.

If the drug has no side effects, that means it doesn't stick to anything else in your body and therefore doesn't affect you in any other way. There are so many molecules in your body that drugs can stick to, and the molecules are all so sticky, that it's completely impossible currently for us to predict this and design drugs that are perfect. Even if we could do that, there could still be side effects. To use the pain example, the cells in your nervous system that process pain also have other jobs, and if you mess up their ability to send messages you might mess up those other jobs as well.

We just don't know enough to truly predict any of this, so drug development essentially works by trial and error.
posted by medusa at 9:53 AM on February 17 [2 favorites]

I would say that people are working on this problem, but it's not an easy one. For example, chemotherapy is commonly colloquially described as an attempt to poison a tumor a bit faster than the rest of the body. That's why chemotherapy has such awful side effects. Basically you inject it into the body and it circulates around the bloodstream affecting everything.

Ideally, of course, what you'd want is for the drug to go right to the tumor and not circulate around the rest of the body. This is Targeted Drug Delivery. It turns out that this isn't a really easy problem; basically you still have to get drug from the bloodstream to the tumor cells and only to the tumor cells.

There are a bunch of approaches that people are trying to take; one common approach is liposomes, where you basically put the drug inside a capsule made of fat. Now the problem is that the drug is inside the fat capsule. So people have made thermally sensitive liposomes that release when the capsule is heated, but this requires the tumor and only the tumor to be heated (easy for say, breast, harder for say, liver), and you still have to get the drug from the blood vessels through the extra-cellular matrix to the cancer cells. Other people have tried putting things on the outside of these capsules which would attach to the cancer cells, but depending on who you're asking, this is hard because these capsules have to still reach the tumor cells.

Anyway, a lot of people are working on it, but it's not an easy problem to non-invasively make something inside your body do what you want it to do when you want it to happen.
posted by Comrade_robot at 9:55 AM on February 17 [4 favorites]

Impossible. Not even a placebo -- which does sometimes work -- is side effect-free. People get side effects from placebos by the same psychosomatic mechanisms that make them occasionally efficacious.
posted by supercres at 9:57 AM on February 17 [3 favorites]

Evolution is conservative, the same biochemical mechanisms end up re-used or copied and slightly modified for completely different purposes.

That means that even if you can target a drug by design to interfere in a precise way with only one biochemical target (which is so far from what we can do that it might as well be magic), you're still going to have off-target effects on the other systems that use the same biochemistry.
posted by atrazine at 10:15 AM on February 17 [5 favorites]

Small molecule drugs are incredibly blunt instruments. For some of them, we don't even fully understand the mechanism of the primary effect, let alone the side-effects.

how deeply we misled ourselves about the value of the human genome project

The products of the human genome project have revolutionized the way biomedical research is done, no hyperbole. Most new developments in the understanding of biology or molecular physiology rely on it.
posted by grouse at 10:20 AM on February 17 [3 favorites]

There are many reasons:

- variability with the patient's ability to metabolize the drug
- variability with where the drug localizes and how much of the drug acts within the patient's body
- variability with how the patient's condition responds to the therapy
- variability with drug constituents (generics and branded drugs are often manufactured with different processes)

Every person has a slightly different metabolism which can affect how much of the drug is absorbed into the body (through the mouth, stomach or intestines), how much is metabolized into non-effective metabolic products (if the drug is administered through IV, say, and is bypassing the digestive tract), and how much passes through the body as waste products.

Once the drug manages to get into the body, where does it go? It's tough to control exactly where the drug goes, once it gets inside you. One classic, recent disaster was gene therapy administered to children born without immune systems, who were administered a modified virus packed with genetic material to help restore a missing class of immune cells. Unfortunately, in rare instances, the virus also acted where it shouldn't have, inserting the good genetic material in the wrong places, a mutation which caused leukemia in a couple patients. So drugs have to not only act on what's causing the disease, they might have to be designed to avoid acting in places of the body where it can cause damage.

As far as variability with response to therapy, cancer is a classic example. We're starting to see that cancer has a lot of genomic variability — even within a tumor, there can be variance in the population of cancer cells contained within. Across patients, such as in those with acute myeloid leukemia, one form of disease is not the same as another, even though they both make the patients sick. The genomic variability of cancerous tumors creates serious challenges for drug research. Some therapies work in some patients, because the therapy can target a weakness in that type of cancer cell. The same therapies do not work in other patients, because they have cancer cells with different genomic constituency, which lack those exploitable weaknesses.

With different companies making the same drug in different formulations, economic considerations can make it difficult to get the same efficacy from all of them. Some respond better to branded drugs than generics (although it is possible that this might be partially due to bias effects), because of a different mixture of fillers and other non-drug components that affect metabolism. Within a company's manufacturing process, also, there can be variance in the quality of product. Lax regulation, as in the case of the FDA not providing adequate oversight of a drug manufacturer, allowed people to contract meningitis from tainted drugs. If it happened there, you can be sure that oversight is equally lax elsewhere, and you are relying on a company to effectively exercise its own quality control measures. Without QC, the drug could, for example, have varied amounts of the active compound.

Some diseases we have drug treatments for, despite not knowing the exact method for the drug's efficacy. Lithium is administered as a psychiatric medicine and works pretty well, but no one is quite sure precisely how it works, although there are theories. It is tougher to control side effects when the drug's mechanism is unknown.

These are all difficult problems to solve, for various reasons.
posted by Blazecock Pileon at 10:43 AM on February 17 [2 favorites]

It's the inherent nature of the system that makes such a thing impossible. Every person is a mutant made from the DNA of two parents. My chemistry traits are nothing like either of my parents. Each had a single copy of few mutations and I drew the short straw (literally - I calced the odds to one in 16 million awhile back). The list of pain killers and antibiotics that work in my parents but have side effects in me is huge as a result. It is because I have a bunch of extra thing floating around in my blood that they don't, the result of that short straw I drew.

Now, I actually like the short straw. There are aspects that give me an edge in certain situations that I would never trade away. But because I have to tell people to stick to the simplest classes of drugs, skin test them first, and give a list of no-gos, it'd be difficult to come up with something for me... let alone a bunch of other folks with similar situations but different mutations.
posted by jwells at 11:05 AM on February 17

Apart from anything else -- there's no such thing as "side effects." There are effects, some of which are desired and some of which aren't. We can get better and better about maximizing desired effects and minimizing undesired ones, but it's an asymptotic problem, you'll never eliminate them completely.
posted by KathrynT at 11:22 AM on February 17 [2 favorites]

One way to think about the whole thing is that if you're working on a drug, you'll hopefully have identified what's called a "receptor" somewhere on (or inside!) a cell somewhere in the body that you'll want this drug to bind to. You can think of the receptor as a lock and the drug molecule as a key: the drug (key) enters the receptor (lock) and turns something on or off inside the cell that the receptor controls.

This receptor could be an enzyme in a pathway somewhere -- think of pathways as sort of like biological production lines, where a enzyme takes an unfinished product (a molecule) from another enzyme and does something to it, adds an atom here or there, shuffles things around, then passes it on to the next enzyme in line. If your body's making too much of the end product of this pathway, for example, you could block a receptor somewhere along this production line with a drug, starting by looking at the molecule that it accepts and then designing a drug that resembles that molecule, but with features tacked-on that allow it to bind more tightly to this receptor than anything else in the vicinity. This effectively "turns off" the receptor, at least until the drug's metabolized and then eliminated by the body.

Unfortunately, this lock-and-key model is kinda really crappy. For one thing, these locks and keys aren't "perfect" matches for each other to begin with, nor are they rigid -- like real-life locks and keys; rather, they're relatively small molecules and proteins, so they'll be bouncing and wiggling everywhere. Before they can bind, though, the drug and the receptor both have to be in the right "shape" (or conformation). So it's better to think of these locks and keys as kind of melty and squishy rather than anything resembling rigid. And the problem is that this melty and squishy behavior is really hard to model and think about, which means we aren't very good at predicting drug behavior a priori.

So, part of the idea's to get your drug to bind to its target receptor as tightly as possible, right? That's all well and good, but there are *so many* other kinds of receptor inside the human body, in all sorts of places, that your drug could possibly bind to and turn on or off to varying extents. Since drug molecules can "melt" to fit almost any lock, it's difficult to predict what other receptors your drug is going to hit, and the side-effect profiles that'll ensue. There are some tests that you can run to (hopefully) make sure your drug isn't hitting up any of the *really bad* receptors, like hERG and some serotonin receptor I can't remember right now, both of which can cause some pretty serious cardiac side effects if your drug turns out to like binding to them. You also have to consider what happens after your drug gets metabolized, too, as its activity profile could change a lot!

In sum,

1) What drugs do, where they go, and what happens to them inside the body is complicated -- "there are just too many variables" and interacting systems and processes;
2) What drug targets ("receptors"/proteins) do *before and after they're bound to a drug* is complicated, too, maybe even far more so, and nobody out there really knows how to model or think about this problem too well at all -- read up on protein folding, protein-protein interactions, etc. if you're interested;
3) You can't really predict what drugs will do inside the body without actually going out and dosing someone, then seeing what happens. However, the risk-averse nature of drug-design work makes it hard to even get to that step; for example, you start out by looking at molecules, and if one looks "bad," as in, it isn't shaped right or has the wrong atoms dangling off of it, you don't even give it a shot. Also, a lot of the data generated by the process is hard to interpret and put into context -- I think that I read somewhere that ~1/3 of drugs on the market bind hERG (which should be a big no-no) but the researchers decided to go ahead anyway since hERG binding isn't necessarily predictive of a bad cardiac side-effect profile.

Hope this helps!
posted by un petit cadeau at 1:13 PM on February 17 [1 favorite]

Drugs generally affect certain neurotransmitters, in certain ways. For example, SSRIs, selective serotonin reuputake inhibitors, keep more serotonin active for longer. The thing is, neurotransmitters do many, many different things, in many parts of the brain, and we can't direct drugs finely enough to only affect specific neurotransmitters in a very specific part of the brain. Even if we could direct them that finely, they still do more than one thing, even in a very specific area, and so - side effects. For example, (just making this up, because I don't feel like looking up stuff right now) say serotonin in the hypothalamus affects depression, but it also affects memory and sleep patterns. So, we made a drug that targets a cause of depression, by making more serotonin available in that part of the brain, but the exact same mechanism does those other things too. There's no way to avoid those side effects.
posted by catatethebird at 1:14 PM on February 17

I'm not sure what you mean by "do not have negative side effects."

I mean, I can walk in CVS and there are aisles and aisles and aisles of useful drugs that I am allowed to just buy and take home without involving a doctor.

I mean, I guess in theory you can OD on too much aspirin, and some people might have allergies, but all those drugs in there have few to no negative side effects to speak of- that's why you can get them without a prescription.
posted by drjimmy11 at 1:31 PM on February 17

I mean, I guess in theory you can OD on too much aspirin, and some people might have allergies, but all those drugs in there have few to no negative side effects to speak of- that's why you can get them without a prescription.

You are deeply confused about the danger level of over the counter medications.
posted by rr at 1:39 PM on February 17 [3 favorites]

Human physiology is legacy spaghetti code, basically. Everything's interconnected and interdependent, and you can't do one thing without having effects on several other system.
posted by WasabiFlux at 1:49 PM on February 17 [1 favorite]

There is always the possibility that a person's immune system will react to a given chemical, so no drug can ever be completely benign to all individuals. Hell, there are people that suffer reactions to water.

But discounting hypersensitivity reactions, I can think of certain types of drugs that have basically no side effects: those that are not absorbed. There are antifungals, antihelminthics, and other anti-infectives that target organisms purely on the skin or in the GI tract. In principle (though often not in practice), these can be effective without causing adverse effects. Off the top of my head, nystatin is a good example. When used as directed, it is essentially without side effects, excepting of course the inevitable rare hypersensitivity reaction.
posted by dephlogisticated at 1:57 PM on February 17

Here are some reported adverse effects for Nystatin in DRUGDEX:

Nausea, vomiting, and diarrhea (including one case of bloody diarrhea) has been reported with nystatin use… Gastrointestinal upset/irritation has been reported with nystatin use… Oral irritation and sensitization has been reported with use of nystatin. Discontinue nystatin therapy if sensitization or irritation occurs… In rare cases, nonspecific myalgia has been reported with nystatin use… Following aerosol nystatin administration, respiratory tract irritation manifested as cough or bronchospasm has been reported on rare occasions.

posted by grouse at 2:07 PM on February 17

grouse, you mostly missed my point. Adverse effects with nystatin are not common, and when they do occur, are largely due to local irritation. That sort of reaction can occur in response to just about anything, even the vehicle carrying a drug. It's not 100% without side effects (nor can anything be), but as far as drugs go, it's very innocuous.
posted by dephlogisticated at 2:30 PM on February 17

A "side" effect depends entirely on what you're taking the drug for. If I take finasteride (Propecia) to shrink my prostate then the extra hair is a side effect. If I'm taking it for baldness then the prostate stuff is a side effect.

Now if I have an enlarged prostate and male pattern baldness, then, horray, I've found a drug with no side effects!*

*Assuming I wanted the cold sweats.
posted by Ookseer at 4:09 PM on February 17

what are the limiting factors that make it difficult to produce pharmaceutical drugs that do not have negative side effects?

Making drugs that target bacteria without hurting humans is relatively easy, because there are pretty significant differences between bacterial biology and human biology. Some of the more powerful, broader-spectrum antibiotics do have other side effects, e.g., sun sensitivity, nephrotoxicity, but most antibiotic side effects have to do with killing beneficial flora in the gut.

Making drugs that target fungi is a bit harder, because fungal biology is closer to human biology. It's still possible to hit just one or the other, but there's some inevitable overlap.

Making drugs that target virii is really hard, because viral biology is human biology. Most antiviral drugs don't actually kill the target organisms, they simply inhibit their reproduction. Going straight after the virus is kind of like going straight after cancer cells: you're killing everything, you just hope that the virus/cancer dies faster than the patient. This is frequently a good bet, but it's No Fun At All.

Drugs that are intended to alter or adjust human vital signs without actively attacking anything almost always have side effects, simply because the human body is an integrated organism. Roughly speaking, push it here and it'll poke out over there, and not always in readily predictable ways. Warfarin, a common blood pressure medication? Works by interfering with the body's use of vitamin K. The primary effect is to reduce coagulation, but it turns out that the body also uses vitamin K to regulate certain bone proteins, leading to an increased risk of osteoporosis and related fractures, simply because of how the drug works.

So the main reason that it's hard to make drugs without side effects? Because the human body is damned complex and interconnected, and you can't mess with it--or things similar to it--for free.
posted by valkyryn at 4:13 PM on February 17 [1 favorite]